Rainfall and groundwater use in rural Kenya

This study examines the relationship between rainfall and groundwater use in rural Kenya, using automatically-transmitted hourly data from handpumps (n = 266), daily rainfall records (n = 19), and household survey data (n = 2508). We demonstrate a 34% reduction in groundwater use during the wet season compared to the dry season, suggesting a large shift from improved to unimproved sources in the wet season. By cross-correlating handpump and rainfall time series, we also reveal substantial short-term changes in groundwater pumping observed immediately following heavy rainfall. Further investigation and modelling of this response reveals a 68% reduction in pump use on the day immediately following heavy rain. We then investigate reasons for this behavioural response to rainfall, using survey data to examine the characteristics, concerns and behaviours of households in the area where the reduction in pump use was most marked. In this area rainwater harvesting was widespread and only 6% of households reported handpumps as their sole source of drinking water in the wet season, compared to 86% in the dry season. These findings shed light on the impact increasing rainfall variability may have on the Sustainable Development Goal of “universal and equitable access to safe and affordable drinking water for all”. Specifically, we suggest a flaw in the water policy assumption that the provision of improved sources of drinking water—in this case community handpumps—translates to consistent use and the associated health benefits. We note that failure to understand and account for actual water use behaviour may results in adverse public health outcomes and maladapted WASH policy and interventions. [Display omitted] •Groundwater use is inversely correlated to rainfall at seasonal and daily timescales.•A large short-term reduction in pumping is observed immediately following heavy rain.•This relationship between rainfall and pumping is modelled and tested.•The existence of improved water supplies does not guarantee their use.•The expected health gains of rural WASH systems may not be realised.