Abstract
The prediction of drought onset and decay in the U.S. Corn Belt region (CBR) on seasonal-to-subseasonal time scales has not been well studied. This study utilizes the subseasonal-to-seasonal prediction archive to assess model errors in large-scale circulation patterns associated with agricultural drought transition periods, targeting models used by the European Centre for Medium-Range Forecasts, National Centers for Environmental Prediction, and Australian Bureau of Meteorology. An analysis of the seasonal cycle of bias for geopotential anomalies at 200 hPa and net radiation at the top of the atmosphere in each model is presented and used to subtract the long-term bias from each model. Model fields are decomposed into three spectral bands—low frequency (periods > 100 days), intraseasonal (periods 20–100 days), and synoptic (periods < 20 days)—to demonstrate each model’s ability to predict patterns associated with agricultural drought transition periods in each band. Results demonstrate that ECMWF and NCEP struggle in predicting the large-scale circulation patterns associated with 20-day agricultural drought and onset transitions, but are more skillful in predicting the patterns associated with 60-day agricultural drought onset and decay events at reforecast hour lead window 360–480 (F360–F480). BoM was not skillful in predicting the circulation patterns associated with either type of drought transition. Results also demonstrate that the errors associated with these events are no worse than historical errors for the target study period.
© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).
