The Effects of Climate Change on Water Management ...

The Effects of Climate Change on Water Management Strategies and Demands in the Central Valley of California Nathan T. VanRheenen1, Andrew W. Wood1, Richard N. Palmer2, Jeffrey T. Payne1, Dennis P. Lettenmaier2 1

Graduate Research Assistant, Department of Civil and Environmental Engineering, University of Washington, Box 352700, Seattle, WA 98195-2700; PH (206) 616-1775; FAX (206) 6859185 2 Professor, Department of Civil and Environmental Engineering, University of Washington, Box 352700, Seattle, WA 98195-2700; PH (206) 685-2658; FAX (206) 685-9185 Abstract This paper provides preliminary results of a study designed to determine the potential impacts of climate change on water resources in the Central Valley of California. A general circulation model is used to derive meteorological forcing functions that, in turn, provide inputs for a macroscale hydrological model which simulates streamflows used to drive a water management model. The water management model characterizes the performance of the coordinated State Water Project and Central Valley Project under altered climates. Preliminary results from these models indicate that both short and long-term climate change will have important implications for future policy and operational decisions in the region. Of particular interest are the effects of the current dynamic operational strategy and design on hydropower generation, water supply, fish endangerment, and flood control. Introduction Global climate change is expected to affect the performance of water resource systems significantly (Frederick and Major, 1997; Gleick, 2000). In many parts of the world, including much of the United States, the demand for consumptive (water supply) and non-consumptive (navigation, hydroelectric power generation, instream flows) uses of fresh water is barely balanced by sustainable surface and groundwater sources. Current indicators, including recent findings from the Intergovernmental Panel on Climate Change (IPCC, 2001) strongly suggest that water resources will respond to global warming in ways that will negatively impact both overall supplies and long-term reliability. The ultimate impact of climate change on water resources will reflect the degree to which adaptive management can be applied effectively. In the case of large, complex, and heavily modified river systems, like those in the western U.S., the potential impacts of future climate change can only be understood in the context of the evolving stresses and conflicts that exist among the multiple uses of a basin’s resources. In the Columbia River basin, for example, average annual flows vary by almost 20% between periods of negative and positive Pacific Decadal Oscillation phases (Hamlet and Lettenmaier, 1999). When considered in the context of increasing demands and changing water policies, the implications of such variations are greater still. Among the western states, California is particularly susceptible to changes in water supply. Bulletin 160-98 of the California Water Plan Update (Department of Water Resources,

VanRheenen et al. – Page 1 of 10

1998) estimates that, at 1995 levels of development, water shortages of 1.6 million acre-ft (maf) occur in average water years. In drought years, shortages as large as 5.1 maf can be expected. By 2020, it is estimated that the shortages will be 2.4 maf in an average water year and 6.2 maf in drought years. A projected increase in population of nearly 50%, coupled with a 36% increase in urban water use are the drivers for this forecasted increase in water demands. These projected shortages do not, however, reflect any change in availability of water due to climate change. Inclusion of climate change effects would likely significantly impact DWR’s water availability projections. The State Water Project (SWP) and the Central Valley Project (CVP) coordinate operations of a system of 20 major dams and reservoirs with a combined storage capacity of 12 maf, 13 major power plants, 500 miles of major canals and aqueducts, and various related facilities (Figure 1). The combined SWP/CVP system is one of the largest water storage and conveyance systems in the world. The SWP/CVP system is operated in accordance with their respective water rights permits and licenses, which are administered by the State Water Resources Control Board (SWRCB). The SWP and CVP are required to meet water quality standards and the demands of senior water rights holders while minimizing the likelihood of jeopardizing endangered and threatened fish species (US Bureau of Reclamation, 1999). The SWP and CVP do not currently operate the reservoirs to optimize power production. In light of the ongoing power shortages experienced in California, it can be expected that this, along with the difficulties in meeting other demands in the system, will become more significant factors as the impacts of global warming are increasingly felt (Lane et al., 1999). In California, as in much of the West, future climate peak seasonal flows are expected to occur earlier in the year, and summer flows are expected to be lower, than under current climate conditions (Cohen et al., 2000). However, these trends obscure the impact of important local characteristics, as well as the interaction of natural and built systems. To better understand local and regionally specific hydrologic impacts of climate change, macroscale hydrologic models, like the Variable Infiltration Capacity (VIC) model (Liang et al., 1994), can be used. The VIC model has been applied to a number of large river basins for assessments of the hydrologic implications of climate change (Nijssen et al., 1997; Hamlet and Lettenmaier, 1999; Nijssen et al., 2000). The U.S. Department of Energy has recently initiated the Advanced Climate Prediction Initiative (ACPI), one part of which is designed to utilize advanced climate prediction capabilities for better assessment of the vulnerabilities of western U.S. water resources systems to climate change. As part of ACPI, the University of Washington (UW) is using the VIC macroscale hydrological model, coupled with a water management model, to assess the implications of climate change for the Central Valley of California. Preliminary results from this study indicate that both short and long-term climate change have important implications to future policy and operational decisions in the region. Of particular interest are the effects of the current dynamic operational strategy and design on hydropower generation, water supply, fish endangerment, and flood control. Future phases of the study will consider alternative operational strategies that might mitigate the short and long-term effects of climate change on water resources in the Central Valley.


