Revising probable maximum precipitation (PMP) estimates under changing climate

Probable Maximum Precipitation (PMP) is a key input in the design and risk assessment of critical infrastructures such as large dams and nuclear power plants. Traditionally, PMP is computed as a fixed upper bound of the precipitation assuming a stationary climate. However, due to climate change, the stationarity assumption may not remain valid in the future. Limited attempts have been made in the past to develop methods for estimating PMP by accounting for non-stationarity in the related hydroclimatic variables. In view of shortcomings associated with those methods, three new nonstationary models are proposed and their potential in determining PMP in a changing climate is illustrated through application to three major flood-prone river basins in India. In this analysis, historical records of precipitation, surface temperature and relative humidity, and their future projections corresponding to eleven CMIP-6 SSPs (Coupled Model Intercomparison Project-6 Shared Socio-economic Pathways) were utilized. The results indicate that PMP estimates obtained using the proposed nonstationary models are significantly higher than those obtained from their underlying conventional stationary model, especially for high-emission scenarios in the near future. The results obtained from this study could be utilized to update historical PMP values and to determine the increase in risk associated with the corresponding probable maximum flood.