Robbing Peter to Pay Paul: Tradeoffs Between Ecosystem Carbon Sequestration and Water Yield

Publication in the Proceeding of the Environmental Water Resources Institute Meeting

August 23-27, 2010. Madison, WI

Steven G. McNulty1, Ge Sun2, Jennifer A. Moore Myers3, Erika C. Cohen4, Peter Caldwell5

1-4 Eastern Forest Environmental Threats Assessment Center, USDA Forest Service 920 Main Campus Dr. Venture II, Suite 300 Raleigh, NC 27606, FAX 919 513-2978 1 email steve_mcnulty@ncsu.edu; PH (919) 515-9489 2 email ge_sun@ncsu.edu; PH (919) 515-9498 3 email eccohen@ncsu.edu; PH (919) 513-3189 4 email jennifer_mooremyers@ncsu.edu; PH (919) 515-9497 Abstract The United States National Forest System supplies much of the nation’s drinking water. However, changes in climate, land use and population are stressing the ability of these forests to provide that ecosystem service. Federal land managers are under increasing pressure to increase ecosystem carbon sequestration in an attempt to partially offset greenhouse gases and slow global warming. Unfortunately, the positive relationship between carbon gain and water use in forests, puts the need for water and increased carbon gain at odds with each other. To assess these tradeoffs, a coupled water supply and demand, carbon sequestration, and biodiversity (WaSSI-CB) model was developed. WaSSI-CB was designed to be run with climate, population, and land use change scenarios to examine the interactions between water, carbon gain and biodiversity change across the 2,100 USGS 8 digit USGS Hydrologic Unit Code watersheds that span the lower 48 US. Results from this model using historic climate and landuse data indicated that the greatest increases in water use conservation may be made through improved irrigation practices, that manipulations in forest cover (i.e., massive harvesting) are an impractical way of increasing water supply, and the that the southeastern US has the highest potential for forest carbon sequestration. Biodiversity was calculated under steady state, historic conditions, with the greatest and mammal biodiversity occurring the southern US. The impact of future climate and population change were not included in this paper due to space limitations, but will be presented at the conference. Introduction Water shortages are not a new phenomena, especially in the western US. Competition for water by agricultural, commercial, industrial, and residential use often places water at a premium across the region. Recently, the often considered “water rich” eastern US has also experienced water shortages as the population, particularly in the southeastern US, has continued to increase. In addition to population growth, climate change is also having an impact on water availability across the US. Much of the interest in climate change justifiably focuses on

103Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

long-term temperature increases. However, in addition to rising air and water temperatures, shifts in the distribution, timing, and rate of precipitation are also expected (IPCC, 2008). Depending on the climate change scenario, some parts of the country will see increasing precipitation while other areas experience reductions. Even if average annual precipitation does not change, the distribution of precipitation within a year may shift, with proportionately more or less rain occurring during the growing season. Climate change may also indirectly impact water availability. Carbon dioxide (CO2) is a major contributor to global warming. As a “greenhouse” gas, CO2 has been steadily increasing in the atmosphere since the start of the industrial revolution in the 1870’s (IPCC, 2008). As the nations of the world explore technological measures to reduce greenhouse gas emissions, they are also exploring non-technological possibilities for remove CO2 from the atmosphere. All plants (including trees) sequester CO2 into tissue (i.e., leaves, stems, branches, roots). The US Forest Service is examining the potential role of the forests for increasing CO2 uptake. If forest CO2 uptake could be increased, then the rate of atmospheric CO2 concentration may decrease, and global warming might be slowed. Most researchers would agree that slowing climate change would be desirable. However, reducing climate change through increased forest growth, and associated increased CO2 uptake will not occur without potentially negative unintended consequences. The general relationship between forest water use and productivity has known for decades (Rosenzweig et al., 1967). This relationship was future developed by Law et al. (2002) using eddy flux data. All other factors being equal, as tree net primary productivity increase, tree water use (i.e., evapotranspiration (ET)) also increases. Forests and grasslands cover over much of the US, and ET is equal to approximately 70% of total annual precipitation, with a higher percentage in the western US and a lower percentage in the eastern US. Therefore, any attempt to increase forest and grassland carbon sequestration will result in reduced water yield (Jackson et al., 2005). If water is already a limited resource in some areas, then land managers should question if efforts to increase carbon sequestration are in the best option for that ecosystem, and if by doing so, they are robbing Peter (the water resource) to pay Paul (carbon sequestration goals). This paper uses the relationships between carbon and water to examine tradeoffs between forest water yield and carbon gain using historic climate data across the lower 48 US. The relationship between ecosystem carbon gain, water use and biodiversity is also presented.