a)

b)

Figure 1 a. Sacramento-San Joaquin River basin and all dams and reservoirs with storage capacities greater than 100 thousand acre-feet. There are a total of 846 dams and reservoirs in this region. b. Sacramento and San Joaquin River basin, including major SWC and CVP dams and reservoirs. These reservoirs represent the primary nodes receiving inflows from the VIC model for use in the water management model.

Climate Prediction and Assessment Approach Over the past decade, numerous analyses of the potential effects of climate change on water resources have been conducted. The most sophisticated of these have employed suites of computer simulations (see, e.g., Lettenmaier et al., 1999). These typically begin with a coupled ocean-atmospheric general circulation model (GCM) that generates future climate scenarios. The future climate scenarios include estimates of future changes in precipitation and temperature that are used to force a rainfall-runoff (hydrologic) model, which produces future streamflow


scenarios. The streamflow scenarios, in turn, serve as input data for a water resources system model that calculates the performance of the water resources system for a given region. The climate prediction and assessment approach adopted for this study generally follows the method of the earlier studies in assembling an offline, one-way linkage between three computer models. The NCAR-DOE Parallel Climate Model (PCM) provides global surface and atmospheric fields, including monthly total precipitation and monthly average temperature, for model integrations of length 20 to 300 years. The PCM output is adjusted statistically to remove regional bias, and the climate scenario anomalies are downscaled for use as forcings at the finer resolution of the Variable Infiltration Capacity (VIC) macroscale physical hydrology model. The resulting streamflow time series produced by the hydrologic model is used as inflow to the water resources system, Central Valley Model (CVMOD). CVMOD is a monthly timestep reservoir model developed by the University of Washington that incorporates the major projects and operational features of the Sacramento-San Joaquin River basin and the Sacramento-San Joaquin Delta, and evaluates the performance of each hydrologic scenario with respect to a comprehensive set of system objectives. The analysis of transient climate change effects on land surface hydrology, hence water resources management, is complicated by the confounding of natural variability (which would occur in the absence of climate change) with the long-term climate change signal. Although most climate change impact assessments to date have been based on transient GCM output that produces a single future climate scenario, such deterministic approaches could well lead to misleading results. To address this issue, we use ensembles of PCM scenarios, each giving rise to a different statistical realization of the surface hydrologic forcings and system performance corresponding to a given emissions scenario. Bias-Correction and Downscaling of Climate Model Scenarios Regional biases in GCM simulations of the observed climate are well documented (Watson et al., 1996; Wilby and Wigley, 1997; Wilby et al., 1999; Maurer et al., 2001). Implementation of an approach to address such biases requires several preliminary steps. First, the bias-correction of GCM outputs is based on a comparison of GCM model climatology (taken from one or more control runs) with observed regional climatology (taken from historical observations). Second, downscaling of multi-decadal climate forecasts from the spatially coarse resolution (~2.5 degrees latitude by longitude), often archived at a monthly timestep, to the resolution of the VIC macroscale hydrology model (daily, 0.125 degree), will be undertaken by two methods: the use of the intermediate resolution Scripps Regional Spectral Model (RSM), and direct downscaling (via interpolation and temporal disaggregation) of output from PCM. These steps create ensembles of VIC daily forcings that can be run from the current date over the forecast horizon. Creating Sacramento-San Joaquin River Basin Future Streamflow Scenarios In this study, the translation of climate information to streamflows makes use of the daily timestep VIC hydrology model, implemented for the Sacramento-San Joaquin River basin at 1/8degree latitude-longitude spatial resolution. The Sacramento-San Joaquin implementation of VIC adopts a set of vegetation and soil parameters described in Maurer et al. (2001). Routing of runoff generated within a grid cell through the channel network is accomplished in two stages. The single runoff time-series produced for each cell is routed to the


