GC53B-1201
Has global warming changed timing of winter-spring streamflows over North America?

Friday, 18 December 2015
Poster Hall (Moscone South)
Jonghun Kam, Geophysical Fluid Dynamics Laboratory, Princeton, NJ, United States; Princeton University, Princeton, NJ, United States, Thomas R Knutson, NOAA Princeton, Princeton, NJ, United States and P C D Milly, USGS, Princeton, NJ, United States
Abstract:
Wherever snowmelt runoff substantially contributes to winter-spring streamflows, warmer winter-spring temperature can accelerate snow melt and reduce later streamflows. These changes can adversely affect human activities and ecological communities (e.g. flood, drought, salmon survival rate, and blooming season). Here we investigate changes in timing of winter-spring streamflows over North America (NA) during 1933-2013 and 1951-2000 using observed streamflow and simulated runoff from pre-industrial (unforced) control and historical (realistically forced) runs from the Geophysical Fluid Dynamics Laboratory Climate Model version 3. The study regions are north of 41˚N in NA. We analyze winter-spring center of volume date (WSCV), the date by which half of the accumulated January through June daily streamflow volume occurs. We first performed a sliding trend analysis of WSCV for time periods starting in various years (1951 through 1984) and ending in 2000. We found that the observed decreasing trends (Theil-Sen slopes) of WSCV over the northeast and northwest U.S. regions are at the edge of detectability (i.e., lie near the edge of the 5th-95th percentile envelope of control runs) for trends beginning any time between 1950 and 1970, but are consistent with the envelope of historical runs for all beginning trend years. Interestingly, for the 1933-2013 analysis, results for the northwest U.S. show that the observed trends of WSCV are positive for periods beginning as early as the mid-1960s, and inconsistent with historical runs for periods beginning in the mid-1950s and later. Aside from this inconsistency, observed trends to 2013 are consistent with both control and historical runs. This study suggests that internal variability has played a major role in timing of winter-spring streamflows to date, despite global warming, and thus that clear detection and attribution of WSCV trends in the study regions may require longer streamflow records than those now available.