Investigating the impact of climate change on New York City’s primary water supply
- First Online:
Future climate scenarios projected by three different General Circulation Models and a delta-change methodology are used as input to the Generalized Watershed Loading Functions – Variable Source Area (GWLF-VSA) watershed model to simulate future inflows to reservoirs that are part of the New York City water supply system (NYCWSS). These inflows are in turn used as part of the NYC OASIS model designed to simulate operations for the NYCWSS. In this study future demands and operation rules are assumed stationary and future climate variability is based on historical data to which change factors were applied in order to develop the future scenarios. Our results for the West of Hudson portion of the NYCWSS suggest that future climate change will impact regional hydrology on a seasonal basis. The combined effect of projected increases in winter air temperatures, increased winter rain, and earlier snowmelt results in more runoff occurring during winter and slightly less runoff in early spring, increased spring and summer evapotranspiration, and reduction in number of days the system is under drought conditions. At subsystem level reservoir storages, water releases and spills appear to be higher and less variable during the winter months and are slightly reduced during summer. Under the projected future climate and assumptions in this study the NYC reservoir system continues to show high resilience, high annual reliability and relatively low vulnerability.
Projected twenty first century changes in air temperature and precipitation patterns due to climate change may alter the availability of water leading to new challenges for water supply planning and management in many regions throughout the world (Bates et al. 2008; Hunt and Watkiss 2011; Vicuna and Dracup 2007; Gleckler et al. 2008) including the New York City (Rosenzweig and Solecki 2001). For mountainous regions of the northeastern U.S. these changes can reduce annual snowpack accumulation, accelerate snowmelt processes and increase water losses due to evapotranspiration (ET) which may lead to more winter flooding and reduced summer flows (Brekke et al. 2009; Milly et al. 2005; Seager et al. 2007; Burns et al. 2007; Blake et al. 2000; Vicuna and Dracup 2007; Frei et al. 2002; Matonse et al. 2011). The potential impact of climate change on the ability to meet future demands for high quality drinking water, and satisfy other competing goals for surface water supplies (Bates et al. 2008; Brekke et al. 2009), is an issue of importance in many regions in U.S. and around the world and of primary concern to New York City (NYC) (NYCDEP 2008). A thorough investigation of climate change impacts on the NYC water supply system (NYCWSS) should include, in addition to climatic variations, the operational constraints of the system and potential responses through adaptive management in order to meet demands and other day-to-day operational goals (NYCDEP 2008).
Storage capacity (106 cu.m)*
Historical dependable yield (cms)*
Drainage area (sq. Km)
Annual average flow (cms)
Coefficient of variation (annual flow)
The Catskill Mountain region is part of the Allegheny Plateau consisting mainly of sedimentary bedrock that rises approximately 1100 m in elevation from the Hudson River, its eastern boundary (Burns et al. 2007). The region is mostly rural and forested with some dairy farms in low-land areas, particularly in the western part of the mountains. About 70 % (2850 square km) of the WOH area is part of the New York State Catskill Forest Preserve. The climate of this region is humid continental with relatively cold winters and cool summers. The temperature is variable among locations in the region and is highly impacted by elevation. Annual average temperatures range between 5.2 °C at higher elevations (Slide Mountain, 807 m elevation) and 7.5 °C at valley locations (Walton, 450 m elevation). Regional hydrology in NYC WOH watersheds is strongly influenced by snowpack and snow melt particularly during March and April (Matonse et al. 2011).