cell outlet using a user-specified unit hydrograph (Lohmann et al., 1998a; 1998b). The hydrographs for the individual cells are routed to the basin outlet through the specified channel network using the linearized St. Venant equations. The routing network for the Sacramento-San Joaquin River basin, which comprises approximately 2,925 cells and a drainage area of about 175,000 square miles, is shown in Figure 1b. Figure 1b also shows the current streamflow simulation points corresponding to the inflow nodes of the water resources management model. Precipitation and daily minimum and maximum temperatures are used to estimate the forcing variables required by the VIC model (e.g., shortwave and longwave radiation, and humidity, in addition to precipitation and surface air temperature). Wind speed is estimated using methods described in Maurer et al. (2001). The observational data typically are taken from NCDC Cooperator Stations, which become available after a lag of several months, and are gridded to the 1/8-degree resolution of the hydrology model. VIC model meteorological forcings for the Sacramento-San Joaquin River have been generated for January 1949 – July 2000. Reservoir Operations CVMOD is a monthly timestep reservoir model that incorporates the major projects and operational features of the Sacramento-San Joaquin basin and simulates the movement and storage of water within the basin given current operational policies. The model domain ranges from Clair Engle Lake and Lake Shasta, near the headwaters of the Trinity and Sacramento Rivers, respectively, to Millerton Reservoir, near the headwaters of the San Joaquin River. It includes many of the major tributaries: the American, Feather, Cosumnes, Mokelumne, Stanislaus, Tuolumne, and Merced River systems. Also included are the Sacramento-San Joaquin Delta and related water supply reservoirs and canals, which most notably include the San Luis Reservoir, Delta-Mendota Canal, and the California Aqueduct. The primary hydrologic input to CVMOD is monthly streamflow, which can be taken either from observed natural or unregulated flows (for studies of past climate) or from the hydrology model driven by output from a climate model (for assessment of future climate). CVMOD can thus be used to explore reliability given various operating policies and alternative climate and operating scenarios. The outputs of the model are reservoir levels and releases, from which the reliability of the system with respect to such operating criteria as water quality, flood control, hydropower production, agricultural and municipal diversions, navigation, and instream flows for fish is computed. Model Testing and Preliminary Runs Initial testing of the hydrologic and water resources model components of the forecast sequence utilized monthly temperature and precipitation anomalies for 2040-2050 from a DOE-PCM3 transient scenario used in the recent U.S. National Assessment and as part of IPCC climate change efforts (Washington et al., 2000). The intent of the testing was to gain an understanding of the sensitivity of the basin to the general shift earlier in the year of seasonal maximum flows due to reduced snowpack that are associated with essentially all climate scenarios. Using the simple perturbation method applied in Lettenmaier et al. (1999), these anomalies were downscaled and a future climate streamflow scenario was created as input to CVMOD. A retrospective hindcast of streamflows for water years 1950-1999 was also prepared and compared to observed streamflow time series from 1950-1999. The initial comparison of


simulated and observed inflows were reasonably accurate, but the comparison did identify areas where improvements in both the hydrologic model calibration and characterization of reservoir system operating rules are needed. Figure 2 shows simulated long-term average storages in Lake Shasta, on the Sacramento River, for current and future climate scenarios. The first scenario uses VIC model streamflow output based on observed conditions during water years 1980 through 1990, while the second scenario uses the same period of record, adjusted to have the same decadal average precipitation and temperature as were predicted by PCM3 for the decade 2040 - 2050. The figure shows averages associated with this ten-year period. These operational regimes simulated within the water resource model are not intended to represent precisely how the Central Valley was managed during the 1980s, which was characterized by constantly changing objectives involving instream flows for fish, temperature regimes, irrigation demands, and power needs. Instead, the simulated operating rules (which are fixed for both climate scenarios) reflect an approximation of current policies. Figures 2 and 3 are examples of typical responses to the climate changes. Average Shasta and Oroville storages show noticeable decreases for the 2040-2050 climate condition. The overall average decrease from the simulated present to the simulated future scenario in the dams was 14.2% for Oroville and 11.6% for Shasta. Because operational regimes and demands were kept constant, the differences in storage reflect only the changes in flow volume and timing.

