92nd American Meteorological Society Annual Meeting (January 22-26, 2012)

Thursday, 26 January 2012
Comparison of the 2011 South Asian Monsoon Onset Characteristics in the GFS and ECMWF Models
Hall E (New Orleans Convention Center )
Siddarth K. Plakkot, Howard University NOAA Center for Atmospheric Sciences, Washington, DC; and V. K. Kumar and D. Carlis

The physical characteristics of the onset of summer South Asian Monsoon (SAM) are examined in the operational NCEP Global Forecast System (GFS) and the ECMWF models. Predicting the large scale monsoon with global models not only provides immense value to the agrarian societies but also works as an underpinning to the understanding of the complex physical processes in the model.

The GFS and ECMWF analyses and 1-7 day forecasts are compared for the monsoon meridional circulation, precipitation distribution and amount, both verified against NCEP/NCAR reanalysis and the Climate Prediction Center (CPC) merged analysis precipitation (CMAP). In 2011, the India Meteorological Department (IMD) defined the onset of SAM on 29 May 2011.The onset phase of the SAM, characterized by strong cross equatorial west to southwesterly flow, is evident in both models' analyses and 5-day forecasts. An abrupt shift of the Intertropical Convergence Zone (ITCZ) from 5°S to 5°N with the associated vertical motions is exhibited in the 26-30 May pentad for both models. The vertical motions strengthen and extend further to 15°N in the 1-5 June pentad.

The CMAP precipitation in the 26-30 May pentad shows a broad ITCZ extending from 55°E-75°E centered on 8°N. The 7-day ECMWF precipitation forecast valid on the onset date exhibits similar characteristics to CMAP whereas GFS precipitation shows a narrow latitudinal band but wider longitudinal extent. The GFS shows enhanced precipitation over Western Ghats and Eastern India compared to ECMWF. Arabian Sea and Indian Ocean rainfall is over predicted for GFS compared to ECMWF. Both models show under prediction over land regions in southern India compared to verification. Precipitation forecasts in both models, in general, show deficiency in predicting the correct location and amount. Furthermore, we will examine the detailed precipitation statistics as well as dynamical aspects of the SAM.

Supplementary URL: