Limitations of NOAA Atlas 14 Rainfall Statistics in Florida

Several approximations are used in deriving the National Oceanic and Atmospheric Administration (NOAA) Atlas 14 depth-duration-frequency (DDF) functions from annual maximum series (AMS) of rainfall amounts over various durations. Analyses reported in this paper show that the return periods in Atlas 14 DDF functions for Florida are likely overestimated at about 10% of the base stations for durations of 15 min, 1 h, and 24 h. This relatively high incidence of return-period overestimation is also evidenced by measurements at several stations having very high Atlas 14-estimated return periods. Comparison of the Atlas 14 generalized extreme value (GEV) distributions of rainfall with sample distributions derived from AMS data show agreement at only 68%, 64%, and 30% of all base stations for durations of 15 min, 1 h, and 24 h, respectively. It is concluded that the expected rainfall amounts for given durations are not always reliable over the Atlas 14 range of return periods.

Automated summary

This paper reveals that the return periods in the NOAA Atlas 14 depth-duration-frequency (DDF) functions for Florida are overestimated by about 10% for durations of 15 min, 1 h, and 24 h. The high incidence of return-period overestimation is further supported by measurements at several stations with very high Atlas 14-estimated return periods. Comparing the Atlas 14 generalized extreme value (GEV) distributions of rainfall with sample distributions from AMS data shows agreement at only 68%, 64%, and 30% of all base stations for durations of 15 min, 1 h, and 24 h, respectively. Therefore, the reliability of expected rainfall amounts within the Atlas 14 range of return periods is not consistent.