i the impacts of land use / land cover … · the impacts of land use / land cover changes on the...

i

THE IMPACTS OF LAND USE / LAND COVER CHANGES ON THE TROPICAL

MARITIME CLIMATE OF PUERTO RICO

A Dissertation

Submitted to the Faculty

of

Purdue University

by

Angel R. Torres-Valcárcel

In Partial Fulfillment of the

Requirements for the Degree

of

Doctor of Philosophy

August 2013

Purdue University

West Lafayette, Indiana

All rights reserved

INFORMATION TO ALL USERSThe quality of this reproduction is dependent upon the quality of the copy submitted.

In the unlikely event that the author did not send a complete manuscriptand there are missing pages, these will be noted. Also, if material had to be removed,

a note will indicate the deletion.

Microform Edition © ProQuest LLC.All rights reserved. This work is protected against

unauthorized copying under Title 17, United States Code

ProQuest LLC.789 East Eisenhower Parkway

P.O. Box 1346Ann Arbor, MI 48106 - 1346

UMI 3605159Published by ProQuest LLC (2013). Copyright in the Dissertation held by the Author.

UMI Number: 3605159

ii

En primer lugar dedico este trabajo al creador por regalarme las destrezas que

poseo y he logrado desarrollar para realizar este proyecto tan importante en mi vida.

También le dedico este trabajo a mi mamá, Angela L. Valcárcel De León quién con

mucha razón me obligó a ir a la escuela todos los días y me enseñó a ser responsable, una

cualidad que parece haberse perdido en nuestra sociedad. Le dedico este trabajo a Puerto

Rico, “Borinquen Bella” mi única amada tierra por la cual trabajo y trabajaré todos los

días de mi vida para que sea un mejor país lleno de paz, armonía y prosperidad. Le

dedico este trabajo además a mi Alma Mater “Universidad de Puerto Rico” (UPR) donde

comenzó mi carrera universitaria y mis sueños profesionales en 1988 como todo prepa

que ahora culmina alcanzando el máximo grado académico existente. Más del 80% de

mis actuales conocimientos y destrezas los adquirí a lo largo de los cursos y experiencias

obtenidas durante el bachillerato y las dos maestrías en el sistema UPR.

En especial dedico este trabajo a la memoria de Ana Colón-Ortiz, mi querida

abuela quién me dio todo su amor durante mi niñez y adolescencia, a Joyce Collins (Titi

Joyce) quien siempre representó alegría y diversión durante mi niñez y adolescencia.

Finalmente, dedico este triunfo a la memoria de mi adorado primo, hermano y amigo

Alexis Valcárcel-Nemerosky a quién nunca olvidaré por todas las amenas experiencias

vividas en mis mejores momentos y por el profundo vacío que su partida ha dejado en mi

vida….descansen todos en paz, NUNCA LOS OLVIDARÉ…

iii

ACKNOWLEDGMENTS

I first thank Dr. Elvia Meléndez-Ackerman for opening the door to science

graduate school in 2004, for what she taught me during my first science graduate school

course, for believing in me and for her continued support from coming back to science all

the way through my PhD defense. Dr. Meléndez-Ackerman got me the first GIS student

copy to keep my work on track, guided me to data and technical sources and is by far the

most influential science graduate professor I have ever had. I also thank Dr. Jon Harbor

for recruiting me back in 2004 to begin my PhD journey, for all his support, trust and

believing in me from the very first day through my defense. Both Dr. Melendez-

Ackerman and Dr. Harbor stood shoulder to shoulder with me through the most difficult

and challenging times during my PhD journey, that I will never forget.

I also thank Dr. Dev Niyogi for supporting me to continue to pursue my PhD

since 2007, Souleymane Fall for his GIS start up help and support, Lei Ming for the

RAMS start up work and Paul Schmid for the final phase of the RAMS. Thanks to Dr.

Gilbert Rochon for staying on my committee despite leaving Purdue and to Dr. Laura

Bowling for giving me a chance to continue my PhD journey by accepting to be on my

committee. My observational analysis was carefully and meticulously done to meet Dr.

Bowling’s high technical expectations.

iv

I also thank Dr. Williams and Dr. Vose from NOAA for providing up to date

weather station data for 1900-2007; Olga Ramos from the Institute of Tropical Forestry

in for providing GIS data and Sigfredo Torres-González from the USGS Caribbean Water

Center office for his support and providing rain gage data. Thanks also to Dr. Chris Daly

for providing PRISM GIS data and professors Larry Theller and Larry Bielh for their GIS

technical support.

Special thanks to my beloved sister Ana L. Torres-Valcárcel and Cesar J.

González-Aviles from COSUAM for believing in this work and for all of the statistical

analysis help and technical support. Having Ana and Cesar on my side was special as I

felt extremely confident and technically supported about data management and statistical

analyses; they truly worked as hard as I did to get all analyses done flawlessly. I thank

Dave and Dianne Williard and all of their family for their help including special moral

and physical support that made my life a bit easier in West Lafayette. Thanks to my

lovely wife Nilda Ortiz-Mercado for her love and close support and for being there

during the most difficult times. Last but not least, I thank all my relatives, former sports

teammates and close friends that helped me with prayers and good wishes encouraging

me to keep going regardless of the challenges and difficult times to finally honor them by

finishing this PhD, the first one in my family, in the name of you all…..I DID IT

v

TABLE OF CONTENTS

Page

CHAPTER 1 INTRODUCTION .........................................................................................1

1.1 Research Questions, Objectives and Hypotheses .......................................................6

1.1.1 Driving Questions .................................................................................................6

1.2 Specific Objectives .....................................................................................................6

1.2.1 Temperature and Precipitation ............................................................................. 6

1.2.2 Storm Event Simulations ..................................................................................... 6

1.2.3 Specific Hypotheses ............................................................................................. 7

1.3 Dissertation Outline ....................................................................................................7

CHAPTER 2 THE IMPACTS OF LAND USE / LAND COVER CHANGES ON

TEMPERATURE PATTERNS IN PUERTO RICO ....................................................... 8

2.1 Abstract .......................................................................................................................8

2.2 Introduction .................................................................................................................9

2.2.1 Global and Regional Synoptic Influences ..........................................................11

2.2.2 Puerto Rico’s Local Climate and Meteorological Conditions ............................12

2.3. Data and Methods ....................................................................................................14

2.4 Temperature Analysis Results and Discussion .........................................................20

2.4.1 Puerto Rico’s Intraregional Climate Variation ...................................................20

2.4.2 HELZ Regional Statistical Analysis ...................................................................22

2.4.5 Land Use / Land Cover (LULC) .........................................................................24

vi

Page

2.4.5.1 ANOVA of Station Temperature Data ............................................................24

2.4.5.2 PCA/EOF Analysis Results .............................................................................26

2.4.5.3 Station Temperature Trends ............................................................................27

2.4.5.4 Temporal and Spatial Frequency Analysis ...................................................29

2.4.5.5 Observation Minus Reanalysis (OMR) ........................................................30

2.4.5.6 Spatial Analysis of Temperatures .................................................................32

2.5 Findings and Conclusions .........................................................................................34

2.5.1 Future Suggestions ..............................................................................................38

2.5.2 Acknowledgements .............................................................................................39

2.6 References .................................................................................................................39


PRECIPITATION PATTERNS IN PUERTO RICO .................................................... 43

3.1 Abstract .....................................................................................................................43

3.2 Introduction ...............................................................................................................44

3.2.1 Study Area ..........................................................................................................46

3.2.2 Previous Precipitation Studies in Puerto Rico ....................................................48

3.2.2.1 Precipitation Studies Related to LULC in Puerto Rico ................................49

3.2.2.2 Rainfall Mapping and Regionalization Studies ............................................50

3.2.2.3 Subregional Precipitation Zones and the Impacts of ENSO and NAO on

Precipitation .......................................................................................................... 51

3.3 Data and Methods .....................................................................................................52

3.3.1 Precipitation and Land Use / Land Cover Data ..................................................52

3.3.2 Puerto Rico Holdridge Ecological Lifezones Data .............................................53

3.3.3 Statistical Methods ..............................................................................................53

3.3.4 GIS Methods ...................................................................................................... 55

vii

Page

3.4 Results and Discussion .............................................................................................57

3.4.1 ANOVA Results .................................................................................................61

3.4.2 Precipitation Trends ............................................................................................62

3.4.3 GIS Interpolated Maps Analysis .........................................................................62

3.5 Conclusions ...............................................................................................................64

3.5.1 Acknowledgments ..............................................................................................66

3.6 References .................................................................................................................66

CHAPTER 4: ASSESSING THE IMPACTS OF LAND USE AND LAND COVER

CHANGES ON PUERTO RICO’S PRECIPITATION USING REGIONAL

ATMOSPHERIC MODELING SYSTEM (RAMS) SIMULATIONS ......................... 70

4.1 Abstract .....................................................................................................................70

4.2 Introduction ...............................................................................................................71

4.2.1 Previous Mesoscale Studies and RAMS Work in Puerto Rico ..........................73

4.3 Methods ....................................................................................................................75

4.3.1 Summary .............................................................................................................75

4.3.2 Numerical Model ................................................................................................76

4.3.2.1 Atmospheric Model: RAMS .........................................................................76

4.3.2.2 Land surface Model: LEAF-3 .......................................................................77

4.3.3 Input Data ...........................................................................................................78

4.3.4 Experimental Design ..........................................................................................79

4.3.4.1 Case Study ....................................................................................................79

4.3.4.2 Land-Surface Scenarios ................................................................................79

4.3.4.3 Control Results and Verification ..................................................................80

4.4 Data ...........................................................................................................................81

4.5 Results and Discussion .............................................................................................82

viii

Page

4.5.1 Precipitation Changes .........................................................................................82

4.5.1.1 Urban Scenarios (UI-A) ................................................................................82

4.5.1.2 Urban Expansion Scenarios (UI-B) ..............................................................83

4.5.1.3 Rain Forest Reserve (RF) .............................................................................84

4.5.1.4 Regenerated Wet Forest (RWF) ...................................................................84

4.6 Conclusions ...............................................................................................................87

4.6.1 Recommendations for Future Work ...................................................................91

4.7 References .................................................................................................................92

CHAPTER 5 CONCLUSIONS .........................................................................................94

5.1 Temperature Findings ...............................................................................................95

5.2 Precipitation Findings ...............................................................................................96

5.3 RAMS Findings ........................................................................................................98

5.4 Synthesis ...................................................................................................................99

5.5 Study Contributions ................................................................................................100

5.6 Study Limitations ....................................................................................................101

5.7 Future Directions ....................................................................................................104

Chapter 2 Temperature Tables .........................................................................................105

Table 2.1. Holdridge Ecological Life Zone (HELZ) relative coverage and number of

temperature stations for Puerto Rico......................................................................... 105

Table 2.2. Characteristics of Major Regions Used in this Study. .................................106

Table 2.3. Seasonal and Annual Temperature statistics for HELZ, Moist Forest Urban

Land Use Areas and Non-Urban, and areas of Regenerated and Unregenerated Forest

1900-2007. ................................................................................................................ 107

Table 2.4. HELZ Temperature Ratios and Differences ................................................108

Table 2.5. Significance of temperature differences between HELZ (ANOVA) ..........109

Table 2.6. Temperatures Variation Explained by HELZ (R2) ....................................110

ix

Page

Table 2.7 Urban versus Non Urban One Way ANOVA ..............................................111

Table 2.8. EOF Modes for all Temperatures ...............................................................112

Table 2.9 Main Locations Top 10% Temperature Stations Summary .........................113

Table 2.10 Main Locations Bottom 10% Temperature Stations Summary ..................114

Table 2.11. Puerto Rico’s Average and Median period trends for all temperatures ....115

Table 2.12. Main Locations Top 10% Temperature Stations Summary .....................116

Table 2.13 Main Locations Bottom 10% Temperature Stations Summary ..................117

Table 2.14. Ranked OMR for Average Temperature Trends of Selected Stations ......118

Table 2.15. Results of the statistical analysis for century average temperature values for

each HELZ from GIS generated maps and each evaluated data base. ......................119

Table 2.16. Difference in Urban versus Non Urban average century or period

temperatures magnitudes from GIS generated maps for each HELZ and data set. .. 120

Table 2.17. Results of the statistical analysis for century average temperature values of

each urban versus non urban evaluated data bases. .................................................. 121

Chapter 3 Precipitation Tables .........................................................................................122

Table 3.1 Summary of previous precipitation research and articles in Puerto Rico .... 122

Table 3.2. Annual effects of ENSO and NAO on Puerto Rico’s Precipitation ........... 124

Table 3.3. Number of stations by Selection Type and Analyzed HELZ and Land Cover

for 1992 Puerto Rico Land Cover Map..................................................................... 125

Table 3.4. Number of stations by Selection Type and Analyzed HELZ and Land Cover

for 2004 Puerto Rico Gap Map . ............................................................................... 126

Table 3.5. Holdridge Ecological Life Zones Distributions and Descriptive Statistics 127

Table 3.6. 1992 LULC Average Monthly Precipitation Ratio 1900-2007 .................. 128

Table 3.7. 2004 LULC Average Monthly Precipitation Ratio 1900-2007 ................. 129

Table 3.8 Yearly Average Total Precipitation for each period and its corresponding

Urban versus Non urban T-test significance values ................................................. 130

Table 3.9 Yearly Average Total Precipitation for each period and its corresponding

Urban versus Non urban T-test significance values. ................................................ 131

x

Page

Table 3.10 270 meter Grid Cell Yearly Average Total Precipitation Trends for each

period and its corresponding Urban versus Non urban T test significance values ... 132

Table 3.11 100 meter Grid Cell Yearly Average Total Precipitation Trends for each

period and its corresponding Urban versus Non urban T test significance values ... 133

Chapter 4 RAMS Tables ..................................................................................................134

Table 4.1 Summary of previous RAMS work about Puerto Rico ............................... 135

Table 4.2. Study objectives, research questions and hypothesis.................................. 137

Table 4.3. Locations of Interest, HELZ and Land Cover ............................................ 138

Table 4.4. Selected Land Cover substitutions for RAMS Simulations........................ 139

Table 4.5. Variables of interest and associated mesoscale rainfall triggering

mechanisms ............................................................................................................... 140

Table 4.6: Table indicating model parameters for each of the three nested grids. ...... 141

Table 4.7. Table of parameters used to define vegetative land-use types in LEAF-3.

( Walko and Tremback 2005) .................................................................................... 142

Table 4.8: Details of land-surface changes for each scenario. ..................................... 143

Chapter 2 Temperature Figures .......................................................................................144

Figure 2.1 Puerto Rico and Global Ocean 1900-2007 Average Temperature Anomalies.

Global data from NOAAA, Puerto Rico data from FILNET 2. ................................ 144

Figure 2.2 Puerto Rico and Global Land 1900-2007 Average Temperature Anomalies.

Global Data from NOAA, Puerto Rico data from FILNENT 2 ................................ 145

Figure 2.3. Puerto Rico 1992 Land Cover Map from Helmer et al, 2002 ................... 146

Figure 2.4. Puerto Rico GAP 2004 Land Cover Map from Gould et al, 2007 ........... 147

Figure 2.5. Puerto Rico Holdridge Ecological Lifezones (HELZ), urban areas and

weather stations. HELZ data from US Forest Service, urban areas data from Puerto

Rico GAP 20014, weather stations data from NOAA Historical Climate Network . 148

Figure 2.6. Distribution of years registering normal (80% frequency), above normal

(>10% frequency) and below normal (<10% frequency) minimum temperature at

each HELZ ................................................................................................................ 149

xi

Page

Figure 2.7 Distribution of years registering normal (80% frequency), above normal

(>10% frequency) and below normal (<10% frequency) average temperatures at each

HELZ ........................................................................................................................ 150


(>10% frequency) and below normal (<10% frequency) maximum temperature at

each HELZ ................................................................................................................ 151

Figure 2.9. Puerto Rico’s SPLINE interpolated Century Maximum Temperature EOF

................................................................................................................................... 152

Figure 2.10. Puerto Rico’s SPLINE interpolated Century Average Temperature EOF

................................................................................................................................... 153

Figure 2.11. Puerto Rico’s SPLINE interpolated Century Minimum Temperature EOF

................................................................................................................................... 154

Figure 2.12. Puerto Rico’s 1900-2007 Maximum Temperature Station Trends ......... 155

Figure 2.13. Puerto Rico’s 1900-2007 Average Temperature Station Trends ............ 156

Figure 2.14. Puerto Rico’s 1900-2007 Minimum Temperature Station Trends .......... 157

Figure 2.15. Puerto Rico’s 1900-2007 Maximum Temperature station trend frequency

distribution ................................................................................................................ 158

Figure 2.16. Puerto Rico’s 1900-2007 Average Temperature station trend frequency

distribution ................................................................................................................ 159

Figure 2.17. Puerto Rico’s 1900-2007 Minimum Temperature station trend frequency

distribution ................................................................................................................ 160

Figure 2.18. Puerto Rico’s Urban 1900-2007 Average Temperature years frequency

distribution ................................................................................................................ 161

Figure 2.19. Puerto Rico’s HELZ 1963-1995 Average Temperature year frequency

distribution ................................................................................................................ 162

Figure 2.20 FILNET GIS interpolated data urban minus non-urban temperature

differences by type of temperature ........................................................................... 163

Figure 2.21 PRISM data urban minus non-urban temperature differences by type of

temperature ............................................................................................................... 164

xii

Page

Figure 2.22. FILNET urban minus non-urban temperatures differences by HELZ .... 165

Figure 2.23. PRISM urban minus non-urban temperatures differences by HELZ ...... 166

Chapter 3 Precipitation Figures .......................................................................................167

Figure 3.1. Puerto Rico’s Holdridge Ecological Lifezones, Areas of Interest & Weather

stations. HELZ data from US Forest Service, urban areas data from Puerto Rico GAP

20014, weather stations data from NOAA Historical Climate Network .................. 167

Figure 3.2 1900-2007 Average and Median Monthly Precipitation for Puerto Rico’s

Holdridge Ecological Lifezones ............................................................................... 168

Figure 3.3. Puerto Rico’s Holdridge Ecological Lifezones Average and Median

Monthly Precipitation through the decades .............................................................. 169

Figure 3.4. Puerto Rico 1992 Land Cover Map from Helmer et al, 2002 ................... 170

Figure 3.5. Puerto Rico GAP 2004 Land Cover Map from Gould et al, 2007 ........... 171

Figure 3.6. Monthly Average and Median Precipitation for Urban stations by HELZ

................................................................................................................................... 172

Figure 3.7. Average & Median Urban versus Non Urban Monthly Precipitation for Wet

Forest selections ........................................................................................................ 173

Figure 3.8. Monthly Average and Median Precipitation for the Moist Forest Urban A

and Non-Urban Selections ........................................................................................ 174

Figure 3.9. Monthly Average and Median Precipitation for the Moist Forest Urban B

and Non-Urban Selections ........................................................................................ 175

Figure 3.10. Average Monthly Precipitation for the Dry Forest Urban 1992 A and Non-

Urban Selections ....................................................................................................... 176

Figure 3.11. Median Monthly Precipitation for the Dry Forest Urban 1992 A and Non-


Figure 3.12. Average Monthly Precipitation for Dry Forest 2004 Urban versus Non-


Figure 3.13. Median Monthly Precipitation for Dry Forest 2004 Urban versus Non-


Figure 3.14. Puerto Rico Annual Cycle Monthly Precipitation by Periods (cm) ...... 180

xiii

Page

Figure 3.15. Wet Forest Annual Cycle Monthly Precipitation by Periods (cm)........ 181

Figure 3.16. Moist Forest Annual Cycle Monthly Precipitation by Periods (cm) ..... 182

Figure 3.17. Dry Forest Annual Cycle Monthly Precipitation by Periods ..................183

Figure 3.18 Seasonal Monthly Total Precipitation by Periods ....................................184

Figure 3.19 Annual Precipitation Quantiles for Wet Forest by Period .........................185

Figure 3.20 Annual Precipitation Quantiles for Moist Forest by Period ......................186

Figure 3.21 Annual Precipitation Quantiles for Dry Forest by Period .........................187

Figure 3.22. 1900-2007 Precipitation Trends by Station ..............................................188

Figure 3.23. 1900-2007 Station Precipitation Trends by period..................................189

Figure 3.24. Number of stations with positive versus negative trends by HELZ and

period ........................................................................................................................ 190

Figure 3.25 Yearly Average Total Precipitation Urban versus Non-Urban Difference

................................................................................................................................... 191

Figure 3.26. Number of study periods receiving higher Yearly Average Urban versus

Non-Urban Total Precipitation ................................................................................. 192

Figure 3.27. Number of study periods recording higher Urban versuss Non-Urban

precipitation trends.................................................................................................... 193

Chapter 4 RAMS Figures ................................................................................................194

Figure 4.1. Map detailing location of each grid for the study. The 50km resolution of

the GFS input data is overlaid on the outermost grid. .............................................. 194

Figure 4.2. Map detailing LEAF-3 land-use types near Puerto Rico. .......................... 195

Figure 4.3. Map of radar derived observed precipitation within the inner grid for

1200UTC 5/23 to 1200UTC 5/24. ............................................................................ 196

Figure 4.1. Map detailing areas of land-use change within the model for each set of

scenarios. Also shown is the region downwind of San Juan analyzed, and the

subdivisions of the island analyzed........................................................................... 197

Figure 4.2. Observed versus simulated temperature during study for a) San Juan

International Airport, b) Arecibo, c) Mayaguez, and d) Yabucoa Harbor. ............... 198

xiv

Page

Figure 4.3. Total simulated precipitation for the inner grid, shown on the same scale as

radar derived precipitation in Figure ........................................................................ 199

Figure 4.4. Changes in sensible and latent heat fluxes at 18UTC 5/23/10 showing an

increase in both gradients.......................................................................................... 200

Figure 4.5. Total accumulated precipitation as a ratio to control for the entire island. 201

Figure 4.6. Total accumulated precipitation as a ratio to control for the western part of

the island. .................................................................................................................. 202

Figure 4.7. Total accumulated precipitation as a ratio to control for the central part of

the island. .................................................................................................................. 203

Figure 4.8: Total accumulated precipitation as a ratio to control for the eastern part of

the island. .................................................................................................................. 204

Figure 4.9. Total accumulated precipitation as a ratio to control for the region

downwind of San Juan. ............................................................................................. 205

Figure 4.10. Total accumulated precipitation as a ratio to control for individual areas of

changed land surface for each scenario. ................................................................... 206

Figure 4.11. Comparison of the change in precipitation between the a) UI5A scenario

and b) UI5B scenario. In UI5A, the surface is changed to forest, reducing the urban

gradient and reducing upwind precipitation. In UI5B, the expanded urban envelope

changes the location of the mesoscale circulation, changing the location of upwind

precipitation. ............................................................................................................. 207

Figure 4.12. Precipitation difference between control and RF1 scenario. Resulting

precipitation represents the combined effects of the changed land surface from forest

to bare soil interacting with the unchanged urban area to the west. ......................... 208

Figure 4.13. Map of precipitation difference between control and a) RWF4 scenario and

b) RWF5 scenario. ................................................................................................... 209

Figure 4.17. Control 6 hour Average Precipitation Time Series ................................. 210

Figure 4.18. Percentage of resulting scenarios with increased versus decreased

precipitation .............................................................................................................. 211

xv

Page

Figure 4.19. Percentage of Increase versus Decrease Precipitation Results by Scenario

................................................................................................................................... 212

Figure 4.20. Precipitation Response ratio for each scenario at each region relative to

control ....................................................................................................................... 213

APPENDICES

Appendix A Figures ......................................................................................................214

Figure A.1. Ecozones Decadal Average Temperature Dry Season Standardized

Anomalies .............................................................................................................. 214

Figure A.2. Ecozones Decadal Average Temperature Wet Season Standardized

Anomalies .............................................................................................................. 215

Figure A.3 Puerto Rico Seasonal Temperature Standardized Anomalies by Decade

................................................................................................................................ 216

Figure A.4 Dry Forest Percentage Decadal Temperature changes ........................... 217

Figure A.5 Moist Forest Percentage Decadal Temperature changes ........................ 218

Figure A.6 Wet Forest Percentage Decadal Temperature changes........................... 219

Figure A.7 1992 A Urban minus Non-Urban Decadal Temperature Difference ...... 220

Figure A.8 1992 B Urban minus Non-Urban Decadal Temperature Difference ...... 221

Figure A.9. 2004 A Urban minus Non-Urban Decadal Temperature Difference ..... 222

Figure A.10. 2004 B Urban minus Non-Urban Decadal Temperature Difference ... 223

Figure A.11 Urban 2004 A versus Urban 2004 B Average Monthly Temperature .. 224

Figure A.12 Urban Stations Minimum Temperature 1900-2007 Trends Distribution

................................................................................................................................ 225

Figure A.13. Urban Stations Average Temperature 1900-2007 Trends Distribution226

Figure A.14 Urban Stations Maximum Temperature 1900-2007 Trends Distribution

................................................................................................................................ 227

Figure A.15 Station Monthly Minimum Temperature by HELZ ............................. 228

Figure A.16 Station Monthly Average Temperature by HELZ ................................ 229

Figure A.17 Station Monthly Average Temperature by HELZ ................................ 230

Figure A.18 Number of Precipitation Stations in Service per year 1900-2007 ........ 231

xvi

Page

Figure A.19 Percentage of Stations Registering Usual versus Extreme Yearly

Average Precipitation for 1900-2007 (Precipitation Station Frequency Distribution)

................................................................................................................................ 232

Figure A.20. Percentage of Decades Registering Usual versus Extreme Yearly

Average Precipitation by HELZ (Decadal Frequency Distribution) ..................... 233


Average Precipitation in the Wet Forest by U/NU Land Cover (Decadal Frequency

Distribution) ........................................................................................................... 234


Average Precipitation in the Moist Forest by U/NU Land Cover (Decadal

Frequency Distribution) ......................................................................................... 235

Figure A.23 Percentage of Decades Registering Usual versus Extreme Yearly

Average Precipitation in the Dry Forest by U/NU Land Cover (Decadal Frequency

Distribution) ........................................................................................................... 236

Figure A.24. 1963-1995 Average Annual Temperature map generated from PRISM

Annual Maximum and Minimum Temperature maps ........................................... 237

Figure A. 25. Holdridge Ecological Lifezones, Temperature Stations and 2004 High

Density and Low Density Urban Areas ................................................................. 238

Figure A.26. 1979-2005 Anomalies Trends from Selected FILNET data stations map

................................................................................................................................ 239

Figure A. 27. 1979-2005 North America Regional Reanalysis Anomalies trends map

................................................................................................................................ 240

Figure A.28. 1979-2005 FILNET selected stations observations anomalies minus

North America Regional Reanalysis trends map ................................................... 241

Figure A. 29. FILNET 1900-2007 Monthly Maximum Temperatures map. ............ 242

Figure A.30. FILNET 1900-2007 Monthly Average Temperatures map ................. 243

Figure A.31. FILNET 1900-2007 Monthly Minimum Temperatures map .............. 244

Figure A.32. 1900-1929 Yearly Average Total Precipitation in centimeters at 100

meter resolution ..................................................................................................... 245

xvii

Page








meter resolution. .................................................................................................... 249



Figure A.38. 1900-1929 Average Total Precipitation Trends at 100 meter resolution

................................................................................................................................ 251

Figure A. 39. 1930-1959 Average Total Precipitation Trends at 100 meter resolution

................................................................................................................................ 252


................................................................................................................................ 253


................................................................................................................................ 254


................................................................................................................................ 255


................................................................................................................................ 256

Appendix B Tables .......................................................................................................257

Table B.1. 1992 LULC Century Average Precipitation Trends (Yearly versus Region)

................................................................................................................................ 257

Table B.2. 1992 LULC PRISM Period Average Precipitation Trends (1963-1995

versus Region) ....................................................................................................... 258

Table B.3. 1992 LULC NARR Period Average Precipitation Trends (1979-2005

versus Region) ....................................................................................................... 259

xviii

Page


................................................................................................................................ 260

Table B.5. 2004 LULC Average Precipitation PRISM Trends (1963-1995 versus

Region) ................................................................................................................... 261

Table B.6. 2004 LULC Average Precipitation OMR Trends (1979-2005 versus

Region) ................................................................................................................... 262

Table B.7. Six Hour Average grid cell precipitation in centimeters for each study

region of the island. The region Downwind of San Juan also includes precipitation

over the ocean. ....................................................................................................... 263

Table B.8. Percentage differences in total precipitation over the modeled period for

each scenario as ratio of the control. Relative changes in precipitation comparing

each scenario to the control by study region. ......................................................... 264

VITA .............................................................................................................................265

xix

ABSTRACT

Torres-Valcárcel, Angel R. Ph.D., Purdue University, August 2013. The Impact of Land

Use / Land Cover Changes on the Tropical Maritime Climate of Puerto Rico. Major

Professors: Jon Harbor and Dev Niyogi.

Previous studies of the influences of Land Use / Land Cover Changes (LULCC) on the

climate of continental areas have provided a basis for our current understanding of

LULCC impacts. However continental climates may not provide complete explanations

or answer specific scientific questions for other regions, such as small tropical-maritime

dominated islands. Here we provide a detailed analysis of century-scale climate change

for Puerto Rico, and assess the degree to which some of this change might be related to

LULCC. We used long-term data, Geographic Information Systems (GIS), statistical

analysis and Regional Atmospheric Modeling Systems (RAMS) to detect and assess the

impact of local urban development on temperature and precipitation. We found strong

evidence of a relationship linking temperature and precipitation magnitudes to local

urban development. Findings for maximum, average and minimum temperature are

robust showing that urbanization has increased local temperatures and levels of impact

found here represent minimum estimates since they were based on data that had some

prior adjustment intended to control for urban signals. Strong evidence of this

relationship was also found in the precipitation data analysis, but no clear correlation was

found in the direction, magnitude, period and location of rain with urban development

implying that other factors dominate or are playing some role in this relationship. RAMS

numerical modeling results were inconclusive suggesting that further tuning of settings

and parameters are needed before model results can be used to guide decision-making.

1

CHAPTER 1 INTRODUCTION

Although weather and climate are complementary concepts, weather refers to the

conditions of the atmosphere at a given time and place, including winds, humidity,

precipitation and temperature, whereas climate refers to the predominant long term

statistics of weather that characterize a given region at a particular time period (Arya,

2001) from decades to geological periods. Understanding the drivers of changes in

climate at a range of scales has emerged as an extremely important scientific goal, given

the impacts of climate changes on human activity and anthropogenic feedbacks on the

atmosphere. The subfield of microclimatology is concerned with atmospheric

phenomena and driving processes within the Planetary Boundary Layer (PBL) and, as

with other scientific disciplines, more processes have been progressively incorporated

into our understanding of the PBL as our research into weather and climate has advanced.

It is well known that surface processes such as radiative energy balance, atmospheric

chemistry and heat fluxes from natural and anthropogenic sources at the surface can alter

local climates (e.g., Chase et al., 2000; Pielke et al., 2002, Kalnay and Cai, 2003; Niyogi

et al., 2004; Vose et al., 2004; Feddema et al., 2005; Christy et al., 2006; Mahmood et al.,

2006; Ezber et al., 2007; Pielke et al., 2007; Hale et al., 2008, Bonan, 2008; Findell et

al.,2009; Pitman et al., 2009; Comarazamy et al., 2010). Microclimatologists are

interested in identifying, understanding and predicting the long term patterns of surface

2

drivers of the atmosphere and the response of local climate to both natural and

anthropogenic drivers, including how changes in surface feedbacks and the intensity of

human activities impact the climate system in the PBL.

Understanding natural surface processes such as evapotranspiration, infiltration,

runoff, erosion, albedo, convection, convergence, advection, heat transfer and energy

fluxes, as well as their magnitudes, is critical for assessing the nature and extent of

anthropogenic impacts on the environment, the atmosphere and climate responses. The

heterogeneous nature of natural surfaces requires that researchers investigate how aquatic

surfaces are different from land surfaces, and how processes vary with land cover and

land use types. The Land Cover concept refers to the physical features that exist on a

surface at a given time, while the Land Use concept adds the anthropogenic realm as it

considers human activity at a given surface covered by a given Land Cover.

Just as natural surface processes are heterogeneous, there are many different

human activities that have particular effects on the climate system by changing the actual

Land Cover and the type and intensity of Land Use. Urbanization and deforestation can

be considered the two most dramatic land cover changes that humans can induce on a

land surface. Deforestation entails the physical removal of the vegetated cover and often

also involves a loss of organic soil, and frequently results in temperature increases,

changes in albedo, reduction of humidity and changing the presence and composition of

organic aerosols. Meanwhile urbanization, often preceded by deforestation, adds

artificial physical features to the land that further alter natural land processes by

decreasing the infiltration, and increasing the heat capacity and thermal conductivity of

natural surfaces.

3

Urbanization is known to increase temperatures but may either increase or

decrease rainfall (Comarazamy et al., 2010; Han et al., 2012). Increased convection,

surface roughness and the presence of some inorganic aerosols in urban areas may

increase rainfall, whereas increased air pollution and inorganic aerosols, as well as

decreased humidity and fragmented latent heat in urban areas may decrease rainfall (Han

et al., 2012). Deforestation and urbanization are dramatic anthropogenic activities that

impact atmospheric conditions.

The modification of surface properties and characteristics through changes in the

Land Use and/or Land Cover are known as Land Use/Land Cover Changes (LULCC).

In some cases, the land use and land cover share the same properties and so one implies

the other; for example, forests, crops and water. However, a region may be defined as

“urban” by its socioeconomic characteristics (Land Use) or by its physical, chemical or

biological properties (Land Cover). Consequently, the type and magnitude of impacts

from these LULCC vary geographically and maybe socially, suggesting that there is

considerable complexity in understanding their impacts on the climate system.

Most studies and models of LULCC impacts on climate have been developed for

continental climates (Comarazamy et al., 2010), such as the continental United States,

where Polar, Continental and Maritime air masses constantly interact. While this may be

representative of conditions in continental climates around the world, and while this work

has advanced basic scientific knowledge about climate and models to predict weather and

long-term patterns, this work has focused on only part of the pool of existing climates.

