Forecasting during the Lake-ICE/SNOWBANDS Field Experiments

Despite improvements in numerical weather prediction models, statistical mo dels, forecast decision trees, and forecasting rules of thumb, human interp retation of meteorological information for a particular forecast situation can still yield a forecast that is superior to ones based solely on automat ed output. While such time-intensive activities may not be cost effective f or routine operational forecasts, they may be crucial for the success of co stly field experiments, The Lake-Induced Convection Experiment (Lake-ICE) a nd the Snowband Dynamics Experiment (SNOWBANDS) were conducted over the Gre at Lakes region during the 1997/98 winter. Project forecasters consisted of members of the academic as well as the operational forecast communities. T he forecasters relied on traditional operationally available data as well a s project-tailored information from special project soundings and locally r un mesoscale models. The forecasting activities during Lake-ICE/SNOWBANDS a re a prime example of how the man-machine mix of the forecast process can c ontribute significantly to forecast improvements over what is available fro m raw model output or even using traditional operational forecast technique s.