Past studies addressing climate change impacts on the NYC water supply have projected changes in climatology and hydrology and found evidence to suggest that some of the projected changes are underway. Burns et al. (2007) applied a non-parametric Mann-Kendall test to study trends in air temperature, precipitation, streamflow and ET for the Catskill Mountain region during the period of 1952–2005. Over the course of their study they found a 0.6 °C increase in yearly air temperature with higher increases in daily minimum temperature during May through September and daily maximum temperature during February through April. Precipitation was observed to increase by 136 mm and ET by 19 mm per 50 years. Peak snowmelt showed a shift from early April to late March with a reduction in snowpack consistent with patterns in streamflow and monthly air temperature. Frei et al. (2002) arrived at similar results when applying a version of the Thornthwaite water-balance model to study inter-annual variations and sensitivity to climate change for the Cannonsville basin (Delaware subsystem). In addition, the results from Frei et al. (2002) indicated a change in mean annual streamflow of approximately 6 % per degree C change in average annual temperature and a 1.5–2 % change of annual streamflow for each percent change in annual precipitation. These changes in hydrology indicate that as a result of climate change more water will become streamflow during winter potentially filling the reservoirs earlier in the spring (Matonse et al. 2011), but also increasing the potential for regional flooding (Burns et al. 2007). These results also suggest that for the NYC WOH watersheds projected increases in precipitation generally surpass the effects of increased water loss due to higher ET rates associated with higher temperatures. Conversely as suggested by Blake et al. (2000), if increases in precipitation are not as large as expected, the resulting reduction in water availability from increased winter spill and summer ET could lead to a greater risk of drought and increased drought severity.
The management of large systems such as the NYC water supply system is complex as it depends on watershed hydroclimatologic characteristics, reservoir capacities, reservoir operating rules, and system demands (Vogel et al. 1995). Operational complexity is a consequence of the large number of interconnected reservoirs and a decision-making process that includes meeting oft-times competing goals to satisfy demands, balance storage in different parts of the system so as to maximize water availability and minimize spills, meet regulatory flow requirements, and maintain water quality standards imposed on the system (NYCDEP 2011). One important aspect in this process is the management of extreme events to mitigate peak flows and drought (Matonse et al. 2011; Yin et al. 2010; Chang and Chang 2001) through the use of control structures and system operations in order to maintain reliability.
A variety of variables or indicators can be used to describe the operational state of the system. These include measures of inflow, reservoir storage and probability of future refill, releases, spills and demands on a daily, weekly, monthly and yearly basis. Some indicators describe the state of the system or its components (e.g. reservoir, subsystem, or other element) at a given point in time and are therefore very important for managing operations on a short-term basis. Other indicators, such as safe yield (NYCDEP 2008; Mayer 1993), system reliability and resilience (Vogel et al. 1995; Vogel and Bolognese 1995; Vogel 1987), storage vulnerability, surface and ground water stress, coefficient of variation of annual inflow (Lane et al. 1999), and standardized net inflow (Vogel et al. 1999) can be used to evaluate the overall performance of the system on a seasonal and/or long-term basis. Most of these indicators have been used in the past in order to: (i) summarize the effectiveness of regional water supply systems to meet demands (Lane et al. 1999), (ii) compare the relative performance of proposed improvements to water supply infrastructure, (iii) evaluate the effectiveness of reservoir operation policies (Yin et al. 2010), (iv) study the impact of new regulatory procedures, and most recently (v) assess the impact of climate change on existing reservoir systems (Lane et al. 1999; Vogel et al. 1999).
Ongoing nationwide infrastructure and system reliability assessments in the United States highlight the importance of system indicators (Harberg 1997; Vogel et al. 1999). Globally, a growing importance is given to local and regional assessment of water systems in balancing water supply and ecological needs, or socio-economic and environmental objectives (Vogel et al. 2007; Lane et al. 1999; Rogers 1999; Gao et al. 2009; Richter et al. 1996) or characterizing the effects of regulation on flow regimes (Black et al. 2005). Vogel et al. (1995) studied the storage-reliability-resilience-yield relationship for four different water supply systems in the northeastern United States, including the NYCWSS. The objective was to integrate the effects of hydroclimatologic characteristics, reservoir system storage capacity, operating rules, and system yield in a systematic manner to answer questions about which water supply systems are most vulnerable to major water supply failure (i.e., inability to satisfy demands) in the region, and what characteristics make one water supply system more vulnerable or more resilient to drought than others. For this study we combined indicators that provide important information for daily operations with those that help evaluate overall reservoir system performance.
2 The OASIS model and the NYCWSS modeling framework
Meeting regulatory release requirements for NYC reservoirs;
Balancing diversions from the Delaware, Catskill and Croton subsystems; and
Meeting demands from both NYC and outside communities (OC).
The operating rules expressed in the OASIS model provide a robust simulation of NYC reservoir system operations under current operating protocols.
In its original form the NYC OASIS model was driven by historical stream flow inputs (Fig. 2 with links in black). As part of this study, the model was configured to use input derived from climate change scenario simulations. Future streamflow projections were obtained from simulations using the Generalized Watershed Loading Functions-Variable Source Area (GWLF-VSA) model (Schneiderman et al. 2007; Schneiderman et al. 2002; Haith and Shoemaker 1987). GWLF-VSA is a lumped-parameter continuous simulation model that simulates daily streamflow, nutrients, and sediment loads on a watershed scale. GWLF-VSA forcing inputs include daily air temperature, precipitation, incoming solar radiation, and daily average relative humidity. Both, the original GWLF (Haith and Shoemaker 1987) and GWLF-VSA treat the watershed as a system of different land areas (Hydrologic Response Units or HRUs) that produce surface runoff, and a single groundwater reservoir that supplies baseflow. GWLF-VSA differs from the original GWLF (Haith and Shoemaker, 1987) in that it simulates saturation-excess runoff on variable source areas, which is considered the primary runoff generation mechanism in WOH watersheds (Walter et al.2003). To do so, GWLF-VSA simulates runoff volumes using the SCS Curve Number Method, as in the original GWLF model, but spatially-distributes the runoff response according to a soil wetness index, based on the TOPMODEL soils-topographic index (Schneiderman et al. 2007; Beven and Kirkby 1979). Potential evapotranspiration in GWLF-VSA is calculated using the Priestley-Taylor method (Priestley and Taylor 1972; Neitsch et al. 2005) with crop parameters varying between growing and dormant seasons. For this study GWLF-VSA is run over the entire reservoir basin area. As this area includes the reservoir surface itself, the model treats precipitation over the water surface as a direct inflow to the reservoir after subtracting potential evaporation.
Future scenarios of precipitation and air temperature derived from different GCM models using a delta-change methodology (Hay et al. 2000; Anandhi et al. 2011) were input into GWLF-VSA for all WOH reservoir watersheds (Fig. 2 with dotted links) to simulate inflows to the reservoir system. The complete integrated modeling framework including GWLF-VSA and OASIS accurately simulate the state of the system over time and provide data to evaluate the performance of the system under projected changes in climate.
In the following sub-sections we describe briefly each of the indicators used in this study to evaluate possible effects of climate change on the NYC water supply. These indicators were selected based on their importance (i) for the NYC water supply operation and (ii) for assessing system performance and subsequent comparison with other systems at a regional level.
3.1 Reservoirs storage, inflow, release, spill, and diversion
Comparing patterns of storage, inflow, release and spill helps describes how the system functions. We compare indicators for the baseline and future projections simulated using different GCM climate scenarios. Inflows are related to hydro-climatic conditions, while the relationship between inflow, storage, release, and spill are governed by the rules within OASIS. Releases (i.e., controlled outflows from the reservoirs outlets) in the Delaware subsystem, for example, are governed by minimum downstream flow objectives and, to a lesser extent, peak flow mitigation and habitat protection. Delaware subsystem reservoirs are typically drawn down from November through March to create storage void equal to half the equivalent water volume in the snowpack and during this time diversions from the Catskill subsystem may be minimized. The Catskill subsystem includes rules to regulate water diversions from Schoharie reservoir through the Shandaken Tunnel to Esopus Creek. These rules account for allowable flow rates, temperature and turbidity limits for the tunnel discharges. Diversions of water from and between reservoirs are based on individual reservoir storage volumes, drought conditions, water quality, demands and local downstream requirements.