We cannot assume that it provides adequate explanations or predictions for tropical

climates, maritime climates or combined tropical-maritime climates. Although

4

continental conditions may provide a theoretical basis to understand, project and predict

atmospheric responses and climate conditions at other locations, scientific rigor requires

us to treat these findings as hypotheses for other settings that must be tested in different

environments. Unfortunately, climate studies from diverse regions, including small

tropical islands, are scarce because of the lack of long-term data, high-resolution maps

and the low density of climate stations in many of these regions (Fall et al., 2006).

Puerto Rico provides a valuable opportunity to test if findings, theories, and

methods developed in continental areas hold in a tropical maritime environment. Puerto

Rico is a small tropical island located in the Caribbean basin and has a maritime climate.

The island has considerable macro and micro climatic variability, and has undergone

dramatic LULCC over the past century (Chinea & Helmer, 2003; Grau et al., 2003).

Fortunately for climate studies, Puerto Rico has an extensive, high-density network of

weather stations with long term data for temperature and precipitation, as well as high

resolution Land Use / Land Cover (LULC) digital maps. Climate variability in Puerto

Rico ranges from a Rain Forest with cool temperatures, high precipitation and humidity,

to Dry Forests with warm temperatures and low precipitation and humidity. Land cover

ranges from a large sprawling urban area dominated by one-story buildings with a well-

studied Urban Heat Island effect, to areas of naturally regenerated forest. The growth of

the major urban center accelerated when an agricultural economy was replaced by an

industrial economy in the mid-twentieth century, with abandonment of agricultural lands

and migration of the rural population into the capital city (Grau et al., 2003). These

conditions provide a great opportunity to study local climate extremes as well as land

cover opposites under the same subtropical maritime environment.

5

Global and Synoptic scale events such as Global Warming, sea level rise, El Niño,

The North Atlantic oscillation and tropical storms naturally have local effects in Puerto

Rico. Short-term variations in temperatures in Puerto Rico are dominated by synoptic

scale seasonality, and long-term trends follow broad global patterns of temperature

change (Chapter 2). Puerto Rico’s topography triggers orographic precipitation,

particularly in the eastern region where trade winds collide with two mountain ranges

(Daly et al., 2003). A geographical location in the path of tropical storms, hurricanes and

cold fronts from the continental United States brings additional moisture into the island at

different times during the year. Analyses of long-term rainfall records show a

predominantly decreasing trend for most stations in Puerto Rico (Chapter 3), and this

may present future challenges for water resources and may drive ecological changes on

the island. Despite strong global, regional and synoptic influences in temperature and

precipitation, smaller scale surface processes also have the potential to provide forcing

feedbacks in local climate. The main purpose of the work presented in this dissertation is

to study whether smaller scale surface processes related to land use change are important

variability factors in the dominant Tropical Maritime climate of Puerto Rico, by

identifying, measuring, and understanding land use - climate feedback responses in local

climatological records.

6

1.1 Research Questions, Objectives and Hypotheses

1.1.1 Driving Questions

Does current knowledge of the impacts of Land Use / Land Cover Changes on

climate in a continental climate hold for a tropical maritime environment?

Is it possible to detect and measure the impact of any Land Use / Land Cover

Changes in Puerto Rico?

Could any changes in long-term temperature and precipitation records in Puerto Rico

be associated with local Land Use / Land Cover Changes?

1.2 Specific Objectives

1.2.1 Temperature and Precipitation

Study long term local spatial and temporal patterns in temperature and precipitation;

Use different methods to determine if it is possible to detect and measure local

temperature or precipitation changes; and

Explain any observed links (or lack of a links) between local Land Use / Land

Cover Changes and precipitation and temperature patterns.

1.2.2 Storm Event Simulations

Examine whether simulations of individual storm events with and without land use

changes can provide insight in to the mechanisms behind any long-term patterns in

temperature and/or precipitation linked to Land Use / Land Cover Changes.

7

1.2.3 Specific Hypotheses

Urban temperatures are higher than non-urban temperatures in Puerto Rico.

Urban precipitation is different than non-urban precipitation in Puerto Rico.

Main Forested and Urban areas are altering the dynamics of weather events.

1.3 Dissertation Outline

This work is divided into two major parts. The first part (chapters 2 and 3)

consists of studies of long-term climate data observations from stations across Puerto

Rico that are designed to assess the spatial and temporal variability of temperature and

precipitation in Puerto Rico. This work reveals underlying natural patterns of magnitudes,

variability and controls that provide a benchmark for comparisons that are used to assess

anthropogenic impacts on local climate associated with Land Use / Land Cover Changes.

A particular focus in these chapters is on detecting climate impacts associated with urban

areas and formerly disturbed areas that are undergoing natural regeneration. The second

part (chapter 4) focuses on computational experiments using the Regional Atmospheric

Modeling System (RAMS) to test hypotheses based on the findings in the first part of the

thesis. By simulating different Land Use / Land Cover Change scenarios for a set of

storm events that occurred over Puerto Rico this work provides a basis for theoretical

explanations of local patterns and responses observed in the long-term data sets. In

chapter 5 the main conclusions from the total body of work are discussed, along with

suggestions for future research that would be a logical extension of this work. Chapters 2,

3, and 4 are written in manuscript format for submission to peer-reviewed scientific

journals.

8


TEMPERATURE PATTERNS IN PUERTO RICO

2.1 Abstract

Land Use / Land Cover Changes (LUCC) are land processes that affect local atmospheric

phenomena and have become increasingly important for modern climate studies. Puerto

Rico has experienced major LUCC in recent decades and there is considerable scientific

and practical interest in understanding the effects this might have had on local climate.

This study provides an analysis of observational data designed to examine potential

LUCC impacts on temperature in Puerto Rico. The primary data were FILNET-adjusted

temperatures from the century-long Historical Climate Network (HCN) version 2

database, generated from climate station data across Puerto Rico, and high resolution

digitalized land use/land cover maps. Analysis of variance and trend analysis were used

to examine differences in historical climate data for sites that were grouped by Holdridge

Ecological Life Zone (HELZ) and subdivided by land use type. We also explored the use

of Empirical Orthogonal Function (EOF) analysis to examine trends in the spatial

structure of temperature patterns and the use of Observation Minus Reanalysis (OMR) to

test for local Land Use / Land Cover (LULC)-driven changes versus intraregional

changes in temperature.

9

We found that: (1) in Puerto Rico urban development has impacted maximum,

average and minimum temperatures with a statistically significant difference in all of

them between urban and non-urban areas in FILNET adjusted data that virtually

eliminated urban signals, hence, our findings represent a minimum level of impact; (2)

The highest temperatures on the island are not occurring in Urban Heat Island (UHI)

areas, including the capital city, San Juan; (3) The highest temperature trends were

detected for maximum and minimum temperatures in the locations with most dramatic

LUCC; (4) Methodologically, stratifying data using the HELZs is a useful approach for

geographic climate analysis aimed at comparing urban versus non urban or rural stations;

(5) OMR can be performed at small tropical scales but some unexpected results raised

reliability questions; (6) Statistical analysis maybe more effective in detecting geographic

differences in small scale tropical climates than OMR; (7) EOF yielded results that were

most consistent with conventional expectations about the location, magnitude, direction

and scale of local LUCC impacts; (8) GIS tools are useful and effective to infer

temperature impacts beyond station data observations; (9) The impact of urban

development on temperatures is detectable across the entire island, regardless of HELZ.

2.2 Introduction

Land Use / Land Cover Change (LULCC) reflects socioeconomic patterns of

human activity and is one way in which people can impact ecological systems and

threaten vulnerable human populations and communities. LUCC also plays a role in

climate feedbacks, particularly influencing regional and local temperatures and

precipitation (Jarengui and Ramales 1996; Kalnay and Cai, 2003; Niyogi et al., 2004;

10

Velazquez-Lozada et al., 2006; Ezber et al., 2007, Pielke et al., 2007; Ji-Young and

Jong-Jin 2008, Fall et al., 2009, Imhoff et al., 2010; Murphy et al 2011; Oleson, 2012).

However, the level of understanding of surface and atmosphere exchanges in the Tropics,

where local surface interactions are expected to dominate boundary-layer processes, is

very limited compared to the mid latitudes (Niyogi et al., 2004). Climate feedbacks could

be different at different latitudes and Pielke et al., (2011) reviewed several studies where

similar LUCC in different geographic locations led to different forcing values and in

some cases altered the sign of the forcing. To advance our understanding of LUCC

impacts on climate it is important to study a range of settings and scales, including

smaller-scale areas such as islands, especially if these sites have long term data sets and a

good distribution of observational stations (Fall et al., 2006). Climate change studies on

small islands are also particularly important because of the vulnerability of small islands

to severe natural phenomena and unique sociological challenges (IPCC WGII, 2007). In

addition, small islands have a higher degree of endemism (number of local and unique

species) that could be threatened by synoptic and global changes.

Assessing the magnitude of past impacts of LUCC, as the basis for assessing and

responding to future impacts, is critical for resource management and conservation,

vulnerability assessment and emergency planning. However, few such analyses have

been attempted to date for small tropical islands. In the work reported here we undertake

an assessment of historical temperature changes and LUCC in Puerto Rico, an island that

offers good opportunities for studying climate change and land use because of its size and

land use change history. Moreover, Puerto Rico has a high density of climate stations that

11

have a century-long record of temperature and precipitation, and high spatial resolution

digital land cover maps.

The work presented here was designed to assess changes in temperature patterns

over time in Puerto Rico’s major ecological life zones, and to assess whether temperature

records include variations related to LUCC. This work is structured as follows: in the

first section we discuss how global and regional synoptic phenomena influence Puerto

Rico’s climate. We show that despite Puerto Rico’s small size, tropical location and

maritime influences, where climate might be expected to show very limited spatial

variations, there is enough intraregional variability to require an approach that subdivides

the study area into ecological life zones. Third we analyze a century of data with different

methods to test hypotheses that, after controlling for potential variability related to

ecological life zones, there are significant differences in temperature trends between

urban and rural areas, with higher absolute values and warming trends in urban areas.

2.2.1 Global and Regional Synoptic Influences

Global land and sea-surface average temperature anomalies over the last century

(Figures 2.1 and 2.2) show three distinctive phases: decreasing negative anomalies

(warming) for about thirty years after 1910; then alternating positive and negative

anomalies around a long term average from the 1940s to the 1970s, and; a period of

notably increasing positive anomalies (warming) since the 1970’s. Puerto Rico’s average

temperature anomalies for the century follow similar trends (Figures 2.1 and 2.2),

suggesting that global drivers play a major role in the large-scale trends in Puerto Rico’s

12

temperature record. Yet the variability of the Puerto Rican data record around these

broad trends is quite large, suggesting that local influences might also be important.

Regional synoptic phenomena also influence Puerto Rico’s climate. The island is in

the path of tropical cyclones and the El Niño Southern Oscillation (ENSO) and the North

Atlantic Oscillation (NAO) are major synoptic scale atmospheric phenomena that

potentially influence climate in the Caribbean and Puerto Rico. Jury et al. (2007) found

that ENSO has no observable effects on Puerto Rico’s yearly precipitation. However,

Malmgren et al. (1998) concluded that ENSO has a positive effect on temperatures in the

southeastern Caribbean including the eastern half of Puerto Rico. Malmgren et al. (1998)

also observed that after 1970 there were increasing local average temperatures regardless

of ENSO strength. Malmgren et al. (1998) found no primary impact of the NAO on

Puerto Rico’s temperature. Thus, global climate changes and regional phenomena

influence temperature patterns over Puerto Rico.

2.2.2 Puerto Rico’s Local Climate and Meteorological Conditions

Puerto Rico is about 180 km wide from east to west, and 60 km from north to

south. The center of the island is dominated by the Cordillera Central mountain range

and there are plains to the north and south (Malmgren and Winter, 1999). Puerto Rico has

a maritime subtropical climate typical of Caribbean islands (Daly et al., 2003). The

climate is generally humid with warmer temperatures along the coastline, decreasing

temperatures with increasing elevation, and small seasonal temperature variations (Daly

et al., 2003). Trade winds blowing east—northeast from the Atlantic have a large

influence on the island’s climate but local land surface characteristics and topography

13

drive the climate on synoptically calm days (Velazquez-Lozada et al., 2006). The

mountains at the center of the island generate orographic precipitation (Malmgren et al.,

1998), shielding the southern part of the island from the Atlantic moisture of the trade

winds and causing higher precipitation and lower temperatures in the north and a dryer

and warmer climate in the south. Temperatures are higher at the coastlines and lower in

the central mountains according to topography-corrected PRISM datasets (Daly et al.,

2003). The Parameter-elevation Regressions on Independent Slopes Model (PRISM)

climate mapping system is explained in more detail in the Data and Methods section.

Average temperature anomalies for Puerto Rico computed from FILNET data for all

stations track consistently with global changes in land and ocean temperatures with an

overall increase of ~1.52 o

C in average temperature over the past century (Figs 2.1 and

2.2).

Puerto Rico has two primary seasons, a five month dry season (winter) from

December to April and a seven month wet (summer) season from May to November

(Malmgren and Winter, 1999) Temperatures begin to rise in February, are highest from

June to September, and then decline from October to January. Precipitation trends

generally resemble the temperature trend, with low precipitation in the winter and

increasing precipitation in March through May when the wet season starts. The driest

period of the wet season is during June and July, and then precipitation rises from

September to October to a peak in November before the winter dry season starts in late

December.

During the dry season, cold fronts occasionally produce orographic rain

(Malmgren and Winter, 1999) and the thermal equator is farthest south with the

14

intertropical convergence zone (ITCZ) located south of the Caribbean Sea between 00 and

50

S (Malmgren and Winter, 1999). During the wet season, the ITCZ is located over the

Caribbean Sea between 60 and 10

0 N (Etter et al., 1987). Atlantic trade winds carry

ITCZ moisture inland causing orographic rain (Malmgren and Winter, 1999) and the

island is subject to tropical systems and storms (Lopez-Marrero and Villanueva- Colón,

2006).

2.3. Data and Methods

The goal of this study was to assess changes in temperature patterns over time in

Puerto Rico’s major ecological life zones, and to assess whether temperature records

include variations related to LUCC. Temperature datasets used include: 1) National

Climatic Data Center (NCDC) FILNET adjusted maximum and minimum temperatures

from all 57 Historical Climatology Network (HCN) temperature stations in Puerto Rico;

the FILNET adjusted data are a complete set of records that include estimates for missing

values based on data from highly correlated neighboring stations controlling for

inconsistencies in measurement instruments, station placement and ground sources of

variation (Menne et al., 2009); 2) The Parameter-elevation Regressions on Independent

Slopes Model (PRISM) temperature (Daly et al., 2003); 3) The North American Regional

Reanalysis (NARR) mean temperature at a monthly time scale (Mesinger et al., 2006).

The land use and land cover datasets used in this study originate from the Institute

of Tropical Forestry of the United States Forest Service in Puerto Rico: (1) the Puerto

Rico Forest Type and Land Cover 1992 (30 meter resolution and 33 LULC classes; Fig.

2.3) from Helmer et al. (2002), and (2) the Puerto Rico Gap Analysis Project 2004 15 x

15

15 meter grid and seventy two (72) land use / land cover classes digitized map (Fig. 2.4;

Gould et al., 2007). In addition, we used a Holdridge Ecological Life Zones (HELZ)

dataset created using a vegetation mapping system based on ecological and

ecophysiological tolerance of plant communities to temperature, humidity, precipitation

and elevation (Fig. 2.5; Lugo et al., 1999). Six HELZs are found in Puerto Rico (Table

2.1) however, three HELZs cover less than 1% of the island and are mostly limited to the

Rain Forest reservation, and so were merged with the more similar HELZ in the

geographical analysis. One HELZ’s was further subdivided into Eastern (Unregenerated)

and Western (Regenerated) for comparative analysis. Code names, extension area and #

of stations for each and land cover sub regions annual and seasonal temperature statistics

for the corresponding HELZ in Puerto Rico are provided in Table 2.2 and Table 2.3.

Five geographic areas of interest were identified for particular focus in this study

because of their climatic properties or because of the large scale of historical LUCC:

- San Juan Urban Area: the most dense and extensive urban landscape in the island

and representative of urban conditions and impacts where Urban Heat Island (UHI)

effects have been detected in previous studies (Velazquez-Lozada et al., 2006)

- Rain Forest Reserve: a forest climate that has Puerto Rico’s highest rainfall totals

and coldest temperatures, and which is currently under pressure from urban

development and expansion

- Regenerated Forest: a wet mountainous region with evidence of dramatic LUCC

consisting of a transition from agriculture to forest, which is opposite to the impacts

typical of development related to human activity

16

- Unregenerated Forest: a wet mountainous region without significant LUCC that

serves as a baseline for comparison with the Regenerated Forest.

- Dry Forest: the warmest and driest HELZ or region on the island

The FILNET HCN 2 data used in this study include an adjustment using an

algorithm designed to control for urban signals in the temperature record. However, at

the time of this study this was the first time that such data were available for long-term

analysis, and so we examine here whether it is possible to detect and quantify impacts of

urbanization on local temperatures in the adjusted data set. Average temperatures were

computed directly from the FILNET monthly data by averaging monthly maximum and

minimum temperatures for the 1900 to 2007 period. Monthly temperature averages were

then computed for each region for the period 1900 to 2007. For the decadal temperature

analysis, the years were grouped by chronological decades starting with 1990 to 1909;

however the final decade in the analysis consists of only 8 years (2000 – 2007). Seasonal

temperatures were computed by averaging monthly temperatures corresponding to the

dry season from December to April and the wet season from May to November

(Malmgren and Winter, 1999).

The FILNET data were selected for sites across the different HELZ and LULC

landscapes and analyzed statistically to test for decadal, monthly and seasonal differences.

Temperature stations from HELZ and LULC regions in Puerto Rico were grouped

together in this study to examine temperature changes by region with particular focus on

urban versus non-urban LULC in each HELZ. Regional temperatures were computed by

averaging values of all stations inside each HELZ, the subdivided HELZs and the urban

17

areas from the 1992 and 2004 LULC maps. Temporal variation was analyzed on

monthly, seasonal and decadal scales for each geographical region.

Only stations located inside the main three HELZ (which account for ~99% of the

island) were considered for statistical analysis but all (56) stations were used to generate

interpolated maps. The “urban” areas were selected from the 1992 and 2004 LULC maps

based on their physical characteristics (Land Cover) as defined by each data set. The

1992 map urban area was defined as “urban and barren” while the 2004 map had two

types of urban classes, “High Density Urban” representing the most built and developed

lands and “Low Density Urban” that represented urban population based on its density

(Gould et al., 2007). The “High Density Urban” classification from the 2004 LC map

was therefore selected for study and only the stations located within this area were

selected for analysis (Figs. 2.3 and 2.4). All stations considered “urban” by our GIS

selection method are located within the Moist Forest (MF) HELZ. The two urban area

extents derived from the 1992 and the 2004 maps were coded U1992 and U2004

respectively. The MF was subdivided in to Moist Forest Overall (MFO) consisting of all

stations including those selected as “urban” for analysis purposes between the three

HELZ, and the Moist Forest Non Urban (MFNU) that excluded all urban stations from

the 1992 LULC map.

Two data sets were identified for the analysis of urban regions, reflecting different

selection strategies to control for possible definition and selection method bias of ARC

MAP 10. Method A involved selecting stations contained inside the 1992 and 2004 areas

that were classified as urban land cover. Method B included all of the stations in Method

A plus additional stations known to be in the San Juan urban area and surrounded by built

18

up areas but which were excluded from the urban classification in Method A because

they do not fall inside the urban area as derived using a default automated method based

on a traditional “Urban Land Use” definition. We were cautious about using such a

double definition of the urban landscape, but felt that it was potentially significant and

worth investigating as “urban” is both a “Land Cover” and a “Land Use” and these are

not necessarily identical. The Urban Land Cover refers to the physical environment of a

landscape (such as roads, parking lots, roofs, medians) while the Urban Land Use refers

to a set of activities (uses) that take place in association with an urban area, and may

include features such as parks within a city. Consider, for example, a weather station

located in the middle of Central Park in New York City. In Method A this weather

station would be considered non-urban because it is in a large forested and grassy area.

In Method B it would be considered urban because it is used as an integral part of the city.

Our sub-goal here was to determine if results were independent from the selection

method by testing if there was any significant difference in using these two ways to group

station sites. Urban areas from A selection were labeled 1992 A (U1992A) and 2004 A

(U2004A), urban areas from the B selection were labeled 1992 B (U1992B) and 2004 B

(U2004B). The WF was subdivided into Unregenerated Wet Forest at the east (UnWF)

and Regenerated Wet Forest in the West (RWF), to analyze the temperature patterns

between the two subdivisions separately (Table 2.2 and Table 2.3).

Analysis of Variance (ANOVA) was performed with maximum, average and

minimum temperatures to examine differences between HELZ as well as between urban

versus non urban LULC in addition to monthly, seasonal and decadal variations. One-

way ANOVA analyzes the differences of a dependent quantitative variable against one

19

independent categorical variable such as temperature against HELZs, regions, months,

seasons or decades while Two-Way ANOVA analyzes the differences of a dependent

quantitative variable against two independent categorical variables (Sincich, 1990; Daniel,

1998). When normal distribution requirements were not met but it was nearly Gaussian,

a student’s t-test was used as this test is less sensitive to deviations from normality

(Gosset, 1908, Hogg & Tanis, 1997; Daniel, 1998; Wigley et al., 2006, Montgomery &

Runger, 2010; Laerd Statistics, 2013). The significance level for all statistics was set at

the conventional 95% (α =0.05). Additional variability analysis involved computing a

coefficient of variation, (CV, standard deviation divided by mean) for each station at

monthly, seasonal and decadal time scales, as a measure of variability relative to the

magnitude of the data.

In addition we used Empirical Orthogonal Function (EOF) / Principal

Components Analysis (PCA) to examine spatial structure in the data. EOF/PCA analysis

was performed for the monthly average temperature of each station. The loadings of the

first mode (which expresses most of the variance) were interpolated to display spatial

patterns (Björnsson and Venegas, 1997; Wilks, 2006) . We also used Observation Minus

Reanalysis (OMR) to test for possible local effects in temperature trends. The main

HELZ and LULC areas of interest (urban area, regenerated forest area, unregenerated

forest area, the rain forest reserve and the dry forest HELZ) were the focus of OMR

analysis. OMR analysis uses monthly temperature anomaly trend differences between

surface observations and upper air estimates from North America regional Reanalysis

(NARR) to detect changes in land surface conditions that may affect the near-surface

climate (Kalnay and Cai, 2003; Zhou et al., 2004; Frauenfeld et al., 2005; Lim et al.,

20

2008; Kalnay et al., 2006, Pielke et al., 2007b; Uppala et al., 2007; Nuñez et al., 2008,

Fall et al., 2009).

Geographic Information Systems (GIS) tools were used to select stations from

across the island and the areas of interest for a various analyses. In addition, GIS was

used to generate maps based on interpolated station data and to extract generated map

values for areas of interest for further statistical analysis.

2.4 Temperature Analysis Results And Discussion

2.4.1 Puerto Rico’s Intraregional Climate Variation

Holdridge Ecological Life Zones (HELZ) are based on temperature, humidity,

precipitation and elevation (Lugo et al., 1999) and although there are six HELZ in Puerto

Rico, three HELZ cover 98.5% of the island (Figure 2.5). Figure 2.5 also shows the

locations of the HCN stations used in this study; the HCN is a subset of the Cooperative

Station Network and includes stations that were selected on the basis of having the most

complete, long-term temperature records (Menne et al., 2008). Given that HELZ are

defined in part based on temperature, it is not surprising that temperature ratios and

differences between HELZ are distinct (Table 2.4), but it is interesting to note that the

magnitude of differences between HELZs is on the same order as differences between

urban and rural areas in Urban Heat Island (UHI) studies in the continental United States;

yearly average changes between urban and rural areas are 2.9 oC in U.S. continental UHIs

(Imhoff et al., 2010) while in tropical locations similar to Puerto Rico changes of around

21

2.0 oC are considered sufficient to qualify as an UHI effect (Velazquez-Lozada et al.,

2006, Murphy et al., 2011). Ecological context and seasonality is also important in the

determination of UHI intensity because different biomes respond differently to

impervious surfaces (Imhoff et al., 2010). Considering that differences in magnitudes of

≥ 2.0 oC are important, it becomes evident that any accurate LUCC analysis must

consider ecological context and control for microclimate variability, and in this study we

achieve this by using HELZ as an underlying classification scheme. Table 2.4

summarizes temperature characteristics for HELZ in Puerto Rico, presented as ratios and

differences from Puerto Rico overall averages. Analysis of variance (One Way ANOVA)

for temperatures across HELZ showed statistical differences between HELZs for most

decadal, seasonal and monthly time periods (Table 2.5).

A Geographic Information System (GIS) was used to create subsets of stations

with respect to the HELZs, 1992 and 2004 LULC maps. Table 2.2 details the area and

number of stations, Table 2.3 shows and maximum, average and minimum temperatures

seasonally and annually for all of these stations from all regions under study for the

period of analysis (1900-2007).

Considering 80% of the years as usual temperatures and 20% as extreme (<10%

and >10%) for Puerto Rico as a whole, the frequency of years with usual versus extreme

temperatures during the century for maximum, average, minimum temperatures followed

a consistent 80% usual temperatures to 20% extreme temperature distribution for the

island (Figs. 2.6, 2.7 and 2.8). Breaking up the temperatures by HELZ, we found that

Moist Forest frequencies were similar across maximum, average and minimum

22

temperatures. However the Wet Forest had a higher frequency of years with minimum

extremes while the Dry Forest had a higher frequency of years with maximum extremes.

Puerto Rico has undergone dramatic LUCC, over the past century mainly

characterized by rapid urban growth and development in combination with a large decline

in agricultural activities (Grau et al., 2003; Helmer, 2004). This has resulted in the

regeneration of forest in some areas that were formerly used for agriculture (Grau et al.,

2003) and the intense development of a coastal tropical city (San Juan) (Helmer, 2004).

Given local variability and statistically distinct life zones a main question addressed here

is whether there are differences in temperature changes related to LUCC (in particular

urbanization) that are distinct from differences between HELZ or, in other words, are

impacts of urbanization on temperature observable when evaluated against temperature

changes in non urban areas in the same HELZ.

2.4.2 HELZ Regional Statistical Analysis

We hypothesized that HELZs would have significantly different temperature

statistics, meaning that differences in temperatures across Puerto Rico can be explained

by HELZ as well as LULC. We found significant differences in wet season, dry season

and decadal average temperatures as function of HELZ at a 95% significance level or

above (Table 2.6). Monthly data shows higher variability as compared to the seasonal

and decadal datasets and therefore the data need to be assessed at different temporal

scales to identify regional differences. One-way ANOVA shows that HELZs are

significantly different from each other for most temperature measures and time periods

(Table 2.5) and this validates our use of HELZ as an important organizing structure

23

within which to examine LULC impacts. The Wet Forest was different from the other

HELZs in all temperature parameters and all time periods. However this was not the case

for all HELZ and all time periods. For example, the Moist Forest and Dry Forest showed

no statistically significant variation in monthly temperature by HELZ (α =0.05) meaning

that they are not different from each other, or that all months are the same when

comparing both. All maximum temperatures for the Dry and Moist Forests are

statistically similar. Minimum and average temperatures for the three main HELZs (Wet

Forest, Moist Forest and Dry Forest) are significantly different at the 95% (α = 0.05).

Average and minimum temperatures of the three sampled HELZs are significantly

different; the Dry Forest and Moist Forest maximum temperatures were statistically

similar (not significant differences). This suggests that it is important when comparing

urban with rural stations to determine if they are in the same HELZ. If not, then

temperature differences will reflect a combination of LULC differences and HELZ

differences.

We also found that HELZs have significantly different temperature variability, as

indicated by the Coefficient of Variation, CV, across monthly, seasonal and decadal data.

The three HELZ regions all had highest average monthly temperature variability during

March (dry season) and lowest during September (wet season), and had highest decadal

variability during the 1950s and lowest decadal variability during the 1940s. The Wet

Forest had the highest average monthly temperature and decadal average variability while

the Dry Forest had the lowest monthly and decadal average variability indicating that

temperatures are more consistent in the warmest regions while the colder regions have a

wider range of temperatures.

24

2.4.5 Land Use / Land Cover (LULC)

Once the existence of HELZ variability was established we addressed the

question of whether there were temperature differences between the urban and non-urban

landscapes within each HELZ using Analysis of Variance (ANOVA) and Student’s T-test

where appropriate.

2.4.5.1 ANOVA of Station Temperature Data

We tested for significant differences between urban and non-urban sites in the

context of differences between HELZ regions and different approaches to selecting

stations that were defined as “urban”. We also compared regenerated and unregenerated

Wet Forest areas. The patterns of statistical differences in the results have the following

key features (Table 2.7):

1. Dry Season, Wet Season and Decadal minimum and maximum temperatures are

significantly different between urban and rural areas, and this holds for almost all

alternative ways of selecting urban stations.

2. Dry Season, Wet Season and Decadal average temperatures are not significantly

different between urban and rural areas, except for the case where urban areas are

selected based on the 2004B land cover selection method.

3. Minimum, maximum and average temperatures are not significantly different

between urban and rural areas in almost all cases when monthly data are used.

4. Regenerated and Unregenerated Wet Forest areas are significantly different for Dry

Season minimum, maximum and average temperatures. They are also significantly

25

different for wet season minimum temperature, and for decadal minimum and

average temperature.

Differences in minimum and maximum temperatures between urban and rural

areas are the most consistently statistically significant. This suggests that focusing on

average temperatures may not capture the most important impacts and relationships. For

example, in a case where one area has a higher minimum and a lower maximum than

another area, there are large differences in minimum and maximum, but because the

increase in the minimum offsets the decrease in the maximum, the average may have

very little change. For example urban average annual temperatures (U1992A) have a

minimum value 1.22 oC higher than comparable non-urban sites (MFNU) and a

maximum value that is 1.35 oC lower. These compensating changes help to explain why

the difference in the average temperature for the U1992A to MFNU is only 0.06 oC.

Monthly temperatures showed the least difference between urban and non-urban

sites. This is expected because of the high variability in the monthly dataset. Meanwhile

seasonal differences are significant primarily because each season groups the highest and

lowest temperatures in the yearly cycle.

The large differences between Urban 2004B and the other urban selections shows

that the method used to select which stations are “urban” and the specific stations

included in each selection can have a major influence on the results, especially when the

number of stations is small. Despite the considerable number and density of temperature

stations in Puerto Rico for its size, a maximum of 9 stations (16%) and just 4 (7%) in the

main city were defined as urban, thus the inclusion or exclusion of particular stations

26

with extreme values can have a large impact in averaged temperature computations and

so careful decision making and objective criteria for inclusion or exclusion become

important.

2.4.5.2 PCA/EOF Analysis Results

EOF is a useful tool for characterizing long weather data series by looking for

dominant modes that allows for a classification of climate patterns. EOF was used here

to examine the spatial patterns of trends in the structure of temperature datasets. The aim

was also to examine if the EOF-based spatial structures align with the land use land cover

boundaries and thus assess the control on some aspects of this structure. We assumed that

the spatial pattern of the dominant mode would match the general pattern of types and

scales of LULCC. EOF first modes results explained 60% to 77% of the temperature

variation while the second mode explained 4% to 7% of the variation (Table 2.8).

Average temperature yielded the largest first mode value followed by maximum

temperature while the lowest first modes resulted for minimum temperature first mode

with the lowest value.

The spatial pattern of the first mode (Figs. 2.9, 2.10 and 2.11) reflected results

matching the most notable spatial patterns of LUCC in Puerto Rico, such as urban heat

island effects represented by higher EOF values in the heavily urban San Juan area and

the rural land use change (forest regeneration) of the Regenerated Wet Forest. Minimum

and Average Temperature EOF maps were consistent with our expectations for Heavy

Urban to have greater warming than the Wet Forest. However, the largest warming

detected was not from the heavily urban San Juan area, and surprisingly the maximum

27

temperature EOF indicated less warming in the San Juan area than other areas around the

northwest of Puerto Rico and east of San Juan, perhaps reflecting increased development

(Figs. 2.9, 2.10 and 2.11). The stations with higher century EOF’s were evenly located

between the Regenerated Forest, the Urban Area and the Dry Forest. Tables 2.9 and 2.10

quantify the number of stations that registered the top and bottom 10% temperatures and

EOFs. Stations at the Regenerated Forest dominated the century maximum temperatures

EOF’s (Table 2.9) and lowest temperatures (Table 2.10). Urban stations unexpectedly

dominated the lowest temperatures century EOF (Table 2.10).

2.4.5.3 Station Temperature Trends

Temperature trend analysis for the places considered most important because of

their dramatic LUCC, or that represent climate opposites, provides insight into

temperature patterns related to LUCC over time. The trends analysis was useful to

indicate sites with the most rapid change in temperature, and thus the largest potential

impact of LUCC. Average and median temperatures, based on all stations in Puerto Rico,

have annual trends for minimum, average and maximum temperatures of 0.01 oC / year

between 1900 and 2007, but with the higher rates of increase in more recent time periods

(Table 2.11); The period 1970-2007 has the highest yearly temperature increases among

the selected periods. The 1900-2007 increases in temperatures were largest at Urban and

Regenerated Forest stations (Table 2.12), the locations with the biggest LUCC. The Dry

Forest has the largest number of stations with positive trends in maximum, average, and

minimum temperature values, followed by Urban then Regenerated Forest stations.

Stations from the Dry Forest HELZ had some of the largest temperature increases,

28

however this is not a location of major LUCC. Nonetheless, abandoned, irrigated

cropland dominates this location and has been frequently overlooked in local climate

studies.

Stations with the top temperature trends are found in Urban, Regenerated Forest

and Dry Forest (Table 2.12), and the highest trends were from maximum and minimum

temperatures in the locations with most dramatic LUCC. However, more urban stations

had very low yearly temperature trends than high yearly temperature trends suggesting

increased variability or an increased amplitude in temperature range at urban stations.

