Forecast Skill of the South American Monsoon System


Charles Jones
Earth Research Institute, University of California, Santa Barbara, Santa Barbara, California

Leila M.V. Carvalho
Earth Research Institute, and Department of Geography, University of California, Santa Barbara, Santa Barbara, California

Brant Liebmann
Cooperative Institute for Research in Environmental Sciences Climate Diagnostics Center, Boulder, Colorado

(Manuscript received 1 October 2011, in final form 1 December 2011)

Abstract

The South American monsoon system (SAMS) is the most important climatic feature in South America and is characterized by pronounced seasonality in precipitation. This study uses the National Centers for Environmental Prediction Climate Forecast System, reforecasts version 2 (CFSRv2), to investigate the skill of probabilistic forecasts of onset and demise dates, duration, and amplitude of SAMS during 1982-2009. A simple index based on the empirical orthogonal function of precipitation anomalies is employed to characterize onsets, demises, durations, and amplitudes of SAMS. The CFSv2 model has useful skill to forecast seasonal changes in SAMS. Probabilistic forecasts of onset and demise dates have 16.5% and 43.3% improvements, respectively, over climatological forecasts. Verification of hindcasts of durations and amplitudes of SAMS shows relatively small biases and root-mean-square errors.