The sensitivity of snowmelt processes to climate conditions and forest cover during rain-on-snow: a case study of the 1996 Pacific Northwest flood

A warm, very wet Paci®c storm caused signi®cant ̄ooding in the Paci®c Northwest during February 1996. Rapid melting of the mountain snow cover contributed to this ̄ooding. An energy balance snowmelt model is used to simulate snowmelt processes during this event in the Central Cascade Mountains of Oregon. Data from paired open and forested experimental sites at locations at and just below the Paci®c Crest were used to drive the model. The event was preceded by cold, stormy conditions that developed a signi®cant snow cover down to elevations as low as 500 m in the Oregon Cascades. At the start of the storm, the depth of the snow cover at the high site (1142 m) was 1.97 m with a snow water equivalent (SWE) of 425 mm, while at the mid-site (968 m) the snow cover was 1.14 m with a SWE of 264 mm. During the 5±6 day period of the storm the open high site received 349 mm of rain, lost 291 mm of SWE and generated 640 mm of runo€, leaving only 0.22 m of snow on the ground. The mid-site received 410 mm of rain, lost 264 mm of SWE to melt and generated 674 mm of runo€, completely depleting the snow cover. Simulations at adjacent forested sites showed signi®cantly less snowmelt during the event. The snow cover under the mature forest at the high site lost only 44 mm of SWE during the event, generating 396 mm of runo€ and leaving 0.69 m of snow. The model accurately simulated both snow cover depth and SWE during the development of the snow cover prior to the storm, and the depletion of the snow cover during the event. This analysis shows that because of the high temperature, humidity and relatively high winds in the open sites during the storm, 60±90% of the energy for snowmelt came from sensible and latent heat exchanges. Because the antecedent conditions extended the snow cover to very low elevations in the basin, snowmelt generated by condensation during the event made a signi®cant contribution to the ̄ood. Lower wind speeds beneath the forest canopy during the storm reduced the magnitude of the turbulent exchanges at the snow surface, so the contribution of snowmelt to the runo€ from forested areas was signi®cantly less. This experiment shows the sensitivity of snowmelt processes to both climate and land cover, and illustrates how the forest canopy is coupled to the hydrological cycle in mountainous areas.# 1998 John Wiley & Sons, Ltd.