Past Climate Change and GW Pumping Influences on Historic Streamflow Traces and the Consequences of Using Historic Streamflow Traces in Operations Planning
Every winter and spring, reservoir operators gamble on the timing and quantity of runoff. They rely on experience and statistics to achieve a balance between creating space for flood control and filling reservoirs for power, municipal, agricultural, and instream flow uses. The statistical analysis--which is key to successful operations--is nearly always constructed from historic (past) streamflows.
* Are there consequences to using historic streamflows to evaluate operational decisions?
Researchers have determined that temperature and precipitation changes over the last century have shifted the timing of runoff and increased the frequency and severity of droughts and winter flooding in the Pacific Northwest (PN) Region. The cumulative effects of irrigation practices and ground water pumping have also altered streamflows.
* Have climate change, irrigation practices, and ground water pumping influenced historic streamflows to the point that applying historic values does not provide a valid assessment of risk?
Need and Benefit
Reclamation operators evaluate volume streamflow forecasts, starting in winter or early spring, to predict the quantity of water expected during the upcoming flood control and fill season. Historic (past) daily streamflows are compared to the volume forecast (future) to help the operators judge the timing and shape of the anticipated runoff and prepare release schedules. Volume forecasts are developed by several groups including Reclamation, the National Weather Service (NWS), and the Natural Resources Conservation Service (NRCS). This research will investigate the consequences of using historic daily streamflows, as opposed to streamflows adjusted for past climate change and irrigation development, to determine future operations given future volume forecasts.
Volume forecasts are continually updated through the season and represent the residual anticipated reservoir inflows and natural streamflows from the current date through an anticipated fill date. The forecasts are often presented as probability of exceedance, say at the 10, 50 and 90 percentile levels. The volume forecasts are seasonal, however, and do not provide a daily distribution of streamflows.
Reclamation operators use several approaches to estimate the daily timing and shape of the runoff predicted by the volume forecasts and to prepare reservoir releases. One approach is to select historic daily streamflow patterns from years which experienced runoff volumes similar to the selected forecast volume level. An operations schedule can then be developed assuming each selected year's historic daily streamflows are equally likely to occur (again) during the upcoming forecast season. Another approach is to shape the forecast volume to produce a daily streamflow hydrograph that is similar to the selected historic daily streamflows. Other statistical approaches produce synthetic daily streamflows, but they also rely on the historic daily streamflows as a foundation.
But historic daily streamflows do not represent present level conditions. Embedded in historic daily streamflow values are the gradual effects of past climate change and the cumulative effects of past irrigation development. Observed temperature and precipitation patterns over the last century have produced less winter snow accumulation, higher winter streamflows, earlier spring snowmelt and earlier peak spring streamflow across the Pacific Northwest Region. Ground water pumping and changes in surface water irrigation practices have reduced and shifted streamflows, sometimes significantly. In the Snake River basin, ground water depletions have gradually grown to over one million acre feet annually, changing the quantity and timing of streamflows.
The 2007 operations studies for the Upper Snake will be used as a case study in the proposed research. Reclamation's Palisades Reservoir in the Upper Snake River basin failed to fill in 2007. Operators had predicted a probability of refill greater than 90 percent and consequently released additional water throughout the winter for maximum channel and riparian habitat support. Even though the streamflows eventually experienced fell well below the 10th percentile of the volume forecast, the calculated probability of reservoir refill may have been more conservative if the historic daily streamflows applied to that volume forecast had been adjusted for the past effects of climate change, irrigation practices, and ground water depletions.
Contact the Principal Investigator for information about these documents.