3.2 Drought occurrence
Drought occurrence is an indicator that is important for operations because it leads to temporary changes in the criteria for moving water from reservoirs and acts to trigger demand reductions. There are three drought levels in the NYCWSS: Watch, Warning, and Emergency. The classification and call for a particular drought level is executed by comparing the total storage of each subsystem with average yearly storage patterns associated with each drought level and each subsystem (Catskill and Delaware). The combination of the drought conditions in Delaware and Catskill subsystems determines the drought status for the entire NYCWSS.
3.3 Standardized net inflow and reservoir system resilience
3.4 Storage ratio
The storage ratio (S/μ), where S is the reservoir storage capacity has been used in the past to characterize system operations. Vogel et al. (1999) analyzed storage-yield curves for different regions across the United States and found that storage-yield curves may lead to different values depending on whether mean annual or monthly flows are used in the computation. Monthly flow based curves will generally show a higher value of storage ratio compared to annual based curves except when Cv ≥ 0.3 and m < 1; when both annual and monthly flows based ratios provide similar (within 30 %) results of storage-yield curves. Though this indicator is less informative to system operations than other indicators (Vogel et al. 1999) it was included in order to compare this study with past reservoir system analyses.
3.5 Reservoir system reliability
For this study this equation was solved to estimate Ra for N = 50 and 100 years. As in Vogel et al. (1999) we assumed a no-failure reliability RN = 0.5. Although reservoirs in the system are interconnected, this assumption is reasonable since we considered the entire system as a unit.
3.6 Reservoir system vulnerability
4 Data and modeling assumptions
4.1 Climate data
Observed climate and streamflow data for this study include historical measurements of daily air temperature and precipitation covering a period from 1927 to 2004. These data used to drive the GWLF-VSA watershed model were obtained from up to nineteen stations for precipitation and four stations for temperature distributed throughout the WOH watershed region.
To create historical daily precipitation time series for each reservoir’s watershed that could be input into the GWLF-VSA model, individual station precipitation values were spatially averaged using Thiessen polygon weighting (Burrough 1987). For air temperature inverse distance weighting was used and lapse rates were applied to account for variations in temperature with elevation (Zion et al. 2011).
List of GCM, emission scenarios and time slices used to develop climate change scenarios for this study. 20C3M represent the baseline (from the 20th century experiment model runs)
Future time slices
A2, A1B, B1
A2, A1B, B1
Future climate projections were constructed using a delta change method (e.g. Hay et al. 2000; Gleick 1986) also known as change factor methodology (Anandhi et al. 2011). In delta change method monthly factors are calculated from the difference between GCM baseline and future simulated using pooled monthly data for the two time slices (Anandhi et al. 2011). The calculated delta change factors were then combined with records of observed data (additively for air temperature and multiplicatively for precipitation) to generate future climate projections for each scenario. In this manuscript we will refer to the 2046–2065 and 2081–2100 periods as the 2055s and 2090s, respectively. The baseline scenario represents model runs using the observed forcing data.
4.2 Modeling assumptions
OASIS simulations were conducted for the entire NYCWSS, which includes the upper and lower Delaware, Catskill, and Croton subsystems (Fig. 1). However, only the WOH portion (the upper Delaware and Catskill basins) was simulated using both current and future modeled inflows. The remaining Croton (East of Hudson (EOH)) portion of NYCWSS and the lower Delaware (LD) were simulated using historical flow inputs for all baseline and future scenarios. Though these two portions have an impact on the water routing processes in WOH, we choose to maintain this assumption for this analysis because EOH contributes about 10 % of all NYCWSS requirements and the LD does not contribute any flow to NYC. NYC and OC average demand levels and system operation rules across the system were considered stationary and remained unchanged between baseline and future simulations.
