Delineation of Groundwater Dependent Ecosystems based on NDVI Variations in the Dry Season

Keisha Baxter1, Isabel Perez2, Dr. Nir Krakauer2 1 LaGuardia Community College, 2The City College of New York

INTRODUCTION!Groundwater is a vital source of water throughout many parts of the world making up

98% of the Earth's available freshwater [1]. It can be extracted and used by

communities, plants, mining, and aquatic life. In fact, there are entire ecosystems that

depend on groundwater for their survival these include wetlands, swamps, springs,

coastal mangroves, and some terrestrial vegetation (phreatophytes) [2].

!Increasing use of groundwater by humans has lessened the amount of water available

to the ecosystems which depend on it for their survival. Due to this groundwater

usage imbalance, measures are being established to protect these groundwater

dependent ecosystems (GDEs) by identifying and mapping their location throughout

the study area.

!These maps will be useful to further serve as a guide for water resources allocators

and managers so they are cautious of the areas they choose to develop and extract

groundwater from, to not cause disturbance to an area which strongly depends on

groundwater. This is analyzed from Normalized

Vegetation Difference Index (NDVI), which shows the greenness of vegetation over an

area.

ABSTRACT!The purpose of this research is to identify and map areas in south central Texas that

shows high potential for Groundwater Dependent Ecosystems (GDEs). A GDE is an

biological life that may be accessing groundwater i.e. water below the earth’s surface

that is stored in an aquifer.

!To identify these GDEs, two types of analyses were conducted.

1) Seasonal Analysis

2) Inter-Annual Analysis

!The analyses yielded two NDVI images each which were subtracted. The difference

NDVI pixels determined whether the area could be classified GDE or non GDE. K-

means unsupervised classification was used to define the thresholds for the groups.

The inter-annual analysis showed that there is a great deal of locations that have

potential to be accessing groundwater in the area.

SITE STUDY!

The study area is located in south-central Texas, about 30 miles from San Antonio, on

the Edwards Plateau, at the geographical coordinate 29.85˚N and 98.48˚W. With

those coordinates at the center, the range of study extended over a 1.2 km2 area that

includes Honey Creek State Natural Area, Guadalupe River State Park, and Beek

Spring. The Plateau extends west to the Pecos River, about 250 miles west of Austin.

METHODOLOGY

Inter Annual Analysis

• The images for this analysis were

chosen using Palmer Drought

Severity Index (PDSI) which

highlighted the wettest and driest

year.

!• These years were chosen and

compared for analysis; a difference

NDVI image was obtained and

analyzed for GDEs

Seasonal Analysis

• To select the dates for analysis,

precipitation (PPT) based approach was

used to identify the years of the highest

and lowest PPT value within a dry season

from 2000 to 2012.

!• The consecutive months with the largest

difference were chosen to be compared

individually as well as the resulting

difference NDVI images of the two.

K-Means Clustering was used to classify difference NDVI as GDE, non GDE and partial GDEs. These clusters are identified in Figures labeled K-means

Clustering.

CONCLUSIONS AND FUTURE WORK !This research has shown that there is a great deal of locations that have

potential to be accessing groundwater. Extreme case testing, which was

the method that results were based upon, also showed that within the 1.2

km2 location that consists of Honey Creek State National Park has high

results of GDEs. Ground based research has proven that the area has

GDEs and remote sensing has provided the same results, showing the

efficiency of the tool and satellites. This research will be furthered with

analysis using an average of all the years in the 2000 to 2012 time series

and creating a map of location of GDEs, and analysis of the partial GDE

location will identify why they have a potential to be accessing

groundwater. Also research will be continued to find GDE in other

location throughout the USA.

ACKNOWLEDGEMENTS!1.This project was made possible by the Research Experiences for

Undergraduates (REU) in Satellite and Ground-Based Remote Sensing at

CREST_2 program.

!2.The National Oceanic and Atmospheric Administration – Cooperative

Remote Sensing Science and Technology Center (NOAA-CREST) for

supporting this project.

!3.This research is supported by the National Science Foundation's

Research Experiences for Undergraduates (NSF REU) under Dr. Reginald

Blake, Dr. Janet Liou-Mark, Ms. Laura Yuen-Lau.

!4.My mentors Dr. Nir Krakauer and Isabel Perez for their patience,

guidance and supportive mentorship throughout this research.

REFERENCES![1] Commission, N. W. (2012). Groundwater Essentials (pp. 7–37).

Canberra ACT: National Water Commission.

![2] Fowler, N. (2005). An Introduction to the Vegetation and Ecology of

the Eastern Edwards Plateau (Hill Country) of Texas. Retrieved from http://

www.sbs.utexas.edu/fowler/generalinfo/epveg/epgeol.htm

![3]Cure, S. (2014). Texas Plant Li fe. Retr ieved from http://

www.texasalmanac.com

.

NDVI COMPARISON -SEASONAL ANALYSIS

INTER-ANNUAL COMPARISON OF NDVI

DIFFERENCE NDVI

FIGURES AND GRAPHS

K-MEANS CLUSTERING OF INTER-ANNUAL ANALYSIS

DIFFERENCE NDVI

DISCUSSION!The inter-annual analysis provides a map of pixels referring to locations, of

the NDVI value for each of the 400 pixels based on the years that had the

highest and lowest Palmer Drought Severity Index (PDSI). K-means

clustering separated the 400 NDVI difference values into 2 distinct groups

labeled high potential for Groundwater Dependence and low potential of

Groundwater Dependence. Here K-means was used as an unsupervised

classification method that statistically defines the threshold for each pixel

as low and high potential to be GDEs.

VEGETATION AND TERRAIN - TEXAS

Precipitation of Study Area in Texas from 2000 to 2012

Prec

ipita

tion

Valu

e (m

m)

0

100

200

300

400

500

TIME (Month- Year)

Apr-00

Aug-00

Apr-01

Aug-01

Apr-02

Aug-02

Apr-03

Aug-03

Apr-04

Aug-04

Apr-05

Aug-05

Apr-06

Aug-06

Apr-07

Aug-07

Apr-08

Aug-08

Apr-09

Aug-09

Apr-10

Aug-10

Apr-11

Aug-11

Apr-12

Aug-12

LOW

Figure -Hydrological Cycle and The effect of Root Depth on Groundwater

Groundwater areas mapped against map of the study area

HIGH

RESULTS

The seasonal analysis was show to be inappropriate for this study due to soil moisture

being present from the previous month. Inter-annual analysis showed that the areas in mid

to lower part of the study area are have high potential to be accessing groundwater.