Assessment of operational forecasts of heavy precipitation in the northern Sierra Nevada mountains

The North American Mesoscale (NAM) model forecasts of heavy precipitation events along the northern Sierra Nevada mountains are assessed by comparing the area-average precipitation over a region that covers three watersheds on the west slope. An event is defined as a 24 h period wherein 38 mm or more fell in the region of interest. The average error (over all events) is an over prediction of 11, 7, and 3 mm for lead times of 24, 48, and 72 h. Individual events can have much greater errors, ranging from -166 to 66 mm. Cases with large dry biases have higher stability, faster low-level winds, higher precipitable water, and higher freezing levels than cases with large moist biases. The large dry-bias cases also have strong upstream gradients in humidity, stability, and wind speed, consistent with the differential deflection mechanism identified by previous authors. One of the root causes for errors in the dry-bias events is an under prediction in the upstream humidity. Causes for errors in the moist-bias events are still being determined at the time of this writing, however, the relatively low melting levels appear to be important as the greatest lifting and production of hydrometeors occurs at subfreezing temperatures. The microphysical parameterization scheme used by the NAM model tends to be rather aggressive at riming frozen particles. This acts to increase fall velocities and allow for less downstream advection and higher accumulations.