Scaling Up: Future Intensity-Duration-Frequency Curves for Extreme Rainfall in North Carolina Using Downscaled Climate Projections

Extreme precipitation and its associated flooding have detrimental impacts including loss of life and property. To mitigate these impacts, water resource managers use intensity-duration-frequency (IDF) curves to plan and operate critical infrastructure. IDF curves show rainfall depth or intensity for various storm durations and recurrence intervals. Currently, NOAA Atlas 14 provides point frequency rainfall estimates that can be visualized as IDF curves. However, Atlas 14 is based on historical data and does not account for changing precipitation characteristics as the climate warms. By scaling Atlas 14 precipitation totals based on future time periods and emissions scenarios, engineers can better prepare for extreme precipitation.

This study provides future IDF curve estimates for North Carolina using downscaled climate projections under RCP4.5 and RCP 8.5 emissions scenarios at mid-century (2040 - 2069) and end-century (2070 -2099). Future rainfall increases are expressed with scaling factors derived from downscaled global climate models within the Localized Constructed Analogs dataset for NC's eight climate divisions. An ensemble average of the regional scale factors is calculated for each climate division and for different rainfall recurrence intervals. Atlas 14 precipitation depths are scaled, creating future IDF curves for various storm duration at mid- and end-century for both emission scenarios.

The largest scaling factors, from RCP8.5 at end-century, increase daily precipitation by up to two feet for rare events (e.g., 100-year storm) and over an inch for more frequent storms. Precipitation depths have the greatest response to this emissions scenario and time frame. RCP4.5 and RCP8.5 mid-century and RCP4.5 end-century all have similar scaling factors, with daily precipitation increasing by at least half an inch for the 25-, 50-, and 100-yr storms. While NC's extreme precipitation is less sensitive to these scenarios, future projections still differ greatly from Atlas 14. Furthermore, storm frequencies are projected to change; a 25-yr historical storm could occur at a 10-yr frequency under RCP8.5 for both the mid and end-century. Continuing to use historical Atlas 14 precipitation estimates may severely underestimate future rainfall and jeopardize the ability of current infrastructure projections to withstand future flooding.