Downscaling river discharge to assess the effects of climate change on cholera outbreaks in the Bengal Delta
Date of Original Version
Endemic cholera in the Bengal Delta region of South Asia has been associated with asymmetric and episodic variability of river discharge. Spring cholera was found to be related to intrusion of bacteria-laden coastal seawater during low flow seasons. Autumn cholera was hypothesized to result from cross-contamination of water resources when high river discharge causes massive inland inundation. The effect of climate change on diarrheal diseases has not been explored, because of the difficulties in establishing linkages between coarse-resolution global climate model outputs with localized disease outbreaks. Since rivers act as corridors for transport of cholera bacteria, the first step is to understand the discharge variability that may occur with climate change and whether it is linked to cholera. Here, we present a framework for downscaling precipitation from global climate models for river discharge in the Ganges-Brahmaputra-Meghna basin. Using a data-mining method that includes particle swarm optimization-based support vector regression, precipitation was downscaled for a statistical multiple regressive model to estimate river discharge in the basin. Key results from an ensemble of HadCM3, GFDL, and ECHAM5 models indicated 8 and 7.5% increase in flows for the IPCC A1B and A2 scenarios, respectively. The majority of the changes are attributable to increases in flows from February through August for both scenarios, with little to no change in seasonality of high and low flows during the next century. The probability of spring and autumn cholera is likely to increase steadily in the endemic region of the Bengal Delta.
Nasr-Azadani, Fariborz, Avinash Unnikrishnan, Ali Akanda, Saiful Islam, Munir Alam, Anwar Huq, Antarpreet Jutla, and Rita Colwell. "Downscaling river discharge to assess the effects of climate change on cholera outbreaks in the Bengal Delta." Climate Research 64, 3 (2015): 257-274. doi:10.3354/cr01310.