Methods Previously we developed a water supply stress index (WaSSI) model (McNulty et al., 2007; Sun et al., 2008a, 2008b). It was designed to predict anthropogenic and environmental factors impacting anthropogenic water stress). The necessary databases are compiled and reformatted to fit into the U.S. Geologic Survey (USGS) 8-digit Hydrologic Unit Code (HUC) watershed as the working scale. There are approximately 2100 8-digit HUC watersheds in the lower 48 U.S. The database includes historic water use and return flow rates by water use sectors, groundwater withdrawal, historic and projected climate, population, and land use. This paper expands on the basic WaSSI model to include ecosystem level relationships with carbon storage (C) and biodiversity

104Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

(B) change in WaSSI-CB. WaSSI-CB is water-centric in that all of the databases used to calculate anthropogenic water stress, are also the only databases needed to calculate both forest carbon sequestration and biodiversity. Therefore, we will first present the databases needed to run WaSSI-CB, and then examine the algorithms that apply to those

databases to predict carbon and water relationships across the US. The relationships and equations between ecosystem carbon gain, water use and yield and ecosystem biodiversity are within WaSSI-CB are presented in Figure 1. Figure 1. Ecological relationship relationships, primary data bases, equations and correlations coefficients used for WASSI-CB. Databases Historic Water Withdrawals and Use The 1995 and 2000 national anthropogenic water use survey datasets published by the USGS were used to determine historic US water demand. Overall, the two survey periods recorded similar water use (Solley et al., 1998; Hutson et al., 2004; Roy et al., 2005). Therefore, we used 1995 datasets as our baseline. The USGS water survey grouped water users into one of seven categories: Commercial, Domestic, Industrial, Irrigation, Livestock, Mining, and Thermoelectric. At the national scale, these sectors represent 3%,

105Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

7%, 8%, 41%, 1%, 1%, and 39% of the total use, respectively. In the eastern U.S. (east of the Mississippi River), thermoelectric water withdrawal dominates (74% of all use), followed by irrigation centered in the Mississippi valley and western Texas regions. However, because the return flow rates from power plants are high (> 90%), irrigation is the largest sector in terms of consumptive (i.e., water that is withdrawn but not returned to the ecosystem) water use (74% of total) followed by thermoelectric use (17%). Over half of the water withdrawal is derived from groundwater in the Mississippi valley, western Texas, and coastal regions. In the western U.S. (west of the Mississippi River), over 80% water withdrawal is used for irrigation, and about 14% is for thermoelectric use. Climate Data Historic monthly climate data (i.e., precipitation and air temperature) compiled by the VEMAP group (Kittel et al., 1997) from 1972-1993 were used as the baseline. The climate data were in a gridded 0.5o by 0.5o (about 50 km by 75 km) format for the continental U.S. Then, the gridded climate datasets were overlaid on 8-digit watersheds.

