A hallmark of MAP (the Mesoscale Alpine Programme) was the availability of control simulations on a 350 x 300 mesh at 3 km horizontal resolution centered over the Alps using the Canadian MC2 model (Benoit et al. 1997). Output from these runs was made available in near real-time and used to guide design and scheduling of research aircraft and ground-network operations during the experiment. This experience led to a widespread impression that MC2 possessed considerable skill in anticipating the timing, location and mode of significant orographic precipitation and associated disturbances in the mesoscale flow field. Statistical evaluations of MC2 performance for the entire MAP experiment period have revealed an overall tendency for MC2 to under-estimate precipitation rates over mountainous terrain (Arena et al., 2001). A notable exception to this bias, however, was observed during MAP IOP 8 (20-21 October 1999). We find that a marked underestimate of low-level upstream orographic blocking in the MC2 control simulation led to significant over-prediction of precipitation over the Alpine slopes in this case, as observed rainfall was relatively light and maximized upstream of the Alps over the Po basin. Horizontally extensive airborne Doppler radar measurements, stretching from the higher south-facing Alpine slopes southward to the Mediterranean Sea, are particularly well suited for evaluation of model performance in a barrier-scale sense. Moreover, airborne Doppler radar data provide a remarkable extension of vector wind observations into deep Alpine valleys (including the Toce, Maggia and Val d'Aosta) surrounding the LMTA.

In this paper we focus on a quantitative evaluation of MC2 wind and precipitation fields for MAP IOP 8. Sensitivity to boundary conditions (viz. as specified by the operational Swiss Model vs. ECMWF) and recent improvements in handling of the vertical coordinate are also evaluated. Frequent realizations of airflow and S-band reflectivity patterns over the Lago Maggiore Target Area (LMTA) by ground-based Doppler radars allow for a more complete comparison of secular trends in the modeled vs. observed winds and precipitation over this long-duration event. A pivotal aspect of our approach is the re-interpolation of hourly MC2 output to a common terrain-relative grid shared by radar analyses accomplished via the University of Washington's "Mountain Zebra" data processing and display system.