Warming climate brings reduction of key water supply in the western United States

As water becomes an increasingly scrutinized resource for California and the western United States, new research by scientists at Scripps Institution of Oceanography at the University of California, San Diego, estimates that vital water resources derived from the Sierra Nevada may suffer a 15- to 30-percent reduction in the 21^st century.

These conclusions are based on both an analysis of historical river streamflow records and simulations of future snowmelt runoff for the next 100 years. According to these studies, the peaks of snowmelt and river flows from the Sierra Nevada are beginning to occur progressively earlier each season due to warming climate conditions.

This hastening of the seasonal cycle of outflows from the Sierra Nevada creates a tendency for more of the flow to occur before reservoirs and water users can safely capture and use it for water-supply purposes. The likely net effect is a decrease in the amount of usable surface water in addition to an extended summer drought period for California, as well as other states in the western United States.

California’s 400-mile-long Sierra Nevada Mountains, the state’s largest single source of water, supply water for agricultural, industrial, and residential uses. The new study simulates responses to climate change in three representative Sierra Nevada rivers, the Merced, Carson, and American. "April through July is the key period in terms of real water supply each year. In the simulations we’ve developed, climate warming causes the river flow to start a month early, resulting in a 15- to 30-percent reduction in the useable part of the water runoff," said Michael Dettinger, a research hydrologist with Scripps Institution and the Untied States Geologic Survey.

The simulation is the latest product derived from the Accelerated Climate Prediction Initiative (ACPI), a multi-institution effort including government laboratories and universities working to create state-of-the-art climate-modeling capability for the United States (funded by the Office of Biological and Environmental Research, U.S. Department of Energy).

The new study comes at a time when California’s water resource managers are increasingly accessing research from climate studies to guide future management directions. The shift to early peaks in snowmelt and streamflow highlights the state’s need to adapt to changing conditions.

"It’s not just that the snowmelt is shifting earlier, we’re seeing increases in winter runoff because winter storms are warmer," said Dan Cayan, director of the Scripps Climate Research Division and a co-researcher on the study. "Reservoirs serve a dual-purpose in California and the west, one for water storage and two for flood protection. So this work helps to elucidate changes associated with a different mix of rain and snow. We will have to deal with new conditions such as a shorter winter and not as much snowpack, in addition to a longer summer when there may be lower streamflows along with increases in the demand for water."

A separate study conducted by Scripps post-doctoral researcher Iris Stewart, with Cayan and Dettinger, has found that the shift to an increasingly earlier runoff timing that is predicted by the hydrologic models for the coming century can already be documented in the historical records of the last 50 years. Based on an analysis of 300 very complete snowmelt stream gauge records throughout western North America, Stewart found a shift similar to the one in the future-looking model studies.

The culprit behind these observed changes appears to be the same as that in simulations–a trend toward warmer winters and springs–although in the Stewart, Cayan, and Dettinger study it is not clear whether these can be ascribed to natural variations or anthropogenic change.

"From this analysis, the change in snowmelt is true not only for California and the western United States, but for western Canada and western Alaska as well," said Stewart. "The results indicate an important part of the variability of streamflow timing derives from changes in the pattern of atmospheric circulation over the Pacific and North America, resulting changes in temperature, and to a lessor extent, precipitation."

A key issue in this research is the question of whether the observed trends are a product of a natural climate variability or the result of temperature and precipitation changes due to global warming.

To address this question, two future climates were analyzed in Dettinger’s study using ACPI models developed under Warren Washington at the National Center for Atmospheric Research. In the "business-as-usual" case, greenhouse gases and aerosols are added and factored at a pace consistent with recent trends. In this simulation a noticeable warm pattern emerges in the 1980s and 1990s. But Dettinger and his colleagues wondered if such a warming could be caused by a natural fluctuation of the climate system. Thus they used a second simulation that kept greenhouse and aerosol factors constant at 1995 levels. In this case the climate and streamflow conditions remain at levels of the 80s and 90s for the next 50 years.

"Having fixed the gases and aerosols the warming stays there despite the natural variability," said Dettinger. "This suggests, at least in the modeling world, the unusually warm character of the last decade or so is greenhouse warming."

"I think the science provided by these kinds of experiments is quite useful for very real problems," said Cayan. "These are the situations our children’s generation and those that follow are going to have to contend with."

# # #

The Accelerated Climate Prediction Initiative (ACPI) Pilot effort was supported largely by the Office of Biological and Environmental Research, US Dept. of Energy (DOE) through numerous contracts and subcontracts to the participants. The production runs of global and regional models were largely accomplished with resources of the Center for Computational Sciences (CCS) at Oak Ridge National Laboratory, which is supported by the Office of Science of the U.S. Department of Energy. The San Diego Supercomputer Center provided dedicated machine resources for the crucial ocean initialization component of the ACPI Pilot. These ocean data were provided by ECCO state estimation activity at Scripps Institution of Oceanography. Los Alamos National Laboratories provided computer resources for PCM work and additional feasibility tests associated with the ocean initialization. In addition to DOE support, many of the participating organizations also supported the effort. These include the Scripps Institution of Oceanography, the U.S. Geological Survey, Department of Defense, NCAR, the University of Washington, and Battelle Pacific Northwest National Laboratories.

Stewart is supported by a University Corporation for Atmospheric Research postdoctoral fellowship and Department of Energy climate research funding. In addition to ACPI funding, Cayan is funded by the National Oceanic and Atmospheric Administration Office of Global Progams via the California Applications Program.