Incorporating CMIP5 Precipitation Projections into IDF Estimates for the Kansas City Area

Abstract

There is building evidence that climate change is causing an intensification of precipitation patterns. Locations around the world can expect to experience more intense precipitation events. Engineers must be able to account for this future climatic uncertainty in their designs in order to develop sustainable and resilient systems. Intensity-Duration-Frequency (IDF) estimates, developed by the National Oceanic and Atmospheric Administration (NOAA), are used across the United States for engineering design. These estimates were developed under the assumption of a stationary climate (with respect to precipitation intensity). However, research has shown that this assumption may lead to the underestimation of extreme events. The objective of this study is to characterize projected changes in storm intensity, duration, and frequency for the Kansas City area, including: 1) Identifying precipitation trends and 2) Develop IDF estimates incorporating projected climate trends. To achieve these objectives, precipitation data was analyzed from six NOAA gages and the Coupled Model Intercomparison Project version 5 (CMIP5) ensemble. Annual and monthly precipitation was analyzed for both the observational gage data and CMIP5 model data using the Mann Kendall trend test. Increasing trends were identified in the winter (Dec.-Feb.) and spring (Mar.-May) months, while decreasing trends were identified for July-September, indicating a potential shift in seasonal precipitation patterns. Increasing trends were identified for annual precipitation for both the gage data and climate model data. Partial Duration Series (PDS) were developed for the six gages using the Peak-over-threshold (POT) method. Significant increasing trends were identified for the frequency of PDS events. A strong correlation was identified between PDS event frequency and annual precipitation. This relationship was used to develop a novel approach for incorporating climate model projections at the monthly scale into gage-based PDS events used to derive IDF curves. In this methodology, the PDS annual exceedance rate for the future time period was determined based on the CMIP5 model projected annual precipitation. IDF estimates incorporating projected climate model trends were then developed using the adjusted PDS data sets. Results showed an increase in event magnitude from the original for most durations and recurrence intervals, across all gages, with the 2-year and 100-year events increasing the most. The increase in the event magnitude has serious implications for engineering design. Critical infrastructure, such as bridges, roads, and overflow channels, that are designed for using stationary methods may be under designed.