Subseasonal variations in spatial signatures of ENSO on the Indian summermonsoon from 1901 to 2009

摘要

Correlations of 1° by 1° seasonal rainfall with Pacific sea surface temperatures (SSTs) reveal spatially distinct teleconnections between El Ni?o–Southern Oscillation (ENSO) and Indian summer monsoon rainfall over the full monsoon season, as well as three subseasons. Over the full season (June–September), Pacific SSTs correlate with rainfall in Western India more than that in Eastern India. This spatial signature shifts as the monsoon progresses through early (June), middle or peak (July–August), and late (September) subseasons. Specifically, a 1°C cooling of the central equatorial Pacific (i.e., La Ni?a conditions) can result in the following: ～70–100 increase in precipitation in north central Indian and the Indo-Gangetic Plains during the early season, ～30–80 increase peak season precipitation in south central India and northwestern Rajasthan, and ～60–100 increase in late season precipitation in northern, northwestern, and central India. Furthermore, the spatial signatures between La Ni?a and El Ni?o are asymmetric in that for a particular location, the enhancement and suppression of rainfall associated with La Ni?a and El Ni?o conditions, respectively, are not equal. El Ni?o suppresses peak season rainfall in the south central and northwestern Rajasthan regions more than La Ni?a enhances it, but the opposite occurs during the late season in northern, northwestern, and central India. Additionally, the correspondence of minima (maxima) in anomalies of velocity potential aloft with maxima (minima) at 925 mb and with positive (negative) surface pressure anomalies suggests that anomalous subsidence (ascent) occurs in July–September during El Ni?o (La Ni?a) times. In the early season, however, patterns of velocity potential composites suggest a region of descent (ascent) over the western equatorial Indian Ocean, along with a region of ascent (descent) over the Indian subcontinent that exists only during the early season but not during the peak or late season. These patterns are consistent with the hypothesis that local Hadley cell circulation affects pressure and thus rainfall during the early season but that a larger-scale mechanism, such as eastward or westward shifts in the Walker circulation, may be more responsible for teleconnections seen throughout the remainder of the season. These findings indicate that focusing monsoon forecasting efforts on these regions and on subseasonal periods while incorporating ENSO asymmetries will yield useful and skillful regional forecasts, compared to the declining utility and skill of all-India summer monsoon rainfall.