Population Data The U.S. Census Bureau records indicate that population increased about 30% between 1980 and 2000. Population projections at the census block level were aggregated to the 8-digit HUC watershed scale for each year from 1972-1993 (U.S. Census, 2002). Use Change The 1992 National Land Cover Dataset (NLCD) was used for the land use and land cover baseline (http://edc.usgs.gov/glis/hyper/guide/mrlc) at a 1km2 spatial resolution. All land use and land cover classes were aggregated into five major categories according to their hydrologic properties. These include forests (i.e., conifers or hardwoods), croplands, urban/residential, and water bodies. Land use is a major driver for the hydrologic model estimates of watershed scale evapotranspiration (as described in the next section). Land use change data provided by Plantinga et al. (2007) does not include any federal lands and excludes some landcover types (e.g., water bodies, roads, barren lands), so those land use and land cover types were derived from the 1992 MRLC data, aggregated to the 8-digit HUC watershed scale, and then added to the Plantinga land use dataset. Model Equations The WaSSI-CB model can calculate anthropogenic water stress, forest carbon sequestration and biodiversity under changing population, landuse and climate (Figure 1). All equations used by WaSSI-CB are presented in Figure 1, but due to space limitations, this paper will only focus on the water and carbon sequestration components of the WaSSI-CB model, but biodiversity interactions will be presented at the meeting. Water Supply, Demand, and Stress Index Monthly water supply was defined as the total potential water available for withdraw from a basin, expressed by the following formula:

WS = SS + GS + ∑RFi

106Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Where, WS = Total water supply volume (m3) for each 8-digit HUC watershed. SS = Total surface water supply for each 8-digit HUC watershed. SS is estimated at the monthly time scale using a water balance model (Zhou et al., 2008). This hydrological model predicts water yield at the HUC level as a function of monthly potential evapotranspiration, land use type, canopy interception capacity, plant rooting depth, soil moisture content, and precipitation received. GS = Total groundwater supply as represented by USGS annual historic (1995) groundwater withdrawal records (Solley et al., 1998). RF = Return flow from each of seven water users i including commercial, domestic, industrial, irrigation, livestock, mining, and thermoelectric sectors. RF is calculated as the USGS historic (i.e., 1995) (Solley et al., 1998) return flow rate (RFR) multiplied by the water use (WU). Return flow rates vary among watersheds and water use sectors. Water demand (WD) represents the sum of all water use (WU) by each of the seven sectors (i), plus public (PB) use and losses representing water transfer between basins and the difference between water withdrawn and delivered by public suppliers (Solley et al., 1998): WD = ∑WUi + ∑PBi i= 1-7

Given these relationships, we proposed a new hydrologic term; the Water Supply Stress Index (WaSSI) (Equation 1). The term WaSSI is used to quantitatively assess the relative magnitude of water supply and demand at the 8-digit HUC watershed level.

x

xx WS

WDWaSSI = (1)

Where x represents either historic or future water supply and demand from environmental and anthropogenic sectors. Once annual values of water use are determined, a monthly function is applied to redistribute annual water use to each month across the 18 water resource regions. Currently, such monthly redistribution schemes are only applied to the irrigation and domestic sectors. We used state-wide water use data to derive the monthly water use functions (Sun et al. 2008b). Forest Carbon Sequestration Rosenzweig (1967) established the relationship between forest water use and carbon gain. Although soil fertility, soil water holding capacity, tree species type and seasonal weather patterns can all impact forest growth, generally, as precipitation increases, so does forest carbon sequestration. The advent of the eddy flux measurements has significantly increased the ability to accurately measure both forest carbon gain and water use during the 1990’s. Law et al (2002) assembled the carbon and water flux data from dozens of forests and grasslands across the globe to develop universal evapotranspiration to carbon sequestration relationships for forests, grasslands and agricultural areas. Law et al. (2002) related ET to various measures of productivity (e.g. gross ecosystem productivity, and net primary productivity). We used the Sun et al. (2008b) estimated ET for the 8-digit HUC watersheds across the conterminous US as the basis for applying the Law et al. (2002) relationships between ecosystem water use and carbon gain.

107Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

First, monthly ecosystem ET was calculated for each watershed from 1972-1994 by multiplying the ET rate for each ecosystem within a watershed by the total area of that ecosystem within each watershed. Similarly, the calculated rate of ET for each ecosystem within each watershed was used as input for Law et al. (2002) calculations of net ecosystem Exchange (NEE). Individual estimates of NEE for each ecosystem were weighted by the total area of each ecosystem within the watershed to estimate total watershed NEE. Biodiversity A similar weighting process was used convert predicted ET and potential evapotranspiration (PET) into measures of biodiversity using equations Currie et al. (1991). Combinations of ET and PET could be used to predict tree, mammal, reptile, amphibian, bird, and vertebrate diversity across the conterminous US (Figure 1). Other measures of biodiversity use estimates of net primary productivity (which can also be calculated from WaSSI-CB) but were not included in this paper due to space limitations.

Results Water Stress and Yield Water availability at the monthly and annual scale is controlled by precipitation and actual evapotranspiration (ET). Evapotranspiration is largely constrained by potential evapotranspiration, precipitation, and vegetation water use efficiency (Sun et al., 2005). The combined spatial and temporal patterns of precipitation, ET, and topography resulted in a complex pattern of water availability at the continental scale (Figure 2).

Figure 2. WaSSI-CB estimated water stress from 1974-1993. Increasing values signify increasing anthropogenic water stress. Values equal or greater than 1 mean that all available water resources are being used for that watershed.

Water Demand / Water Supply (WaSSI) (1974-1993)

WaSSI0 - 0.010.01 - 0.050.05 - 0.10.1 - 0.20.2 - 0.40.4 - 0.60.6 - 0.80.8 - 1>1.0

States

108Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

WaSSI, unlike the Palmer Drought Severity Index or other ecosystem based measures of water stress, factors in anthropogenic water demand. For this reason, even areas with high annual levels of precipitation can have high WaSSI values. For example northern Georgia and northern Virginia are not normally drought stressed, but both areas have watersheds with high WaSSI values due high water demand in and around Altanta (Georgia) and Washington (DC) (Figure 2). Areas that received lesser amounts of precipitation, have higher rates of ET, and/or high population or irrigation demands had the highest WaSSI values. High WaSSI values in the Great Plains were irrigation driven while most of California WaSSI values were largely a function of high population density (and to a lesser extent irrigation. The highest WaSSI values occurred in Arizona where a combination of increasing population, and low precipitation resulted in major imbalance between water supply and demand, (and the country’s highest predicted WaSSI values). Ecosystem Carbon Sequestration WaSSI-CB predicted that most of the western US (with the exception of the Pacific northwest coastal zones) was a slight carbon source during the 20 year study period (Figure 3).

Figure 3. WaSSI-CB predicted net ecosystem exchange (NEE) for the 20 year period between 1972-1993. High temperatures and limited precipitation were not conducive to carbon gain. The grasslands of the prairie states were predicted to be slight carbon sinks, while the wetter midwest grasslands were predicted to be moderate sinks of carbon. By far, the largest carbon sinks occurred in the southeastern US, were long growing seasons, warm temperatures and abundant precipitation maximized forest productivity.

109Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Forest Removal Impacts Scenario on Water Yield A national comparison water stress and carbon gain revealed that the western US is largely unsuitable for facilitating significant increases in ecosystem carbon sequestration. Water is already limiting and further attempts to increase forest growth (e.g. increase stand stocking, fertilization) would result in additional stress to the water resource. Conversely, when a 20% reduction in forest cover was applied to the conterminous US, WaSSI-CB predicted moderate increases in water yield across the region (Figure 4).

Figure 4. WaSSI-CB predicted change in water yield resulting from a 20% reduction in forest cover. While it is not practical to maintain a 20% reduction in forest cover (i.e. leaf area) due to vegetative re-growth, the model does illustrate the maximum potential impact that such a practice could have on a national scale. The maximum increase in water supply could be calculated by multiplying the increase in water yield (mm) by the total area each watershed. This value assumes that all available water could be harvested from within a watershed. Given that is not possible, realized gains in water yield would be proportionally less. Biodiversity Ecosystem biodiversity can be defined both as a function of ecosystem productivity (Constanza et al. 2006) or by water (or potential) water use (Currie, 1991). Both approaches are useful, but due to space limitations we only present the relationships between ecosystem water use and biodiversity (i.e. Currie approach). Both ET and PET are related to biodiversity, with PET being a measure of ecosystem water use if water were not limiting, and ET being a measure of actual water use in the presence or available water. For this reason, ecosystems with high air temperatures (i.e., tropic and deserts) have similar PET, but very different ET. If component is very dependent on ample water supplies for existence (e.g. tree biodiversity), then ET will be a

110Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

better measure of biodiversity (Figure 1). However, if abundant water is not a major factor determining species suitability to an ecosystem (e.g. reptiles), than PET may be a better indicator or biodiversity (Figure 1). For this reason, tree biodiversity is highest across the southern US, and lowest in the dry desert southwestern US.

Figure 5. WaSSI-CB predicted biodiversity across the lower 48 US under steady state conditions during the period from 1972-1993. Conversely, reptiles are prefer warm, dry climates. Therefore, it should not be surprising that WaSSI-CB would predict that southern California, Arizona, Texas and Florida would host the highest levels of reptile biodiversity (Figure 6). All of these estimates of biodiversity assume that the ecosystem is given sufficient time to maximize the potential habitat needed to achieve these levels of species diversity. The complex niche structure often require hundreds of years to become established (e.g. old growth habitats). Climate change is occurring at a much faster rate so it is not likely that these levels of biodiversity will be achievable is ecosystem shift occur in the decades to come.

111Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Figure 6. WaSSI-CB predicted reptile biodiversity under historic 1972-1993 climate. Conclusions This work represents the first step towards examining water supply and demand tradeoffs with the desire to increase ecosystem carbon sequestration at a watershed scale across the conterminous U.S. Relationships between carbon gain and ecosystem water use, with biodiversity are also presented. Historically, the eastern US has the both the highest potential for increasing carbon sequestration and the greatest demand for water from both natural and anthropogenic sectors. It is important to consider these tradeoffs between natural resource sectors if unintended consequences (e.g. water shortages in areas of increase carbon sequestration) are to be avoided. Several areas need improvement and should be considered for future studies in modeling water stresses at large extents including; 1) flow routing and basin water transfer; 2) limitations on water withdrawal due to ecosystem needs. Large amounts of water are needed to satisfy environmental flows, thus greatly limiting water available for human use; 3) the impact of both climate change and land use change and urbanization on water quality and how reduced water quality can impact potable water availability; and 4) validation of water yield, carbon sequestration, and changing biodiversity predictions at the watershed scale across the conterminous US. The Intergovernmental Panel on Climate Change (IPCC) warns that freshwater resources are vulnerable and are to be strongly impacted by climate with a wide range of consequences on human societies and ecosystems. Research gaps exist in understanding and modeling the impacts of climate change at the scales relevant to decision making (IPCC, 2008). The model developed by this study can be used as a framework to examine implications of climate change mitigation and adaptation measures

112Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

References Costanza, R, B.Fisher, K. Mulder, S. Liu, T. Christopher. 2006. Biodiversity and

ecosystem services: A multi-scale empirical study of the relationship between species richness and net primary production. Ecological Economics 61:478-491.

Currie, D.J. 1991. Energy and Large-Scale Patterns of Animal- and Plant-Species Richness. The American Naturalist 137:27-49.

Hutson, S.S., Barber N.L., Kenny J.F., Linsey K.S., Lumia D.S., and Manpin M.A., 2004. Estimated Use of Water in 2000. U.S. Geological Survey Circular 1268, Alexandria, Virginia. Available at http://water.usgs.gov/pubs/circ/2004/circ1268/pdf/circular1268.pdf. Accessed in September 2004.

IPCC, 2008. Technical Paper on Climate and Water. http://www.ipcc.ch/meetings/session28/doc13.pdf

Jackson, R.B., E.G. Jobba´gy, R. Avissar, S.B. Roy, D.J. Barrett, C.W. Cook, K.A. Farley, D.C. le Maitre, B.A. McCarl, and B.C. Murray. 2005. Trading Water for Carbon With Biological Carbon Sequestration. Science, 310:944-1947.