Most stations registered positive or increasing trends for minimum, average and

maximum temperatures for the century (Figs. 2.12, 2.13 and2.14). However, minimum

temperatures (Fig 2.14) have the highest number of stations registering negative or

decreasing yearly trends.

For the 1970 to 2007 warming period most stations yielded yearly trends an order

of magnitude higher than those for the entire century. Overall, 18%, 27% and 57% of

stations had minimum, average and maximum temperatures respectively that increased at

rates > 0.01 oC / year (0.1

oC / decade), but for the 1970 to 2007 period these increased to

95%, 100% and 95% of stations respectively. Thus during the 1970 to 2007 period

minimum, maximum and average temperatures all had higher rates of increase, and the

change was most marked for station average and minimum temperatures.

29

2.4.5.4 Temporal and Spatial Frequency Analysis

Frequency Analysis allows the identification of patterns and helps in the detection

of internal variability. Here we focus on patterns associated with 80% usual versus the

10% higher and 10% lower extreme years distribution (Figs. 2.6, 2.7 and 2.8). This

allows a relative comparison of patterns in each HELZ in terms of the frequency of usual

versus extreme temperatures or trends. Temperature magnitudes and trends were

analyzed for stations of each HELZ using number of years for the temporal analysis and

number of stations for the spatial analysis, exceeding the base usual values.

The highest maximum, average, and minimum temperatures occur in the Dry

Forest, and the Urban areas have the next highest minimum and average temperatures.

The lowest maximum, average, and minimum temperatures are found in the Regenerated

Wet Forest and Rain Forest Reserve stations, and all the lowest temperatures on the

island consistently occur in stations from the Wet Forest HELZ. The Dry Forest HELZ

dominated higher extreme temperatures (>90%) followed by the Moist Forest HELZ

while the Wet Forest HELZ dominated lower extreme temperatures (<10%) (Figs. 2.6,

2.7 and 2.8). The Wet Forest stations had the fewest extreme values while the Moist

Forest dominated with the most stations registering the largest percentage of average and

maximum temperature trends (Figs. 2.15, 2.16 and 2.17). Urban stations consistently had

a larger percentage of years registering higher extreme temperatures than the Moist

Forest HELZ where urban stations are located (Fig 2.18). This observation suggests that

Moist Forest warming maybe related to the higher extreme values that urban stations are

contributing. For the OMR period the Dry Forest registered the largest percentage of

years with higher extreme temperatures (Fig. 2.19).

30

2.4.5.5 Observation Minus Reanalysis (OMR)

We used OMR as a method to detect, evaluate and quantify the magnitude of

impact of LUCC in the tropical maritime island environment. Thirteen stations were

selected for OMR analysis representing LULC differences in Puerto Rico from urban

regions to forest; the 2004 San Juan urban area, the 1992 urban area, the Regenerated

Wet Forest and Unregenerated Wet Forest. At the time of this study there were no

reanalysis database grids available for maximum and minimum temperature, so OMR

analysis was limited to average temperatures.

The station (FILNET) average and median yearly trend (0.02 oC / year) indicates

a higher surface temperature warming trend than the average NARR trend (0.005 oC) for

the OMR selected stations. This higher rate of warming for ground observations

(FILNET) than for the high atmosphere reanalysis grids (NARR) suggests that local

influences are driving local temperatures in addition to larger-scale changes. Stations

with the higher OMR trends were from the Regenerated Wet Forest and these values

were higher values than those from stations in the San Juan Urban region where UHI

have been detected (Table 2.14). This suggests that land processes associated with forest

regeneration in the Wet Forest are causing more intense than that occurring in the urban

region.

However, the high OMR trends in the Unregenerated Wet Forest stations are

counterintuitive because no dramatic LULCC has been documented at this location.

Because the OMR method is expected to detect LULCC, we expected to get the highest

trend values at the San Juan urban area stations and the Regenerated Wet Forest where

we know LULCC has occurred. For the OMR period (1979-2005), the Wet Forest shows

31

a different pattern than other locations in Puerto Rico. During this period all evaluated

HELZ and LC, except for the Wet Forest, have a notably higher frequency of years

exceeding the 10th

percentile magnitudes for average temperature compared to the

century frequency (Figs 2.18 and2.19). However, the higher, OMR trends in the Wet

Forest maybe indicative of higher sensitivity to land change processes. Averaging and

computing the median for the selected stations for OMR, the Unregenerated Wet Forest

tops the urban area and the Regenerated Wet Forest. It should be noted that reanalysis

grids have small variability while surface stations have higher variability and trend range

so any station with high trends for the 1979-2005 period will drive OMR trends to higher

values, which seems to have happened here. However, to pinpoint the particular factors

affecting the values a more detailed analysis or additional studies are needed to determine

the source of the higher OMR trends, particularly in the virtually undisturbed

Unregenerated Wet Forest region.

Urban internal stations in San Juan yielded higher OMR trends than urban coastal

stations suggesting higher warming at the center of the city where the urban area is more

dense and a coastal urban cooling effect. This is represented by the urban station

SAN_JUAN_WSFO with the lowest FILNET trends and the only negative OMR in the

sampled stations. Another urban station (CAYEY_1_E), located far away from the San

Juan Urban area, had higher OMR values than all other more heavily urban stations.

Possible explanations for OMR scores of this urban station could be:

a) the ecological context of this particular urban station may be different than the rest

of the San Juan urban stations. Although the station in question falls in the same

Moist Forest HELZ as the other San Juan stations, it is located in a rural setting in

32

the central interior part of the island with higher elevation and more abundant forest

like vegetation than the other coastal urban stations.

b) the location of this particular urban station may have undergone more dramatic

LUCC than the other coastal urban stations which are in areas that underwent

development long ago or development rate has slowed, and thus the land use has

not undergone much change during the 1979-2005 OMR period.

c) a combination of a) and b)

Given the unexpected nature of some OMR results we are cautious about

extending them more broadly because of the small domain size, possible land sea breezes

that can affect reanalysis for the grid, and the possibilities that other factors beyond land

use may be affecting the local temperatures such as aerosols and station changes.

2.4.5.6 Spatial Analysis of Temperatures

Geographic Information System (GIS) tools were used for map generation and to

further assess urban impacts in areas where there were no local station data.

Geoprocessing tools (SPLINE) from ARC MAP 10 and 10.1 were used to interpolate

station temperature data and assess temperature patterns and changes related to urban and

non-urban areas in each HELZ. FILNET adjusted data (1900-2007) and PRISM raw data

(1963-1995) were independently processed for analysis as a baseline from an

independent method in the search for distinctive temperature patterns that may or may

not occur in both data sets.

33

FILNET maximum and minimum temperatures for the century have a wider range

than PRISM maximum and minimum temperatures ranges between 1963 and 1995 by

over 2 oC. However, while the difference between maximum temperatures in both time

periods is less than 1 oC, the difference between both minimum temperatures is over 2

oC.

This suggests that although maximum temperatures are very similar for both periods,

temperatures during 1963-1995 were warmer because of higher minimum temperatures.

PRISM maps were generated for a period of globally and locally increasing positive

temperature anomalies (warming pattern) while FILNET century data includes colder

temperatures. However FILNET and PRISM generated maps show high consistency in

several locations, particularly the warm coastlines and cooler sites at the mountainous

center of the island. Maps generated from both data sets were further processed with

GIS tools to extract temperature values from the areas of interest and to assess Urban

versus Non Urban temperatures at each HELZ. Average century data for all temperatures

from GIS generated maps for the three HELZ under study were all statistically different

(Table 2.15). In other words, the results of interpolated maps kept the expected

corresponding temperature ranges and magnitudes for each particular HELZ.

Despite the very small and scattered pattern of urban development in the Wet

Forest, this HELZ produced the highest urban to non-urban differences in temperatures

independent of the data set (FILNET or PRISM) (Table 2.16). There is a clear pattern in

the magnitude of urban to non-urban temperatures depending on the HELZ in which they

are located; the colder the HELZ the larger the magnitude of temperature differences

between urban versus non urban regions, conversely, the warmer the HELZ the smaller

the magnitude of temperature differences between urban versus non-urban. This is well

34

established for Maximum and Average temperatures, however, minimum temperatures in

the Moist Forest unexpectedly resulted with higher magnitude of temperature differences

than even the Wet Forest, suggesting an increase in temperature ranges from urban areas

in this particular HELZ (Figs. 2.20, 2.21, 2.22 and2.23).

Two tailed student’s t test from FILNET and PRISM GIS generated maps

extracted data confirms that all urban temperatures (maximum, average and minimum) in

all HELZs (Dry Forest, Moist Forest and Wet Forest) across the island are significantly

different from all non- urban temperatures at each corresponding HELZ (Table 2.17).

This indicates that urban development has increased all temperature magnitudes across

the whole island. This finding is remarkably important because it corroborated results

from two separate datasets (FILNET and PRISM) each generated by different methods

and covering differing time periods.

2.5 Findings and Conclusions

The role of land use and land use change related to urban development in

controlling temperatures in Puerto Rico was examined in several ways. Our work

showed that controlling for HELZ was important to avoid drawing erroneous conclusions

regarding temperature differences between urban and non-urban areas from stations that

might be in different regional settings represented by HELZs.

The most important findings were that, even though FILNET HCN 2 data were

adjusted in part to control for urban signals, urban versus non urban temperature

differences were detected with an ANOVA analysis of surface stations data from the

Moist Forest and t- tests of GIS interpolated data. The magnitude of the differences

35

between urban and non-urban areas are from around 0.5 oC to around 2

oC depending on

the HELZ but since they were found in adjusted data that was intended to suppress the

urban signal we expect the differences to be higher in raw data and so these represent

minimum estimates of the magnitude of urban effects. Stations in the main urban area of

San Juan (Moist Forest) had the most significant changes in maximum and minimum

temperatures while average temperature differences were not statistically different,

however, analysis of GIS generated data did yield statistically significant differences in

average temperatures between Urban and Non Urban across the island. Urban land

use/land cover changed maximum and minimum temperatures in Moist Forest stations,

and minimum temperatures were impacted the most. This is illustrated in the urban

versus non-urban maximum temperature values in Table 2.3: for the 2004 Land Cover

Map data, non-urban area values are higher than the urban area, but part of this is because

higher maximum temperature values occur in the southwest of the island where several

non urban stations are located thus increasing the average values of maximum non-urban

temperatures. However, urban average minimum temperatures in the main city of San

Juan were higher than non-urban minimum temperatures. Also, forest regeneration in

the Wet Forest had a larger impact on maximum temperatures than urbanization in the

Moist Forest. Analysis of GIS interpolated data from climate stations showed that all

temperatures (maximum, average and minimum) are impacted by urban development

across the entire island, regardless of HELZ. Also colder-wetter regions such as the Wet

Forest are more impacted by urban development than warmer dryer regions like the Dry

Forest, regardless of the extension of urban development.

36

The EOF method produced results more consistent with expectations that

locations with higher LULCC would have warmer locations and positive trends. The

trends analysis showed the highest trends were detected in maximum and minimum

temperatures from the locations with most dramatic LUCC in Puerto Rico. The OMR

method results showed highest trends for the Regenerated Wet Forest but it was also

surprising to find a very high trend in a location with no documented LULCC. Overall,

the set of techniques and methods used indicate that urban land use change does impact

local temperatures across Puerto Rico.

Over the last century temperatures in Puerto Rico increased, with rates averaged

for all stations in Puerto Rico of 0.67 oC/year in minimum temperature, 1.11

oC/year in

average temperature, and 1.51 oC/year in maximum temperature. Changes over the last

century show distinct patterns related to Puerto Rico’s HELZ, but temperatures produced

warming trends in all sampled HELZs (Dry Forest, Moist Forest and Wet Forest).

Analysis of changes in temperature stations between Urban versus Non Urban areas from

Puerto Rico’s major urban setting in San Juan found significant differences in minimum

and maximum temperatures between Urban and Non Urban, but found no significant

differences in average temperatures. Differentiating between Land Use and Land Cover

was found to be important in this study as we assessed the impact of differentiating

between remote sensing based classification and a classification that also included local

knowledge of land use. Some results were sensitive to the selection method, suggesting

that definitions of urban areas need to be considered carefully to avoid having

conclusions dependent on selection method. This becomes particularly important when

the number of stations is small.

37

Holdridge Ecological Life Zones are related to natural landscapes and can be a

useful tool in studies of LUCC impacts on climate. HELZ’s must be considered when

comparing urban to rural temperature stations because some HELZs have significantly

different temperatures and not accounting for this may lead to result misinterpretation and

spurious conclusions. The highest temperatures in Puerto Rico are in the Dry Forest

HELZ and the lowest temperatures are in the Wet Forest, but surprisingly the warming

trends in these HELZ are comparable to those in urban regions.

Limitations of this work included data accuracy for map and station locations and

climate data. As with any study that relies on a small number of data sites, the statistical

results were sensitive to the inclusion or exclusion of individual stations for the HELZ’s

and urban regions. Although Puerto Rico has a great density of temperature stations,

unfortunately only a handful (a maximum of 17% for the 1992 map) are located in urban

developed areas, making the inclusion or exclusion of stations as “urban” an important

issue.

As no urban stations from the Dry Forest and the Wet Forest were available, the

urban station analysis was limited to the Moist Forest. In addition, it is unfortunate that

no current NARR grids allow for OMR analysis for maximum and minimum

temperatures for urban and non-urban areas. At least at this stage, OMR should be

complemented with other methods such as ANOVA.

Station location and GIS selection method play a big role in determining where

the LULC categories are located and impact the results of the statistical analysis and the

conclusions. Stations that lie close to HELZ borderlines could be considered on one

HELZ or another depending on its coordinates and map accuracy. Therefore, station

38

placement historic documentation could become essential for future systematic data

adjustment.

2.5.1 Future Suggestions

The potential importance of the distinction between Land Use and Land Cover

has been highlighted here and may be important for future studies in Puerto Rico and

elsewhere. To advance the analysis further in Puerto Rico, urban stations representing

the Wet Forest and Dry Forest HELZ are needed. Varying spatial resolution and LULC

classes between maps makes explaining patterns and results problematic, so it would be

helpful to standardize LULC classes for Puerto Rico across maps in future work. More

detailed studies of Non Urban LULC temperatures may further increase understanding of

temperature patterns and impacts that may be important for land managers, and more

detailed subdivisions of urban land use into residential, industrial and commercial may

help in refining our understanding of the details of urban impacts.

Performing OMR on more stations and including maximum and minimum temperatures

is important as urban impacts are expected to be reflected in these temperatures;

unfortunately current reanalysis datasets only provide average temperatures. Future

comparison of raw data versus FILNET data is critical to assess how the use of the

algorithm may change the conclusions of this work. Also, using remote sensing images

to estimate ground temperatures may allow spatial coverage and reduce uncertainty for

data adjustment.

39

2.5.2 Acknowledgements

We thank Dr. Vose and Dr. Williams at NOAA for providing station temperature

adjusted data for 1900-2007 from 57 stations in Puerto Rico. Also thanks to Olga Ramos

from the Institute of Tropical Forestry for providing HELZ and other local GIS data and

Dr. Chris Daly for providing PRISM GIS data.

2.6 References

Arya, P. S. (2001). Introduction to micrometeorology (Vol. 79). Access Online via

Elsevier.

Björnsson, H., and S. A. Venegas, (1997): A manual for EOF and SVD analyses of

climatic data. CCGCR Rep. 97-1, 52 pp.

Chinea, J. D., & Helmer, E. H. (2003). Diversity and composition of tropical secondary

forests recovering from large-scale clearing: results from the 1990 inventory in Puerto

Rico. Forest Ecology and Management, 180(1), 227-240.

Comarazamy, D. E., González, J. E., Luvall, J. C., Rickman, D. L., & Mulero, P. J.

(2010). A land-atmospheric interaction study in the coastal tropical city of San Juan,

Puerto Rico. Earth Interactions, 14(16), 1-24.

Daly, C., E.H. Helmer, and M. Quinones. (2003). Mapping the climate of Puerto Rico,

Vieques, and Culebra. International Journal of Climatology, 23: 1359-1381.

Daniel, Wayne W. (1998). Biostatistcs: A Foundation for Analysis in the Health

Sciences. Seven Edition. Jonh Wiley and Sons, Inc.

Ezber, Yasemin, Omer Lutfi Sen, Tayfun Kindap and Mehmet Karaca. (2007). Climatic

effects of urbanization in Istanbul: a statistical and modeling analysis. Int. J. Climatol.

27: 667–679 (2007)

Fall, S., Niyogi, D., & Semazzi, F. H. (2006). Analysis of mean climate conditions in

Senegal (1971-98). Earth Interactions, 10(5), 1-40.

Fall Souleymane, Niyogi Dev, Gluhovsky Alexander, Pielke Roger A. Sr, Kalnay

Eugenia and Rochon Gilbert. (2009). Impacts of land use land cover on temperature

trends over the continental United States: assessment using the North American

Regional Reanalysis. Int. J. Climatol. (2009). DOI: 10.1002/joc.1996

40

Gosset, William S. “Student”. (1908). “The Probable Error of a Mean.” Biometrika 6

(1908): 1-25. http://www.jstor.org/stable/2331554

Gould, W., Alarcón, C., Fevold, B., Jiménez, M.E., Martinuzzi, S., Potts, G., Solórzano,

M., and Ventosa, E. (2007). Puerto Rico Gap Analysis Project – Final Report. USGS,

Moscow ID and the USDA FS International Institute of Tropical Forestry, Río Piedras,

PR. 157 pp. and 8 appendices

Grau, H. R., Aide, T. M., Zimmerman, J. K., Thomlinson, J. R., Helmer, E., & Zou, X.

(2003). The ecological consequences of socioeconomic and land-use changes in

postagriculture Puerto Rico. BioScience, 53(12), 1159-1168.

Han, J. Y., Baik, J. J., & Khain, A. P. (2012). A numerical study of urban aerosol impacts

on clouds and precipitation. Journal of the Atmospheric Sciences, 69(2), 504-520.

Helmer E. H., Ramos O., Del M. López, T., Quiñones M., And Diaz W. (2002). Mapping

the Forest Type and Land Cover of Puerto Rico, a Component of the Caribbean

Biodiversity Hotspot. Caribbean Journal of Science, Vol. 38, No. 3-4, 165–183, 2002

Helmer, E.H. (2004). Forest conservation and land development in Puerto Rico.

Landscape ecology. Vol. 19 (2004): Pages 29-40

Hogg, R.V., Tanis, E. (1997). Probability and Statistical Inference, Fifth Edition.

Prentice-Hall, Inc.

Imhoff Marc L., Zhang Ping, Wolfe Robert E. and Bounoua Lahouari. (2010). Remote

sensing of the urban heat island effect across biomes in the continental USA. Remote

Sensing of Environment 114 (2010) 504–513.

IPCC Working Group II. (2007). Fourth Assessment Report. Climate Change 2007:

Impacts, Adaptation and Vulnerability.

Jauregui Ernesto and Romales, Ernesto. (1996). Urban Effects on Convective

Precipitation In Mexico City. Atmospheric Environment Vol. 30, No. 20, pp. 3383-3389

Ji-Young Han and Jong-Jin Baik. (2008). A Theoretical and Numerical Study of Urban

Heat Island–Induced Circulation and Convection. American Meteorological Society

Vol.65 June 2008. 1859-1877.

Jury, Mark; Malmgren, Bjorn A. and Winter Amos’. (2007). Subregional precipitation

climate of the Caribbean and relationships with ENSO and NAO. Journal of

Geophysical Research, Vol. 112: 1-10.

Kalnay, E., and M. Cai, (2003): Impact of urbanization and land-use change on climate,

Nature, 423, 528– 531

http://www.jstor.org/stable/2331554

41

Laerd Statistics. (2013). Independent T-Test Using Spss. [Statistics Website]. Retrieved

From Https://Statistics.Laerd.Com/Spss-Tutorials/Independent-T-Test-Using-Spss-

Statistics.Php

Lopez-Marrero and Villanueva- Colón. (2006). Atlas Ambiental de Puerto Rico.

Conservación de Recursos, Capitulo 11. p 128. Editorial de la Universidad de Puerto

Rico

Lugo, Ariel E., Sandra L. Brown, Rusty Dodson, Tom S. Smith, and Hank H. Shugart.

(1999). The Holdridge Life Zones of the conterminous United States in relation to

ecosystem mapping. Journal of Biogeography, 26:1025-1038.

Oleson, Keith. (2012). Contrasts between urban and rural climate in CCSM4 CMIP5

climate change scenarios. Journal of Climate, March 2012, Vol. 25, No. 5 : pp. 1390-

1412

Malmgren B, Winter A, Chen D. (1998). El Niño–Southern Oscillation and North

Atlantic Oscillation control of the Puerto Rico climate. Journal of Climate 11: 2713–

2717.

Malmgren B, Winter A. (1999). Climatic Zonation in Puerto Rico Based on Multivariate

Statistical Analysis and an Artificial Neural Network. Journal of Climate, 12(4): 977–

985.

Menne, M. J., Williams Jr, C. N., & Vose, R. S. (2009). The US Historical Climatology

Network monthly temperature data, version 2. Bulletin of the American Meteorological

Society, 90(7), 993-1007.

Mesinger F, DiMego G, Kalnay E, Mitchell K, Shafran PC, Ebisuzaki W, Jovic D,

Woollen J, Rogers E, Berbery EH, Ek MB, Fan Y, Grumbine R, Higgins W, Li H, Lin

Y, Manikin G, Parrish D, Shi W. 2006. North American regional reanalysis: A long-

term, consistent, high-resolution climate dataset for the North American domain, as a

major improvement upon the earlier global reanalysis datasets in both resolution and

accuracy. Bulletin of American Meteorology Society 87: 343–360.

Montgomery, D. C., & Runger, G. C. (2010). Applied statistics and probability for

engineers. Wiley. com.

Moran, J. M. (2002). Online weather studies. American Meteorological Association.

Murphy, D. J., Hall, M. H., Hall, C. A., Heisler, G. M., Stehman, S. V., &

Anselmi‐Molina, C. (2011). The relationship between land cover and the urban heat

island in northeastern Puerto Rico. International Journal of Climatology, 31(8), 1222-

1239.

https://statistics.laerd.com/spss-tutorials/independent-t-test-using-spss-statistics.php

https://statistics.laerd.com/spss-tutorials/independent-t-test-using-spss-statistics.php

42

Pielke R.A, J . Adegoke, A. Beltran N-Przekurat, C. A. Hiemstra, J . Lin, U. S. Nair, D.

Niyogi and T. E. Nobis. (2007). An overview of regional land-use and land-cover

impacts on rainfall Tellus, 59B, 587–601.

Sincich, Terry. (1990). Statistics by Example. Fourth Edition. Dellen Publishing

Company

Uppala S., A. Simmons, D. Dee, P. Kållberg and J.-N. Thépaut. (2007). Atmospheric

Reanalyses and Climate Variations. Climate Variability and Extremes during the Past

100 Years. Volume 33. 103-117

Velazquez-Lozada A, Gonzalez JE, and Winter A . (2006). Urban Heat Island Effect

Analysis for San Juan, Puerto Rico, Journal of Atmospheric Environment, 40 (9): 1731-

1741.

Wigley, T. M., Santer, B. D., andLanzante, J. R. (2006). Appendix A: Statistical issues

regarding trends. Temperature trends in the lower atmosphere: steps for understanding

and reconciling differences, 129-139.

Wilks, Daniel, S. (2006). Statistical Methods in the Atmospheric Sciences, Volume 91,

Second Edition

http://www.springerlink.com/content/n22227/?p=a2662a5e335846548598b63be62cfd57&pi=0

http://www.springerlink.com/content/n22227/?p=a2662a5e335846548598b63be62cfd57&pi=0

43


PRECIPITATION PATTERNS IN PUERTO RICO

3.1 Abstract

Water is critical for life and the sustaining of natural and managed ecosystems,

and precipitation is a key component in the water cycle. To understand the controls on

long-term changes in precipitation characteristics for scientific and environmental

management applications it is necessary to examine both the impacts of global climate

change on local and regional precipitation, and whether local land use and land cover

change (LUCC) have played a significant role in changing precipitation. For the small

tropical island of Puerto Rico where maritime climate is dominant we used long-term

precipitation and land use and land cover data to assess whether there were any detectable

impacts of LUCC on precipitation over the past century. Particular focus was given to

detecting and quantifying impacts from the urban landscape on mesoscale climates across

Puerto Rico. We found no statistical evidence for differences between average monthly

precipitation from urban and non-urban areas directly from surface stations but GIS

generated maps analyzed data did produced statistical differences (α = 0.05) in yearly

average total precipitation and its corresponding trends across the island. In addition, we

found that generally precipitation in Puerto Rico has been decreasing for the entire

century because of the sharp decrease in periods (months or years) with low rain.

44

However precipitation trends at particular stations contradict synoptic-scale long-

term trends, which suggests that local land use/land cover effects are driving precipitation

at particular locations.

3.2 Introduction

As understanding and awareness of global climate changes have become more

widespread, the local context of climate change has become an increasingly important

issue for local governments, communities and institutions concerned about water supply,

extreme climate events, and the economic and social consequences of changes in

seasonal climate conditions and variability. In addition to large-scale global and regional

drivers, there is an increasing body of literature that demonstrates that land use / cover

changes (LUCC) associated with anthropogenic activities such as urbanization and

deforestation can impact local climates (Chase et al., 2000; Pielke et al., 2002, Kalnay

and Cai, 2003; Niyogi et al., 2004; Vose et al., 2004; Feddema et al., 2005; Christy et al.,

2006; Mahmood et al., 2006; Ezber et al., 2007; Pielke et al., 2007; Hale et al., 2008,

Bonan, 2008; Findell et al.,2009; Pitman et al., 2009).

The effects of urbanization on local climate were first observed in Europe

centuries ago (as cited in Velazquez-Lozada et al., 2006) leading to the recognition of

what is now known as the Urban Heat Island (UHI) effect. Although there has been

considerable interest in evaluating temperature differences between urban areas and their

rural surroundings (Jauregui and Romales 1996, Ezber et al., 2007, Oke, 2009, Imhoff et

al., 2010, Murphy et al., 2011; Chapter 2), studies of variations in precipitation due to

urbanization (Neelin et al., 2006; Niyogi et al., 2011) or other changes in LULC such as

45

deforestation / afforestation are limited (Pielke et al., 2007). Deforestation can result in

albedo increases, reduction of evapotranspiration (which changes sensible and latent heat

partitioning), and rainfall interception (van der Molen, 2002). Such changes resulting

from deforestation have been linked to reductions in cloud cover and cloud formation

height that could reduce precipitation (van der Molen, 2002, Pielke et al., 2007). On the

other hand, afforestation (the creation of forests in places where they did not previously

exist), although considered desirable in some respects, may lead to unintended results

depending on local conditions and processes (Pielke et al., 2007). This reinforces the

importance of studying climate responses to LUCC in a variety of environmental settings.

Much of the existing work examining the impacts of LUCC on local climates has

focused on mid latitude continental sites. However the complex dynamics of LUCC-

climate mechanisms vary from place to place and from land cover to land cover (Pielke et

al., 2007). Further, it is known that similar land cover changes induce different climate

feedbacks at different latitudes (Claussen et al., 2001). There is comparatively little

climate research in tropical settings and on small tropical islands (Van der Molen, 2002,

Fall et al., 2006). Such work is particularly challenging because of a typical scarcity of

long-term data and low data densities (Fall et al., 2006). Further, the traditional coarse

resolution grids designed for climate assessment in continental settings, including

generalized assumptions of similar vegetation and land cover types, are not well suited

for studying the scale and heterogeneity of small regions. Yet LUCC may be an

important component of climate change on small tropical islands, particularly given

large-scale historical vegetation changes associated with agricultural transitions and rapid

urbanization, including coastal development related to tourism. However the weather and

46

climate of tropical islands is also affected by strong maritime influences as well as

synoptic scale phenomena such as El Niño and the North Atlantic Oscillation.

Fortunately, the existence of long-term databases and a relatively high density of

observation stations in Puerto Rico may provide unique opportunities to assess climate

variations on a small tropical island and to detect and isolate any regional drivers of local

climate conditions. A variety of climate studies have been undertaken in Puerto Rico

addressing temperature regionalization (Ewel and Whitmore, 1973; Daly et al., 2003),

precipitation regionalization (Ewel and Whitmore, 1973; Carter and Elsner, 1996; Carter

and Elsner, 1997; Malmgren and Winter 1999; Daly et al., 2003; Jury et al., 2007),

rainfall classification (Pagán-Trinidad, 1984; Carter and Elsner, 1996; Ramírez Beltrán,

2007), regional synoptic influences (Malmgren et al., 1998), tropical storm patterns

(Nyberg et al., 2007), urban heat islands (Velazquez-Lozada, 2006; Murphy et al., 2011)

and using observation and numerical experiments (Velazquez-Lozada, 2006;

Comarazamy and González, 2008; Murphy et al., 2011) (Table 3.1). The work presented

here characterizes precipitation patterns in Puerto Rico and provides a first attempt to

assess in detail whether precipitation changes reflect variations related to local land use

and land cover changes. We discuss how global and regional synoptic phenomena

influence Puerto Rico’s climate and then analyze a century of data with a range of

methods to test hypotheses relating LUCC to precipitation changes and differences.

3.2.1 Study Area

Puerto Rico is the smallest of the Greater Antilles located at 18o N latitude and

66o W longitude in the eastern part of the Caribbean basin, and one of the world’s

47

biodiversity hotspots (Helmer et al., 2002). Puerto Rico is 160 km long by 50 km wide

and includes several smaller islands (Gould et al., 2007). The island has a Central

Mountain Range running east-west, the Luquillo Mountains in the northeast and karst

topography dominates in the northwest. Fifty-three percent of the island’s terrain is

mountainous, 25% are plains and 20% hills (Gould et al., 2007). Wetter regions occur on

the northern side of mountains that shield the southern drier region from Atlantic

moisture. Precipitation in Puerto Rico shows a yearly cycle with a bimodal distribution

(two maxima) peaking first in May as the wet season starts and then a second and biggest

peak in October-November, showing consistency with patterns in the Caribbean Basin

(Jury et al., 2007; Jury, 2009). Mean annual temperatures range from 22 ºC to 25 ºC.

(Gould et al., 2007). Six distinctive Holdridge Ecological Lifezones (HELZ; Holdridge,

1967) are found in Puerto Rico (HELZ are defined by humidity, annual precipitation and

potential evapotranspiration) ranging from Rain Forest (precipitation over 4000 mm/year)

to Dry Forest (precipitation below 900 mm/year) (Gould et al., 2007); however 99% of

the island is covered by Moist Forest, Dry Forest and Wet Forest HELZ (Figure 3.1).

Seasonal temperature trends and long-term trends in mean annual temperature in Puerto

Rico generally track equivalent trends in Global Land and Sea Surface Temperatures

(Chapter 2).

Pagán-Trinidad (1984) identified several major forcings for precipitation in

Puerto Rico that are a function of season and location:

• Orographic – related to mechanical uplift of air caused by mountains. Mostly

associated with persistent easterly Trade Winds in eastern Puerto Rico during

the Dry Season.

48

• Convection – caused by differential land heating, including triggering by urban

landscapes

• Tropical Systems – easterly waves and synoptic scale systems bring

precipitation for all or most of the island, especially during Hurricane season

(July- November)

• Cold fronts – westerly systems from northern latitudes dominate western Puerto

Rico during late Wet Season and Dry Season

This basic classification is useful for understanding the primary synoptic settings

for precipitation episodes that underlie the spatial and temporal variations in precipitation

discussed in this study. In addition, other land biological, chemical and physical features

or processes can affect variables such as temperature, humidity, surface roughness and

aerosols that are related to precipitation. Temperatures affect vertical velocity and

convective potential related to cloud formation and rain intensity. Aerosols affect water

droplet formation and also cloud formation potential. Surface roughness can increase

convergence and cloud formation potential inducing local precipitation. Land processes

such as evapotranspiration, energy fluxes and cloud formation can also drive local

precipitation. Evapotranspiration, water content and humidity affect the availability of

water for cloud formation.

3.2.2 Previous Precipitation Studies in Puerto Rico

Most long-term studies of Puerto Rico’s climate have used a limited number of

stations, because relatively few stations have a long-term record (Malmgren et al., 1998;

Larsen 2000), while shorter-term studies make use of the fact that more stations have

49

data available for specific shorter periods, especially since 1960 (Ray 1933; Pagán-

Trinidad 1984; Carter and Elsner, 1996, 1997; Malmgren and Winter 1999;

Comarazamy, 2001; Van der Molen, 2002; Daly et al., 2003; Harmsen et al., 2009;

Ramírez-Beltrán et al., 2007; Jury et al., 2007, Comarazamy and González, 2008; Jury

and Sanchez, 2009). Long-term studies suggest that precipitation has been decreasing in

the Caribbean since the 1970s and that droughts in Puerto Rico are periodic (Larsen,

2000). Some studies predict that global warming should result in an increase in negative

precipitation anomalies during the summer (June-August), increased dry season duration

and more frequent heavy rain events in the Caribbean (Angeles, et al., 2006; Neelin et al.,

2006; Harmsen et al., 2009). More generally, rainfall in most subtropical areas, including

the Caribbean, is projected to decline by around 20% over the next 100 years (Jury,

2009). Other studies suggest that hurricane frequency in the Caribbean is returning to a

long-term average level instead of increasing due to global warming (Nyberg et al., 2007).

3.2.2.1 Precipitation Studies Related to LULC in Puerto Rico

Observational as well as computer-modeling studies have been used to assess

impacts of LULC on Puerto Rico’s precipitation. Pagán-Trinidad (1984) assessed

precipitation origin and rain intensity variation across different landscapes of the island,

including urban settlements, and attempted to classify different rainfall origins and

associate them with island regions and landscapes. More recently, climate models and

numerical experiments have focused primarily on the impacts of urbanization on

meteorological variables around San Juan (Comarazamy, 2001; Comarazamy and

González, 2008; Comarazamy et al., 2010) and on the impacts of coastal deforestation

50

(Van der Molen, 2002). Comarazamy et al. (2010) identified localized precipitation

increases caused by urban effects from San Juan. However these modeling studies have

been limited primarily to validation efforts and have large errors and low accuracy for

urban areas (Comarazamy, 2001; Comarazamy and González, 2008).