5 Evapotranspiration, snow and reservoir inflows in the study area
During the growing season, part of the increased precipitation is offset by increased ET, resulting in only a slight increase in the inflow to the reservoirs. It is interesting to note that even though there is little change in flow during the growing season period, the average unsaturated zone soil moisture (not shown here) decreases slightly during the spring months. This is because the soil has a limited storage, and the increased precipitation cannot increase the maximum soil moisture storage. The increased ET rate in the future climate simulations therefore tends to enhance the decline in soil moisture storage from its maximum value during inter-storm periods.
6 Results for system indicators
To describe the state of the water supply system and help assess performance a number of indicators were selected and evaluated in terms of their ability to capture and display system changes associated with projected future climate change.
6.1 Reservoirs storage, spill and release
While differences in storage among the two projected future climate time periods are minimal, changes in spill and release appear more pronounced for the end of the century. It is important to note that, although the total volume of simulated future releases and spills during late fall and winter increase somewhat, the volumes of spill and release never reach the highest levels encountered in the baseline period during April.
6.2 Drought conditions occurrence
6.3 Reservoir system performance
6.3.1 Demand ratio α, standardized net inflow, coefficient of variation and storage ratio
The coefficient of variation of annual inflow (Cv) remains below 0.3 for most future simulations except for a few in the 2090s, while the storage ratio shows a decrease under future simulated climate. These results are consistent with findings by Vogel et al. (1999) who found that systems showing pronounced within-year behavior are associated with Cv < 1 and Cv ≤ m ≤ (1/Cv). Also, Vogel et al. (1999) showed how for systems with Cv less than 0.3, the storage ratio based on annual flow differs from the storage ratio based on monthly inflows. Our results are consistent with those from Vogel et al. (1999) indicating slightly higher storage ratios when using monthly flows rather than annual flows; a sign of the impact of seasonality associated with monthly flows. Project changes in the coefficient of variation and storage ratios are within the range found by Vogel et al. (1999) for this region, while standardized net flow values are slightly higher, probably due to the fact that their estimations of yield and mean annual inflow were made at a larger regional scale.
6.3.2 System resilience, reliability and vulnerability
Annual reliability of reservoirs across the United States is generally between 0.97 and 0.985 (Vogel et al. 1999), with annual reliability being the steady-state probability that a reservoir will deliver the required yield for a particular year without failure. Our baseline simulation (Fig. 10) suggests that WOH reservoirs are in the upper limit of this range and that future changes in climate will result in no substantial difference in annual reliability. The variability among the different model scenario simulations is less than 0.0015. A high positive correlation between resilience and reliability for the NYC reservoir system is not surprising; Vogel et al. (1999) arrived at similar results for reservoir systems in the eastern United States.
As under the present conditions projected D for future climate simulations indicate that the NYC reservoir system will continue showing high resilience and low vulnerability (D ≈ 0.1). As D is an indicator of average number of consecutive years a reservoir system fails to deliver an expected yield a unity vulnerability (D = 1) can be interpreted as a failure state that will last one year on average (Vogel et al. 1999). This is important because based on our results any eventual failure in the NYC reservoir system will last for far less than a year. Our D, r, and Ra values are within the range found in previous studies for this region (e.g., Lane et al. 1999; Vogel et al. 1999).
7 Summary and conclusions
This study focuses on the potential impacts of climate change on NYC water supply with regards to reservoir operations and water supply system performance. An ensemble of three GCMs, three emission scenarios and two future time periods were used to simulate a present baseline and 16 different future climate change projections for the west of Hudson (WOH) region of the NYCWSS, which contributes more than 90 % of all water needed to meet NYC water demands. Air temperature and precipitation derived from these scenarios were used as inputs to the GWLF watershed model to simulate inflows required to run OASIS and simulate reservoir operations.
Results from this study comparing baseline inflow estimations with simulations for two future time slices representing the 2055s and 2090s suggest that the trend of increasing streamflow identified in historical records from Catskill basins (Burns et al. 2007) will continue so that on average annual streamflow will increase by approximately 97 mm in the next fifty year period. No statistically significant difference was detected in the annual streamflow between the future 2055s and 2090s.