3,000 2,800

Simulated Present

2,600

Simulated Future

2,400 2,200 2,000 1,800 1,600 1,400 1,200 oct

nov

dec

jan

feb

mar

apr

may

jun

jul

aug

sep

Fig. 2. Composite monthly storages, 1980-1990, in Lake Shasta generated from inflows produced by the VIC base (Simulated Present) and PCM3 Decade 4 (Simulated Future) runs. 1900 taf is the minimum storage desired by the Bureau of Reclamation (U.S. Bureau of Reclamation, 1999). Dead storage is 1200 taf.


3,000

Simulated Present Simulated Future

2,500 2,000 1,500 1,000 500 0 oct

nov

dec

jan

feb

mar

apr

may

jun

jul

aug

sep

Fig. 3. Composite monthly storages, 1980-1990, in Lake Oroville generated from inflows produced by the VIC base case (Simulated Present) and PCM3 Decade 4 (Simulated Future) runs. 1300 taf is the minimum storage desired by the Bureau of Reclamation (U.S. Bureau of Reclamation, 1999). Dead storage is 900 taf.

The overall effect on the SWP/CVP system is consistent with the Lake Shasta results (Figure 4). Overall, the climate changes result in a 13.1% loss in storage over the average year of operation. This has several negative effects on the contracted uses of all the dams in the region, especially on in-stream temperature requirements (which are currently not being modeled). The magnitude of the simulated changes is large enough that it is likely that system operation would have to be altered significantly. We are currently evaluating changes to system operations to determine the extent to which the effects shown in Figure 4 could be mitigated.

9,000 8,500

Simulated Present

8,000

Simulated Future

7,500 7,000 6,500 6,000 5,500 5,000 4,500 4,000 oct

nov

dec

jan

feb

mar

apr

may

jun

jul

aug

sep

Fig. 4. Composite monthly storages, 1980-1990, in all modeled reservoirs north of the Sacramento-San Joaquin Delta generated from inflows produced by the VIC base case (Simulated Present) and PCM3 Decade 4 (Simulated Future) runs.


Figure 5 shows that without changes in reservoir operation, power production would be significantly impacted by climate change. The total annual decrease in Shasta Powerhouse production is 277 megawatt-hours (MWh), an overall decrease in capacity of 14.8%. In the climate change scenario, less power is generated every month of the year when compared to present power production. These decreases are most significant in the months of December, January, March, and June. It is important to note, however, that Shasta Dam currently is contracted only for flood storage, irrigation, navigation, the environment, and municipal deliveries. Although hydropower is an authorized function of dams within the SWP/CVP system, it is subordinate to other operating objectives for both the CVP and the SWP. 350,000 300,000

Simulated Present

250,000

Simulated Future

200,000 150,000 100,000 50,000 0 oct

nov

dec

jan

feb

mar

apr

may

jun

jul

aug

sep

Fig. 5. Composite monthly energy production (MWh), 1980-1990, in Lake Shasta generated from inflows produced by the VIC base (Simulated Present) and PCM 3 Decade 4 (Simulated Future) runs.

Summary and Conclusions This paper summarizes preliminary results of a two-year study to evaluate the impacts of a global climate change on the Sacramento and San Joaquin River watersheds. These results were generated by integrating the outputs from global circulation models and physically based hydrologic models into a water resource simulation model. The initial findings indicate that, for climate conditions predicted to occur in 2040-2050, impacts on performance of the CVP/SWP system would be substantial. These impacts result from a decrease in total annual runoff and a temporal shift in peak flows. In our early analysis of the northern Sacramento system, both Lake Shasta and Lake Oroville experience decreases in average annual storage of more than 10% and decreases in annual energy produced of approximately 15%. Predicted increases in water demands have not been included in the calculation of future reservoir storages and hydropower production. As a result, it is likely that the decreases in both storage and hydropower production by the middle of the 21st century would, in fact, be much higher than estimated here. We are currently extending our analysis to the entire Sacramento-San Joaquin River complex. To accomplish this, we are reviewing performance of the VIC model, and are improving CVMOD to better characterize recent operations and water regulations. Because the Sacramento-San Joaquin Delta plays a significant role in the operations of Shasta, Oroville, and