3.2.2.2 Rainfall Mapping and Regionalization Studies

Several studies have attempted to map precipitation around the island or the

Caribbean basin using a variety of methods and techniques (Table 3.1). Ewel and

Whitmore (1973) used long-term station data, vegetation characteristics and forest types

to define climate provinces for Puerto Rico. Carter and Elsner, (1996, 1997), used factor

analysis with Partially Adaptive Classification Trees to regionalize precipitation.

Malmgren and Winter (1999) combined artificial neural networks with Principal

Components Analysis (PCA) to map precipitation regions in Puerto Rico. Unfortunately,

no stations from the western half of Puerto Rico were included in the study, and this is

where both the driest region and one of the wettest regions of the island are located.

Meanwhile, Jury et al. (2007) used the same method to regionalize rainfall for the

entire Caribbean basin. An alternative approach, Parameter-elevation Regressions on

Independent Slopes Model (PRISM) used elevation models, upslope exposure to winds

carrying moisture, distance to the coastline weather station data and physical parameters

for climate mapping simulation (Daly et al., 2003). The work reported in Daly et al.,

2003 was the most recent attempt to map the climate (temperature and precipitation) of

Puerto Rico using modern sophisticated methods based on natural landscape properties,

but not LULC features.

51

3.2.2.3 Subregional Precipitation Zones and the Impacts of ENSO and NAO

Several studies have suggested the existence of sub-regional precipitation zones

or clusters based on rain patterns around the Caribbean. Puerto Rico is consistently

placed in the southeastern cluster, characterized by bimodal seasonal precipitation with

80% of the precipitation falling during summer (May – December) (Jury et al., 2007).

Two major regional atmospheric phenomena are known to have an important influence

on the Caribbean climate and Puerto Rico: El Niño Southern Oscillation (ENSO) and the

North Atlantic Oscillation (NAO) (Malmgren et al., 1998) (summarized in Table 3.2).

According to Malmgren et al. (1998) ENSO has no observable effects on Puerto Rico’s

yearly precipitation. Seasonally, ENSO seems to have a positive effect on the

southeastern region of the Caribbean which includes the eastern half of Puerto Rico (Jury

et al., 2007).

The NAO is more influential than ENSO in the southeastern Caribbean where

Puerto Rico is located (Jury et al., 2007) and the stronger the NAO the lower the

precipitation (Malmgren et al., 1998). Monthly and seasonal precipitation respond

differently to NAO in the southeastern Caribbean, although particular months may show

positive correlation with NAO and receive higher precipitation. In general, there is a

negative correlation between NAO and precipitation in most Caribbean subregions (Jury

et al., 2007). Seasonal influence is critically important in the southeastern Caribbean

because most of the precipitation in the Caribbean falls during the summer (Jury et al.,

2007). Simulations with a mesoscale model using the Parallel Climate Model (PCM) to

project future climate changes in Puerto Rico under the IPCC’s Business as Usual (BAU)

52

Scenario showed that SOI and NAO have important controls on annual Caribbean rainfall

variability (Angeles et al., 2006).

3.3 Data and Methods

The purpose of this study was to assess whether LULC and changes in LULC

have a significant impact on precipitation statistics in Puerto Rico over the past century.

Precipitation data for LULC types of key local interest, such as “urban” and “regenerated

forest” areas, were evaluated against data for nonurban areas within the same Holdridge

Ecological Life Zone (HELZ) (Figs 3.2 and 3.3); comparing data within HELZ was

found to be a useful approach in understanding LULCC impacts on temperature (Chapter

2) and thus it is reasonable to apply a similar approach for the study of precipitation. To

examine how various climate study methods help in understanding the role of LULC on

precipitation we used both simple and sophisticated research methods, including

descriptive and inferential statistics (Analysis of Variance; ANOVA), traditional climate

research methods like trends analysis, and Geographic Information Systems (GIS).

3.3.1 Precipitation and Land Use / Land Cover Data

Monthly raw total precipitation observation data for 1900-2007 from 139 stations

in Puerto Rico were provided by Dr. Williams and Dr. Vose from NOAA and were used

for geographical analysis. The Helmer et al. (2002) Puerto Rico Forest Type and Land

Cover digitized map (c.1992) and the Puerto Rico Gap Analysis Project map from Gould

et al., 2007 (c.2004) were provided by the United States Forest Service’s Institute of

Tropical Forestry in Puerto Rico. The 1992 Map used 30 meter grid spacing and 33 land

53

use /cover classes while the 2004 Map used 15 meter grid spacing and 72 land use / cover

classes.

3.3.2 Puerto Rico Holdridge Ecological Lifezones Data

Puerto Rico HELZ digital maps were provided by the United States Forest

Service’s Institute of Tropical Forestry in Puerto Rico from Puerto Rico Gap Analysis

Project (Gould et al., 2007). There are six HELZs in Puerto Rico; Subtropical Dry Forest

(DF), Subtropical Moist Forest (MF), Subtropical Wet Forest (WF), Subtropical Lower

Montane Wet Forest (LMWF), Subtropical Lower Montane Rain Forest (LMRF), and

Subtropical Rain Forest (RF). The main three HELZ are DF, WF and MF which together

cover 99% of the island. The DF is the smallest of the main HELZs, covering 14% of the

island and has the highest temperatures and lowest precipitation. The MF is the largest

HELZ and covers 62% of the territory and has medium level temperatures and

precipitation. The WF covers 23% of the island and has the lowest temperatures and

highest precipitation. The other remaining three HELZ cover less than 1% of the island

and are mostly limited to the Rain Forest reservation. For simplification and convenience

they were not analyzed as independent regions but considered part of the Rain Forest

Reserve or the Wet Forest HELZ, however, all station data were used for map creation.

3.3.3 Statistical Methods

Since average quantities are heavily influenced by extreme values, average and

median precipitation curves were plotted together to track occurrence of higher

precipitation periods. We expect average and median curves to be close and very similar

54

if data follows a symmetric distribution; however if higher precipitation periods dominate

the frequency of periods then the median would be well above the average and if lower

precipitation periods dominate the median would be well below the average.

Statistical analyses of the observational data focused on testing for precipitation

differences between HELZ and between LULC classes within HELZ. To examine for

possible precipitation differences associated with HELZ or land use, ANOVA was used

to detect monthly, seasonal and decadal differences between regions. Digital maps were

generated from individual station records by interpolating precipitation values and were

used to provide a visual representation of spatial patterns of precipitation.

In addition, the coefficient of variation, CV (standard deviation divided by mean)

for each station was computed for different time periods. The CV estimates the variability

of the data relative to its magnitude and is a useful tool to find spatial patterns of

variability and change. The CV was mapped using Arc Map 10 Spatial Analyst Tool using

the Inverse Distance Weighted (IDW) interpolation method to assess spatial patterns of

change. The IDW method interpolates spatial values as a function of the inverse of the

distance between stations and suitable for climate mapping. Preliminary test maps were

generated using different settings of the IDW tool to assess its reliability to represent the

broad island wide patterns of rainfall already known in Puerto Rico such as the regions

with the highest and lowest precipitation. IDW settings were kept in default but several

tries were made at different grid sizes get the most detail by matching the highest

resolution layer at 15 meters by 15 meter. For convenience and processing power

limitations grid size were set at 270 meters and 100 meters.

55

Simple linear regression was used to analyze precipitation time series linear

trends in different time periods using a linear least squares fit model, given that the data

fits a normal distribution. Data were analyzed in different time periods; the entire record

of over a century of data (1900-2007); 30 year periods (1900-1929, 1930-1959, 1960-

1989 and 1990-2007), PRISM (1963-1995) and Reanalysis (1979-2005). The later

period of 1990-2007 is shorter including only the data available for a full year at the

beginning of this study, the PRISM period is frequently used for studies because of the

high amount of station/year data and the Reanalysis period is when atmospheric grid

became available until the latest year available at the beginning of this study. We

considered PRISM and Reanalysis periods to evaluate how quantitatively distinctive they

are from the other periods and how selecting them could have altered our results.

However, only the 30-year periods were considered for ANOVA although all periods

were analyzed for trends.

3.3.4 GIS Methods

ARC MAP GIS 9.2 was used to select climate stations inside the HELZs and

specific land use classes using the 1992 and 2004 LULC maps. Only stations located

inside the main three HELZ were considered for regional ANOVA analysis but all (139)

stations were used to generate GIS interpolated maps. From the 1992 LULC map, the

“urban and barren” land cover class was considered as “urban” while in the 2004 LULC

map, the “High Density Urban” land cover class was selected as “urban” (Figs 3.4 and

3.5). Different biomes respond differently to impervious surfaces and so ecological

contexts are important (Imhoff et al., 2010). To control for any local ecological variation,

56

“urban” regions were analyzed in their HELZ against their respective “non urban” areas

to avoid any misinterpretation of the results due to stations located in different HELZ.

The two urban land covers from the 1992 and the 2004 maps were coded for the

corresponding HELZ as U[HELZ]92 and U[HELZ]04 respectively.

Several data subsets were used for the analysis of urban regions because of the

use of two different station regional selection methods (Type A and Type B). Each data

subset was statistically analyzed in separate groups to meet statistical independence

assumption criteria. The ARC MAP GIS 9.2 default “intersect” data selection method

considered only stations contained inside the urban LULC and was classified as “Type A

Selection”, the other method (Type B Selection) used 30, 60 and 90 meters radius buffers

around each station. Urban areas from A selection were coded 1992 A (U1992A) and

2004 A (U2004A), urban areas from the B selection were coded 1992 B (U1992B) and

2004 B (U2004B). As the number of urban stations increased by the increased buffer

size, they were subtracted from the Non Urban counterpart and new averages were

computed for both, the new urban region with additional stations and the new non urban

station with subtracted stations (Table 3.3 and Table 3.4).

The DF was subdivided into Urban Dry Forest by LULC map and Selection type

into (UDF92A, UDF92B, UDF04A and UDF04B) which included all urban stations from

the DF and Dry Forest Non Urban (DFNU92A, DFNU92B, DFNU04A and DFNU04B)

which excluded all urban stations from the DF. The MF was subdivided into Moist

Forest (MF) consisting of all stations including those selected as “urban” for analysis

purposes between the three HELZ, the Moist Forest Non Urban Selection A (MFNUA)

and B (MFNUB) excluded urban stations for each and 1992 and 2004 LULC maps coded

57

by selection type A or B (MFNU92A, MFNU92B, MFNU04A and MFNU04B). The

WF was subdivided into Wet Forest Reserve (WFR) located at the base of the Rain

Forest Reserve, the Unregenerated Wet Forest at the east (URWF) and the Regenerated

Wet Forest in the West (RWF), where natural reforestation has occurred.

Monthly total average, median and average total precipitation for each year were

computed directly from the monthly data to statistically evaluate decadal, monthly and

seasonal differences. The summarized data were computed by averaging all stations

inside the HELZs, the subdivided HELZs and the urban areas from the 1992 and 2004

LULC maps. The averages represent total averages for each month, and the medians

represent the median of the median values for total precipitation for each month

respectively throughout the 1900-2007 period. Monthly summary data were computed

by averaging or identifying the median precipitation values for each month for the period

1900 to 2007 for each region. Seasonal precipitation was computed by averaging

monthly totals corresponding to the local Dry Season from December to April and the

Wet Season from May to November (Malmgren and Winter, 1999). Finally, ARC MAP

10.1 IDW interpolation tool was used to create precipitation maps for each period to later

extract and assess urban versus non-urban magnitudes with statistical methods.

3.4 Results and Discussion

In Puerto Rico there are two statistically distinctive (α = 0.05), periods of rainfall:

the Dry Season or Winter (December to April) and the Wet Season or Summer (May to

November). Wet months and wet places show higher rainfall variation while dry months

and drier places show less rainfall variation (Fig 3.6). Median and average annual cycle

58

curves tend to separate late in the wet season as a result of larger rainfall events at the end

of the summer months. During the annual cycle a small variation at urban stations seem

to occur as they receive less rainfall during the Wet Season and in general slightly more

precipitation falls in the non-urban region, in particular this is true for the Wet Forest and

Moist Forest HELZs, however the Dry Forest showed a different pattern in some

selection method samples by registering more precipitation in urban stations than in non-

urban. (Figs. 3.7 to 3.13).

Historical monthly cycle graphs show the highest rainfall variability occurs at the

beginning of the Wet Season or at the driest months (June and July) where most of the

precipitation deficits are evident in all HELZs and across climatological periods (Figs

3.14 to 3.17). The peak months of the Wet Season are more consistent showing the least

variability over time. Wet Season Precipitation has been decreasing in all HELZ through

the century while Dry Season precipitation showed slight increases in the Moist Forest

and Dry Forest, but the Dry Forest have been the HELZ with the highest and most

notable increase (Figure 3.18).

Historical trends for the 90th

, 50th

and 10th

percentiles of monthly precipitation for

the 30-year climatological periods in each HELZ have decreased (Figs. 3.19, 3.20 and

3.21). Most of the trends are decreasing for the entire century however, the 90th

percentile in the Wet Forest and the Dry Forest have an increasing trend, suggesting an

increase in frequency of larger precipitation periods (months or years) in these two

HELZs while smaller periods are decreasing in the entire island. Despite the dominant

decreasing trend in the 50th

percentile, it has been increasing since 1989 in all HELZ

suggesting increase in median precipitation in the last period of 17 years. All 10th

59

percentiles have been decreasing rapidly for the entire century in each HELZ, suggesting

a heavy decrease in small precipitation periods, however only in the Dry Forest the 10th

percentile has been increasing in the last period of 17 years. The decrease in precipitation

combined with an increase of heavy precipitation periods of the most recent 17 years

matches projections for the Caribbean Basin under climate change scenarios (Neelin et al.,

2006). Although from the analyzed data the magnitude and intensity of precipitation

events is not evident, periods of larger precipitation may occur by either low frequency of

large precipitation events or high frequency of small precipitation events both yielding

large amounts of total accumulations for a given period (months or years). Climate

change projections for the Caribbean point at higher frequency of dry periods combined

with a lower frequency of high precipitation periods (Neelin et al., 2006).

Monthly precipitation for the period 1900-2007 in most areas of Puerto Rico

follows a Gaussian (Normal) distribution. However, eastern HELZ Subdivision sites

precipitation record fitted the Johnson transformation distribution (Table 3.5), which may

suggest a different dominant synoptic or sea breeze orographic forcing rainfall origin at

these locations. At all Wet Forest Subdivisions median precipitation is higher than the

average (Table 3.5) as they receive more rain events of higher magnitudes than other

HELZ’s. Of the three HELZ studied, the most precipitation falls in the Wet Forest, as

expected, followed by the Moist Forest while the Dry Forest receives the lowest

precipitation and is statistically different from the Moist Forest and the Wet Forest (α =

0.05). The Moist Forest and the Wet Forest precipitation average are higher than the

average precipitation for the main three HELZs.

60

Average annual precipitation in Puerto Rico has been decreasing for the past

century (1900-1990) in all three HELZs with most stations recording negative trends

(Figure 3.3). However, particular stations, zones and time periods, positive trends have

also occurred (Figs. 3.21 and 3.22). Since the 1970s average annual precipitation level

then increased slightly since the 1990s (Figure 3.3). A notable increase in median

precipitation primarily in the Wet Forest from 1920s to the 1950s was followed by a

dramatic drop in median precipitation in all HELZs after 1970, but all HELZs show

increasing median and average precipitation in the last decade of the record (Figure 3.3).

Dry and wet periods in Puerto Rico seem to follow a cyclic pattern consistent in the

Caribbean basin that has been associated with the North Atlantic Oscillation (Larsen,

2000). Average and median decadal precipitation has greater separation since the 1970s,

particularly in the Dry Forest (Figure 3.3). Average and median rainfall has increased

consistently in the Dry Forest since the 1970s while in the Moist and Wet Forests it began

since the 1990s (Figure 3.3). Greater separation between average and median curves

after 1970 is consistent with global warming projections for the Caribbean of an increase

in drier periods combined with bigger precipitation events (Neelin et al., 2006).

In general, urban areas received slighter less average precipitation than the non-

urban areas at each HELZ, with the exception of the 1992 map B selection 30 meter

buffer. Table 3.6 shows urban versus non-urban average monthly precipitation

differences expressed using a ratio in every HELZ and for every selection method. Ratio

values > 1 indicate average monthly precipitation that is above the average for Puerto

Rico as a whole or the HELZ, values, 1 indicate the contrary. The 2004 map station

61

selections consistently showed more rainfall over urban areas than over non-urban in the

Dry Forest in all selections (Table 3.7).

3.4.1 ANOVA Results

Urban regions could induce precipitation because of increased convection and/or

convergence over the city and the presence of favorable aerosols for cloud formation. On

the other hand precipitation could decrease because of the presence of particular

unfavorable aerosols for cloud formation, air pollution and decrease of mixing ratio or

less available moisture. If urban areas in Puerto Rico are somehow affecting local

precipitation then the differences between urban and non-urban regions should be

reflected in any direction. However, no statistical differences were found in average

monthly precipitation between urban areas and non-urban areas in each HELZ for any

selection type or buffer size based on One Way ANOVA (α = 0.05). Thus we reject the

null hypothesis that there are statistically meaningful differences in average monthly

precipitation between urban and non-urban areas. Also our results are independent of the

selection method used to categorize stations between urban and non-urban. In other

words, urban and non-urban areas statistically receive the same amounts of monthly

average precipitation and, based on the used data and methods, the development of urban

landscapes has not statistically changed the amounts of rainfall compared to the average

and variability of the corresponding HELZ. This result contrasts with effects observed in

larger continental cities, and may reflect the overwhelming dominance of maritime or

synoptic conditions in controlling precipitation across Puerto Rico, little contrast between

urban and non-urban areas or the small size of urban areas in the island.

62

3.4.2 Precipitation Trends

In general, decreasing trends dominate average rainfall over Puerto Rico for the

last century regardless of HELZ (Figure 3.22), which is consistent with reports of

decreasing Caribbean precipitation (Larsen, 2000). However, the most recent 17 years

show a different pattern, with most stations having positive trends (Figs. 3.23 and 3.24).

In the Dry Forest a different pattern than the other HELZ occurred, registering increasing

precipitation trends in the urban area for most time periods that were analyzed (century,

PRISM, NARR), consistently contradicting the long term pattern of decreasing trends

that dominate the island. Thus, even on a small island in the Tropics the scale of Puerto

Rico, there are noticeable intraregional climatic differences. HELZ are important to

account for when comparing urban and rural climate stations because detectable

differences may be because of natural differences in microclimates.

3.4.3 GIS Interpolated Maps Analysis

Yearly average total precipitation and its corresponding trends from surface

stations were interpolated with GIS ARC MAP 10.1 to further investigate precipitation

patterns and changes related to urban and non-urban areas at each HELZ. The data were

divided into the study periods previously mentioned and average values were interpolated

applying the IDW method commonly used for precipitation point data. Statistical

analysis was performed on GIS generated grid cells to evaluate urban versus non-urban

magnitudes.

We found statistical differences between urban and non-urban yearly average total

precipitation in most time periods and HELZs (Tables 3.8 and 3.9). These findings were

63

consistent with GIS analysis of PRISM maps generation by a different method not

considering any land covers, using different spatial resolution and time period. However,

some findings were somehow unexpected or at least counter intuitive as there is no clear

correlation between urban and non-urban precipitation with the direction, time, quantity

and location (Figs. 3.25 and 3.26) We expected that any urban impacts on the

precipitation would be evident or limited to the later periods when urban development

and buildup has been notably intense. Statistical differences between urban and non

urban areas from the beginning of the century would suggest that urban impacts have

always existed locally or that rainfall naturally splits statistically differently at urban and

non urban locations.

We also found statistical differences between urban versus non-urban yearly

average total precipitation trends but this relationship does not remain constant across

periods and occurs in both directions in all HELZs (Tables 3.10 and 3.11). Higher urban

trends prevail in most periods in the Wet Forest and Moist Forest HELZs (Figure 3.27).

Higher urban trends in the Wet Forest had a direct relationship consistent with this HELZ

receiving more average total rain than its non urban counterpart as earlier discussed and

may indicate the higher sensibility of this HELZ to urban development. Meanwhile the

dominance of higher urban trends in the Moist Forest contrasts urban versus non-urban

average totals findings that showed more even split across study periods (Figs. 3.26 and

3.27). The Dry Forest also contrasts average totals findings by showing a more even

urban versus non-urban trends split in the opposite case (Figs. 3.26 and 3.27). The

relationships found in the Moist Forest and Dry Forest urban and non-urban trends by

64

periods, may be indicative of a future shift related to urban development in both regions

as they are the two most urbanized HELZs in the island.

3.5 Conclusions

Precipitation has been decreasing in Puerto Rico for most of the century and in all

HELZs, particularly before 1970, a period in which monthly average and median curves

are relatively consistent (Figure 3.3). Seasonally, Wet Season Precipitation has been

decreasing in all HELZ through the century but the Dry Forest have been the HELZ with

the highest and most notable Dry Season precipitation increase (Figure 3.18). However,

a different pattern emerged after 1970 with average monthly and median precipitation

curves showing more separation, particularly in the Dry Forest HELZ. The Dry Forest is

the only HELZ where urban precipitation has been increasing recently (Figure 3.24).

This new pattern could well be the effects of new climate or just the first half of the 30-

year period receiving higher precipitation that precedes the second half of the period of

decreasing similar to the one that dominates all 30 year periods of the analysis.

We found evidence that urban development has impacted local precipitation based

on an urban effect on local precipitation detected by GIS generated data analysis.

However, this impact was not detected directly from station data analysis. Finding urban

impacts from the beginning of the century was unexpected. The found relationship exists

in both directions as some HELZ receive more precipitation over urban areas than over

non-urban areas while others behave the opposite way. Further, the found relationship is

never constant and is reversed in some periods (Tables 3.8 and 3.9). Precipitation over

urban areas dominates in the Wet Forest while precipitation over non-urban areas

65

dominate in the Dry Forest (Figure 3.24). These findings were also unexpected. The

Wet Forest is mostly forested and urban development is virtually nonexistent, however it

may reveal a higher sensitivity or response to urban impacts than other HELZs. On the

other hand, the Dry Forest is the driest, warmest and most lightly forested region of the

island where urban development is not as intensive and widespread as in the Moist Forest.

This condition could increase urban precipitation because of increased surface roughness

and convection.

In addition, this work has provided an effective new approach that could be used

by small islands to assess LULCC impacts in the local climate. This method could be

applied to any climate variable and any land use or land cover type using station dada,

GIS tools and analysis of variance.

Future research should consider rainfall intensity variation across landscapes as

well as distinction between types of or rainfall source across landscapes as well. Also,

radar and satellite studies should be use to complement knowledge due to the scatter

nature of rainfall and the limitation that represent working with station data that may not

accurately capture the wide spread nature of precipitation. In addition, we recommend

the use of adjusted and filtered data to isolate locally generated events from synoptic

events and standardized land cover vegetation classification for climate and ecological

research. Finally, the use of earlier Land Cover digitized maps would help the analysis

since in its absence this study assumed that the current urban developed area was

unchanged throughout the entire century, resulting in an overestimation of the amount of

urban area or pixels that existed at the beginning of the century that may in turn explain

some unexpected results for the periods early in the century.

66

3.5.1 Acknowledgments

We thank Dr. Vose and Dr. Williams at NOAA for providing precipitation

observation data for 1900-2007 from 139 stations in Puerto Rico. Also, Olga Ramos

from the Institute of Tropical Forestry for providing HELZ and other local GIS data, Dr.

Chris Daly for providing PRISM GIS data and Sigfredo Torres-González from the USGS

Caribbean Water Center office for providing rain gage data.

3.6 References

Angeles, M. E., Gonzalez, J. E., Erickson III, D. J., & Hernández, J. L. (2006). An

assessment of future Caribbean climate changes using the BAU scenario by coupling a

global circulation model with a regional model. Proc. 86th Am. Met. Soc. Meet.,

Atlanta, Georgia, USA.

Carter, M. M. and Elsner, J. B. (1996). Convective Rainfall Regions of Puerto Rico.

International Journal of Climatology, Vol. 16, 1033-1043

Carter, M. M., & Elsner, J. B. (1997). A statistical method for forecasting rainfall over

Puerto Rico. Weather and Forecasting, 12(3), 515-525.

Claussen, Martin; Brovkin, Victor; and Ganopolski, Andrey. (2001). Biogeophysical

versus biogeochemical feedbacks of large-scale land cover change. Geophysical

Research Letters, March 15, 2001. Vol. 28, No. 6, 1011-1014.

Comarazamy, D.E., (2001). Atmospheric modeling of Caribbean Region: precipitation

and wind analysis in Puerto Rico for April 1998. Thesis, Master of Science in

Mechanical Engineering, University of Puerto Rico, Mayaguez Campus.

Comarazamy Daniel E. and González, Jorge E. (2008). On the Validation of the

Simulation of Early Season Precipitation on the Island of Puerto Rico Using a

Mesoscale Atmospheric Model. American Meteorological Society. Journal of

Hydrometeorology. Vol. 9. Issue 3. June 2008. 507-520.

67






Ewel J and Whitmore J. (1973). The Ecological Life Zones of Puerto Rico and the US

Virgin Islands. Forest Service Research Paper ITF-18. Institute of Tropical Forestry Rio

Piedras Puerto Rico. Forest Service US Department of Agriculture.

Ezber, Yasemin, Omer Lutfi Sen, Tayfun Kindap and Mehmet Karaca. (2007). Climatic

effects of urbanization in Istanbul: a statistical and modeling analysis. Int. J. Climatol.

27: 667–679 (2007)

Fall S., Niyogi D. , Semazzi F., (2006), Analysis of Mean Climate Conditions in Senegal

(1971–98), Earth Interactions, 10, Paper 5, 1-40.

Gould, W., C. Alarcón, B. Fevold, M.E. Jiménez, S. Martinuzzi, G. Potts, M. Solórzano,

and E. Ventosa. (2007). Puerto Rico Gap Analysis Project – Final Report. USGS,

Moscow, ID and the USDA Forest Service International Institute of Tropical Forestry,

Río Piedras, PR.

Grau, H. Ricardo, T. Mitchell Aide, Jess K. Zimmerman, John R. Thomlinson, Eileen

Helmer, and Xioming Zou. (2003). The Ecological Consequences of Socioeconomic

and Land-Use Changes in Postagriculture Puerto Rico. BioScience. Vol. 53 No. 12

1159-1168

Harmsen, Eric W; Norman L. Miller, Nicole J. Schlegel, J.E. Gonzalez. (2009). Seasonal

climate change impacts on evapotranspiration, precipitation deficit and crop yield in

Puerto Rico. Agricultural Water Management 96 (2009) 1085–1095

Helmer, E. H., O.Ramos, T. Del M. López, M.Quiñones, and W. Diaz. (2002). Mapping

the Forest Type and Land Cover of Puerto Rico, a Component of the Caribbean

Biodiversity Hotspot. Caribbean Journal of Science, Vol. 38, No. 3-4, 165–183,

Jauregui Ernesto and Romales, Ernesto. (1996). Urban Effects on Convective

Precipitation In Mexico City. Atmospheric Environment Vol. 30, No. 20, pp. 3383-3389

Jury, Mark; Malmgren, Bjorn A. and Winter Amos. (2007). Subregional precipitation

climate of the Caribbean and relationships with ENSO and NAO. Journal of

Geophysical Research, Vol. 112: 1-10.

68

Jury, M. R. (2009). An intercomparison of observational, reanalysis, satellite, and

coupled model data on mean rainfall in the Caribbean. Journal of Hydrometeorology,

10(2), 413-430

Jury MR and Sanchez DM. (2009). Composite Meteorological Forcing of Puerto Rican

Springtime Flood Events. American Meteorological Society. Weather and

Forecasting. Volume 24. February. 262-271

Larsen, M.C., (2000), Analysis of 20th century rainfall and streamflow to characterize

drought and water resources in Puerto Rico, Physical Geography, 21, 494–521.



2717.

Malmgren B, Winter A. (1999). Climate Zonation in Puerto Rico Based on Multivariate


985.

Ming Lei, D. Niyogi, C. Kishtawal, R. Pielke Sr., A. Beltrán-Przekurat, T. Nobis, and S.

Vaidya, (2008): Effect of explicit urban land surface representation on the simulation of

the 26 July 2005 heavy rain event over Mumbai, India. Atmos. Chem. Phys., 8: 8773-

8816

Ming Lei and Niyogi Dev. (2010). Simulation of Indianapolis Thunderstorms using

Explicit Urban Parameterization within a Mesoscale Model, J. Appl. Meteor. Climatol.,

in revision.





1239.

Neelin, J. D. , M.Munnich, H. Su, J. E. Meyerson, and C. E. Holloway. 2006. Tropical

drying trends in global warming models and observations. PNAS. April. vol. 103. no.

16. 6110- 6115

Pagan-Trinidad, (1984). Statistical Analyses of Spatial and Temporal Storm Rainfall

Characteristics in Puerto Rico. Project No. 374103 Final Technical Report. United

States Department of Interior

69

Pielke R.A, J . Adegoke, A. Beltran N-Przekurat, C. A. Hiemstra, J . Lin, U. S. Nair, D.

Niyogi and T. E. Nobis. (2007). An overview of regional land-use and land-cover

impacts on rainfall Tellus, 59B, 587–601

Ramirez-Beltran, Nazario D., William k. M. Lau, Amos Winter, Joan M. Castro, Nazario

Ramirez Escalante. (2007). Empirical probability models to predict precipitation levels

over Puerto Rico Stations. American Meteorological Society. Monthly Weather

Review. Volume 135: 877-890.

Ray, C. L. (1933). Wet and Dry Periods in Puerto Rico, 1899 1932. Monthly Weather

Review, 61, 222.

Stewart I, Oke T. (2009). Newly Developed “Thermal Climate Zones” for Defining and

Measuring Urban Heat Island Magnitude in the Canopy Layer. Proceedings of the

American Meteorological Annual Meeting, Phoenix, Arizona. 11–15 January. Available

online at http://ams.confex.com/ams/pdfpapers/150476.pdf. [Access year 2009].

Uppala, S., Simmons, A., Dee, D., Kållberg, P., & Thépaut, J. N. (2007). Atmospheric

reanalyses and climate variations. In Climate Variability and Extremes during the Past

100 Years (pp. 103-117). Springer Netherlands.

Van der Molen, M. K. (2002). Meteorological impacts of land use change in the maritime

tropics. PhD, Vrije Universiteit, Amsterdam, The Netherlands.

Velazquez-Lozada A, Gonzalez JE, and Winter A . (2006). Urban Heat Island Effect


1741.

70

CHAPTER 4: ASSESSING THE IMPACTS OF LAND USE AND LAND COVER

CHANGES ON PUERTO RICO’S PRECIPITATION USING REGIONAL

ATMOSPHERIC MODELING SYSTEM (RAMS) SIMULATIONS

4.1 Abstract

Climate change has global to local consequences; however, local climates are also

affected by boundary layer feedbacks from Land Use and Land Cover Changes (LULCC).

Understanding the role that LULCC play in modifying local climate and weather events

is critical for land use planning and impact mitigation initiatives. Most climate studies

have focused on assessing impacts of global warming and climate change on global and

local temperatures in continental settings, while fewer studies have focused on

precipitation feedbacks from the drivers and internal forcing particularities of tropical

maritime climates feedbacks from LULCC. However, several studies have been

conducted in the tropical island of Puerto Rico regarding LULCC providing good

opportunities for follow up studies. Here we used Regional Atmospheric Modeling

System (RAMS) to assess the feedbacks of the major LULCC in Puerto Rico in a

selected event to observe how land features and processes may be driving, impacting or

playing a role in precipitation events. We also explored possible explanations for the

longer term observed patterns of previous precipitation studies over Puerto Rico during

the past century. We found that (1) the Central and Western parts of the island respond

more to LULCC while the Eastern part is less sensitive and appears to be controlled by

71

other factors (2) all substitutions in urban areas caused decreased precipitation

island wide (3) the frequency of unexpected counter intuitive results stresses the need for

further research to reliably run RAMS in environments and conditions similar to those of

Puerto Rico.

4.2 Introduction

Climate Change has global to local drivers, feedbacks and impacts. However,

local impacts are known to be affected by natural phenomena and anthropogenic

activities that occur within the Planetary Boundary Layer (PBL). The field of

microclimatology is concerned with the study and understanding of atmospheric

phenomena and variables that play a role in weather and climate within the PBL. This

includes land processes such as urbanization, and deforestation. Many existing climate

models address global climate dynamics and variables that operate at very large scales,

however PBL land features and processes interact with local atmospheric variables and

can alter the local climate.

Local natural and anthropogenic land features and processes can alter

precipitation events by changing related variables such as humidity, surface roughness,

temperature, vertical velocity, aerosols, and processes such as evapotranspiration rates

and cloud formation (Chapter 3). Surface temperatures in turn have an impact on vertical

velocity and convective potential related to cloud formation and rain intensity.

Evapotranspiration, water content and humidity control the availability of water for cloud

formation. Aerosols effect water droplet formation and also cloud formation potential.

Surface roughness increases convergence and cloud formation potential. With such a

72

wide range of interconnected processes, a particularly effective way to examine the

potential impacts of LULCC on the climate system at a local level is through numerical

modeling.

The Regional Atmospheric Modeling System (RAMS) is a meteorological

computer simulation application developed by the Atmospheric Science Department from

Colorado State University to conduct computational experiments of regional and local

atmospheric circulation at high resolutions. This model is of particular importance for

regional and local meteorological studies because it considers terrain properties and

landscape energy fluxes and dynamics often unaccounted for in global scale circulation

models. RAMS uses observed meteorological data to perform simulations that allow us

to study potential land features and land process change scenarios that would otherwise

be impossible, or at least extremely expensive and time consuming, real world

experiments.

