Implications of Changes in Precipitation Amount and Pattern for Water Resources: Global Study of Rainwater Harvesting Systems

Climate change is expected to change the distribution, frequency and intensity of precipitation events, which can affect the reliability of renewable water resources. We compare the historical (1951-2010) changes in annual-mean and annual-maximum daily precipitation in a global set of weather station observations (GHCN-Daily) and bias-corrected precipitation projections of 5 global climate models from the Inter-Sectoral Impact Model Intercomparison Project (ISI-MIP), temporally and spatially subsampling the models as observations, and develop the study to the period 2011-2099 for model projections under the RCP8.5 forcing scenario. We also develop a Rainwater Harvesting System (RWHS) model and drive it with observational and modeled daily precipitation data to study the changes in reliability of the RWHS for each station in order to study the impact of changes in precipitation pattern on reliability of precipitation-based water supply. Results show that historical mean and maximum precipitation has increased at a rate of 7.64 and 10.14 % per K global warming, respectively, which is higher than the subsampled ISI-MIP models' average trends of 1.36 and 7.34 % per K, respectively. Despite the faster increase in maximum precipitation than mean precipitation, the reliability of the RWHS driven by observed precipitation has increased by an average of 0.2% per decade over 1951-2010. By contrast, all 5 ISI-MIP model daily precipitation series imply decreases in mean reliability over the station locations, for an average 0.15% per decade decrease over 1951-2010. By scaling the decadal precipitation to the initial decade's average, to factor out the impact of change in precipitation amount on the RWHS reliability, observations show an average 0.11%/decade increase while models show an average 0.13%/decade decrease in the reliability. Our results show that, compared to station observations, climate models underestimate the increasing trends in mean and maximum precipitation and also show the opposite direction of average change in the reliability of a model water supply system. These systematic mismatches for the recent decades suggest the need for caution in using precipitation trend scenarios derived from climate models as a basis for designing water supply systems.