On a seasonal basis monthly inflows will rise for almost all months with the greatest changes during winter and early spring due to a combined effect of more rainfall and snowmelt associated with higher temperatures (Matonse et al. 2011). Increases in winter inflows are more pronounced during the 2090s compared to 2055s. During summer projected changes in streamflow are relatively small suggesting that increased evapotranspiration (ET) rates associated with higher summer temperatures largely counteracts increased precipitation at this time. This result is not surprising; in the past Burns et al. (2007) and Blake et al. (2000) suggested that higher ET rates will impact inflows in a warmer climate. Future summer flows remained basically unchanged as both ET and precipitation increased; effectively canceling the individual effects of each component. It should be noted that the uncertainty of future precipitation is much greater than that of future temperature. If the future precipitation estimates are over-predictions, then summer low flows could decrease more dramatically and annual inflow into the reservoir system will be less than expected.
To better gauge the impact of climate change on reservoir system operation, indicators such as reservoir inflow, storage, release, spill, and drought level were examined. Other indicators, such as the coefficient of variation of inflow, standardized net inflow, reservoir system resilience, reliability and vulnerability which have been used in the literature to measure and compare reservoir system performance were also examined.
Reservoirs filling earlier with inflows more evenly distributed during winter and early spring and a reduction in the historically observed April runoff peak. Under these conditions releases and spills will become higher during late fall and winter and less during April.
A decrease in the average number of days the Catskill and Delaware reservoirs will be under drought emergency, warning and watch conditions.
An increase in average standardized net inflow m (m > 1). Average coefficients of variation Cv for future simulations that are similar to baseline and less than 0.3, and decreased average storage ratios (D < 1), but with storage ratios based on monthly flows being slightly higher than storage ratios based on annual flows, indicating the effect of seasonality present in monthly flows. Our results support a previous regional analysis that characterizes reservoirs systems for the eastern United States region by Cv < 1 and Cv ≤ m ≤ (1/Cv) and are consistent with a reservoir system dominated by within-year variability; also found in previous studies to be characteristic for the eastern United States region.
Historical meteorology and model simulated baseline and future meteorology and streamflow were used in this study of the effects of climate change on NYC water supply. For this study only historical water demands and operational rules were available and these were considered stationary during future simulations. Maintaining current rules and demands helps evaluate the system effectiveness in responding to changes in climate alone (Matonse et al. 2011). However, population and other socio economic changes in the future may alter the current level of water demand (IPCC 2001). In addition historical data analysis indicates a positive correlation between water demands and high temperatures suggesting a potential direct impact of climate change on demands. These factors and any changes in regulatory flow requirements and water quality standards may have an impact on system operations and need to be accounted for in future studies. Also, the variability of inflows associated with future climate simulations are not directly obtained from the GCMs variability, but reflect the variability of the historic climate, a limitation of the delta change method. As water supply systems are sensitive to the frequency and magnitude of extreme hydrological events the use of model ensembles including other (dynamical and/or statistical) downscaling methods can potentially provide additional information that is important for the impact assessment and adaptation processes (Stainforth et al. 2007).
In conclusion, according to the climate models and scenarios applied in this study the NYC reservoir system will most likely continue to show high resilience, high annual reliability and relatively low vulnerability. As in previous studies reliability and resilience show a positive correlation. These conclusions will continue to be evaluated as updated climate model scenarios and future demand projections become available.
The authors are grateful to New York City Department of Environmental Protection (NYC DEP) who provided funding for the research that led to this manuscript. This paper does not necessarily reflect the official views of NYC DEP, and no endorsement should be inferred; to Grantley W Pike for the valuable review; and to David Lounsbury and Donald Kent of the NYC DEP water quality modeling group for their inputs. Also, the authors want to thank the Columbia University Center for Climate Systems Research and the NASA Goddard Institute for Space Studies for providing the NCAR, GISS and ECHAM GCM data.
