On the spatial extent of extreme rainfall events over India during the summer monsoon

Abstract

The spatial extent of extreme rainfall events directly affects the damages caused and impacts of such events. However, it has been less explored in the literature. This thesis examines extreme rainfall events of different sizes over central India during the Indian summer monsoon for the period 1951-2015. It utilizes a 2D connected component labelling algorithm to identify the sizes of events in the daily 1 1 gridded rainfall dataset provided by the India Meteorological Department. In the first part, this work revisits the observed trends in extreme rainfall events with the improved definition of events that accounts for their sizes. It shows that while 60% of the fractional increase in the area covered by daily extreme rainfall during the study period is due to a rise in the number of events, the rest is contributed by their increasing size. The increase after 1990 is, however, mainly due to the increase in the average size and not the frequency of their occurrences. This reveals the changing spatial characteristics of rainfall extremes over the study region that was not noted before. To get a mechanistic view of extremes of different sizes, we classify them as small (area . 104 km2 ), medium (104 > area . 7 104 km2) and large events (area > 7 104 km2 ). A majority of these events (>80%) are associated with synoptic-scale monsoon low-pressure systems (LPSs) but are accompanied by different background conditions. In the second part, I propose a physical mechanism for large-sized extreme rainfall events. All of the largesized events are produced by LPSs within 400 km of their center, with a clear preference to the south-western sector. Another consistent synoptic feature for large-sized events are the presence of secondary cyclonic vortices (SCVs) to the west of LPSs and the extratropical forcing. The interaction of two cyclonic vortices forms conditions favourable for long-lived, organized, and slow-moving convective systems that produce large-sized extreme rainfall events in the region between them. These events are preceded by the extratropical upper-level trough intrusion a week before. This trough moves eastward and is replaced by a stationary high. This midlatitude stationary high comes in phase with the equatorial monsoon trough. It provides eddy momentum, dynamic forcing, and static instability that strengthens the equatorial monsoon trough. The SCV and LPSs embedded in the monsoon trough are strengthened subsequently. In the third part, the thesis examines the characteristics and background conditions of LPSs that produce extreme rainfall events of different sizes and when LPSs do not produce extreme rainfall. The analysis reveals that medium and large event-producing LPSs tend to occur during the positive phase of monsoon intra-seasonal oscillations (MISO) when an active monsoon trough is present over central India. The LPSs that produce small events or no extreme rainfall event occur mainly during the neutral or negative MISO phases. The monsoon trough sets up large scale dynamic forcing and strengthens the low-level cyclonic circulation through the diabatic generation of potential vorticity. The LPSs of the positive MISO phase intensify along their track by interacting with the low-level jet. These conditions help in the organization of convection and lead to medium and large size events. The LPSs of the negative or neutral phases of MISO do not intensify much and trigger scattered convection, leading to small-size events or no extreme rainfall. The extreme-rainfall-producing LPSs are slower, moister, and stronger than the LPSs that do not produce extreme rainfall. These results connote a modulation of the precipitation response of monsoon lows by MISO. We explore the possible factors responsible for the observed changes in the final part. We find that changes in high-frequency intraseasonal oscillations and synoptic variability coincide with the changes in EREs. Over the western part of the Indo-Gangetic plains, their variance is decreasing, concurrently reducing small and medium EREs over this region. On the other hand, their increase over central India facilitates the rise in medium and large EREs. Large-scale circulation patterns are becoming more conducive for medium and large EREs. In addition, the southward shift of LPS tracks partly explains the increasing trend of medium and large events over the southern latitudes (south of 20 N) of central India and the decrease of small and medium EREs over the northern latitudes (north of 20 N). This work shows that the consideration of spatial collocation of heavy rainfall events while defining them gives an improved understanding of the underlying physical processes and observed trends than using the traditional grid-based definition