San Luis Reservoirs, and others, it will be explicitly represented in the enhanced modeling system. Finally, we are developing a set of flexible operating rules and alternatives that considers the shift in runoff volume and timing that occurs with climate change and optimizes operations to both meet demands and maximize energy produced, given different climatic and demand scenarios. By successfully implementing these modeling strategies in the Central Valley of California, we intend to demonstrate the potential for use of linked climate, hydrology, and reservoir operations models to improve understanding of the implications of climate variability and change on water resources in the western United States. References Cohen, S. J., Miller, K. A., Hamlet, A. F., and Avis, W. (2000). “Climate change and resource management in the Columbia River basin.” Water International, 25(2), 253-272. Department of Water Resources. (1998). California Water Plan Update. Bulletin 160-98. California Department of Water Resources, Sacramento, CA, December, 1998. Frederick, K. D., and Major, D. C. (1997). “Climate change and water resources.” Climatic Change, 37, 7-23 Gleick, P. H. (2000). “Water: The potential consequences of climate variability and change for the water resources of the United States”, Report of the Water Sector Assessment Team of the National Assessment of the Potential Consequences of Climate Variability and Change, Pacific Institute, Berkeley, CA. Hamlet, A. F, and Lettenmaier, D. P. (1999). “Columbia River streamflow forecasting based on ENSO and PDO climate signals.” Journal of Water Resources Planning and Management, 125(6), 333-341. IPCC (Intergovernmental Panel on Climate Change). (2001). Climate Change 2001: The Scientific Basis: Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change. http://www.ipcc.ch/pub/spm22-01.pdf. Lane, M. E., Kirschen, P. H., and Vogel, R. M. (1999). “Indicators or impacts of global climate change on U.S. water resources.” Journal of Water Resources Planning and Management, 125(4), 194-204. Lettenmaier, D. P., Wood, A. W., Palmer, R. N., Wood, E. F., and Stakhiv, E. Z. (1999). “Water resources implications of global warming; a U.S. regional perspective.” Climatic Change, 43, 537-579. Liang, X. D., Lettenmaier, D. P., Wood, E. F., and Burges, S. J. (1994). “A simple hydrologically based model of land surface water and energy fluxes for GCMs.” Journal of Geophysical Research, 99(D7), 14415-14428. Lohmann, D., Raschke, E., Nijssen, B., and Lettenmaier, D. P. (1998a). “Regional scale hydrology I: formulation of the VIC-2L model coupled to a routing model.” Hydrological Sciences Journal, 43(1), 131-141. Lohmann, D., Raschke, E., Nijssen, B., and Lettenmaier, D. P. (1998b). “Regional scale hydrology II: application of the VIC-2L model to the Weser River, Germany.” Hydrological Sciences Journal, 43(1), 143-157. Maurer, E. P., O'Donnell, G. M., Lettenmaier, D.P., and Roads, J.O. (2001). “Evaluation of the land surface water budget in NCEP/NCAR and NCEP/DOE AMIP-II reanalyses using an off-line hydrologic model.” J. Geophys. Res., in press.


Nijssen, B. N., Lettenmaier, D. P., Liang, X., Wetzel, S. W., and Wood, E. F. (1997). “Streamflow simulation for continental-scale river basins.” Water Resour. Res., 33(4), 711-724. Nijssen, B. N., Schnur, R., and Lettenmaier, D. P. (2000). “Global retrospective estimation of soil moisture using the VIC land surface model, 1980-1993.” J. Clim. (in press). U.S. Bureau of Reclamation. (1999). Central Valley Project Improvement Act: final programmatic environmental impact statement. U.S. Department of the Interior, Sacramento, CA, October, 1999. Washington, W. M., and coauthors. (2000). “Parallel Climate Model (PCM) control and transient simulations.” Climate Dynamics, in press. Watson, R. T., Zinyowera, M. C., and Moss. R. H. (1996). Climate Change: Impacts, adaptations, and mitigation of climate change, 1995, 879 p., Cambridge University Press. Wilby, R. L. and Wigley, T. M. L. (1997). “Downscaling General Circulation Model output: a review of methods and limitations.” Prog. Phys. Geog., 21, 530-548. Wilby, R. L., Hay, L. E., and Leavesley, G. H. (1999). “A comparison of downscaled and raw GCM output: implications for climate change scenarios in the San Juan River basin, Colorado.” J. Hydrol., 225, 67-91.