Law BE, Falge E, Gu L, Baldocchi DD, Bakwin P, Berbigier P, Davis K, Dolman AJ, Falk M, Fuentes JD, Goldstein A, Granier A, Grelle A, Hollinger D, Janssens IA, Jarvis P, Jensen NO, Katul G, Mahli Y, Matteucci G, Meyers T, Monson R, Munger W, Oechel W, Olson R, Pilegaard K, Paw U KT, Thorgeirsson H, Valentini R, Verma S, Vesala T, Wilson K, Wofsy S. 2002. Environmental controls over carbon dioxide and water vapor exchange of terrestrial vegetation. Agricultural and Forest Meteorology 113: 97–120. doi:10.1016/S0168-1923(02)00104-1.

McNulty, S.G., G. Sun, E.C. Cohen, and J.A. Moore Myers. 2007. Change in the Southern U.S. Water Demand and Supply over the Next Forty Years. Book Chapter 5. In W. Ji (Editor) Wetland and Water Resource Modeling and Assessment: a Watershed Perspective. P 43-56. CRC Press.

Plantinga, A.J., Alig, R.J., Eichman, H., and D.J. Lewis, 2007. Linking Land-Use Projections and Forest Fragmentation Analysis. Res. Pap. PNW-RP-570. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Station. 41 p.

Rosenzweig 1967. Roy, S.B., P.F. Ricci, K.V. Summers, C-F. Chung, and R.A. Goldstein, 2005. Evaluation

of the sustainability of water withdrawals in the United States, 1995 TO 2025. Journal of the American Water Resources Association 41(5):1091-1108.

Solley, W.B., R.R. Pierce, H.A. Perlman, 1998. Estimated Use of Water in 1995. U.S. Geological Survey Circular 1200, Alexandria, Virginia. Available at http://water.usgs.gov/watuse/pdf1995/pdf/summary.pdf. Accessed in July 2007.

Sun, G., S.G. McNulty, J. Lu, D.M. Amatya, Y. Liang, and R.K. Kolka, 2005. Regional annual water yield from forest lands and its response to potential deforestation across the Southeastern United States. Journal of Hydrology 308:258-268.

Sun, Ge, Steven G. McNulty, Jennifer A. Moore Myers, and Erika C. Cohen. 2008a. Impacts of multiple stresses on water demand and supply across the southeastern United States. J. American Water Resource Assoc. 44(6): 1441-1457.

113Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.

Sun, G., S.G. McNulty, J.A. Moore Myers, and E.C. Cohen 2008b. Impacts of Climate Change,Population Growth, Land Use Change, and Groundwater Availability on Water Supply and Demand across the Conterminous US. Watershed Update Vol. 6, No. 2

U.S. Geologic Survey Water Resources Division, 1994. Hydrologic unit maps of the coterminous United States: 1:250,000 scale, 8-digit hydrologic unit codes and polygons. Reston, Virginia.

U.S. Global Change Program, 2000. Climate Change Impacts on the United States: The Potential Consequences of Climate Variability and Change. Overview. A Report to the National Assessment Synthesis Team. Cambridge University Press. 154 p

Zhou, G., G. Sun, X. Wang, C. Zhou, S.G. McNulty, J. M. Vose, and D.M. Amatya, 2008. Estimating Forest Ecosystem Evapotranspiration at Multiple Temporal Scales with a Dimension Analysis Approach. Journal of the American Water Resources Association 44(1): 208-221.

114Watershed Management 2010 © ASCE 2011

Watershed Management 2010

Dow

nloa

ded

from

asc

elib

rary

.org

by

Nor

th C

arol

ina

Stat

e U

nive

rsity

on

12/1

0/14

. Cop

yrig

ht A

SCE

. For

per

sona

l use

onl

y; a

ll ri

ghts

res

erve

d.