Biennial characteristics of Indian monsoon rainfall

Abstract

Basic diagnoses of summertime all-India rainfall (AIR) are conducted in an effort to evaluate the hypothesis that asserts the biennial alternation between successive strong and weak Indian monsoon summer seasons. The associations between subsets of the 10, 20, and 30 strongest and weakest monsoon seasons from 1871 to 2001, the mean rainfall in the summer seasons that precede and follow them, and the frequency with which the seasons are preceded or followed by strong (+10%)/weak (-10%) or surplus (above normal)/deficient (below normal) summer monsoon rainfall are assessed. The relationships are also stratified by the presence or absence of anomalous El Niño-Southern Oscillation (ENSO) conditions. Spectral and spatial comparisons between the monsoon's bienniality and ENSO teleconnection are then conducted using Indian district rainfall and wavelet decomposition. For the strongest monsoon seasons, statistically significant associations with the preceding summer rains are found that are consistent with the biennial hypothesis. In contrast, for years following the strongest monsoon seasons, and for years that both precede and follow the weakest monsoon seasons, significant biennial associations are largely absent. When the relationship with ENSO is considered, significant associations are found that act both in opposition to, and in support of, the biennial hypothesis. Often, strong La Niña seasons are preceded by weak or deficient summer rains, and weak El Niño seasons are followed by strong summer rains. The sequences identified are associated with transitions between El Niño and La Niña conditions. However, in contradiction to the biennial hypothesis, it is also found that following strong La Niña seasons, average monsoon rainfall lies significantly above its climatological average, and the frequencies with which both deficient and weak monsoon seasons occur are significantly below their expectation values. During years in which ENSO conditions are near neutral, evidence of bienniality is poor. Mean anomalies in summer rainfall preceding and following anomalous monsoon seasons during neutral ENSO conditions are either very weak or are not statistically discernable from chance. In addition, the frequency with which weak/neutral monsoon seasons are preceded by strong rains lies significantly below its expectation value, suggesting interannual persistence rather than bienniality. Only the frequency with which strong monsoon seasons are followed by deficient rains is found to be both statistically significant and consistent with the biennial hypothesis. Comparisons of the spectral and spatial characteristics of rainfall's bienniality and the correlation between rainfall and Niño-3 SST further illustrate the strong correspondence that exists between biennial monsoon variability and ENSO. The temporal evolution of biennial power in the monsoon is found to evolve closely with that of ENSO over the past 50 yr and the spatial distribution of bienniality in India is shown to mirror that of ENSO's teleconnection. The analysis suggests that theories attempting to describe the physical mechanisms that underlie monsoon bienniality must also begin to address the asymmetric and episodic nature of the monsoon's year-to-year variability. Moreover, the results bring into question the hypothesis that bienniality is an inherent mode of variability for a closed Indian Ocean-monsoon system. © 2004 American Meteorological Society.