Forecasting during the Lake-ICE/SNOWBANDS Field Experiments

Abstract

Despite improvements in numerical weather prediction models, statistical models, forecast decision trees, and forecasting rules of thumb, human interpretation of meteorological information for a particular forecast situation can still yield a forecast that is superior to ones based solely on automated output. While such time-intensive activities may not be cost effective for routine operational forecasts, they may be crucial for the success of costly field experiments. The Lake-Induced Convection Experiment (Lake-ICE) and the Snowband Dynamics Experiment (SNOWBANDS) were conducted over the Great Lakes region during the 1997/98 winter. Project forecasters consisted of members of the academic as well as the operational forecast communities. The forecasters relied on traditional operationally available data as well as project-tailored information from special project soundings and locally run mesoscale models. The forecasting activities during Lake-ICE/SNOWBANDS are a prime example of how the man?machine mix of the forecast process can contribute significantly to forecast improvements over what is available from raw model output or even using traditional operational forecast techniques.