The Use of Snow-Level Observations Derived from Vertically Profiling Radars to Assess Hydrometeorological Characteristics and Forecasts over Washington’s Green River Basin

Two vertically pointing S-band radars (coastal and inland) were operated in western Washington during two winters to monitor brightband snow-level altitudes. Similar snow-level characteristics existed at both sites, although the inland site exhibited lower snow levels by ~70 m because of proximity to cold continental air, and snow-level altitude changes were delayed there by several hours owing to onshore translation of weather systems. The largest precipitation accumulations and rates occurred when the snow level was largely higher than the adjacent terrain. A comparison of these observations with long-term operational radiosonde data reveals that the radar snow levels mirrored climatological conditions. The inland radar data were used to assess the performance of nearby operational freezing-level forecasts. The forecasts possessed a lower-than-observed bias of 100–250 m because of a combination of forecast error and imperfect representativeness between the forecast and observing points. These forecast discrepancies increased in magnitude with higher observed freezing levels, thus representing the hydrologically impactful situations where a greater fraction of mountain basins receive rain rather than snow and generate more runoff than anticipated. Vertical directional wind shear calculations derived from wind-profiler data, and concurrent surface temperature data, reveal that most snow-level forecast discrepancies occurred with warm advection aloft and low-level cold advection through the Stampede Gap. With warm advection, forecasts were too high (low) for observed snow levels below (above) 1.25 km MSL. An analysis of sea level pressure differences across the Cascades indicated that mean forecasts were too high (low) for observed snow levels below (above) 1.25 km MSL when higher pressure was west (east) of the range.