The influence of interyear and interdecadal variability on the hydrograph at the Dalles Dam from 1900 to 1998 have been summarized by Miles et al. (2000; see also Hamlet and Lettenmaier, 1999). A dominant source of the inter-year variability in flows has been driven by the climate variability associated with El Niño-Southern Oscillation (ENSO) and La Niña conditions. The Pacific Decadal Oscillation (PDO) also drives variability of flows (Miles et al., 2000). These two large-scale climatic drivers (ENSO and PDO) can interact to affect the lowest and the highest flows. Although these climate change drivers are important and must be noted, a detailed analysis of them was beyond the scope of this report.

Prospective future climate changes (driven by greenhouse gas emissions) have been simulated, with many simulation model results suggesting that the water supply of the Columbia River may be reduced in the next half century. Scenarios of future changes in the Columbia River hydrograph suggest that future warming will move the river toward conditions, on average, that closely resemble conditions observed during the warm phases of ENSO and PDO during the last century (Hamlet and Lettermaier, 1999; Miles et al., 2000). These simulations were generated with two general circulation models (GCMs) for the years 2025, 2045, and 2095 using expected rates of carbon dioxide emissions. One model was from the Max Plank Institute in Germany and the other was the Hadley 2 model from the Hadley Center in the United Kingdom. Both models indicate warming in all months relative to historical air temperature from 1961 to 1997. For 2045 the projected air temperature increases in individual months range from about 1° to about 4°C. The fact that the Hadley 2 model projects wetter conditions than observed historically especially in summer and fall, while the Max Planck model projects dryer conditions in the summer and fall, demonstrates the uncertainties associated with climate change model projections of changes in precipitation associated with temperature increases. As noted, the models are more consistent in projecting temperature increases.