This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.
- Bates B, Kundzewicz C, Wu W, Palutikof P (eds) (2008) Climate change and water. Technical Paper of the Intergovernmental Panel on Climate Change. Geneva, Switzerland, IPCC Secretariat, p 210Google Scholar
- Burrough PA (1987) Principles of geographical information systems for land resources assessment. Clarendon, Oxford, 186Google Scholar
- Cole TT, Wells SA (2002) CE-QUAL-W2: a two-dimensional, laterally averaged, hydrodynamic and water quality model, version 3.1. Instruction Report EL-2002-1., U.S. Army Engineering and Research Development Center, Vicksburg, MSGoogle Scholar
- Harberg RJ (1997) Planning and managing reliable urban water systems. American Water Works Association, DenverGoogle Scholar
- Hardison CH (1972) Potential United States water-supply development. J Irrig Drain Div ASCE 98(3):479–492Google Scholar
- Hay LE, Wilby RL, Leavesley GH (2000) A comparison of delta change and downscaled GCM scenarios for three mountainous basins in the United States. J Am Water Resour Assoc 36(2)Google Scholar
- HydroLogics Inc (2007) User manual for OASIS with OCL, Columbia, MDGoogle Scholar
- IPCC (2001) Climate Change 2001: the scientific basis. In: Houghton JT, Ding Y, Griggs DJ, Noguer M, van der Linden PJ, Dai X, Marskell K, Johnson CA (eds) Contribution of working group I to the third assessment report of the intergovernmental panel on climate change. Cambridge University Press, p 881Google Scholar
- Joint Editorial Board Representing APHA, ASCE, AWWA, and WPCF (1969) Glossary: Water and Wastewater Control EngineeringGoogle Scholar
- Mays LW, Tung Y-K (1992) Hydrosystems engineering and management. McGraw-Hill Inc, New YorkGoogle Scholar
- Mayer RA (1993) Safe yield study of the New York City reservoir system. Progress report: Catskill/Delaware filtration. Strategic Services Division, NYC DEPGoogle Scholar
- Neitsch SL, Arnold JG, Kiniry JR, Srinivasan R, Williams JR (2005) Soil and water assessment tool: theoretical documentation, version 2005. Agricultural Research Service, Temple, 476Google Scholar
- New York City Department of Environmental Protection (NYCDEP) (2011) 2011 Watershed protection program summary and assessment. March 2011. http://www.nyc.gov/html/dep/pdf/reports/fad51_wmp_2011_fad_assessment_report_03-11.pdf. Accessed 27 April 2012
- New York City Department of Environmental Protection (NYCDEP) (2008) Climate change program: assessment and action plan. May 2008. Report 1. A Report based on the ongoing work of the DEP climate change task force. http://www.nyc.gov/html/dep/pdf/climate/climate_complete.pdf. Accessed 27 April 2012
- Rosenzweig C, Solecki WD (2001) Climate change and a global city: learning from New York. Environment 43(3):8–18Google Scholar
- Vogel RM (1985) The variability of reservoir storage estimates, Ph.D. thesis, Cornell University, Ithaca, NYGoogle Scholar
- Vogel RM, Sieber J, Archfield SA, Smith MP, Apse CD, Huber-Lee A (2007) Relations among storage, yield and instream flow. Water Resour Res 43. doi:10.1029/2006WR005226
- Vogel RM, Lane M, Ravindiran RS, Kirshen P (1999) Storage reservoir in the United States, Journal of Water Resources Planning and Management. ASCE 125(5):245–254Google Scholar
- Zion MS, Pradhanang SM, Pierson DC, Anandhi A, Lounsbury DG, Matonse AH, Schneiderman EM (2011) Investigation and modeling os winter streamflow timing and magnitude under changing climate conditions for the Catskill Mountain region, New York, USA. Hydrol Process 25:3289–3301. doi:10.1002/hyp.8174 CrossRefGoogle Scholar