Extreme rainfall in Florida; local climatologies revisited

For 62 cooperative and first-order observing sites in Florida, daily rainfall values for the period 1948-2008 were acquired from National Climatic Data Center (NCDC) files and archived at the Florida Climate Center. These daily values were used to compile totals for periods of 24 and 48 hours, with emphasis placed on a 48-hour period. Both annual maximum and partial duration rainfall amounts were computed. Generalized Extreme Value (GEV) distributions whose moments were derived via the method of L-moments were fitted to the rainfall data. Amounts having return periods from two to 100 years were derived from these distributions.

In general, partial duration data produced larger return values at shorter return periods and annual maxima produced larger return values at longer return periods. In order to focus on maximum rainfall, derived values were combined to produce distributions having the larger return value for each return period. Further, return values from these combined distributions were applied via the National Resource Conservation Service (NRCS) method to estimate local surface runoff. Such runoff estimates are considered to have the same return periods as the applied rainfall. Examples of rainfall and runoff curves derived from the combined distributions for both 24 and 48h periods are discussed.

Identification of added risk beyond 24 hours can be aided by considering both the 24h and 48h climatologies. An index to quantify any added risk in the 48h period was constructed from (1) the 100-year 24h rainfall and (2) the difference in return periods when that 24h amount is considered a 48h amount. Increased values of this index reflect continued risk in a 48h period and suggest added value in consideration of the local 48h climatology. Comparative results are discussed for various sites. In particular, the effect of hurricanes is evident.