In order to assess impacts of global-scale changes on local weather and climate,

the internal variability of local climate and weather must be first assessed (IPCC, 2007).

Observational climate and weather studies are useful to assess internal variability and

long- term changes (e.g., chapters 2 and 3) but allow little control over forcing variables.

In contrast, numerical computational experiments rely on many assumptions but can

complement observational studies by simulating possible scenarios by controlling values

of key variables and parameters . Modeling systems, such as RAMS have been used to

perform numerical computational experiments that help us understand and predict the

possible feedbacks and impacts of natural changes and anthropogenic activities. Thus

this type of model has great potential to increase understanding of the role that LULCC

73

has played in the changes revealed by the studies of long-term climate data presented in

chapters 2 and 3.

RAMS has previously been used to study the weather of Puerto Rico with an

emphasis on particular places and specific phenomena, such as the impacts of the Urban

Heat Island (UHI) over San Juan, the temperature impacts of different natural forests, and

CO2-forced climate change scenarios (Table 4.1). Although these studies provided an

important base for further research, and some of them addressed rainfall questions, most

of them were focused on understanding or predicting specific phenomena over the San

Juan region or the impacts of the urban area, or addressed synoptic influences or forcings.

Island-wide spatial variations in precipitation induced by mesoscale natural and

anthropogenic land process feedbacks are relatively understudied, and there are

unanswered questions about the local drivers, feedbacks, impacts and internal variability

that need to be addressed to effectively assess the impacts of global- to local-scale

changes on Puerto Rico’s precipitation.

4.2.1 Previous Mesoscale Studies and RAMS Work in Puerto Rico

Previous studies using RAMS in Puerto Rico have focused primarily on the Urban

Heat Island (UHI) effect in the San Juan area and the future impacts of expanding the city

(Table 4.1; Comarazamy, 2001; Velazquez-Lozada A et al., 2006; Comarazamy and

González, 2008; Comarazamy et al., 2010). Van der Molen (2002) investigated the

meteorological effects of deforestation of the coastlines with an emphasis on Rain Forest

dominant vegetation types; no urban conditions were considered and the precipitation

analysis was short and inconclusive. In contrast, Angeles et al. (2006) focused on

74

synoptic impacts of larger-scale driving forces by simulating an IPCC Business as Usual

CO2 increase scenario as well as El Niño and the North Atlantic Oscillation impacts on

Puerto Rico’s temperature and precipitation. They found that surface air temperatures

will be 2.5°C above monthly average for 2048 and that mesoscale rainfall is strongly

influenced by land dry areas .

In this work we used RAMS to assess particular LULCC feedbacks in Puerto

Rico and how they drive precipitation changes in different locations. We focus on

particular areas that represent Holdridge Ecological Life Zones (HELZ) and regions with

most dramatic LULCC impacts over the past century, as indicated by past work. This

includes the San Juan urban area and a regenerated forest region in western Puerto Rico.

The main objectives of this study are (see also Table 4.2) to:

1. Assess the impacts of LULCC feedbacks on local precipitation events

2. Examine how major LULCC and dominant HELZ play a role in driving or

modifying local precipitation events

3. Provide possible explanations for observed long term precipitation patterns between

urban and non urban areas in different HELZs

4. Assess possible long-term precipitation response to changes in different land uses

and land covers in Puerto Rico.

75

4.3 Methods

4.3.1 Summary

In this research we examined the potential use of the RAMS model to examine

how individual storm events are affected by changes in land use in selected areas of

Puerto Rico. A map of the island using 2001 satellite-derived land classes (Friedl et al.,

2002) was modified to conform with the latest land cover map from the Puerto Rico 2004

Gap Analysis Project by substituting RAMS land cover classes where needed. The

model was configured to replicate control runs and then selected land cover types were

substituted and the model run was repeated to study the response of the model to selected

land cover changes (Table 4.4). Such an approach is a standard experimental process for

LULCC modeling studies (Pielke et al. 2011). Model simulations involved

systematically substituting the actual land covers at locations of interest with alternative

RAMS land cover classes, while keeping remaining areas the same so that we could

isolate the responses for specific land use changes. Additional parameter adjustments

and modifications of the initial conditions were used to study model output responses to

different environmental conditions.

To select candidate events for RAMS modeling, radar reflectivity data

corresponding to different precipitation events were evaluated. The intent was to identify

relatively locally triggered rain event around areas representing the major natural

variability microclimates and LUCC patterns of the island (Table 4.3). NOAA daily

storm reports were examined to verify and understand the non-synoptic nature of events.

Station hourly data were compiled wherever available for verification. Model

76

verification was performed by qualitative evaluation comparing radar loops with

equivalent output from RAMS for control runs simulating May 23, 2010,. Simulated

radar loops for the control runs were evaluated for (1) precipitation totals, (2) rain

location, (3) time of occurrence, (4) rain duration and (5) rain intensity against observed

radar loops from the mesoscale triggered rain events from around the island.

Simulated weather data and variables of interest associated with mesoscale

precipitation triggering mechanisms were evaluated around areas of interest representing

major LUCC locations and the HELZs in the island (Table 4.5). Rainfall quantities and

related variables were observed, mapped, tabulated and plotted across the locations of

interest so that differences could be evaluated.

4.3.2 Numerical Model

This numerical experiment uses the RAMS model to evaluate the model

sensitivity to the surface land type. In order to achieve this, a series of scenarios are run

within the mode for a thunderstorm on 5/23/10 over Puerto Rico.

4.3.2.1 Atmospheric Model: RAMS

The Regional Atmospheric Modeling System (RAMS) is a sophisticated, cloud-

resolving mesoscale model capable of simulating thunderstorms at city-scale (Cotton et

al., 2003; Pielke et al., 1992). It solves the non-hydrostatic atmospheric equations of

motion on a polar-stereographic C-grid (Arakawa and Lamb, 1981). The version run for

this study utilizes the Harringon radiative parameterization (1997), Klemp-Wilhelmson

boundary conditions (1978), and the Mellor-Yamada turbulence closure scheme (1982).

77

The cloud microphysics parameterization is a two-moment bin-emulating scheme

(Meyers, Walko, Harrington, and Cotton, 1997; Saleeby and Cotton, 2004). It utilizes

two cloud-condensation nuclei sizes based on observations by Hobbs et al. (1980) to

represent two modes of cloud nucleation. This is relevant as within urban areas, such as

San Juan, urban aerosols fit such observations.

For this study, RAMS was run in a three-way interactive nested grid configuration

(Walko, Tremback, Pielke, andCotton, 1995). From coarsest to finest, the grid spacing

was 64km – 16km – 4km with a timestep of 90s – 15s – 2.5s respectively. Figure 4.1

shows the relative locations of each grid within the study domain. The vertical spacing

was 40m with a 1.1 stretching ratio at each vertical grid level (i.e. Δz1=40m, Δz2=44m,

Δz3=48.4m etc.) The Kain-Fritch convective parameterization (1993) was used in the

two outmost grids. The inner grid has sufficient resolution such that the cloud

microphysics alone is capable of resolving convective (Kain, 2004). Table 4.6 details the

size of the grid domain and initialization parameters.

4.3.2.2 Land surface Model: LEAF-3

The land-surface in RAMS is parameterized by the fully coupled Land

Ecosystem-Atmosphere Feedback model (LEAF-3, Walko et al., 2000; Walko and

Tremback 2005). It uses a finite number of vegetation parameters, measurable by satellite,

to classify the land-surface vegetation type into one of twenty-one classes. The vegetation

parameters are green vegetation albedo, brown vegetation albedo, emissivity, maximum

simple ratio (defined as the ratio of 1+NDVI to 1-NDVI), maximum total area index,

stem area index, vegetation clumping fraction, vegetation fraction, vegetation height, root

78

depth, dead fraction and minimum stomatal resistance. Table 4.7. Table of parameters

used to define vegetative land-use types in LEAF-3. ( Walko and Tremback 2005)

The model uses a one-to-one lookup table to create the land-surface classification

based on the land classes from the Olson Global Ecosystem updated to 2001 MODIS

land-cover classes at 1km resolution (Friedl et al., 2002; Olson et al., 2001). The Olson

classes map the biosphere at approximately 1km resolution. The lookup table condenses

the 96 Olson classes into the 21 model classes. It should be noted that while more classes

are possible in the model, each model land class must be sufficiently different from the

others to make a border between land classes significant. Figure 4.2 shows the model

land-classes in the Olson tile containing Puerto Rico. There is no differentiation between

different types of tropical forest because of the insignificantly small measurable

differences between these types, as far as the model is concerned. Therefore the model

will not distinguish between a wet forest and a dry forest, but rather classify all as

“Evergreen broadleaf forest”, and this is discussed further in the limitations section of the

paper. The San Juan urban area was parameterized with the Town Energy Budget urban

canopy parameterization (Masson, 2000).

4.3.3 Input Data

The model is initialized from the NCEP GFS analysis (Kalnay, Kanamitsu, and

Baker, 1990) of temperature, moisture, and wind vectors and reinitialized every six hours

at the outer grids, with the inner grid nudged. The GFS analysis has 1.0° resolution

(approximately 50km near Puerto Rico) and represents the finest available resolution in

this vicinity. The GFS analysis is upscaled to the outer grid resolution of 64km, then

79

downscaled within RAMS to each of the two inner grids. The dynamic downscaling for

this experiment is consistent with the conclusions made by Castro et al. (2005) to retain

the observed atmospheric conditions. Other variables are computed within the model to

retain dynamic balance.

4.3.4 Experimental Design

4.3.4.1 Case Study

The case for this experiment is a series of weakly forced (or air-mass)

thunderstorms over the island of Puerto Rico on May 23, 2010. At approximately 17UTC,

a thunderstorm forms upwind of San Juan, weakens over the city, then slightly

reintensifies downwind. Later that hour, thunderstorms form downwind of the reforested

wet forest, on the west end of the island. Both events demonstrate some modification

over the land-surfaces under investigation: the reforested wet-forest, the San Juan urban

area, or the rain forest reserve. Figure details observed precipitation on the island as

derived from combined radar and rain-gauge observations (following Seo, 1998).

4.3.4.2 Land-Surface Scenarios

The experimental design for this modeling study simulates the same initial

conditions from 5/22 – 5/24, 2010, changing the underlying land-surface conditions to

determine how they affected thunderstorm development. The goal of the control

experiment is to produce a reasonable simulation of observations, not a replication. Each

land-surface scenario is then compared to the control to determine how individual

80

elements impact storm development and total precipitation. In each case, the scenario

compares a changed land-surface to either urban or evergreen broadleaf forest.

Land surfaces considered are bare soil, grassland, shrubland, crops, and an

expansion of existing land surface types. Bare soil removes all vegetation or buildings,

simulating a null case. Grassland and shrubland represent a less intensely vegetated

surface than broadleaf forest. Crops demonstrate how storm development changes with

agricultural land-use change. Expansion of the areas covered by existing land surface

types may demonstrate either a green-planning scenario, as in the case of expanded forest,

or a future urbanization development scenario.

Table 4.8. provides details for each scenario run for this experiment, including

how the land surface was changed. For the forest scenarios (RF and RWF), the bare soil,

grassland, and shrubland changes simulate the effects of lessening vegetation intensity,

while the spatial expansion does the opposite. For the urban scenarios, the changes

simulate how different urban boundaries or an expanded urban envelope can impact

storm development. Figure 4.1 shows the regions of changed land-surface for each set of

scenarios. Also shown is the downwind region of San Juan analyzed for precipitation

change and the three subdivisions of the island analyzed.

4.3.4.3 Control Results and Verification

The simulation reproduced the observations reasonably well. As previously stated,

the purpose of the control run was not to replicate observations, but produce a similar

precipitation pattern to study the land surface impacts. Figure 4.2 shows the model and

observed temperatures for different points on the island. Temperatures patterns are

81

smoother than observed data because the relatively coarse resolution of the model cannot

replicate microscale features. Otherwise, the temperatures agree quite well with

observations. Figure 4.3 details the model-produced precipitation during the period of

radar observation. The model produces surface precipitation in the same general area as

observed, making this useful for the sensitivity study. The precipitation is less than

observed due to the convective parameterization in the coarser grids. Table 4.9 compares

the selected event with the verified RAMS control run.

4.4 Data

Five minute reflectivity radar loops were obtained from the National Weather

Service office in San Juan. Land Use / Land Cover as well as Holdridge Ecological Life

Zones, geological, topographic and hydrologic digital maps were provided by the United

States Forest Service, International Institute of Tropical Forestry. Radiosonde data were

downloaded from the Wyoming Radiosonde global database, CMORPH data were

downloaded from the CMORPH database website, TRMM data were downloaded from

the TRMM database website. Hourly weather data were downloaded from NOAA NCDC

to generate time series of the control runs. Total precipitation satellite distribution maps

where downloaded from NOAA’s Advance Hydrologic Prediction Service.

82

4.5 Results and Discussion

4.5.1 Precipitation Changes

The first method by which precipitation is compared for this study compares total

accumulated water volume at the surface compared with that simulated in the control. For

example, 1mm of surface rain over a 4km grid represents 16,000m3 of total water, or 1.6

x 109 kg of water. When compared to a different value of the same magnitude, it is

equivalent to comparing precipitation amounts. Figures 4.8 to 4.13 show the percent of

surface water in each region compared to control.

4.5.1.1 Urban Scenarios (UI-A)

For each region on the island, changing the San Juan urban surface to another

land surface type reduced overall precipitation. The important forcing driving

precipitation in San Juan for this event is the mesoscale boundary between the urban and

rural environments. By changing it from urban to any other land surface, the gradient is

reduced and precipitation at the boundary is reduced. Notably, changing it from urban to

bare soil (1), grass (2), shrubs (3), and crops (4) reduces the surface temperature gradient

in each case. Changing the surface to broadleaf forest (5) would reduce the surface

gradient the most, but also adds more potential evapotranspiration, increasing moisture

and not reducing precipitation as much as the other scenarios.

In the downwind region, the precipitation change behaves differently. Downwind

intensification is governed primarily by atmospheric aerosols affecting drizzle formation.

Therefore, urban aerosols or some other source of aerosols are necessary for downwind

83

intensification. For all scenarios except the change to bare soil, the precipitation in the

downwind region is reduced, presumably because of the reduction in aerosols. In UI1A

(Fig. 4.12), the storm remains intense downwind not because of reintensification, but

because it did not weaken at the urban boundary.

4.5.1.2 Urban Expansion Scenarios (UI-B)

The urban expansion scenarios modify the location of the upwind mesoscale

circulation. Therefore precipitation upwind of the urban area increases in the central

region of the island (Figure 4.7) and is greatly reduced over the now-larger urban center

(Fig. 4.13). Overall precipitation on the entire western third of the island is increased due

to higher overall rain rates. The greatest changes come from those scenarios (UI-2B, UI-

5B) where the urban area is expanded south. This tightens the gradient between the

broadleaf forest and the urban area and modifies the atmospheric water availability

upwind.

Figure 4.11 demonstrates the effect of the upwind mesoscale circulation. In the

UI-5A scenario, the urban area is replaced with broadleaf forest, reducing the gradient

and thus upwind precipitation. In UI-5B the expanded urban area changes the location of

the gradient, thus increasing precipitation in the new upwind boundary. This particular

change, between eliminating and intensifying the gradient demonstrates that the observed

change to the thunderstorm on radar in the vicinity of San Juan was due to the urban-rural

boundary.

84

4.5.1.3 Rain Forest Reserve (RF)

Changes to the land surface of the rain forest reserve produce precipitation

changes both in the immediate vicinity and upwind of the San Juan urban area, due to

combined effects of urban-rural boundaries and forest boundaries. Changing the land

surface of the rain forest reserve from forest to bare soil (RF-1) creates another surface

mesoscale boundary capable of modifying storm behavior. The precipitation downwind

of the San Juan urban area is increased in the RF-1 scenario because of secondary

boundaries formed at the forest-bare soil boundary.

In the RF-1 scenario precipitation is increased upwind in the western and central

portions of the island. The total decrease in evapotranspiration keeps more water in the

atmosphere, leading to more efficient precipitation formation upwind of the rain forest

reserve. Figure 4.15 shows a map of precipitation changes as compared to the control. It

shows a large expanded area of increased precipitation around the changed land surface

resulting from both: greater atmospheric moisture availability and a change in surface

gradient. The combined surface gradients of sensible and latent heat (Fig. 4.7) between

forest to bare soil and urban to rural increases the precipitation upwind of the urban area.

4.5.1.4 Regenerated Wet Forest (RWF)

When considering the changes noted in the RWF scenarios, one should consider

the difference between a realistic change and a false teleconnection. The control

produced the lowest surface precipitation compared to observation on the western third of

the island, causing a comparison to show very large percent differences compared to the

control. In the central part of the island (Fig. 4.10), changing the RWF to bare soil (RWF-

85

1) or shrubs (RWF-3) increases precipitation, while changing it to grassland (RWF-2) or

crops (RWF-4) decreases precipitation. These precipitation changes are due to the change

in surface land gradient and surface moisture availability. The crops and grassland

decrease the gradient, while the bare soil and shrubs increases it. By expanding the RWF

(5), the forest becomes more complete, akin to the rainforest reserve. Therefore

precipitation increases greatly, especially downwind and in the region of land use change.

Figure 4.13 shows the mapped precipitation difference between the RWF-4 and RWF-5

scenarios. The RWF-4 increases precipitation in the central part of the island, making it

more realistic than the control. The RWF-5 scenario increases precipitation especially in

the changed areas of new forest.

This study provides insight into the land surface interactions behind a typical

thunderstorm system on Puerto Rico. As noted in Figure 4.11, the observed thunderstorm

modification at the urban boundary can be replicated in the model via eliminating and

intensifying the boundary. However, the highly heterogeneous nature of the land surface

makes it difficult to conclude how other regions modify the thunderstorm. Especially

difficult is the regenerated wet forest, which appears poorly represented in the study.

As simulated in the model, the eastern part of the island received the most

precipitation the day of the event (Fig. 4.17) but this region responded to each expansion

scenario and LC substitution reducing precipitation. Certain regions showed stronger

sensitivity to LULCC than others. Changing the land cover resulted in decreased

precipitation for 75% of the cases (Fig. 4.18). The primary reason for this is that changes

in the land surface caused the peak precipitation from the control simulation (Fig. 4.6) to

move farther offshore and thus decrease for all the island subregions. The fact that all but

86

one urban scenario decreased precipitation suggests that the urban land surface effect is

dependent on the location of the urban area relative to the location of precipitation.

Substituting shrubs for the existing land type caused the most instances of

increased precipitation in the scenarios (46.7%) followed by expansion of existing forests

(30%), and bare soil. Substituting forests for the control land surface decreased

precipitation in all cases (100%), with decreases in precipitation for grassland (93.3%),

crops (86.7%), and the city expansions (80%) (Fig. 4.19). The shrub substitutions

increased precipitation in the most cases because it was heterogeneous compared to the

most other types of land surface. Therefore it created the upward motions at the boundary

necessary to increase precipitation. While counterintuitive, substituting forest decreased

precipitation because most of the island is broadleaf forest (Fig. 4.2). Therefore changing

other land surfaces to forest decreases any land surface heterogeneity, reduces the effect

of the land-sea breeze driving the precipitation, and moves the peak precipitation further

offshore.

Expanding the rain forest reserve reduced precipitation island-wide, while

expanding the regenerated wet forest increased precipitation. This is because expanding

the rain forest reserve decreased the urban heat island effects on upward motion. The

regenerated wet forest is far enough away from the urban area that it is not directly linked

to its effects. Therefore by expanding it, the moisture availability on the island increases

and there is more precipitation. Expanding the urban area westward resulted in the

greatest decrease in precipitation. This is due to the particular location of the urban-rural

gradient. By moving it westward, it particularly effects the distribution of precipitation,

moving it offshore and decreasing it over the entire island.

87

The particular location of the urban-rain forest boundary in the control was

important to how much precipitation fell on the eastern portion of the island. Any

changes moved the precipitation farther offshore and affected the entire island. Other

parts of the island were more locally affected by changing the land surface (Fig. 4.20).

The central part of the island was highly variable to changes in the land surface. Both

affected by the urban-rural boundary and moisture from the western half, it was

particularly susceptible to the movement of the peak precipitation further on land or

offshore. The western part of the island was also susceptible to this.

4.6 Conclusions

The modeling component of this study simulated a single event, May 23, 2010,

where possible thunderstorm modification was observed in the vicinity of San Juan.

Using the RAMS model, several scenarios were run to study the impact of individual

regions of land-surface heterogeneity on the event. The control land-cover produced a

reasonable, but not exact, simulation of the observed precipitation event. The land surface

was then varied in the vicinity of the San Juan, Rain Forest Reserve, regenerated wet

forest, and unregenerated wet forest ecoregions, and new scenarios were simulated with

each.

Changes to modeled precipitation were most intense over and downwind of the

San Juan as shown in the changes to the urban area. Eliminating the urban area increased

precipitation of the urban center, while expanding it reduced precipitation. Expanding the

urban area also changed how much precipitation fell downwind. Relatively large changes

88

to precipitation were also simulated by changing the Rain Forest Reserve. Eliminating the

rain forest reserve reduced precipitation over the area. However, the Rain Forest Reserve,

for this particular event lies downwind of San Juan. Therefore there is a combined effect

of the San Juan urban area and the Rain Forest Reserve, which makes it difficult to

attribute the changes in precipitation to either region individually. Changing the Rain

Forest Reserve to bare soil to crops, respectively, decreases and increases precipitation in

the area. The decrease with bare soil is expected due to the reduced moisture. The crop

surface would theoretically also decrease surface moisture. However, the crop surface

also interacts with the unchanged urban land surface, and causes an invigorated storm

downwind of San Juan. This dual interaction makes it difficult to attribute observed

changes to the thunderstorm only to the urban area, or only to the urban-forest boundary.

Furthermore, the storm was not in the vicinity of the unregenerated and

regenerated wet forest ecoregions; yet, the change in precipitation upwind of the urban

area was on the same order of magnitude as the changes downwind with similar land

surface changes. This false teleconnection in the model likely does not represent a

realistic change and is not corroborated by other studies. Therefore it is difficult to

attribute significance to any of the individual land-surface interactions with the

thunderstorm. That is to say if any changes to the thunderstorm are produced near the

urban area, they are not significantly different from changes produced in a scenario away

from the urban area. This is not to say we cannot draw more generalized conclusions

about potential urban-rural and other land-surface heterogeneity interactions with

thunderstorm. It just cannot easily be shown through single-scenario based modeling

studies as shown here.

89

There are several limitations associated with this type of study:

1. The multiple collocated land-surface heterogeneities.

2. The scale of precipitation forcing on Puerto Rico.

3. The time scale of the study performed.

4. Data availability and the model's ability to initialize.

There are several urban-rural heterogeneities on Puerto Rico, and each could

individually modify a thunderstorm. San Juan has crops, forest, and most importantly

water (see next paragraph) on its boundary. A thunderstorm will be modified differently

depending on which of these boundaries it interacts with. In considering the Rain Forest

Reserve, changes to the urban area and changes to the forest may have competing

impacts. For example, changing either the urban area or the forest to crops increases

precipitation downwind, but for different reasons. Additionally, expanding the urban area

has the same impact as reducing the forest. Given that expanding the urban area in this

particular case may involve also reducing the forest, the impacts can also be falsely

intensified. Based on this, it is not easily possible to perform a true single-variable

scenario study with San Juan and the Rain Forest reserve.

The biggest problem with studying land-surface interaction with precipitation in

any way on Puerto Rico is the scale of the precipitation forcing. The primary mode of

precipitation on the island is seasonality (time). Precipitation will be heavier during the

wet season, and lighter during the dry season regardless of the land surface. The

secondary mode, more local than seasonality, is land-sea boundary and terrain. Local

precipitation in Puerto Rico is governed most strongly by the land-sea breeze combined

90

with local terrain forcing convection during the day and suppressing it at night. The

urban-rural heterogeneity is the strongest on the island, but it still remains a weaker

overall forcing than the synoptic and fixed-land based forcings. In order to observe the

nature of the urban-rural precipitation change on the island, it must be filtered from the

larger scale processes. Beyond the urban-rural, other heterogeneities (i.e. forest-crops)

represent an even weaker forcing and require further filtering to observe.

While RAMS is capable of simulating local-scale precipitation, it will also

simulate the larger scale processes at the proper scale. A case-scenario approach, as

shown here, is not the best way to remove these large-scale forcings to demonstrate the

land-surface heterogeneity based changes. The time-scale of the study becomes important

in order to filter the land-surface forcing from the larger scale forcings. For a single-event,

the largest-scale forcings dominate, even in a mesoscale model. At a longer time-scale

(weeks to months), the larger scale forcings become close to an average value and can be

removed to reveal the local forcing. However, the particular shortcomings of the model

prevent such a study here.

The best available input data near Puerto Rico is the GFS analysis, at 0.5-1.0 deg

resolution (50-100km). There exist some finer scale products near Puerto Rico, but given

the downscaling necessary to achieve 1km resolution near San Juan, they cannot be

applied to the entire outer domain of the model. Downscaling 50km input data to

represent a 1km feature is feasible, but is poorly represented here. We are forced to use a

single-event control because other events do not produce a control simulation close to

observations. These studies, not shown, do not void nor validate any previous

assumptions on land-surface interaction, but rather demonstrate a true limitation of the

91

RAMS model in Puerto Rico. The model has successfully simulated land-surface

interactions in other regions, even those using the same downscaling. However, the

particular problems present in Puerto Rico prevent RAMS in its current form from

successfully simulating cases here. Given the modeling experiment was only a portion of

the greater study, we cannot make any conclusions regarding the specific interaction of

the different ecoregions on the island with thunderstorms based only on this portion.

4.6.1 Recommendations for Future Work

Based on the results of this study, we can propose several future experiments,

beyond the scope of this paper. First, the model must be better calibrated to handle

simulations over Puerto Rico. This may involve assimilating surface and upper-air

observations along with reanalysis to improve model initialization. Additionally, NAM

analysis at a finer resolution, not readily available for research, could be used to initialize

the model. When reasonable and consistent control simulations can be produced, week

long to seasonal simulations can be performed to isolate the land-surface signal from

larger scale signals. The single-event runs are useful, but do not produce statistical results

for this particular type of event. Most importantly, the physics of the model may need to

be modified to handle precipitation on the island. The cloud physics, with the exception

of the GCCN, are parameterized based on continental clouds. The sea salt, and especially

ice-nucleating dust in the tropics are different enough to be significant. Furthermore for a

seasonal run, observational assimilation of aerosols, a feature only present in the most

theoretical of research at the current time, may be necessary to best simulate the seasonal

92

precipitation. Only with those improvements can we more fully extract the land-surface-

precipitation interaction on Puerto Rico.

4.7 References

Angeles, M. E., Gonzalez, J. E., Erickson III, D. J., & Hernández, J. L. (2006). An

assessment of future Caribbean climate changes using the BAU scenario by coupling a

global circulation model with a regional model. Proc. 86th Am. Met. Soc. Meet.,

Atlanta, Georgia, USA.

Carter, M. M., & Elsner, J. B. (1997). A statistical method for forecasting rainfall over

Puerto Rico. Weather and Forecasting, 12(3), 515-525.

Comarazamy-Figueroa, D.E., (2002). Atmospheric modeling of Caribbean Region:

precipitation and wind analysis in Puerto Rico for April 1998. Thesis, Master of

Science in Mechanical Engineering, University of Puerto Rico, Mayaguez Campus.

Comarazami, Daniel E. and González Jorge E. (2008). On the validation of Early Season

Precipitation on the Island of Puerto Rico Using a Meso Scale Atmospheric Model.

Journal of Hydrometeorology. Vol 9. Issue 3. 507-520.






Fall S., Niyogi D. , Semazzi F., (2006), Analysis of Mean Climate Conditions in Senegal

(1971–98) , Earth Interactions, 10, Paper 5, 1-40.

IPCC 2007. Fourth Assessment Report. Climate Change (2007): Chapter 9:

Understanding and Attributing Climate Change



2717.

Malmgren B, Winter A. (1999). Climate Zonation in Puerto Rico Based on Multivariate


985.

93





1239.

Ramirez-Beltran, Nazario D., William k. M. Lau, Amos Winter, Joan M. Castro, Nazario

Ramirez Escalante. (2007). Empirical probability models to predict precipitation levels

over Puerto Rico Stations. American Meteorological Society. Monthly Weather Review.

135: 877-890.

Velazquez-Lozada A, Gonzalez JE, and Winter A. (2006). Urban Heat Island Effect


1741.

Van der Molen, Michiel K. (2002). Meteorological Impacts of Land Use Change in the

Maritime Tropics. VRIJE UNIVERSITEIT . PhD Thesis.

Walko, Robert L. and Tremback, Craig J. INTRODUCTION TO RAMS 4.3/4.4

RAMS: The Regional Atmospheric Modeling System. Technical Description (Draft)

94

CHAPTER 5 CONCLUSIONS

The field of microclimatology is concerned with phenomena and surface

processes related to long-term weather within the Planetary Boundary Layer (PBL).

Natural sources as well as anthropogenic activities alter and modify surface processes in

ways that feedback in to the local climate by changing physical, biological and chemical

properties. Land Use / Land Cover Changes are known to alter land surface

characteristics and fluxes that can alter local weather and long-term climate. Most

knowledge about these impacts has been derived from studies in continental climates and

less is known about other types such as tropical – maritime climates. This dissertation

provided an opportunity to expand our knowledge of tropical climates dominated by

maritime conditions. Small tropical islands with maritime climates face unique

challenges to Global Warming/Climate Change because of sea level rise and loss of

coastlines in addition to any synergistic forcing effects of local LULCC.

The primary objective of this study was to determine the nature and pattern of

long-term changes in temperature and precipitation in Puerto Rico, a tropical maritime

island, and to determine if components of these changes in climate may be attributable to

local Land Use / Land Cover Changes. Long-term patterns for temperature and

precipitation were assessed by geographic region using Holdridge Ecological Life Zones

95

(HELZs). HELZs provide a useful tool to assess local climate variation as they

integrate ecological and environmental variables. The analyses included specific

evaluations of whether urbanization and reforestation are significant drivers of local

climate variation. A detailed numerical model that simulates individual storm events was

then used to simulate impacts of Land Use / Land Cover changes as part of building

explanations of how Land Use / Land Cover change alters the local weather events that

are the basis for long term climates.

5.1 Temperature Findings

On a scale of many decades, temperature changes in Puerto Rico are broadly

consistent with Global Sea and Land Temperature variations over the same time periods;

however there is considerable local temperature variability as represented by HELZ. A

variety of techniques were used to determine if there were significant differences in

temperature records between urban and non-urban areas and as a function of forest

regeneration. PCA/EOF identified urban areas and regenerated forest areas as being

distinct from areas not impacted by these land use changes, with higher values in stations

from the San Juan Urban area and the Regenerated Forest. The results of the OMR

analysis were counterintuitive, with the highest OMR trends in stations in the

Unregenerated Forest where no Land Use / Land Cover Changes have been documented.

However, the OMR results were limited to average temperatures because there are no

NARR maximum and minimum temperatures available and this may be one cause of the

anomalous results.

96

ANOVA was found to be a reliable method to assess urban versus non-urban

differences in Puerto Rico. Average temperatures in urban and non-urban areas were not

statistically distinct, and this finding is consistent with previous studies. However

maximum and minimum temperatures were significantly different between urban and

non-urban areas, with a confidence interval of 95% (α = 0.05). Statistical analysis of

GIS generated maps at each HELZ also detected a significant difference between all

urban and non-urban temperatures. These results were consistent with the results from

the statistical analysis of PRISM generated maps for all temperatures at each HELZ with

the same confidence interval and critical value. These findings provide strong evidence

that urban development have impacted temperatures around the whole island or that

urban temperatures signals were detected at each HELZ.

Temperature results represent minimum impacts since the analysis was done on

adjusted temperature data formulated to eliminate or reduce urban signals from the record.

Environmental impacts of findings may result in increased energy consumption because

of the increase in population use of air conditioner. Ecological impacts may result in

displacement of species and changes in habitat extension or boundaries. Mitigation may

be possible through urban greening and reforestation practices to increase albedo.

5.2 Precipitation Findings

Precipitation in Puerto Rico has been decreasing over the past century, consistent

with trends in the Caribbean basin as a whole. However, there is considerable

geographical variability in precipitation in Puerto Rico as represented by HELZ. Most

stations in Puerto Rico across all HELZs display the dominant pattern of decreasing

97

precipitation for the past century. However for the last 17 years a notable pattern of

increasing stations trends suggests that precipitation patterns maybe changing. Average

and median precipitation have been relatively close in all HELZs over the past century,

however from the 1990’s a new pattern has emerged with greater separation between

average and median precipitation magnitudes. This suggests an increase in distance or

difference between months with higher precipitation and those of lower precipitation that

is consistent with climate change projections for the Caribbean basin in which dryer

periods are expected with occasional periods of heavy rain.

Precipitation in non-urban stations was constantly higher than urban precipitation

across all HELZs. The slight reduction in precipitation that falls over urban areas may

relate to reductions in the mixing ratio and/or because of downwind advection of low

pressures. However, the small magnitude of the difference between urban and non-urban

precipitation may be because any effects of the urban settlements are too small to be seen

within the larger variability and trend associated with the predominant humid maritime

climate. The magnitude of the differences of monthly average precipitation on surface

stations is not statistically significant using a confidence interval of 95% (α = 0.05).

However, statistical analysis of GIS generated maps of yearly average total precipitation

detected statistical differences between urban and non-urban precipitation at each HELZ

using the same confidence interval and critical value. Further analysis of GIS generated

maps for monthly average precipitation trends confirmed statistical differences between

urban and non urban magnitudes. In addition, PRISM generated maps and confirmed

these findings using the same confidence interval of 95% (α = 0.05). However, finding

statistical differences between urban and non-urban areas from the beginning of the

98

century is surprising and may point to the assumptions of the unchanged land cover

through time or interpolation function hyper sensitivity.

