CMIP3 projected changes in the annual cycle of the South American Monsoon

Nine models from the Coupled Model Intercomparison Project version 3 dataset are employed to examine projected changes in the SouthAmericanMonsoon System annual cycle by comparing the 20th Century and SRES A2 scenarios. The following hypotheses are examined: (1) the warm season climate responses in the Southeast, Continental South Atlantic Convergence Zone (CSACZ) and Monsoon regions are related by regional circulation and moisture transport changes which, in turn, must be consistent with robust large-scale changes in the climate system, and (2) an increased threshold for convection in a warmer world may affect the timing of warm season rains. The present analysis reaffirms that the Southeast region is likely to experience increased precipitation through the warm season. Additional results exhibit more uncertainty due to large inter-model variance and disagreement in the A2 scenarios. Nevertheless several statistically significant results are found. In the Monsoon and to a lesser extent in the CSACZ region, the multi-model median suggests reduced precipitation during spring (Sep–Nov). These continental precipitation changes are accompanied by a southward shift of the maximum precipitation in the South Atlantic Convergence Zone. Changes in circulation include a poleward displaced South Atlantic Anticyclone (SAAC) and enhanced moisture transport associated with a strengthened northerly low level flow east of the Andes during spring. Moisture transport divergence calculations indicate unchanged divergence in the Monsoon region during spring and increased convergence in the Southeast throughout the warm season. The circulation and moisture transport changes suggest the increased precipitation in the Southeast during spring may be related to changes in the SALLJ and SAAC, which both enhance moisture transport to the Southeast. The seasonally dry Monsoon region is further affected by an increased threshold for convection in the warmer, more humid and stable climate of the 21st century, which combined with the circulation changes may weaken the onset of the rainy season. Although there is substantial variability among the models, and the results are represented by small changes compared with the multi-model variance, the statistical significance and consistency with expected robust large scale changes provide a measure of confidence in otherwise tentative results. Further testing of the relationships presented here will be required to fully understand projected changes in the South American Monsoon