90 likes | 187 Views
Tom.Pagano@por.usda.gov. Long-term Trends in Water Supply Forecast Skill. Are there any long-term trends in April 1 st water supply forecast skill? If so, where, when and why?. Average skill of all sites over 20-yr moving window For the West, as a whole,
E N D
Tom.Pagano@por.usda.gov Long-term Trends in Water Supply Forecast Skill Are there any long-term trends in April 1st water supply forecast skill? If so, where, when and why?
Average skill of all sites over 20-yr moving window For the West, as a whole, skill peaked in ~1965-1985 then slumped afterwards Good Bad
Average skill of all sites over 20-yr moving window For the West, as a whole, skill peaked in ~1965-1985 then slumped afterwards Good Bad Good Bad “Synthetic” hindcasts reproduce this feature almost exactly Therefore it’s not the human forecaster
Recent skill anomalies Where has skill been going down? Mostly in Oregon, Colorado, AZ/NM Doing fine in CA/NV/UT Better than usual +.15 0.0 -.15 -.30 Worse than usual
20-year moving window Spring precipitation “irregularity” More than 1 = more extreme than usual Less than 1 = Calm, reliably near-normal Westwide average of 29 basins Calm Extreme
Calm Typical Extreme 1961-80 1981-00 Where is spring precip more irregular? Now, especially in PNW and Southwest, whereas before it was very calm This matches decline in forecast skill
Of course, the $100,000 question: Will trend continue or return to normal?
Issues of climate non-stationarity Are rising temperatures a problem for water supply forecasts? Or just a problem for water managers?
Summary Skill trends evident, trends dominated by climate Flows more variable, persistent Spring precip is more extreme Need to reconsider assumptions about climate stationarity, in context of “soil moisture” Need to know how a warmed climate affects regression forecasts