5.3 RAMS Findings

Regional Atmospheric Modeling System (RAMS) is a weather event software

capable of performing numerical experiments related to land features and processes and

their effects on atmospheric phenomena. RAMS was used in this work to help us

understand how the major land use land cover changes and ecological regions in Puerto

Rico play a role changing or modifying precipitation events. A local thunderstorm event

with little synoptic influence was selected for the simulation. The most common and

probable land covers types and scenarios were systematically substituted to analyze

responses to these changes.

Some patterns seem to confirm that precipitation in the eastern part of the island is

more controlled by other drivers than Land Use / Land Cover Changes while the central

part showed the most sensitivity to simulation scenarios. However, the counterintuitive

nature of some results raised questions about the accuracy of RAMS and its reliability to

produce credible results at this scale. For instance, the fact that the model cannot yet

distinguish between Dry Forest and Wet Forest vegetation is an important limitation give

that the Dry Forest is the second largest HELZ in Puerto Rico and also the second more

impacted by urban development. Regarding results, the model predicts decreased

precipitation in the whole island with every single urban area substitution including forest.

It also it predicts increasing precipitation in other parts of the island is the Rain Forest

was substituted by bare soil. Although Puerto Rico currently developing “Land Use Plan”

99

may benefit from the use numerical modeling as well as other small jurisdictions, more

studies are needed to fine tune parameters and settings to improve performance.

5.4 Synthesis

Strong evidence for the impacts of urban development in Puerto Rico on

temperatures and precipitation have been provided to address an important climate

question relevant for assessing Climate Change impacts locally and around the world. A

simple method using station data and statistical methods without the use of mathematical

transformations or Reanalysis databases was developed and demonstrated suitable to

assess land use / land cover impacts for other small areas on any climate variable of

interest. Station location changes may be critical for some results that depend on small

number of stations, although adjusted data should account for that, the use of large

samples (many climate stations) should lessen any possible bias.

Computer model simulations are potentially useful in helping us reveal,

understand and predict how land process and features can change weather events but the

pilot study done in this dissertation suggests the much more fine-tuning is needed to

adequately model weather dynamics on small islands. Spatial resolution and

parameterization issues need to be addressed to accurately reproduce weather events at

small scales in order to produce credible results that can be instrumental for decision

making policies and practices. Puerto Rico offers great opportunities for follow-up

studies and developing better tools and methods for climate research of small islands.

100

5.5 Study Contributions

o Scientific Knowledge and Decision Making

Temperature findings increase our understanding of warming trends and temperatures

impacts of urban versus non-urban areas. Temperature maybe represent additional

stress or displacement for some species, higher temperatures may increase health

hazards and local energy consumption. Precipitation findings present important

challenges to be addressed locally as rainfall show a general decreasing trend and

local reservoirs are experiencing storage reductions because of sedimentation while

population and consumption increase. However, analysis from this study is

consistent with precipitation projections for the Caribbean region expecting dryer dry

periods and wetter more intense precipitation in wet periods pointing at critical water

conservation and management policy needed for adaptation.

o Statistical differences

ANOVA and t-test were used to assess statistical differences in temperatures

magnitudes and precipitation quantities between the urban and non urban areas using

a confidence interval of 95% or significance level of 0.05.

o Considered operational definition of Land Use versus Land Cover

This work methodologically addressed the difference between Land Use and Land

Cover by performing a replicate analysis for both definitions (Selection A and

Selection B).

o First use of FILNET 2 adjusted data in Puerto Rico

The temperatures data used for this study was first available in the summer of 2008.

o Use of HELZ’s for microclimate studies

101

Holdridge Ecological Life Zones, is a useful geographical tool created and used by

ecologists that generates climate provinces based on biotemperature, humidity, and

precipitation representing different conditions at different latitudes and altitudes. In

other words, climate variables are already contemplated and embedded with

vegetation.

o Multiple methods comparison

Traditional and contemporary climate methods (ANOVA, T-Test, PCA/EOF, OMR

and trends analysis) were combined to test the hypothesis of urban areas impacting

the local climate by altering local temperatures and precipitation.

o Regenerated Forest and Dry Forest Analysis

So far, related climate studies in Puerto Rico have focused on the urban impacts of

the San Juan urban area and the Rain Forest while studies about land processes in the

Regenerated Forest and the Dry Forest have been lacking.

o Characterization and Parameterization

Findings and results could provide valuable information to characterize the climate of

Caribbean islands and parameterize similar climates to feed climate models and

simulations.

5.6 Study Limitations

o Long term analysis is limited to the past century and the use of monthly summarized

temperature and precipitation data from land surface stations, unfortunately no

additional climate variables observations in Puerto Rico, such as winds, humidity,

pressure, radiation etc have such long term databases.

102

o Temperature analysis made on adjusted temperature continuous database while

precipitation analysis was made on raw data containing many gaps and missing

values.

o HELZs borders may be questionable as zones are assumed to be steady and not have

changes for many decades, it is currently unknown how HELZ borders and

extensions have changed.

o Station locations were assumed to remain unchanged for the period of the study,

however, temperature adjusted data should have account for that.

o FILNET 2 algorithm uncertainties are clear since it effectiveness to remove urban

signals from the temperatures are evident since the statistical analysis was able to

detect urban signals.

o Urban station temperature analysis was limited to the Moist Forest, while non-urban

temperature stations were located in the Dry Forest and Wet Forest to do a

comparative analysis.

o Too few temperature stations are located at urban areas and near natural reservations.

More urban temperature stations are needed even at the Moist Forest since currently

only the San Juan Urban area is represented and more stations are needed near natural

reservations to monitor changes and impacts . Some critical locations of research

interest, either have no stations or very little numbers such as east of San Juan, north

of the Rain Forest, the Rain Forest and the Unregenerated Wet Forest.

o No altitudinal variation was considered for temperature, however HELZ’s already

controls for any variation related to altitude through its climate regionalization

scheme.

103

o Main focus was on urban land use / land cover change; feedbacks from other land

uses / land covers, ecological & environmental variables were not considered.

However, the method used here may be use to study any other climate variable and

land use/land cover.

o No NARR maximum and minimum temperature data available to do OMR

o OMR analysis performed on selected stations and not all areas represented and no

Dry Forest stations were included.

o Raw precipitation raw data uncertainties, gaps and discontinuities.

o RAMS uncertainties and difficulties to parameterize or distinguish between local

vegetation types BATS may not accurately represent all existent Puerto Rico’s Land

Covers.

o Simulated events were limited rain events, no calm days were simulated.

o Control event verification data was limited to a handful of stations with hourly data

leaving out areas of research interest.

o There is only one radio sounding station in Puerto Rico to cover the entire island.

o Selected simulated weather events may not represent long term patterns

o Study was limited to local effects; no association or impacts of larger climatic

forcings combined with local land processes have been studied.

o The role and effects of aerosols and environmental pollution in air are not considered.

o No surface fluxes databases available

o Land Covers were limited to the most recent maps assuming that urban regions

remained static from actual to the past overestimating the amount and extent of urban

areas.

104

5.7 Future Directions

Longer term simulations runs are important to understand how changes in short

term weather conditions may alter the longer-term climate patterns. Higher time

resolution data for long term analysis and more climate variables such as humidity, winds,

pressures, radiation etc as well as surface fluxes are desirable. Satellite based

temperature and precipitation data may be useful for complementing station data

discontinuities and geographic coverage. Also, radar data may help in the study of

precipitation. Since maritime conditions dominate, sea surface data should be linked to

local land surface processes. Further studies should address other land covers types like

different agriculture lands, different levels of urbanization and different land uses that

represent different combinations of albedo and aerodynamic roughness levels. The

impacts of aerosols and environmental pollution should also be considered since they

play an important role in precipitation. Past Land Use / Land Cover reconstructions

would be useful to link surface processes with early century climate. Parameterization

studies of local land covers are needed so that they become more accurately represented

in BATS. Studies about the Regenerated Forest and Dry Forest represent unique and

important opportunities for further research.

105

CHAPTER 2 TEMPERATURE TABLES

Table 2.1. Holdridge Ecological Life Zone (HELZ) relative coverage and number of

temperature stations for Puerto Rico

Puerto Rico HELZ Coverage (% of

Puerto Rico’s land area)

Number of HCN temperature

stations

Subtropical Dry Forest (DF) 14 7

Subtropical Moist Forest (MF) 62 35

Subtropical Wet Forest (WF) 23 10

*Subtropical Lower Montane Wet Forest < 1 0

*Subtropical Lower Montane Rain Forest < 1 1

*Subtropical Rain Forest < 1 1

*= recoded as WF zones excluded from regional analysis breakdown. HELZ data from

International Institute of Tropical Forestry, weather stations from NOAA Historical

Climate Network .

106

Table 2.2. Characteristics of Major Regions Used in this Study.

a Area as a percentage of Puerto Rico’s total area

* One station was excluded from all preliminary analysis because of data errors

** All urban temperature stations were located in the Moist Forest HELZ

*** Consists of all Moist Forest stations excluding all stations coded as urban for 1992

Region CODE areaa # stations

Puerto Rico PR 100% 57*

Dry Forest DF 14% 7

Moist Forest MF 62% 35

Moist Forest Non Urban *** MFNU n/a 27

Urban LC 1992 A** U1992A n/a 7

Urban LC 1992 B** U1992B n/a 9

Urban LC 2004 A** U2004A n/a 3

Urban LC 2004 B** U2004B n/a 5

Wet Forest WF 23% 11

Unregenerated Wet Forest East UnWF n/a 9

Regenerated Wet Forest West RWF n/a 2

107

Table 2.3. Seasonal and Annual Temperature statistics for HELZ, Moist Forest

Urban Land Use Areas and Non-Urban, and areas of Regenerated and

Unregenerated Forest 1900-2007.

Temperature, oC

Region

Annual Dry Season Wet Season

Min Ave Max Min Ave Max Min Ave Max

DF 21.58 26.12 30.66 20.03 24.86 29.69 22.70 27.04 31.38

MF 20.41 25.41 30.41 18.83 24.05 29.27 21.56 26.41 31.25

WF 17.33 22.26 27.19 15.85 20.98 26.11 18.41 23.20 27.98

RWF 17.17 22.16 27.16 15.67 20.85 26.02 18.27 23.13 27.99

UWF 18.03 22.68 27.33 16.64 21.56 26.48 19.05 23.51 27.97

MFNU 19.51 24.63 29.76 17.97 23.32 28.67 20.63 25.59 30.56

*U1992A 20.73 24.57 28.41 19.20 23.21 27.23 21.83 25.55 29.27

*U1992B 20.95 24.93 28.90 19.44 23.59 27.73 22.04 25.90 29.76

*U2004A 20.84 24.92 29.00 19.20 23.50 27.81 22.03 25.95 29.87

*U2004B 21.24 25.26 29.28 19.70 23.90 28.11 22.36 26.25 30.14

* Stations located in the Moist Forest HELZ

10

8

Table 2.4. HELZ Temperature Ratios and Differences

DF = Dry Forest, MF = Moist Forest, WF = West Forest, PR = Puerto Rico

Temperature Ratio

(HELZ Temp / PR Temp; oC)

Temperature Difference

(HELZ Temp – PR Temp; oC)

Remarks HELZ Min. Ave. Max. Min. Ave. Max.

DF 1.10 1.07 1.05 1.93 1.74 1.45 At least 1.5 degrees warmer than PR

MF 1.04 1.04 1.04 0.76 1.03 1.21 Around 1 degree warmer than PR

WF 0.88 0.91 0.93 -2.32 -2.12 -2.02 Over 2 degrees colder than PR

PR 1 1 1 0 0 0 Over 1 degree of temperature differences between HELZ’s

10

9

Table 2.5. Significance of temperature differences between HELZ (ANOVA)

Decadal Seasonal Wet Seasonal Dry Monthly

HELZ Min Ave Max Min Ave Max Min Ave Max Min Ave Max

DF 0.000 0.000 0.315 0.000 0.000 0.000 0.000 0.000 0.000 0.145 0.365 0.836

MF 0.000 0.000 0.315 0.000 0.000 0.768 0.000 0.000 0.032 0.145 0.365 0.836

WF 0.000 0.000 0.000 0.000 0.000 0.768 0.000 0.000 0.032 0.000 0.000 0.000

Bold values are significant (α = 0.05)

11

0

Table 2.6. Temperatures Variation Explained by HELZ (R2)

HELZ Monthly Variation Dry Season Wet Season Decadal Variation

ANOVA Min Ave Max Min Ave Max Min Ave Max Min Ave Max

% explained (R2) 61.7 66.0 70.8 97.4 97.8 96.7 96.4 95.5 94.0 97.72 96.59 94.72

11

1

Table 2.7 Urban versus Non Urban One Way ANOVA

Urban versus Non Urban One Way ANOVA, for urban classification by selection method

Urban

Code

Monthly Seasonal Dry Seasonal Wet Decadal

Min Ave Max Min Ave Max Min Ave Max Min Ave Max

*U1992A 0.245 1.000 0.022 0.000 0.928 0.000 0.000 0.998 0.000 0.000 0.985 0.000

*U1992B 0.114 0.977 0.281 0.000 0.281 0.000 0.000 0.180 0.000 0.000 0.208 0.000

*U2004A 0.185 0.980 0.408 0.000 0.666 0.000 0.000 0.092 0.003 0.000 0.242 0.000

*U2004B 0.037 0.726 0.799 0.000 0.001 0.005 0.000 0.000 0.152 0.000 0.000 0.045

Regenerated Forest versus Unregenerated Forest One Way ANOVA

UWF vs RWF 0.591 0.842 0.995 0.000 0.000 0.040 0.000 0.061 1.000 0.000 0.003 0.832

Bold values are significant (σ = 0.05)

11

2

Table 2.8. EOF Modes for all Temperatures

Temperature Type First Mode Second Mode Total

Minimum 60.13% 6.88% 67.01%

Average 77.73% 4.27% 81.99%

Maximum 72.71% 5.14% 77.85%

11

3

Table 2.9 Main Locations Top 10% Temperature Stations Summary

Century Temperatures Century EOF

Location HELZ Min Ave Max Min Ave Max

*San Juan Urban MF 2/5 1/5 0 1/5 1/5 0

Regenerated Forest WF 0 0 0 1/5 0 3/5

Unregenerated Forest WF 0 0 0 0 0 0

Rain Forest WF 0 0 0 0 0 0

Dry Forest DF 3/5 4/5 2/5 0 0 1/5

Numbers from the total Top 10%

*Includes urban stations outside San Juan area classified as urban in 1992 and/or 2004

11

4

Table 2.10 Main Locations Bottom 10% Temperature Stations Summary

Location HELZ

Century Temperatures Century EOF

Min Ave Max Min Ave Max

*San Juan Urban MF 0 0 0 2/5 2/5 2/5

Regenerated Forest WF 4/5 4/5 2/5 1/5 0 0

Unregenerated Forest WF 0 0 0 0 0 0

Rain Forest WF 1/5 1/5 2/5 0 0 0

Dry Forest DF 0 0 0 1/5 0 0

Number from the total Bottom 10%


11

5

Table 2.11. Puerto Rico’s Average and Median period trends for all temperatures

Trend

Century

1900-2007

PRISM

1963-1995

OMR

1979-2005

Warming

1970-2007

Min Ave Max Min Ave Max Min Ave Max Min Ave Max

Average (oC / year) 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.03

Median (oC / year) 0.01 0.01 0.01 0.02 0.02 0.02 0.02 0.02 0.02 0.03 0.03 0.03

11

6

Table 2.12. Main Locations Top 10% Temperature Stations Summary

Location HELZ

Century Temperatures Century Trends


*San Juan Urban MF 2/5 1/5 0 1/5 0 1/5

Regenerated Forest WF 0 0 0 1/5 0 1/5

UnRegenerated Forest WF 0 0 0 0 0 0

Rain Forest WF 0 0 0 0 0 0

Dry Forest DF 3/5 4/5 2/5 1/5 1/5 0

Numbers from the total Top 10%


11

7

Table 2.13 Main Locations Bottom 10% Temperature Stations Summary

Location HELZ

Century Temperatures Century Trends


*San Juan Urban MF 0 0 0 2/5 2/5 1/5

Regenerated Forest WF 4/5 4/5 2/5 1/5 1/5 0

UnRegenerated Forest WF 0 0 0 0 0 0

Rain Forest WF 1/5 1/5 2/5 0 0 0

Dry Forest DF 0 0 0 1/5 0 0

Numbers from the total Bottom 10%.


118

Table 2.14. Ranked OMR for Average Temperature Trends of Selected Stations

Station Name

HELZ/LC

1979 – 2005 Yearly Trends oC 1979 – 2005 Decadal Trends

oC

FILNET NARR OMR FILNET NARR OMR

MARICAO_2_SSW Regenerated

Forest 0.033 0.007 0.026 0.331 0.070 0.261

CERRO_MARAVILLA Regenerated

Forest 0.027 0.005 0.023 0.274 0.048 0.227

CARITE_DAM UnReg.

Forest 0.026 0.005 0.021 0.255 0.049 0.206

CAYEY_1_E U1992 0.025 0.004 0.021 0.249 0.044 0.205

SAN_LORENZO_ESPINO UnReg.

Forest 0.025 0.005 0.020 0.246 0.051 0.196

RIO_PIEDRAS U2004 0.021 0.005 0.016 0.210 0.048 0.162

SAN_SEBASTIAN_2_WNW Regenerated

Forest 0.022 0.006 0.016 0.217 0.058 0.158

GUINEO_RESERVOIR Regenerated

Forest 0.018 0.004 0.014 0.185 0.045 0.140

LARES_2_SE Regenerated

Forest 0.018 0.005 0.013 0.177 0.048 0.129

SAN_JUAN_CITY U1992B &

U2004B 0.011 0.006 0.006 0.114 0.056 0.058

GARZAS Regenerated

Forest 0.010 0.005 0.005 0.102 0.052 0.050

ADJUNTAS_SUBSTN Regenerated

Forest 0.007 0.005 0.001 0.066 0.052 0.014

SAN_JUAN_WSFO U2004 0.005 0.006 -0.001 0.048 0.059 -0.011

119

Table 2.15. Results of the statistical analysis for century average temperature values

for each HELZ from GIS generated maps and each evaluated data base.

HELZ

PRISM maps

1963-1995 t -test (2 tailed)

FILNET SPLINE maps

1900-2007 t - test (2 tailed)

Max T Ave T Min T Max T Ave T Min T

Wet Forest 0.00 0.00 0.00 0.00 0.00 0.00

Moist Forest 0.00 0.00 0.00 0.00 0.00 0.00

Dry Forest 0.00 0.00 0.00 0.00 0.00 0.00

Results at 95% confidence interval (σ = 0.05)

120

Table 2.16. Difference in Urban versus Non Urban average century or period

temperatures magnitudes from GIS generated maps for each HELZ and data set.

Wet Forest

FILNET Temperature oC PRISM Temperature o

C

Temperature U NU U-NU U NU U-NU

Maximum 30.16 28.13 2.02 30.21 27.97 2.24

Average 24.22 23.05 1.17 24.02 22.75 1.27

Minimum 18.28 17.98 0.31 17.88 17.58 0.31

Moist Forest


C


Maximum 29.62 29.15 0.47 30.15 29.68 0.48

Average 25.00 24.27 0.73 25.22 24.54 0.68

Minimum 20.37 19.70 0.68 20.33 19.49 0.84

Dry Forest


C


Maximum 30.13 29.80 0.34 31.08 30.78 0.30

Average 25.56 25.32 0.24 25.80 25.53 0.27

Minimum 20.99 20.85 0.15 20.55 20.32 0.23

121

Table 2.17. Results of the statistical analysis for century average temperature values of

each urban versus non urban evaluated data bases.

* Two Way ANOVA done on FILNET Station data from Urban Land Cover from 2004B and

1992B maps.

** Student’s t-test done with 95% confidence interval (σ = 0.05) on 2004 Urban Land Cover

map.

HELZ

*FILNET station U/NU

Century 2 Way ANOVA

**PRISM maps U/NU

1963-1995 t test (2 tailed)

**FILNET SPLINE maps U/NU

Centuryt test (2 tailed)

Max T Ave T Min T Max T Ave T Min T Max T Ave T Min T

Wet Forest N/A N/A N/A 0.00 0.00 0.00 0.00 0.00 0.00

Moist Forest *0.00 >0.05 *0.00 0.00 0.00 0.00 0.00 0.00 0.00

Dry Forest N/A N/A N/A 0.00 0.00 0.00 0.00 0.00 0.00

122

CHAPTER 3 PRECIPITATION TABLES

Table 3.1 Summary of previous precipitation research and articles in Puerto Rico

Article # stations Period Method Remarks

Ray, 1933

46

1899-1932

% departures from

normal

The smaller the average the greater the year by year variation and vice versa. Gradual decrease in precipitation preceded drought years. Wet years come in pairs

Ewel and Whitmore,

1973

143

1900-1969

Holdridge Ecological

Life Zones

Microclimate classification based on biotemperature, humidity and precipitation. Various lengths of data, longest individual period 15 years

Pagan-Trinidad,

1984 10 1971-1983

Statistics, frequencies, probability

Spatial and temporal variability of storm rainfall (storm duration, rain intensity) Geographic consideration

Carter and Elsner, 1996

22 1973-1988 EOF

Diurnal rainfall regionalization (6 regions), No land cover change considered. The eastern part showed low hourly variability while the western part showed high hourly variability.

Carter and Elsner, 1997

22 1973-1988 EOF, Statistical

Classification Tree Regionalization (6 regions), No land cover considered

Malmgren et al., 1998

5 1901-1995 Station data, SOI and

NAO indexes, Burnaby test

Regional synoptic phenomena rainfall and temperature influence over PR

Malmgren and Winter

1999 18 1960-1990

Rotated PCA and Neural Networks

Seasonal Rainfall Regionalization (4 zones) In three zones precipitation increased

consecutively at each season and peaked in the Fall, in one zone it peaked in the summer.

Larsen 2000

12

1900-2000

Drought Index, rainfall vs stream flow

comparison

Interregional and intraregional analysis. Precipitation decreasing in the Caribbean and PR

Comarazamy, 2001

15 April 1998 Mesoscale model

April 1998 wind and rainfall simulation San Juan city precipitation was under predicted, The most accurate results occurred at higher elevations were uplift from northeasterly winds interacts with steeper slopes

Van der Molen, 2002

1 portable + various on site

instruments

May 1997 - May 1998

Field observations, mesoscale model

Analyzed deforested areas, measured forest reservation forests, Urban land cover not considered

123

Table 3.1 cont.

Daly et al., 2003

47 Temp; 108 Precip.

1963-1995 PRISM

No land cover change considered. Spatial variation of rainfall was associated to elevation, upslope exposure to winds carrying moisture and distance to coastline.

Neelin et al., 2006

Gridded data 1970-2003 1950-2002 1951-2000

Satellite databases, gridded station

data, Precipitation models

Higher model agreement in the Caribbean and Central American of Summer dying trend and increase in heavy rain events

Ramírez-Beltrán et al.,

2007 6 1901-2001 Statistical models

Logistic Regression; Categorical Classification of precipitation events, no land cover considered

Jury et al., 2007

35 regional (7 from PR)

1951-1981 Factor Analysis Caribbean basin rainfall regionalization

Nyberg et al., 2007

4 cores (paleo study)

1730-2005

Coral cores, wind shear record

reconstruction, artificial neural

networks

Historical Caribbean basin hurricane activity

Comarazamy and

González, 2008

15 1993 and

1998 Mesoscale model

Early wet season simulation, No land cover change considered

Harmsen et al., 2009 3 1960-2000

GCM downscaling, trends, linear

regression

precipitation deficits scenarios resulted in crop yield reduction and wetter wet seasons and drier dry seasons

Jury, 2009

Interpolated data from

17,000 Caribbean

stations to grids

1979-2000

Observations, Reanalysis,

Satellite, and Coupled

Model Data

Evaluated ability of different products to represent mean annual Caribbean rainfall. Caribbean rainfall is projected to decline around 20% over the next 100 yr.

Jury and Sanchez,

2009 60 (rain gages) 1979-2005

NCAR, statistics of daily rainfall

Most flood events in Puerto Rico occurred in May, August and September

Comarazamy et al., 2010

N/A Atlas Mission

Observations

February 10-20, 2004

Mesoscale model

Mixed Urban and Natural adjustments yielded more accurate results. U-R temperature up to 2.5o C difference Increased precipitation downwind Southwest San Juan

12

4

Table 3.2. Annual effects of ENSO and NAO on Puerto Rico’s Precipitation

a) Malmgren et al., 1998; b) Jury et al., 2007 Effects are based on observations unless otherwise indicated.

* are effects deduced from the paper narrative, and + are effects described as expected in the paper narrative.

Climate Variable

ENSO Effect

NAO Index

High Average Low High Average Low

1990-1995a a a a a a a

Total Precipitation No effect+ No effect+ No effect+ < average* No effect* > average*

Average Precipitation No effect No effect No effect < average No effect > average

Median Precipitation No effect+ No effect+ No effect+ < average+ No effect+ > average+

1951-1981b b b b b b b

Monthly Total Precipitation No effect+ No effect+ No effect+ > average > average < average

Monthly Average Precipitation No effect* No effect* No effect* > average* > average* < average*

Seasonal Precipitation > average > average > average < average < average > average

12

5

Table 3.3. Number of stations by Selection Type and Analyzed HELZ and Land Cover for 1992 Puerto Rico Land

Cover Map.

Each selection data set was evaluated in an independent replicated analysis.

1992 Puerto Rico Land Cover Map Stations by Analyzed Selection Type and Study Regions

HELZ Total

stations

HELZ

subdivisions

Selection A Selection B 30 m Selection B 60 m Selection B 90 m

U NU U U NU NU NU NU

Wet Forest 27

Regenerated 0 22 0 22 4 18 7 15

Unregenerated 0 3 0 3 0 3 0 3

Rain Forest Reserve 0 2 0 2 0 2 0 2

Moist Forest 75 N/A 13 62 21 54 N/A N/A N/A N/A

Dry Forest 24 N/A 2 22 4 20 11 13 N/A N/A

12

6

Table 3.4. Number of stations by Selection Type and Analyzed HELZ and Land Cover for 2004 Puerto Rico Gap Map .

Each selection data set was evaluated in an independent replicated analysis.

2004 Puerto Rico GAP Map # of Stations by Analyzed Selection Type and Study Regions

HELZ Total

stations

HELZ

subdivisions

Selection A Selection B 30 m Selection B 60 m

U NU U NU U NU

Wet Forest 27

Regenerated 0 22 0 22 0 22

Unregenerated 0 3 0 3 0 3

Rain Forest Reserve 0 2 0 2 0 2

Moist Forest 75 N/A 7 68 11 64 15 60

Dry Forest 24 N/A 3 21 4 20 5 19

12

7

Table 3.5. Holdridge Ecological Life Zones Distributions and Descriptive Statistics

HELZ Data

distribution Mean

(cm/month) Standard Deviation

Median (cm/month)

Maximum (cm/month)

Remarks

WFR Johnson

Transformation 21.89 6.198 24.95 29.12

Rain Forest Reservation. Wettest region in Puerto Rico, different to most of Puerto Rico. Median

above Average

WF Gaussian 18.29 8.165 17.76 30.27 Center Mountains. Highest rainfall variation and

maximum

UnWF Johnson

Transformation 19.79 6.733 23.10 27.32 Eastern Mountain, Median above Average

RWF Gaussian 18.52 8.484 18.88 30.19 Regenerated Wet Forest, Median Above Average,

MF Gaussian 14.20 4.773 14.34 20.18 Most of Puerto Rico. Mean and median very similar

DF Gaussian 8.19 4.447 7.80 15.33 Driest region in Puerto Rico. Significantly different

to ALL others in every test

12

8

Table 3.6. 1992 LULC Average Monthly Precipitation Ratio 1900-2007

Total HELZ / PR Selection A Selection B 30 meters Selection B 60 meters

Remarks 1992 cm/month cm/month Urban Non Urban Urban Non Urban Urban Non Urban

WF 18.29 1.32 N/A N/A 1.07 0.99 0.91 1.01 Urban B 30 More precipitation over urban

MF 14.20 1.02 0.90 1.01 0.91 1.02 0.91 1.02 Less precipitation over urban areas vs no urban

DF 7.80 0.56 0.97 1.06 1.02 1.06 1.02 1.08 Less precipitation over urban areas vs no urban. Smaller difference

PR 13.88 1.00 0.87 1.02 0.89 1.00 0.85 1.03 Less precipitation over urban areas vs no urban

12

9

Table 3.7. 2004 LULC Average Monthly Precipitation Ratio 1900-2007

Total (cm/month)

HELZ / PR (cm/month)

Selection A Selection B 30 meters Selection B 60 meters

Remarks 2004 Urban Non Urban Urban Non Urban Urban Non Urban

WF 18.29 1.32 N/A N/A N/A N/A N/A N/A Less precipitation over urban areas vs no urban

MF 14.20 1.02 0.96 0.99 0.94 1.00 0.95 1.00 Less precipitation over urban areas vs no urban

DF 7.80 0.56 1.26 1.02 1.17 1.03 1.08 1.05 More precipitation over urban areas vs no urban

PR 13.88 1.00 0.901 0.910 0.882 0.914 0.877 0.917 Less precipitation over urban areas vs no urban

13

0

Table 3.8 Yearly Average Total Precipitation for each period and its corresponding Urban versus Non urban T-test

significance values

270m 1900-1929 cm/year 1930-1959 cm/year 1960-1989 cm/year 1990-2007 cm/year 1963-1995 cm/year OMR cm/year PRISM cm/year

U NU Sig. U NU Sig. U NU Sig. U NU Sig. U NU Sig. U NU Sig. U NU Sig.

WF 375.16 341.81 0.000 429.40 398.76 0.000 470.76 406.70 0.000 202.75 214.86 0.000 504.77 433.56 0.000 361.72 319.98 0.000 165.62 172.84 0.000

MF 347.26 340.77 0.000 279.59 338.75 0.000 364.01 362.77 0.027 170.64 183.18 0.000 386.65 385.55 0.057 297.06 299.43 0.000 99.45 104.12 0.000

DF 226.24 220.29 0.000 266.67 240.35 0.000 244.45 255.27 0.000 151.52 153.16 0.001 267.08 273.51 0.000 232.89 233.05 0.802 207.74 216.95 0.000

270 meter grid cell

13

1

Table 3.9 Yearly Average Total Precipitation for each period and its corresponding Urban versus Non urban T-test

significance values.

100m 1900-1929 cm/year 1930-1959 cm/year 1960-1989 cm/year 1990-2007 cm/year 1963-1995 cm/year OMR cm/year PRISM cm/year


WF 375.33 342.11 0.000 434.80 398.95 0.000 479.00 407.02 0.000 199.31 214.37 0.000 513.57 434.07 0.000 362.57 320.32 0.000 165.62 172.84 0.000

MF 347.45 341.00 0.000 260.05 341.21 0.000 364.02 362.77 0.000 169.90 183.31 0.000 387.81 385.43 0.000 298.75 299.43 0.000 99.45 104.12 0.000

DF 225.54 219.91 0.000 271.75 238.65 0.000 240.31 255.44 0.000 149.74 153.06 0.000 263.21 273.52 0.000 229.67 233.04 0.000 207.74 216.95 0.000

100 meter grid cell

13

2

Table 3.10 270 meter Grid Cell Yearly Average Total Precipitation Trends for each period and its corresponding

Urban versus Non urban T test significance values

270m

1900-1929 cm/year 1930-1959 cm/year 1960-1989 cm/year 1990-2007 cm/year 1963-1995 cm/year OMR cm/year Century cm/year


WF -0.008 -0.015 0.000 0.133 0.026 0.000 -0.132 -0.195 0.000 -0.710 -0.451 0.000 -0.104 -0.135 0.000 -0.139 -0.664 0.000 -0.041 -0.043 0.000

MF -0.013 -0.018 0.000 0.005 0.006 0.413 -0.009 -0.090 0.000 -0.120 -0.163 0.000 -0.048 -0.115 0.000 -0.109 -0.187 0.000 -0.044 -0.050 0.000

DF -0.007 -0.016 0.000 -0.040 -0.052 0.000 -0.274 -0.270 0.389 0.025 -0.002 0.000 -0.182 -0.162 0.000 -0.066 -0.064 0.467 -0.042 -0.069 0.000

13

3

Table 3.11 100 meter Grid Cell Yearly Average Total Precipitation Trends for each period and its corresponding

Urban versus Non urban T test significance values

100m 1900-1929 cm/year 1930-1959 cm/year 1960-1989 cm/year 1990-2007 cm/year 1963-1995 cm/year OMR cm/year Century cm/year


WF -0.008 -0.019 0.000 0.134 0.029 0.000 -0.133 -0.198 0.000 -0.762 -0.464 0.000 -0.100 -0.136 0.000 -0.142 -0.675 0.000 -0.039 -0.071 0.000

MF -0.015 -0.023 0.000 -0.003 0.009 0.000 0.011 -0.094 0.000 -0.120 -0.166 0.000 -0.030 -0.121 0.000 -0.098 -0.191 0.000 -0.040 -0.045 0.000

DF -0.011 -0.023 0.000 -0.038 -0.053 0.000 -0.287 -0.273 0.000 0.019 -0.003 0.000 -0.194 -0.163 0.000 -0.072 -0.069 0.052 -0.045 -0.052 0.000

134

CHAPTER 4 RAMS TABLES

13

5

Table 4.1 Summary of previous RAMS work about Puerto Rico

Article Objectives Assumptions Remarks Findings

Comarazamy, 2001

Simulate early wet season precipitation (April 1998)

April 1998 RAMS circulation & rainfall simulation

urban vegetation and green areas not considered,

inland water bodies were not represented and soil type

soil moisture seem inadequate

City very inaccurate

higher elevations most accurate at higher elevations northeasterly winds interacts

with steeper slopes increasing precipitation.

