OBSERVED LEAD-LAG RELATIONSHIPS BETWEEN INDIAN-SUMMER MONSOON AND SOME METEOROLOGICAL VARIABLES

Lagged relationships between the Indian summer monsoon and several cli mate variables are investigated. The Variables examined are gridded fi elds of snow cover (14 years), sea surface temperature (41 years) and 500 hPa geopotential height north of 20 degrees N (42 years). We also used series of global air temperature (108 years) and Southern Oscilla tion index (112 years). Precipitation over all India during June-Septe mber over a 112 year period are used as Indian monsoon index. Emphasis is put on early monsoon precursors. In agreement with the tendency fo r a low frequency oscillation in the ocean-atmosphere system, several precursor patterns are identified as early as the year preceding the m onsoon. The most important key regions and seasons of largest correlat ions are selected and the corresponding series are used to perform a m onsoon prediction. The prediction shows however a relatively moderate score mainly due to the not highly significant correlations. To improv e the predictions we filtered the variables into their biennial (1.5-3 .5 years) and low frequency (3.5-7.5 years) modes. Correlations betwee n the monsoon and the filtered variables are higher than those obtaine d without filtering especially for the biennial mode. The two modes ar e out-of-phase before the monsoon and in-phase during and after. This phasing is found in all variables except for snow cover for which the two modes are in-phase before the monsoon and out-of-phase during and after. It is suggested that such phasing may be important for the form ation of snow and could explain the higher correlations when variables are concomitant or are lagging the monsoon. Early predictions of the monsoon based on those two modes show improved scores with highly sign ificant correlations with the actual monsoon.