Role of climate change in urban flood-relevant sub-daily rainfall extremes in India

Sub-daily rainfall extremes that drive rapid urban flooding are expected to intensify under anthropogenic climate change. Yet, their attribution remains uncertain due to limited observation records, lack of adequate representation of relevant physical processes and coarse spatio-temporal resolution of climate models. For a rapidly-urbanizing highly-populated country such as India with ambitious growth targets, such extremes are critical as urban flooding is often associated with significant loss of lives, environment and socio-economic damage. We assess the contribution of human-induced climate change to sub-daily extreme rainfall and its implications for urban flooding over the high-density heritage city of Ahmedabad. Two Inter-Sectoral Impact Model Intercomparison Project (ISIMIP3b) ensemble outputs, Hist-Nat (historical natural, a counterfactual driven only by solar and volcanic forcing) and Hist (factual, natural plus anthropogenic forcing) are bias-adjusted and statistically downscaled using ISIMIP3-BASD on six most-recent generation models at 0.25° resolution. Temporal disaggregation of rainfall from daily to hourly scales is carried out using a simple, yet effective k-nearest neighbour (kNN) approach evaluated against observations. Rainfall Intensity-Duration-Frequency (IDF) curves are derived for various return periods relevant to urban flood management. Observations show significant increases in short-duration rainfall intensities for Ahmedabad, ranging from 2.9% to 49.1% across different return periods with rarer events showing larger intensifications.

However, model simulations aren’t consistent with each other in terms of nature of change in rainfall extremes, resulting in equivocal attribution conclusions. While the multi-model mean suggests anthropogenic forcing has intensified short-duration rainfall extremes (1-13 hours) by 5-10% and reduced long-duration events (14-24 hours) by approximately 15%, individual models show divergent responses. These findings highlight limitations of current global climate models in attributing sub-daily rainfall extremes to climate change in the Indian monsoon region where fidelity of such models have been questioned by earlier regional studies on seasonal means. It is interesting to note, however, that based on observations alone, short duration high intensity rainfall extremes are found to be rising in this city, concurrent with expansion of built-up areas, thereby increasing exposure of urban population and environment to the risk of flooding.