Precipitation-snowmelt timing and snowmelt augmentation of large peak flow events, western Cascades, Oregon

摘要

This study tested multiple hydrologic mechanisms to explain snowpack dynamics in extreme rain-on-snow floods, which occur widely in the temperate and polar regions. We examined 26, 10 day large storm events over the period 1992-2012 in the H.J. Andrews Experimental Forest in western Oregon, using statistical analyses (regression, ANOVA, and wavelet coherence) of hourly snowmelt lysimeter, air and dewpoint temperature, wind speed, precipitation, and discharge data. All events involved snowpack outflow, but only seven events had continuous net snowpack outflow, including three of the five top-ranked peak discharge events. Peak discharge was not related to precipitation rate, but it was related to the 10 day sum of precipitation and net snowpack outflow, indicating an increased flood response to continuously melting snowpacks. The two largest peak discharge events in the study had significant wavelet coherence at multiple time scales over several days; a distribution of phase differences between precipitation and net snowpack outflow at the 12-32 h time scale with a sharp peak at /2 radians; and strongly correlated snowpack outflow among lysimeters representing 42% of basin area. The recipe for an extreme rain-on-snow event includes persistent, slow melt within the snowpack, which appears to produce a near-saturated zone within the snowpack throughout the landscape, such that the snowpack may transmit pressure waves of precipitation directly to streams, and this process is synchronized across the landscape. Further work is needed to understand the internal dynamics of a melting snowpack throughout a snow-covered landscape and its contribution to extreme rain-on-snow floods.