Evaluating the performance of short-term climate forecasts

The economic impacts of weather and climate extremes have generated increased interest in the feasibility of medium to long-range forecasts of weather and climate on the scale of weeks, months and seasons. The TASC Center for Atmospheric, Space, and Oceanographic Modeling and Forecasting has established a climate modeling capability and is currently evaluating the accuracy and skill of climate forecasts on these time scales.

We will present recent results of inter-model comparisons and comparisons between observed weather data and forecasts of variable lead times. Keys areas of focus will be: 1) What are the best time averages to forecast (i.e., weekly, bi-weekly, monthly, or seasonal means); 2) What is the greatest forecast lead time for a given time average; 3) Are certain years more accurately forecasted (i.e., El Nino vs. La Nina); and 4) What are the most effective means for downscaling global gridded climate forecasts to specific point locations.