Van der Molen, 2002

Simulate Rain Forest types

to study meteorological differences between natural land covers and deforested

locations in Puerto Rico.

coastal deforestation covered with pasture and

lowland forest

Analyzed deforested areas, measured forest reservation vegetation, Urban

land cover not measured, considered or simulated

data gathered from a field campaign results not verified against actual

observations

conversion of coastal forests to pasture would result in decrease precipitation

Angeles et al., 2006

Simulate IPCC Bussines as Usual scenario towards

2048

Transient CO2 increase

RAMS and PCM coupled model

Sea Level (SL) will increase in 0.35 cm, SOI and NAO drives the 15 years annual rainfall

variability. The most accuracy achieved at synoptic scales for

the dry and late rainfall seasons. Mesoscale rainfall strongly influenced by the land dry areas

late rainfall season mesoscale rainfall in the Caribbean is driven by the vertical wind shear &

dry/moisture advection. RAMS finer grid predicts future warmer areas over Puerto Rico

surface air temperatures 2.5°C above monthly average for 2048. Early rainfall season shows

sudden rainfall increases dry season will have more intense and abrupt rainfall

13

6

Table 4.1 cont.

Velazquez-Lozada et al.,

2006

Assess current UHI over San Juan and simulate ad quantify future UHI caused

by development

Potential native vegetation-

scenario replacing

urban needle leaf trees

replaced

concrete with island

predominant soil texture

urban expansion

based on population

increase for 2050

Topography was not modified for any of the simulations

impact of urban LCLU in the upper atmosphere is related to the sensible heat fluxes

from the surface

80°F potential vegetation scenario

82.5 °F actual scenario

84.75 °F future scenario

Comarazamy and González, 2008

Simulate early wet season precipitation (April 1998)

Year 1998 considered

representative of climatology

April considered Wet Season

1998 very warm and hurricane active

year No land cover considered Early wet season precipitation is characterized

by convective rainfall, differential heating and local moisture transport

greatest accuracy at higher elevations urban area very inaccurate

Comarazamy and González, 2010

Simulate current urban conditions, potential

vegetation and a mixture urban and natural

Assess reliability of RAMS to study LULC changes in

tropical coastal areas

Validated control runs against ATLAS 2004 aerial observation mission.

Mixed Urban & Natural adjustments yielded more accurate results.

UHI showed about 2.5 C intensity Increased precipitation downwind/southwest of San

Juan

13

7

Table 4.2. Study objectives, research questions and hypothesis

Objectives Research Sub questions Hypotheses

Discover and measure LUC/C and dominant HELZ feedbacks on local precipitation events

Is there any LUC/C and dominant HELZs based feedback on local precipitation events?

Local LUC/C and dominant HELZs feedbacks on local precipitation events can be effectively detected and measured through RAMS simulations.

Explain how mayor LUC/C and dominant HELZ play a role driving or modifying local precipitation events.

What are the major LUC/C feedbacks on local precipitation events? What are the dominant HELZs feedbacks in local precipitation events? What mechanism, environmental variables and conditions related to major LUC/C and dominant HELZs are controlling or modifying local precipitation events? How are the identified mechanism, environmental variables and conditions controlling or modifying local precipitation events?

Major LUC/C are modifying local precipitation events by changing magnitudes of variables of interest and/or altering convection, convergence and/or cloud formation mechanisms Dominant HELZs are controlling local precipitation events trough the variables of interest and/or convection, convergence and/or cloud formation mechanisms Variables of interest and convection, convergence and/or cloud formation mechanisms acting on local precipitation events can be identified and measured trough RAMS simulations

Explain the observed long term precipitation patterns between urban vs non urban areas in the different HELZs

Why are Urban & Non Urban monthly precipitation statistically similar across Puerto Rico? What mechanisms, environmental variables and conditions are causing that monthly precipitation is statistically similar across Puerto Rico?

Although individual precipitation events may be modified by mayor LUC/C, monthly quantities are balanced by different responses throughout the month. The same mechanisms, environmental variables and conditions identified and measured trough RAMS simulations are causing similar urban vs non urban monthly precipitation

Asses possible long term precipitation response to changes in the different land use and land covers in Puerto Rico.

What are the expected impacts potential of LUC/C feedbacks on local precipitation events? What is the magnitude potential of LUC/C feedbacks on local precipitation events?

13

8

Table 4.3. Locations of Interest, HELZ and Land Cover

Main Location Description HELZ Actual Land Cover

San Juan Urban The largest urban/city like setting in the island

where UHI effect has been detected Moist Forest Urban

Rain Forest Reserve Natural forest reserve with the lowest temperatures

and highest precipitation Wet Forest Broadleaf Forest / Non Urban

Regenerated Forest Former crop lands where natural forest re-growth

have occurred Wet Forest

Broadleaf Forest / Non Urban / extremely

very low scattered Urban

Unregenerated Forest Similar environment to Regenerated Forest but no

crops or forest re-growth Wet Forest Broadleaf Forest / Non Urban

Dry Forest Natural environment with the highest temperatures

and lowest precipitation Dry Forest Irrigated Crops / Some Urban

Dry Forest Reserve Natural forest reserve with the highest

temperatures and lowest precipitation Dry Forest Shrubs / Non Urban

13

9

Table 4.4. Selected Land Cover substitutions for RAMS Simulations

Land Cover Substitutions Justification

Evergreen Broadleaf Dominates higher altitudes in the island

Bare soil Base comparison to urban development

Forest & Urban Expansions Forecast potential changes of forest growth and impacts of urban development

Grass Dominates lower altitudes in the island

Shrubs Dominates driest regions in the island

Cropland Hint to a past dominated land cover and forecast possible impacts of potential agriculture development

policy

14

0

Table 4.5. Variables of interest and associated mesoscale rainfall triggering mechanisms

Variable Trigger Mechanism

Justification Remarks Hypothesis

Precipitation All local & synoptic

Researched state variable / dependent variable

Focused on local mechanisms, can either increase or decrease with impacts

Expected lower in urban and air polluted locations

Temperature Convection May increase convection Triggers raising air from land processes Expected higher at urban and

regions with intense land processes

Surface heat flux

Convection; Convergence

Increased convection & may facilitate air mass discrepancies

Caused by raising air from land processes Expected higher at urban and

regions with high land processes

Latent heat flux

Convergence; cloud nuclei

May facilitate air mass discrepancies and cloud

formation

Evaporation caused by temperature also related to plant transpiration; water vapor

available for clouds,

Expected higher at urban and regions with intense land processes

Relative humidity



formation

Water vapor available for clouds Caused by evaporation and plant transpiration


Mixing ratio Convergence; cloud nuclei


formation

Water content available for clouds related to evaporation and plant transpiration


Soil Moisture Convection Fuel storms; May increase

convection Associated with temperature, evaporation

and plant transpiration Expected lower in urban and high

temperature locations

Vertical velocity

Convection Implies convection Associated with temperature and raising

air from land processes Expected higher in urban and high

temperature locations

Rain water content


May facilitate air mass discrepancies

Associated with evaporation and plant transpiration


Cloud fraction Convection;


Increased convection & may facilitate air mass discrepancies

and cloud formation

Associated with evaporation, plant transpiration, natural aerosols and air

pollution


Surface convergence


Implies air mass discrepancies and may facilitate cloud formation

Associated with physical features on land causing uplift of air masses

Expected higher at urban and regions with intense land processes

14

1

Table 4.6: Table indicating model parameters for each of the three nested grids.

Grid 1 Grid 2 Grid 3

NX x NY 48 x 32 50 x 34 70x 38

Center Lat./Lon. (18.23N, 66.45W) (18.23N, 66.45W) (18.23N, 66.45W)

NZ 48 48 48

Δx and Δy 64km 16km 4km

Unstretched Δz 40m 40m 40m

Δt 90 15 2.5

Initialization GFS Analysis G1 & 80% GFS nudging G2 & 50% GFS nudging

Convective scheme Kain-Fritch Kain-Fritch Explicit

Cloud Microphysics Explicit Explicit Explicit

14

2

Table 4.7. Table of parameters used to define vegetative land-use types in LEAF-3. ( Walko and Tremback 2005)

Green

veg.

albedo

Brown

veg.

albedo

Emiss.

Max.

simple

ratio

Max.

total

area

index

Stem

area

index

Veg.

clumping

fraction

Veg.

Fraction

Veg.

height

Root

depth

Dead

fraction

Min.

stomatal

res.

0 - Ocean 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0.00 0.0 0.0 0.0 0

1 - Lakes, rivers,

streams 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0.00 0.0 0.0 0.0 0

2 - Icecap, glacier 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0.00 0.0 0.0 0.0 0

3 - Desert, bare soil 0.00 0.00 0.00 0.0 0.0 0.0 0.0 0.00 0.0 0.0 0.0 0

4 - Evergreen needle

leaf tree 0.14 0.24 0.97 5.4 8.0 1.0 1.0 0.80 20.0 1.5 0.0 500

5 - Deciduous needle

leaf tree 0.14 0.24 0.95 5.4 8.0 1.0 1.0 0.80 22.0 1.5 0.0 500

6 - Deciduous broadleaf

tree 0.20 0.24 0.95 6.2 7.0 1.0 0.0 0.80 22.0 1.5 0.0 500

7 - Evergreen broadleaf

tree 0.17 0.24 0.95 4.1 7.0 1.0 0.0 0.90 32.0 1.5 0.0 500

8 - Short grass 0.21 0.43 0.96 5.1 4.0 1.0 0.0 0.75 0.3 0.7 0.7 100

9 - Tall grass 0.24 0.43 0.96 5.1 5.0 1.0 0.0 0.80 1.2 1.0 0.7 100

10 - Semi, desert 0.24 0.24 0.96 5.1 1.0 0.2 1.0 0.20 0.7 1.0 0.0 500

11 - Tundra 0.20 0.24 0.95 5.1 4.5 0.5 1.0 0.60 0.2 1.0 0.0 50

12 - Evergreen shrub 0.14 0.24 0.97 5.1 5.5 1.0 1.0 0.70 1.0 1.0 0.0 500

13 - Deciduous shrub 0.20 0.28 0.97 5.1 5.5 1.0 1.0 0.70 1.0 1.0 0.0 500

14 - Mixed woodland 0.16 0.24 0.96 6.2 7.0 1.0 0.5 0.80 22.0 1.5 0.0 500

15 - Crop, mixed

farming, grassland 0.22 0.40 0.95 5.1 5.0 0.5 0.0 0.85 1.0 1.0 0.0 100

16 - Irrigated crop 0.18 0.40 0.95 5.1 5.0 0.5 0.0 0.80 1.1 1.0 0.0 500

17 - Bog or marsh 0.12 0.43 0.98 5.1 7.0 1.0 0.0 0.80 1.6 1.0 0.0 500

18 - Wooded grassland 0.20 0.36 0.96 5.1 6.0 1.0 0.0 0.80 7.0 1.0 0.0 100

19 - Urban and builtup 0.20 0.36 0.90 5.1 3.6 1.0 0.0 0.74 6.0 0.8 0.0 500

20 - Wetland evergreen

broadleaf tree 0.17 0.24 0.95 4.1 7.0 1.0 0.0 0.90 32.0 1.5 0.0 500

143

Table 4.8: Details of land-surface changes for each scenario.

Scenario San Juan City Rain Forest Reserve Regenerated Wet Forest

Urb

an

Sce

na

rio

s

UI-1A Replace w/ bare soil - - - - - -

UI-2A Replace w/ grassland - - - - - -

UI-3A Replace w/ shrubs - - - - - -

UI-4A Replace w/ crops - - - - - -

UI-5A Replace w/ forest - - - - - -

Ra

in F

ore

st R

eser

ve

Sce

na

rio

s

RF-1 - - - Replace w/ bare soil - - -

RF-2 - - - Replace w/ grassland - - -

RF-3 - - - Replace w/ shrubs - - -

RF-4 - - - Replace w/ crops - - -

RF-5 Expand in all dir. - - -

Reg

ener

ate

d W

et F

ore

st

Sce

na

rio

s

RWF-1 - - - - - - Replace w/ bare soil

RWF-2 - - - - - - Replace w/ grassland

RWF-3 - - - - - - Replace w/ shrubs

RWF-4 - - - - - - Replace w/ crops

RWF-5 - - - - - - Expand in all dir.

Urb

an

Ex

pa

nsi

on

Sce

na

rio

s UI-1B Expand west - - - - - -

UI-2B Expand south - - - - - -

UI-3B Expand east - - - - - -

UI-4B Expand west & east - - - - - -

UI-5B Expand in all dir. - - - - - -

144

CHAPTER 2 TEMPERATURE FIGURES

Figure 2.1 Puerto Rico and Global Ocean 1900-2007 Average Temperature Anomalies.

Global data from NOAAA, Puerto Rico data from FILNET 2.

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

19

00

19

10

19

20

19

30

19

40

19

50

19

60

19

70

19

80

19

90

20

00

Puerto Rico Global Ocean

Ce

lc

ius

145

Figure 2.2 Puerto Rico and Global Land 1900-2007 Average Temperature

Anomalies. Global Data from NOAA, Puerto Rico data from FILNENT 2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

0.4

0.6

0.8

1

19

00

19

10

19

20

19

30

19

40

19

50

19

60

19

70

19

80

19

90

20

00

Puerto Rico Global Land

Ce

lc

ius

146

Figure 2.3. Puerto Rico 1992 Land Cover Map from Helmer et al, 2002

147

Figure 2.4. Puerto Rico GAP 2004 Land Cover Map from Gould et al, 2007

.0 10 20 30 405

Miles

148

Figure 2.5. Puerto Rico Holdridge Ecological Lifezones (HELZ), urban areas and

weather stations. HELZ data from US Forest Service, urban areas data from

Puerto Rico GAP 20014, weather stations data from NOAA Historical Climate

Network

.0 10 20 30 405

Miles

Legend

Stations

Heavy Urban 2004

HELZ

Holdridge Ecological Lifezones

Dry Forest

Moist Forest

Lower Montane Rain Forest

Subtropical Rain Forest

Lower Montane Wet Forest

Subtropical Wet Forest

costa

Legend

Stations

Heavy Urban 2004

HELZ


Dry Forest

Moist Forest





costa

149


(>10% frequency) and below normal (<10% frequency) minimum temperature at

each HELZ

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

< 10% 11-89% > 90%

Puerto Rico Wet Forest Moist Forest Dry Forest

% o

f y

ears

150

Figure 2.7 Distribution of years registering normal (80% frequency), above normal

(>10% frequency) and below normal (<10% frequency) average temperatures at

each HELZ

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

< 10% 11-89% > 90%


% o

f y

ears

151


(>10% frequency) and below normal (<10% frequency) maximum temperature at

each HELZ

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

< 10% 11-89% > 90%


% o

f y

ears

152

Figure 2.9. Puerto Rico’s SPLINE interpolated Century Maximum Temperature

EOF

Legend

urban_2004_area

lifezones83_transparent

<all other values>

ECOZONE

df-S

mf-S

rf-LM

rf-S

wf-LM

wf-S

Stations

FILNET_Max_T_EOF

ValueHigh : 167045

Low : 5690

.0 10 20 30 405

Miles

153

Figure 2.10. Puerto Rico’s SPLINE interpolated Century Average Temperature

EOF

.0 10 20 30 405

Miles

Legend

urban_2004_area


<all other values>

ECOZONE

df-S

mf-S

rf-LM

rf-S

wf-LM

wf-S

Stations

FILNET_Ave_T_EOF

ValueHigh : 165077

Low : 3146

154

Figure 2.11. Puerto Rico’s SPLINE interpolated Century Minimum Temperature

EOF

.0 10 20 30 405

Miles

Legend

urban_2004_area


<all other values>

ECOZONE

df-S

mf-S

rf-LM

rf-S

wf-LM

wf-S

Stations

FILNET_Min_T_EOF

ValueHigh : 181039

Low : -33286

155

Figure 2.12. Puerto Rico’s 1900-2007 Maximum Temperature Station Trends

-0.040

-0.030

-0.020

-0.010

0.000

0.010

0.020

0.030

Ce

lciu

s/ye

ar

156

Figure 2.13. Puerto Rico’s 1900-2007 Average Temperature Station Trends

-0.030

-0.020

-0.010

0.000

0.010

0.020

0.030

Ce

lciu

s/ye

ar

157

Figure 2.14. Puerto Rico’s 1900-2007 Minimum Temperature Station Trends

-0.030

-0.020

-0.010

0.000

0.010

0.020

0.030

Ce

lciu

s/ye

ar

158

Figure 2.15. Puerto Rico’s 1900-2007 Maximum Temperature station trend

frequency distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

< 10% 11-89% > 90%

Total Wet Forest Moist Forest Dry Forest

% o

f st

atio

ns

159

Figure 2.16. Puerto Rico’s 1900-2007 Average Temperature station trend frequency

distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

< 10% 11-89% > 90%


% o

f st

atio

ns

160

Figure 2.17. Puerto Rico’s 1900-2007 Minimum Temperature station trend

frequency distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

< 10% 11-89% > 90%


% o

f st

atio

ns

161

Figure 2.18. Puerto Rico’s Urban 1900-2007 Average Temperature years frequency

distribution

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

< 10% 11-89% > 90%

Moist Forest Urban 92 A Urban 92 B Urban 04 A Urban 04 B

% o

f ye

ars

162

Figure 2.19. Puerto Rico’s HELZ 1963-1995 Average Temperature year frequency

distribution

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

80.0%

90.0%

100.0%

< 10% 11-89% > 90% Moist Forest Urban 92 A Urban 92 B Urban 04 A Urban 04 B

% o

f ye

ars

163

Figure 2.20 FILNET GIS interpolated data urban minus non-urban temperature

differences by type of temperature

0.00

0.50

1.00

1.50

2.00

2.50

Maximum Average Minimum

Wet Forest Moist Forest Dry Forest

Tem

per

atu

re (

Cel

ciu

s)

164

Figure 2.21 PRISM data urban minus non-urban temperature differences by type of

temperature

0.00

0.50

1.00

1.50

2.00

2.50

Maximum Average Minimum


Tem

pe

ratu

re (

Cel

ciu

s)

165

Figure 2.22. FILNET urban minus non-urban temperatures differences by HELZ

0.00

0.50

1.00

1.50

2.00

2.50


Minimum Average Maximum

Tem

pe

ratu

re (

Cel

ciu

s)

166

Figure 2.23. PRISM urban minus non-urban temperatures differences by HELZ

0.00

0.50

1.00

1.50

2.00

2.50


Minimum Average Maximum

Tem

per

atu

re (

Cel

ciu

s)

167

CHAPTER 3 PRECIPITATION FIGURES

Figure 3.1. Puerto Rico’s Holdridge Ecological Lifezones, Areas of Interest &

Weather stations. HELZ data from US Forest Service, urban areas data from

Puerto Rico GAP 20014, weather stations data from NOAA Historical Climate

Network

.0 10 20 30 405

Miles

Legend

Stations

Heavy Urban 2004

HELZ


Dry Forest

Moist Forest





costa

Legend

Stations

Heavy Urban 2004

HELZ


Dry Forest

Moist Forest





costa

168

Figure 3.2 1900-2007 Average and Median Monthly Precipitation for Puerto Rico’s


2.0

6.0

10.0

14.0

18.0

22.0

26.0

30.0

Wet Forest Average Wet Forest Median Moist Forest Average Moist Forest Median Dry Forest Average Dry Forest Median

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

169

Figure 3.3. Puerto Rico’s Holdridge Ecological Lifezones Average and Median

Monthly Precipitation through the decades

3.0

8.0

13.0

18.0

23.0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Wet Forest Average Wet Forest Median Moist Forest Average Moist Forest Median Dry Forest Average Dry Forest Median

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

170

Figure 3.4. Puerto Rico 1992 Land Cover Map from Helmer et al, 2002

171

Figure 3.5. Puerto Rico GAP 2004 Land Cover Map from Gould et al, 2007

.0 10 20 30 405

Miles

172

Figure 3.6. Monthly Average and Median Precipitation for Urban stations by

HELZ

1.5

4.5

7.5

10.5

13.5

16.5

19.5

22.5

25.5

January February March April May June July August September October November December

Urban RWF 92B 60m Average Urban RWF 92B 60m Median Urban MF 92B 30m Average Urban MF 92B 30m Median Urban MF 2004B 60m Average Urban MF 2004B 60m Median Urban DF 92B 60m Average Urban DF 92B 60m Median Urban DF 2004B 60m Average Urban DF 2004 B 60m Median

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

173

Figure 3.7. Average & Median Urban versus Non Urban Monthly Precipitation for

Wet Forest selections

4.0

9.0

14.0

19.0

24.0

29.0

Wet Forest Urban 92B 60m Average Wet Forest Urban 92B 60m Median Wet Forest Non Urban 92B 60m Average Wet Forest Non Urban 92B 60m Median Urban RWF 92B 90m Average Urban RWF 92B 90m Median WF Non Urban 92B 90m Average WF Non Urban 92B 90m Median

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

174

Figure 3.8. Monthly Average and Median Precipitation for the Moist Forest Urban

A and Non-Urban Selections

5.0

7.0

9.0

11.0

13.0

15.0

17.0

19.0

21.0


Moist Forest Total Average Moist Forest Total Median Urban MF 92A Average Urban MF 92A Median Urban MF 2004A Average Urban MF 2004A Median MNFU 92A Average MFNU 92A Median MFNU 2004A Average MFNU 2004A Median

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

175

Figure 3.9. Monthly Average and Median Precipitation for the Moist Forest Urban

B and Non-Urban Selections

5.0

7.0

9.0

11.0

13.0

15.0

17.0

19.0

21.0


Moist Forest Total Average Moist Forest Total Median

Urban MF 92B 30m Average Urban MF 92B 30m Median

Urban MF 2004B 30m Average Urban MF 2004B 30m Median

Urban MF 2004B 60m Average Urban 2004B 60m Median

MFNU 92B 30m Average MFNU 92B 30m Median



Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

176

Figure 3.10. Average Monthly Precipitation for the Dry Forest Urban 1992 A and

Non- Urban Selections

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0

Jan Feb Mar Apr May Jun Jul Aug Spt Oct Nov Dec

Urban Dry Forest 92 A Average DF No Urban 92 A Average

Urban Dry Forest 92 B 30m Average DF No Urban 92 B 30m Average

Urban Dry Forest 92 B 60m Average DF No Urban 92 60m Average

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

177

Figure 3.11. Median Monthly Precipitation for the Dry Forest Urban 1992 A and

Non- Urban Selections

1.5

3.5

5.5

7.5

9.5

11.5

13.5


Urban Dry Forest 92 A Median DF No Urban 92 A Median

Urban Dry Forest 92 B 30m Median DF No Urban 92 B 30m Median


Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

178

Figure 3.12. Average Monthly Precipitation for Dry Forest 2004 Urban versus Non-

Urban Selections

2.0

4.0

6.0

8.0

10.0

12.0

14.0

16.0


Urban Dry Forest 2004 A Average DF No Urban 2004 A Average

Urban Dy Forest 2004 B 30m Average DF No Urban 2004 B 30m Average

Urban Dry Forest 2004 B 60m Average DF No Urban 2004 B 60m Average

Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

179

Figure 3.13. Median Monthly Precipitation for Dry Forest 2004 Urban versus Non-

Urban Selections

2.0

4.0

6.0

8.0

10.0

12.0

14.0


Urban Dry Forest 04 A Median DF No Urban 04 A Median



Mo

nth

ly P

reci

pit

atio

n (

cm/m

on

th)

180

Figure 3.14. Puerto Rico Annual Cycle Monthly Precipitation by Periods (cm)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0


1900-1929 1930-1959 1960-1989 1990-2007

Pre

cip

itat

ion

(cm

/mo

nth

)

181

Figure 3.15. Wet Forest Annual Cycle Monthly Precipitation by Periods (cm)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0


1900-1929 1930-1959 1960-1989 1990-2007

Pre

cip

itat

ion

(cm

/mo

nth

)

182

Figure 3.16. Moist Forest Annual Cycle Monthly Precipitation by Periods (cm)

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0


1900-1929 1930-1959 1960-1989 1990-2007

Pre

cip

itat

ion

(cm

/mo

nth

)

183

Figure 3.17. Dry Forest Annual Cycle Monthly Precipitation by Periods

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0


1900-1929 1930-1959 1960-1989 1990-2007

Pre

cip

itat

ion

(cm

/mo

nth

)

184

Figure 3.18 Seasonal Monthly Total Precipitation by Periods

0.0

5.0

10.0

15.0

20.0

25.0

30.0

1900-1929 1930-1959 1960-1989 1990-2007

Wet Forest Dry Wet Forest Wet

Moist Forest Dry Moist Forest Wet

Dry Forest Dry Dry Forest Wet P

reci

pit

atio

n c

m/m

on

th

185

Figure 3.19 Annual Precipitation Quantiles for Wet Forest by Period

160

180

200

220

240

260

280

1900-1929 1930-1959 1960-1989 1990-2007

10th Percentile 50th Percentile 90th Percentile

Pre

cip

itat

ion

(cm

/yea

r)

186

Figure 3.20 Annual Precipitation Quantiles for Moist Forest by Period

120

130

140

150

160

170

180

190

200

210

220

1900-1929 1930-1959 1960-1989 1990-2007


Pre

cip

itat

ion

(cm

/yea

r)

187

Figure 3.21 Annual Precipitation Quantiles for Dry Forest by Period

60

70

80

90

100

110

120

130

140

1900-1929 1930-1959 1960-1989 1990-2007


Pre

cip

itat

ion

(cm

/ye

ar)

188

Figure 3.22. 1900-2007 Precipitation Trends by Station

-0.7

-0.6

-0.5

-0.4

-0.3

-0.2

-0.1

0

0.1

0.2

0.3

Stations

Pre

cip

itat

ion

Tre

nd

s (

cm/y

ear)

189

Figure 3.23. 1900-2007 Station Precipitation Trends by period

0

10

20

30

40

50

60

70

80

90

100

1900-1929 1930-1959 1960-1989 1990-2007

Stations showing Positive trends Stations showing negative trends

Nu

mb

er o

f St

atio

ns

190

Figure 3.24. Number of stations with positive versus negative trends by HELZ and

period

0

10

20

30

40

50

60

70

80

90

100

1900-1929 1930-1969 1970-1989 1990-2007

Puerto Rico Positive Puerto Rico Negative

Dry Forest Positive Dry Forest Negative

Moist Forest Positive Moist Forest Negative

Wet Forest Positive Wet Forest Negative

Nu

mb

er o

f P

reci

pit

atio

n S

tati

on

s

191

Figure 3.25 Yearly Average Total Precipitation Urban versus Non-Urban

Difference

-100.0

-80.0

-60.0

-40.0

-20.0

0.0

20.0

40.0

60.0

80.0

100.0

1900-1929 1930-1959 1960-1989 1990-2007


Pre

cip

itat

ion

Dif

fere

nce

(cm

/ye

ar)

192

Figure 3.26. Number of study periods receiving higher Yearly Average Urban

versus Non-Urban Total Precipitation

0

1

2

3

4

5

6


Urban Non Urban

Nu

mb

er o

f S

tud

y P

erio

ds

193

Figure 3.27. Number of study periods recording higher Urban versuss Non-Urban

precipitation trends

0

1

2

3

4

5

6

7


Urban Non Urban N

um

ber

of

Stu

dy

Per

iod

s

194

CHAPTER 4 RAMS FIGURES

Figure 4.1. Map detailing location of each grid for the study. The 50km resolution of

the GFS input data is overlaid on the outermost grid.

195

Figure 4.2. Map detailing LEAF-3 land-use types near Puerto Rico.

196

Figure 4.3. Map of radar derived observed precipitation within the inner grid for

1200UTC 5/23 to 1200UTC 5/24.

197

Figure 4.1. Map detailing areas of land-use change within the model for each set of

scenarios. Also shown is the region downwind of San Juan analyzed, and the

subdivisions of the island analyzed.

198

Figure 4.2. Observed versus simulated temperature during study for a) San Juan

International Airport, b) Arecibo, c) Mayaguez, and d) Yabucoa Harbor.

199

Figure 4.3. Total simulated precipitation for the inner grid, shown on the same scale

as radar derived precipitation in Figure .

200

Figure 4.4. Changes in sensible and latent heat fluxes at 18UTC 5/23/10 showing an

increase in both gradients.

201

Figure 4.5. Total accumulated precipitation as a ratio to control for the entire island.

202

Figure 4.6. Total accumulated precipitation as a ratio to control for the western part

of the island.

203

Figure 4.7. Total accumulated precipitation as a ratio to control for the central part

of the island.

204

Figure 4.8: Total accumulated precipitation as a ratio to control for the eastern part

of the island.

205

Figure 4.9. Total accumulated precipitation as a ratio to control for the region

downwind of San Juan.

206

Figure 4.10. Total accumulated precipitation as a ratio to control for individual

areas of changed land surface for each scenario.

207

Figure 4.11. Comparison of the change in precipitation between the a) UI5A

scenario and b) UI5B scenario. In UI5A, the surface is changed to forest, reducing

the urban gradient and reducing upwind precipitation. In UI5B, the expanded

urban envelope changes the location of the mesoscale circulation, changing the

location of upwind precipitation.

(a)

(b)

208

Figure 4.12. Precipitation difference between control and RF1 scenario. Resulting

precipitation represents the combined effects of the changed land surface from

forest to bare soil interacting with the unchanged urban area to the west.

209

Figure 4.13. Map of precipitation difference between control and a) RWF4 scenario

and b) RWF5 scenario. Shown here for clarification of large percentage difference

in Fig. 12.

(b)

(a)

210

Figure 4.17. Control 6 hour Average Precipitation Time Series

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1

1.1

18UTC 5/21 00UTC 5/22 06UTC 5/22 12UTC 5/22 18UTC 5/22 00UTC 5/23 06UTC 5/23 12UTC 5/23 18UTC 5/23 00UTC 5/24 06UTC 5/24 12UTC 5/24

Pre

cip

itat

ion

(ce

nti

met

ers)

Total Island

Western 3rd

Central 3rd

Eastern 3rd

Downwind of San Juan

May 23, 2010

211

Figure 4.18. Percentage of resulting scenarios with increased versus decreased

precipitation

25%

73%

2%

increased decreased equal

212

Figure 4.19. Percentage of Increase versus Decrease Precipitation Results by

Scenario

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

increase decrease

Shrubs Expand Forests Bare Soil Expand City Crops Grassland Forests

Perc

enta

ge o

f Sce

nari

o Re

sults

213

Figure 4.20. Precipitation Response ratio for each scenario at each region relative to

control

0.00

0.50

1.00

1.50

2.00

2.50

Total Island Western 3rd Central 3rd Eastern 3rd Downwind of San Juan

Urban/Bare Soil Urban/Grassland Urban/Shrubs Urban/Crops Urban/Forest Rain Forest/Bare Soil Rain Forest/Grassland Rain Forest/Shrubs Rain Forest/Crops Rain Forest/Expand all Regenerated Forest/Bare Soil Regenerated Forest/Grassland Regenerated Forest/Shrubs Regenerated Forest/Crops Regenerated Forest/Expand all Urban/Expand West Urban/Expand South Urban/Expand East Urban/Expand East & West Urban/Expand all

cm

APPENDICES

214

Appendix A Figures

Figure A.1. Ecozones Decadal Average Temperature Dry Season Standardized

Anomalies

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

19

00

19

10

19

20

19

30

19

40

19

50

19

60

19

70

19

80

19

90

20

00

Dry Forest Dry Moist Forest Dry Urban 1992 Dry Urban 2004 Dry Moist Forest No Urban Dry Wet Forest Dry Wet Forest East Dry Wet Forest West Dry Puerto Rico

Te

mp

era

ture

(C

elc

ius

)

215

Figure A.2. Ecozones Decadal Average Temperature Wet Season Standardized

Anomalies

-2.5

-2

-1.5

-1

-0.5

0

0.5

1

1.5

2

2.5

19

00

19

10

19

20

19

30

19

40

19

50

19

60

19

70

19

80

19

90

20

00

Dry Forest Wet Moist Forest Wet Urban 1992 A Wet Urban 2004 A Wet Moist Forest No Urban Wet Forest Wet

Te

mp

era

ture

(C

elc

ius

)

216

Figure A.3 Puerto Rico Seasonal Temperature Standardized Anomalies by Decade

-2.50

-2.00

-1.50

-1.00

-0.50

0.00

0.50

1.00

1.50

2.00

2.50

19

00

19

10

19

20

19

30

19

40

19

50

19

60

19

70

19

80

19

90

20

00

Average Dry Average Wet

Average Dry Maximum Average Wet Maximum

Average Dry Minimum Average Wet Minimum

Te

mp

era

ture

(C

elc

ius

)

217

Figure A.4 Dry Forest Percentage Decadal Temperature changes

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

2.00%

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

min ave max

Te

mp

era

ture

(C

elc

ius

)

218

Figure A.5 Moist Forest Percentage Decadal Temperature changes

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

2.00%

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

min ave max

Te

mp

era

ture

(C

elc

ius

)

219

Figure A.6 Wet Forest Percentage Decadal Temperature changes

-2.00%

-1.50%

-1.00%

-0.50%

0.00%

0.50%

1.00%

1.50%

2.00%

1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

min ave max

Te

mp

era

ture

(C

elc

ius

)

220

Figure A.7 1992 A Urban minus Non-Urban Decadal Temperature Difference

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Tem

pe

ratu

re (

Ce

lciu

s)

Min Ave Max

221

Figure A.8 1992 B Urban minus Non-Urban Decadal Temperature Difference

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Tem

pe

ratu

re (

Ce

lciu

s)

Min Ave Max

222

Figure A.9. 2004 A Urban minus Non-Urban Decadal Temperature Difference

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Te

mp

era

ture

(C

elc

ius)

Min Ave Max

223

Figure A.10. 2004 B Urban minus Non-Urban Decadal Temperature Difference

-2.0

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

1900 1910 1920 1930 1940 1950 1960 1970 1980 1990 2000

Tem

pe

ratu

re (

Ce

lciu

s)

Min Ave Max

224

Figure A.11 Urban 2004 A versus Urban 2004 B Average Monthly Temperature

18.5

20.5

22.5

24.5

26.5

28.5

30.5


Urban 2004 A Max Urban 2004 B max

Urban 2004 A Ave Urban 2004 B Ave

Urban 2004 A Min Urban 2004 B Min

Tem

pe

ratu

re (

Cel

siu

s)

225

Figure A.12 Urban Stations Minimum Temperature 1900-2007 Trends Distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 10% 11-89% > 90%


% o

f st

atio

ns

226

Figure A.13. Urban Stations Average Temperature 1900-2007 Trends Distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 10% 11-89% > 90%


% o

f st

atio

ns

227

Figure A.14 Urban Stations Maximum Temperature 1900-2007 Trends Distribution

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

< 10% 11-89% > 90%


% o

f st

atio

ns

228

Figure A.15 Station Monthly Minimum Temperature by HELZ

15.0

17.0

19.0

21.0

23.0

25.0

27.0

29.0

31.0

33.0

35.0

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec

Dry Forest

Moist Forest

Wet Forest

Tem

pe

ratu

re (

Cel

siu

s)

229

Figure A.16 Station Monthly Average Temperature by HELZ

15.0

17.0

19.0

21.0

23.0

25.0

27.0

29.0

31.0

33.0

35.0


Dry Forest

Moist Forest

Wet Forest

Tem

per

atu

re (

Cel

siu

s)

230

Figure A.17 Station Monthly Average Temperature by HELZ

15.0

17.0

19.0

21.0

23.0

25.0

27.0

29.0

31.0

33.0

35.0


Dry Forest

Moist Forest

Wet Forest

Tem

per

atu

re (

Cel

siu

s)

231

Figure A.18 Number of Precipitation Stations in Service per year 1900-2007

0

10

20

30

40

50

60

70

80

90

100

1900 1905 1910 1915 1920 1925 1930 1935 1940 1945 1950 1955 1960 1965 1970 1975 1980 1985 1990 1995 2000 2005

Dry Forest Moist Forest Wet Forest

Nu

mb

er

of

Stat

ion

s

232

Figure A.19 Percentage of Stations Registering Usual versus Extreme Yearly

Average Precipitation for 1900-2007 (Precipitation Station Frequency Distribution)

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

Scarse Usual Excess

Puerto Rico (all stations)

Per

cen

tage

of

stat

ion

s

233


Average Precipitation by HELZ (Decadal Frequency Distribution)

0.0%

10.0%

20.0%

30.0%

40.0%

50.0%

60.0%

70.0%

< 10% 11-89% > 90%


Per

cen

tage

of

dec

ade

s

234


Average Precipitation in the Wet Forest by U/NU Land Cover (Decadal Frequency

Distribution)

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 10% 11-89% > 90%

Wet Forest Urban 92B 60m No Urban 92B 60m Urban 92B 90m No Urban 92B 90m

Pe

rcen

tage

of

dec

ades

235


Average Precipitation in the Moist Forest by U/NU Land Cover (Decadal Frequency

Distribution)

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 10% 11-89% > 90%

Moist Forest Urban 92A No Urban 92A Urban 2004A No Urban 2004A

Per

cen

tage

of

dec

ade

s

236

Figure A.23 Percentage of Decades Registering Usual versus Extreme Yearly

Average Precipitation in the Dry Forest by U/NU Land Cover (Decadal Frequency

Distribution)

0%

10%

20%

30%

40%

50%

60%

70%

80%

< 10% 11-89% > 90%

Dry Forest

Urban 92B 60m

NU 92B 60m

Urban 04A

No Urban 04A

Urban 04B 30m

No Urban 04B 30m

Per

cen

tage

of

de

cad

es

237

Figure A.24. 1963-1995 Average Annual Temperature map generated from PRISM

Annual Maximum and Minimum Temperature maps

238

Figure A. 25. Holdridge Ecological Lifezones, Temperature Stations and 2004 High

Density and Low Density Urban Areas

239

Figure A.26. 1979-2005 Anomalies Trends from Selected FILNET data stations map

240

Figure A. 27. 1979-2005 North America Regional Reanalysis Anomalies trends map

241

Figure A.28. 1979-2005 FILNET selected stations observations anomalies minus

North America Regional Reanalysis trends map

242

Figure A. 29. FILNET 1900-2007 Monthly Maximum Temperatures map.

