J12.17
Improved intraseasonal South Asian monsoon forecasts using a GCM by comparison to wavelet analysis of empirical precipitation data
Dan C. Collins, Georgia Institute of Technology, Atlanta, GA; and C. Hoyos and P. J. Webster
Much research has focused on interannual monsoon variability. Through wavelet analysis, it can be shown that intraseasonal monsoon variability is significant in determining total regional precipitation. In a previous study, the intraseasonal spectral band of climate predictor time series and precipitation and runoff predictands was separated, and a simple multivariate linear regression model was used to determine the strength of forcing of monsoon variability for various atmospheric variables and SST. Using this wavelet banded linear empirical model as a description of intraseasonal monsoon forcing, GCM monsoon forecasts can be improved to predict seasonal break and active monsoon periods by application of a model dependent algorithm. Wavelet analysis of model precipitation (and OLR) shows that the relative strength and timing of the intraseasonal band of the dynamical model does not always coincide with empirical precipitation data. Though the GCM may not accurately predict active and break periods in regional precipitation, principal components of GCM data show that some model variables reflect the intraseasonal variability found in empirical data. Criteria for active and break periods for GCM calculated variables are derived by correlating principal components of GCM data to variables of monsoon forcing in the empirical wavelet banded regression analysis. An improved forecast of intraseasonal monsoon rainfall variability is calculated from the GCM forecast and these objectively defined criteria. .
Joint Session 12, Subseasonal forecasting (Joint with 15th Symp. on Global Change and Climate Variations and the Symp on Forecasting the Weathe and Climate of the Atmosphere and Ocean; Room 6C)
Tuesday, 13 January 2004, 8:30 AM-5:30 PM, Room 6C
Previous paper Next paper