Precipitation-Snowmelt Timing and Snowmelt Augmentation of Large Peak Flow Events, Western Cascades, Oregon

Abstract:

Extreme rain-on-snow floods are known to result from snowmelt coincident with precipitation, but comparatively little is known about the relative timing of these factors within storm events. Cumulative net snowmelt (hourly, from a snowmelt lysimeter) was plotted against precipitation for 26 large storms (> 1-yr return period) over the period 1991-2012 in the transient snow zone of the H.J. Andrews Experimental Forest in the western Cascades of Oregon. The relative timing of precipitation and net snowmelt at the hourly time scale was assessed with wavelet coherence.

Five precipitation-net snowmelt response categories were identified: flat; persistent melt; persistent snow accumulation; late melt; and late snow accumulation. Persistent melt events were characterized by increasing cumulative net snowmelt and precipitation and had the highest mean peak flow and water available for runoff. Both the persistent melt and persistent snow accumulation categories had large, contiguous regions of significant wavelet coherence at multiple temporal scales, but pulses of precipitation preceded pulses of snowmelt in the persistent melt events, whereas precipitation was absorbed by the snowpack in the persistent accumulation category.

A dewpoint temperature consistently above 0.5°C, elevated wind speeds, and a high fraction of precipitation falling as rain in the persistent melt category facilitated rapid snowmelt rates. During the two extreme rain-on-snow events in the sample, snowmelt was significantly synchronized with precipitation at 1-h to 64-h time scales throughout the 10-day event duration. Event categorization and analysis of wavelet coherence between precipitation and snowmelt can help predict peak discharge magnitude.