243

Figure A.30. FILNET 1900-2007 Monthly Average Temperatures map

244

Figure A.31. FILNET 1900-2007 Monthly Minimum Temperatures map

245


meter resolution

246


meter resolution

247


meter resolution

248


meter resolution

249


meter resolution.

250


meter resolution

251


252


253


254


255


256


25

7

Appendix B Tables

TABLE B.1. 1992 LULC CENTURY AVERAGE PRECIPITATION TRENDS (YEARLY VERSUS REGION)

Selection A Selection B 30 meters Selection B 60 meters Selection B 90 meters

1992 Urban No Urban Urban No Urban Urban No Urban Urban No Urban Remarks

WF N/A N/A - (100%) - (86%) - (100%) - (83%) - (100%) - (82%) Decreasing precipitation

MF - (85%) - (74%) - (81%) - (74%) N/A N/A N/A N/A Decreasing precipitation

DF 50/50 - (77%) - (75%) - (75%) N/A N/A N/A N/A Urban A behaving differently

Total - (80%) - (77%) - (80%) - (77%) N/A N/A N/A N/A Decreasing precipitation

Values are percentages of the number of stations located in urban or non urban areas respectively.

25

8

Table B.2. 1992 LULC PRISM Period Average Precipitation Trends (1963-1995 versus Region)


1992 Urban No Urban Urban No Urban Urban No Urban Remarks

WF N/A N/A - (100%) - (75%) - (75%) - (77%) Decreasing Precipitation

MF - (77%) - (73%) - (79%) - (72%) N/A N/A Decreasing Precipitation

DF 50/50 + (55%) - (75%) + (60%) N/A N/A Non urban behaving different to others

Total - (73%) - (69%) - (79%) - (67%) N/A N/A Decreasing Precipitation


25

9

Table B.3. 1992 LULC NARR Period Average Precipitation Trends (1979-2005 versus Region)



WF N/A N/A - (100%) - (59%) - (100%) - (53%) Decreasing Precipitation

MF + (57%) - (64%) - (64%) - (59%) N/A N/A Urban A increasing

DF 50/50 - (54%) - (77%) 50/50 N/A N/A Urban A & 30m NU behaving different

Total + (55%) - (62%) - (67%) - (58%) N/A N/A Urban A increasing


26

0




WF N/A N/A N/A N/A N/A N/A No urban stations

MF - (86%) - (75%) - (91%) - (73%) - (80%) - (75%) Decreasing precipitation

DF - (67%) - (76%) - (75%) - (75%) - (80%) - (74%) Decreasing precipitation

Total - (80%) - (77%) - (87%) - (76%) - (80%) - (77%) Decreasing precipitation

Values are percentages of the number of stations located in urban or non urban areas respectively

26

1

Table B.5. 2004 LULC Average Precipitation PRISM Trends (1963-1995 versus Region)




MF 50/50 - (76%) - (70%) - (74%) - (69%) - (75%) Decreasing precipitation

DF - (67%) + (57%) 50/50 + (55%) - (60%) + (58%) Non Urban behaving differently

Total - (56%) - (70%) - (64%) - (70%) - (67%) - (70%) Decreasing Precipitation


26

2

Table B.6. 2004 LULC Average Precipitation OMR Trends (1979-2005 versus Region)




MF - (67%) - (60%) - (83%) - (57%) - (75%) - (58%) Decreasing precipitation

DF + (100%) - (57%) + (100%) - (62%) + (67%) - (58%) Urban increasing

Total 50/50 - (61%) - (63%) - (60%) - (64%) - (60%) Urban A behaving differently


263

Table B.7. Six Hour Average grid cell precipitation in centimeters for each study

region of the island. The region Downwind of San Juan also includes precipitation

over the ocean.

Time Total Island

(cm)

Western 3rd

(cm)

Central 3rd

(cm)

Eastern 3rd

(cm)

Downwind of

San Juan (cm)

18UTC 5/21 0 0 0 0 0

00UTC 5/22 0.087 0.21 0.051 0.000 0.000

06UTC 5/22 0.32 0.58 0.28 0.10 0.084

12UTC 5/22 0.34 0.46 0.38 0.17 0.20

18UTC 5/22 0.46 0.46 0.52 0.38 0.45

00UTC 5/23 0.59 0.49 0.64 0.65 0.62

06UTC 5/23 0.45 0.33 0.40 0.62 0.50

12UTC 5/23 0.50 0.44 0.48 0.57 0.35

18UTC 5/23 0.40 0.32 0.39 0.49 0.36

00UTC 5/24 0.49 0.48 0.46 0.54 0.59

06UTC 5/24 0.58 0.58 0.58 0.56 0.78

12UTC 5/24 0.72 0.79 0.70 0.67 1.00

264

Table B.8. Percentage differences in total precipitation over the modeled period for

each scenario as ratio of the control. Relative changes in precipitation comparing

each scenario to the control by study region.

Scenario Total Island

(%)

Western 3rd

(%)

Central 3rd

(%)

Eastern 3rd

(%)

Downwind of

San Juan (%)

UI1A 0.81 0.90 0.90 0.80 1.03

UI2A 0.75 0.79 0.95 0.70 0.93

UI3A 0.71 0.77 0.67 0.72 0.31

UI4A 0.62 0.76 0.66 0.58 0.64

UI5A 0.75 0.82 0.68 0.79 0.62

RF1 1.35 1.25 2.15 0.96 1.44

RF2 0.83 0.90 1.30 0.69 0.81

RF3 1.14 1.21 1.56 0.90 0.90

RF4 0.77 1.10 1.05 0.65 0.52

RF5 0.60 1.00 0.73 0.47 0.26

RWF1 0.70 1.01 1.10 0.49 0.80

RWF2 0.71 0.76 0.70 0.70 0.61

RWF3 1.20 1.21 1.90 0.80 1.08

RWF4 0.73 0.80 0.76 0.70 0.63

RWF5 1.07 0.78 1.40 0.95 1.24

UI1B 0.70 0.75 0.67 0.70 0.76

UI2B 0.94 1.27 1.30 0.72 0.81

UI3B 0.88 1.26 1.21 0.74 0.96

UI4B 0.82 1.09 0.94 0.75 0.70

UI5B 0.93 0.81 1.00 0.88 0.76

265

VITA

265

Angel Ruben Torres-Valcárcel MPH Ph.D

Education

PhD, Climatology/Environmental Sciences, Natural Resources Sustainability &

Conservation. Purdue University, West Lafayette Indiana 2005-2013 GPA 3.75

MPH, Public Health General Program, 2001, University of Puerto Rico, Medical

Sciences Campus

Graduate studies towards a Master of Planning / Environmental Planning. 1995 – NF,

University of Puerto Rico, Rio Piedras Campus (57 / 60 credit hour completed)

BS, Environmental Sciences, 1995, University of Puerto Rico, Rio Piedras Campus

Other Education & Training

Non profit Organization Management Workshop:

“How to Develop a Volunteer Program” October 23 – November 6, 2009

Non profit Organization Management Workshop:

“How to Apply for Federal Tax Exemption”; “How to Retain Federal tax Exemption

Status”, Puerto Rico Federal Affairs Administration. August 18, 2009

Advance Management Principles, Krannert School of Business Executive Education

Program Purdue University, May 2008

266

Science teachers Meteorological Education Certification Workshop sponsored by

American Meteorological Society (listener) September 2007

Network Plus Certification Training; Network Technician Workshop, 2002

Educational Computer Center, Carolina Puerto Rico

PhD research topics

The Impact of Land Use & Land Cover Changes in Puerto Rico’s Climate (2008-present)

The instrumentation of mangrove’s methane emissions as an ecosystem biomarker (2005-

2007)

.Other research

Volunteer ecological research work to establish a mechanized field sample image

processing system (August - December 2004)

Masters Degree Research Experience:

The Use of a Spatial Criteria in the Management of Ryan White Title I Funds in

San Juan EMA – Final Project GIS Study 2000, Advisor Jose Cobos MD

Undergraduate Research:

Biogeography of Sierra Palm in the Caribbean National Forest – Final Project

GIS study 1995, Advisors: Alberto Sabat PhD, Jose Molinelli Freytes PhD

Other Undergraduate Research Experience:

Sexual Tendencies of Sierra Palm offspring – Ecology Lab, Biology Department

University of Puerto Rico, Rico, Rio Piedras Campus; Research Assistant, June 1994 to

December 1994, PI: Alberto Sabat PhD

Forest Nutrients Dynamics – Advance Experimental Ecology, Biology Department; June

1994 to December 1994, Professor: Ariel Lugo PhD

267

Acknowledgments

July 2005 - 2006 – GPA above 3.50 Dean Graduate School Office of Minority Programs

July 2006 - 2007 - GPA above 3.50 Dean Graduate School Office of Minority Programs

Awarded Fellowships, Scholarships and Grants

Bisland Dissertation Fellowship - January 2010

Kinesis Scholarship - January 2008, 2009, 2010

Henri Silver Graduate Scholarship - May 2008

Fernandez Bjerg Scholarship - January 2007

Purdue Strategic Initiatives Fellowship - January 2007

Purdue Doctoral Fellowship - January 2005

Research Skills & Qualities

Computer literate, GIS expertise, strategic planning, program evaluation design, grant

writing, operational analysis, Environmental and Public Health programs and project

management. Goal oriented, creativity, initiative.

Languages

Fluent in English and Spanish, Reading Writing and Speaking.

Research Interests

Environmental Sciences, Sustainability, Sustainable Management, Biodiversity,

Conservation, Ecology, Public Health, Environmental Planning, Climatology, Renewable

Energy, Bioremediation, Phytoremediation.

Teaching Experience

January 2011 – May 2011

Part Time Professor - Teach undergraduate & graduate courses, evaluate and grade

student’s work and academic performance. Guide and supervise graduate student’s final

268

projects. Biology Department; Environmental Science Program Pontifical Catholic

University of Puerto Rico

Biological Sciences– Undergraduate

Biological Sciences Online Theoretical Course

Environmental Topics – Graduate

Environmental Science Master’s Degree Seminar Coordination

Community Service II – Graduate

Environmental Science Master’s Degree Final Community Project Supervision

Environmental Planning - Graduate

Environmental Science Master’s Degree Environmental Planning Course

August 2010 to December 2010:

Part Time Professor - Teach undergraduate & graduate courses, evaluate and grade

student’s work and academic performance. Guide and supervise graduate student’s final

projects. Biology Department; Environmental Science Program Pontifical Catholic


Environmental Management – Undergraduate

Environmental Science Bachelors Degree Theoretical Conference

Environmental Health – Graduate

Environmental Science Master’s Degree Theoretical Conference

Community Service I – Graduate

Environmental Science Master’s Degree Final Project

269

January 2008 to May 2008:

Teaching Assistant – Grading and academic assistance for two courses. Department of

Earth & Atmospheric Sciences, Purdue University.

Work Experience

May 2011 to present: Program Planner; Soccer Tournament and Competition

Development Program. Designed league’s development plans; reviewed & developed

competition ruling and regulations, designed game calendars, designed on league security

protocols. Academia Nacional de Fútbol - Puerto Rico Soccer Federation.

January 2001 to January 2005: Independent Contractor / Consultant; Program Planning,

Evaluation, Fund Raising, Grant Writing and Information Systems Services – Community

Projects

November 1999 – September 2000: Program Evaluation Consultant – Ryan White Title

IV, Puerto Rico Department of Health

May 1999 to June 2001: Information Center Operator; System Maintenance and Trouble

Shooting, Client Service - Center for Interdisciplinary Studies and Information Systems,

Graduate School of Public Health, Medical Sciences Campus

March 1998 to May 1998: Research Assistant; Grant writing clerical & computer

applications support, Internet Search - Center for Evaluation and Socio Medic Research,

University of Puerto Rico, Medical Sciences Campus.

February 1996- February 1999: Planner / Assistant – GIS projects, Environmental Studies

Planning Projects, Policy Analysis, Planning Studies, Environmental Assessment; G .

Navas & Associates

270

July 1 1994 to July 30 1994: Clerk; Howard Hughes Summer Job Program – Support

Clerical Tasks; Environmental Protection Agency Caribbean Office

September 1989 – February 1996: Office Assistant – Work Study Program – Economic

Assistance Office, University of Puerto Rico, Rio Piedras Campus

Independent Professional Projects

Innovative Strategies Education Program – Proposal for the creation of an Aquatic

Program in a Public School. Written October 2004; Approved April 2005.

Health/Recreational Services Project – Independent Contractor

Planning and Policy Analysis for the Development of a Resort with Recreational Services

for Retired Workers, Private 2002

AIDS Task Force – Volunteer Work; Planning, Evaluation and Grant Writing Assistance;

Ryan White Title I Management Office, Local Government Agency 1997 – present

Ryan White Care Act Title I 2004 narrative, planning.

Ryan White Title I grant writing for FY 2005; Severe Need, 2005 Program Work Plan

Health Services - Ryan White Title IV 1999 Grant - Consultation

Program Progress Report for 1999, Program Work Plan for 2000 Grant, Evaluation Work

Plan for 2000 Grant, Continuous Quality Improvement Initiatives Plan and narrative for

2000 Grant, Information System Protocol, Case Management Guide

Review Assessment 2000, Fathers Advisory Committee Assessment Report 2000

Environmental Assessments – G Navas & Associates Inc.

Housing Project Impact Assessment in Aibonito – Local Government, 1998

271

Fresh Water Well Potential Assessment in Vega Alta – Private, 1998

Flooding Control Alternatives for Vega Alta Municipality - Local Government, 1998

Land Use and Environmental Impact Assessment of BFI Proposed Domestic Land Fill in

Salinas- Local Government, 1998

Corporate Professional Projects

Planning – G. Navas & Associates Inc.

Land Use Policy and Impacts of a Residential Commercial Complex in Vega Alta –

Local Government, 1998

Zoning Maps Updates- GIS Mapping - Local Governments, 1997

Cartographic System for Cement Shipping – GIS Mapping – Private, 1996

Historical Migration Analysis for Mayaguez Municipality - Local Government, 1996

Community Projects

Founded “Corporación para la Sustentabilidad Ambiental” COSUAM de Puerto Rico

(Eng. Environmental Sustainability Corporation) a local private non profit organization

to promote environmental sustainability (August 2007). Elected President - November

2007 – stepped down March 15, 2011.

School Aquatic Program: September 2004 - 2006

Julio Sellés Solá Elementary School; Innovative School Swimming Program; Grant

Writing 2004 & 2005

“Academia Nacional de Fútbol Inc.”; Youth Soccer Development Project: August 2003

– January 2005

Program Coordinator: Program Planning, Fund Raising, Grant Writing, President /

Delegate

Co Vice-president (2003-2004)

First Women’s Open Indoor Soccer Cup - June 2005

272

Puerto Rico Soccer Beach 2005 tournament executive committee- August 2005

Founding member of “Liga Metropolitana de Fútbol” (Metropolitan Youth Soccer

League) -2007

“Centro de Envejecientes Manantial de Amor”; Center for Elderly Services; Program

Consultant, Planning, Fund Raising, Grant Writing 2001

Puerto Rico Lupus Group; Lupus Support Group Program Consultant: Planning, Fund

Raising, Grant Writing 2001

Luis A Señeriz Foundation - MADDPR; Mothers Against Drunk Drivers of Puerto Rico;

Program Consultant: Planning, Fund Raising, Grant Writing 2001

“BioHealth Survey Systems”; Community Health Research Board Secretary; Planning,

Fund Rising, Grant Writing 2001

Metro Emergency Response Team; Fire, Rescue and Medical Emergency Response

Services Program Consultant: Planning, Fund Raising, Grant Writing 2001.

Public Statements & Community Participation

March 21, 2013 – “La Descarga Original” Sports radio talk show participation.

Host: Talk show members Subject: National Soccer League Plans and Tournaments.

December 7, 2012 – Environmental Expert Community Panel. Review and comment on

sixth grade class Environmental Education Project “PRO-CASA”. University of Puerto

Rico Elementary School.

August 5, 2012 – Local Newspaper “El Vocero” Interview

Host: Adriana Vélez Subject: COSUAM’s Sustainable School Program, COSUAM

origins, development, projects and accomplishments

273

June 14, 2012, - Caribbean Landscape Conservation Cooperative Open House

(Participant)

Host: Dr. William Gould / International Institute of Tropical Forestry, USDA Forest

Service

April 16, 2012 – Local Newspaper “El Nuevo Dia” Interview

Host: Iliana Fuentes Lugo Subject: COSUAM’s Sustainable School Program, COSUAM

origins, development, projects and accomplishments

February 9, 2012 – Local station Radio Hit 1250 am radio interview

Host: Marielisa Ortiz Berríos. Subject: Environmental Sustainability, COSUAM’s

development, projects and accomplishments, Sustainable School Program

April 8, 2011 - Newspaper column “La Vía de los Verdes” about the controversial

building of a local gas pipeline”. “El Nuevo Día”.

April 1, 2011 – Local News TV interview about COSUAM and its participation in the

local environmental organization coalition for the “Under Water Press Conference

2011”announcing “Earth Month” activities.

March 14-30, 2011 – University Program Proposal Evaluator for “Consejo de Educación

Superior” (transl. Puerto Rico’s Higher Education Council)

July 17, 2010 - Newspaper column “¿Cual Puerto Rico Verde?” Education on

Environmental Sustainability and critical revision on local “green washing” initiatives”.

“El Nuevo Día”.

June 5, 2010 – San Juan Urban Long Term Research Area (ULTRA) Community Forum

about quality of life and environmental health of the city of San Juan and surrounding

areas.

274

April 13, 2010 – Community focal group evaluation of the Center for Volunteer

Development” about the quality of services provided

April 1, 2010 – News magazine column about COSUAM and its ongoing projects in

Vieques; “Vieques Events”. April 2010 issue

March 30, 2010 – Video Interview -“Te Informa”, Purdue University Hispanic TV

Program.

Host: Isabel Trujillo. Subject: COSUAM’s development, projects and accomplishments

March 26, 2010 - Newspaper column “¿Desarrollo Sostenible?” Contextual revision on

the controversial path of local government initiatives towards environmental

sustainability. “El Nuevo Día”

February 19, 2010 - Radio interview WALO 1240. “A Ciencia Cierta”

Host: Susan Soltero. Subject: COSUAM’s Sustainable Schools Program

December 14, 2009 - Newspaper column “Sustentabilidad Ambiental o Desarrollo

Sostenido” addressing the differences between “Environmental Sustainability” initiatives

vs “ Sustainable Development” initiatives. “Claridad”

December 1, 2009 - Community focus group to validate a questionnaire about the

desirable traits profile of local high school graduates

September 3, 2009 - Radio interview WALO 1240. “A Ciencia Cierta”

Host:Susan Soltero. Subject: The Impacts of Land Use and Land Cover Changes in

Puerto Rico's Climate; Land Surface Impacts on post land-fall storm structure

September 16, 2009 - Radio interview WALO 1240. “A Ciencia Cierta”

Host:Susan Soltero. Subject: Environmental Sustainability; COSUAM

275

May 9, 2009 - Newspaper column “Sustentabilidad en Peligro” on the risks of local

government legislation for Puerto Rico sustainability goals. “El Nuevo Día”

December 18 2008; Newspaper column “Ojo a las Canalizaciones” on the

environmental and public safety concerns of channelizing rivers and streams, “El Nuevo

Día”

October 1, 2008; Newspaper column “El COPUR y las Medallas Olimpicas: La Crisis

Social del Deporte en Puerto Rico” addressing the social importance of sports in Puerto

Rico Claridad

September 4, 2008; Newspaper column “Medallas en Contexto” addressing Puerto Rico’s

Olympic sports issues, “El Nuevo Día”

January 7, 2008; Radio Interview at local radio station as public education activity on

how to reduce individual ecological footprint, Magic 97.3 FM

December 22, 2007 - Newspaper Column “ Ademas de Paseo Caribe” on the problem of

commercialization of residential areas “El Nuevo Día” page 89.

November 2007 – Submitted public declaration with recommendations on the problem of

commercialization of residential areas.

May - July 2007 Local Radio Interviews / public education on global warming/climate

change WKAQ 580, “La Hora Magica” & Radio Isla 1320 & “Si No Lo Digo Reviento”

July 2005; Public Statement before Villa Nevarez Lion’s Club members in behalf of an

aquatic program for a local Elementary Public School.

276

March 2003; Pronouncement before Puerto Rico’s Permits and Regulations

Administration. Represented the Community in opposition to the demolition of empty

residential structures for the development of commercial parking areas.

April 2003; Pronouncement before Puerto Rico’s Planning Board – Represented the

community in opposition to the expansion of commercial activities into the residential

area through changes of zoning districts.

November 2003; Pronouncement before Puerto Rico’s Planning Board - Represented the

community in opposition to the expansion of commercial activities into the residential

area through changes of zoning districts.

Attended Conferences and Workshops

Online Science Seminar: "Spatial and temporal analysis of land cover and landscape

structure change in Zagros forests, Western Iran” by Dr. Henareh. November 28, 2012.

International Institute of Tropical Forestry.

San Juan ULTRA Seminar. “The Study of the Social-Ecology, Resilience,

and Sustainability of Cities” by Dr. Charles Redman from the School of Sustainability,

Arizona State University. January 18, 2011 at Center for Puerto Rico, Sila M. Calderon

Foundation.

Online Course Design Workshop. October 1 & October 15, 2010. Pontifical Catholic


Workshop of Theory & Design of Long Term Research in Socio-Ecological Systems,

December 16-18, 2008. University of Puerto Rico

"Legacies of the Rain Forest Project and the Future of Environmental Sciences in Puerto

Rico" Symposium. November 2007. University of Puerto Rico

277

Invited Guest to UPR Law School Maritime Rights Oral Exam (Court simulation on sea

pollution liability case), May 2007, University of Puerto Rico

GIS symposium: “The Use of Geographic Information Systems in Applied Ecology and

Conservation”, May 28 2005, University of Puerto Rico.

Research Seminar: “How to get a Published Article” by Dr. Ruth E. Zambrana: June 27,

2005, University of Puerto Rico

Offered Conferences, Presentations, Workshops & Talks

Activity: “US Green Building Council SEEDS Program for schools”

Date: June 19, 2013

Location: “Administración de Asuntos Energéticos”, San Juan, Puerto Rico

Topic: Composting Principles

Audience: School teachers and administrators

Activity: Summer Camp Workshop talk

Date: June 12, 2013

Location: “Santo Tomás de Aquino ”, Bayamón, Puerto Rico

Topic: Environmental Sustainability

Audience: Summer camp youngsters

Activity: Summer Camp Workshop talk

Date: June 7, 2013

Location: “Luis Muñoz Marín Foundation”, Trujillo Alto, Puerto Rico


Audience: Summer Camp youngsters

Activity: Sustainable Communities Program introductory talk

Date: May 24, 2013

Location: “Residencial Público La Montaña”, Aguadilla, Puerto Rico


Audience: Public housing kids, youngsters, parents and community leaders

Activity: Sustainable Schools Program Introductory talk

Date: May 17, 2013

Location: “Juan A. Sánchez Dávila Elementary School”, Manatí, Puerto Rico

Topic: Sustainable Schools Program

Audience: Teachers, students parents, and community leaders

278

Activity: Sustainable Schools Program (School Environmental Fair) Composting talk

Date: May 10, 2013

Location: “Jardines del Paraiso Elementary School”, Rio Piedras, Puerto Rico

Topic: Compost basics / How to build a compost

Audience: 2nd

grade students

Activity: Sustainable Schools Program (School Environmental Fair) Composting

workshop

Date: May 8, 2013

Location: “Francisco Felicie Elementary/Middle School”, Vega Alta, Puerto Rico

Topic: Compost basics / How to build a compost workshop

Audience: 5th

grade students and teachers

Activity: Sustainable Schools Program Composting workshop

Date: May 7, 2013

Location: “Juan José Osuna High School”, San Juan, Puerto Rico


Audience: High students from 9th

to 12th

and teachers

Activity: Sustainable Schools Program talk

Date: May 1, 2013

Location: “Luis Llorens Torres High School”, Juana Díaz, Puerto Rico


Audience: 11th

and 12th

students and teachers

Activity: Community Composting Project workshop

Date: April 23, 2013

Location: “Colegio Nuestra Señora Del Carmen”, Trujillo Alto, Puerto Rico


Audience: 8th

& 9th

grade students and teachers

Activity: Community Composting Project demonstrative talk


Location: “Trujillo Alto’s Major office”, Trujillo Alto, Puerto Rico

Topic: Compost & Composting

Audience: Municipal government employees

Activity: Sustainable School Program introductory talk


Location: “Juan José Osuna” High School, San Juan, Puerto Rico

Topic: Sustainable Schools Program principles and benefits

Audience: Middle and high school students and teachers

Activity: “US Green Building Council SEEDS Program for Schools”

Date: March 23, 2013

279

Location: “Interamerican University, Arecibo” , Puerto Rico

Topic: School & Community Works, COSUAM’s Sustainable Schools Program

Audience: School officials and teachers

Activity: Green Week presentation talk


Location: “Four Points Hotel”, Caguas , Puerto Rico

Topic: Compost & Composting

Audience: Hotel administrators and employees

Activity: Green Week presentation talk


Location: “Four Points Hotel”, Caguas , Puerto Rico


Audience: Hotel administrators and employees

Activity: COSUAM’s Community Compost Project presentation talk


Location: “El Comandante”, San Juan” , Puerto Rico

Topic: Composting

Audience: Adults, community leaders and neighborhood members



Location: “Venus Gardens Middle School, San Juan” , Puerto Rico

Topic: Composting

Audience: Ninth grade students



Location: “Antonio S Pedreira Elementary School”, Trujillo Alto , Puerto Rico

Topic: Composting

Audience: Third grade students


Date: March 6, 2013

Location: “Centro Niños en Acción, Trujillo Alto” , Puerto Rico

Topic: Composting

Audience: Pre School students (3 – 4 years old)

Activity: COSUAM’s Sustainable Schools Program presentation talk

Date: January 31, 2012

Location: “Dra. Conchita Cuevas High School, Gurabo” , Puerto Rico


Audience: High School students (Seniors, Juniors & Sophomores)

280


Date: December13, 2012

Location: “University of Puerto Rico Elementary School, San Juan ” , Puerto Rico


Audience: Sixth grade students


Date: November 2, 2012

Location: “Arturo Grant Pardo Middle School, Lajas ” , Puerto Rico


Audience: Seventh grade students


Date: October 26, 2012

Location: “Centro y Colegio Valenciana, Añasco ” , Puerto Rico

Topic: Energy, energy conservation and wind energy harnessing

Audience: First & Second grade students



Location: “Liceo Aguadillano ” Aguadilla, Puerto Rico





Location: “Dr. Conchita Cuevas High School, Gurabo Puerto Rico ”

Topic: Environmental Sustainability and Organic Gardening

Audience: High School students



Location: “Agricultural Extension Service, Gurabo Puerto Rico ”

Topic: COSUAM’s Sustainable Schools Program

Audience: Government officials, community organizations

Activity: “COSUAM’s Sustainable Schools Program Talk”

Date: September 12, 2012

Location: “Saint Francis School” – Carolina, Puerto Rico


Audience: High school senior students


Date: August 30, 2012

281

Location: “Puerto Rico College of Architects ”




Date: May 3, 2012




Activity: COSUAM’s Sustainable Schools Program Talk”


Location: “University of Puerto Rico Secondary School – San Juan Puerto Rico




Date: February 3, 2012

Location: “Laura Mercado” Middle & High School – San Germán Puerto Rico


Audience: School teachers



Location: “Padre Pablo Gutierrez Elementary School” – Aguada Puerto Rico

Topic: Environmental Sustainability for teachers

Audience: School teachers


Date: May 7, 2011




Activity: “School Composting Conference”


Location: “German Rieckehoff High School” – Vieques, Puerto Rico

Topic: Composting Theory for students


Activity: “School Composting Workshop”


Location: “German Rieckehoff High School” – Vieques, Puerto Rico

Topic: Composting Theory and Practice for students


282



Location: “20 de septiembre de 1988 Middle School” – Vieques, Puerto Rico


Audience: Middle School students & teachers



Location: Epifanio Estrada Middle School – Aguada, Puerto Rico


Audience: Middle School students

Activity: Environmental Science Seminar


Location: Pontifical Catholic University of Puerto Rico – Ponce, Puerto Rico

Topic: The Impacts of Land Use/Land Cover Changes in Puerto Rico’s Climate

Audience: Science Graduate Students

Activity: “Recyclable Containers Conversion Workshop”


Location: Vieques, Puerto Rico

Topic: Conversion of used containers into recycling containers

Audience: School students, teachers, community organizations & local government

officials

Activity: Energy Conservation Week


Location: Loíza County, Puerto Rico

Topic: “Energy Conservation”

Audience: Local government employees

Activity: Liberal Arts Week


Location: National University College, Arecibo Puerto Rico

Topic: “Sustainable Development or Environmental Sustainability” for students

Audience: College students

Activity: NYU-FRN Winter 2009-Carbon and Climate - Local Research Symposium


Location: University of Puerto Rico, Rio Piedras Campus

Topic: The Impacts of Land-Use and Land-Cover Changes in the Climate of Puerto Rico

Audience: Scientists, university professors and students

283

Scientific Publications

Rochon et al., (2010). “Real-Time Remote Sensing in Support of Ecosystem Services &

Sustainability” in Peter Liotta, et al., eds. Ecosystems Services & Environmental

Security. Thousand Oaks, CA: Sage Publications, (volume 69, 2010).

Rochon, G. L., Quansah, J. E., Fall, S., Araya, B., Biehl, L. L., Thiam, T., ... &

Maringanti, C. (2010). Remote Sensing, Public Health & Disaster Mitigation. In

Geospatial Technologies in Environmental Management (pp. 187-209). Springer

Netherlands.

Rochon, G. L., Niyogi, D., Fall, S., Quansah, J. E., Biehl, L., Araya, B., ... & Thiam, T.

(2010). Best management practices for corporate, academic and governmental transfer of

sustainable technologies to developing countries. Clean Technologies and Environmental

Policy, 12(1), 19-30.

i the impacts of land use / land cover … · the impacts of land use / land cover changes on the...

Documents