Submitted to: International Conference on Water Resources Engineering Proceedings
Publication Type: Proceedings
Publication Acceptance Date: August 12, 1999
Publication Date: N/A
Interpretive Summary: High intensity rainfalls may produce floods. The dimensions of culverts and other flood prevention structures are set according to the likelihood that rainfalls of certain intensity and duration occur in a given time interval. When a civil engineering firm designs these structures, it usually obtains these rainfall characteristics from a publication by the National Weather Service in the National Oceanic and Atmospheric Administration (NOAA) Atla 2 series. A previous study in 1988 found that the estimates of short- duration precipitation intensities, based on NOAA Atlas 2, were substantially lower than estimates based on data from a dense rain gages network in Walnut Gulch, an Agricultural Research Service experimental watershed in Southeastern Arizona. Since the previous study, more data have been gathered and these data were used to compute new estimates of short- duration precipitation intensities. The estimates obtained in this study are closer to the values reported by the NOAA than those reported in 1988, however, early records on Walnut Gulch suggest greater intensities, and that the true frequency of high intensity rainfall events may lie between the mean values in this study and the estimates in the 1988 study.
Technical Abstract: In 1988, Osborn and Renard published a paper with the intensity-duration- frequency values for 1 hr and less, based on data from a dense network of rain gages in southeastern Arizona. The authors compared their findings wit the values that would be derived by extrapolation from the maps of 6-hr and 24-hr rainfall amounts with different frequencies published by the National lWeather Service in the NOAA Atlas 2 series. Osborn and Renard (1988) found that the estimates of short-duration precipitation intensities, based on NOAA Atlas 2, were lower than estimates based on data from the rain gages network for 50- and 100-year events. In this work, five sets of four gages each were created. Each set of gages has a total record length of 140 years as opposed to 90 plus years in the original study. Two-, 5-, 10-, 25-, 50-, and 100-yr rains for 5-, 15-, 30-, and 60- min durations were estimated. Th mean point frequencies of the 5 sets of gages are very close to the NOAA values for the 5-, 15-, and 30- min durations and all the return periods (between 0 and 7%), whereas the mean point frequencies for all durations fo the 25-, 50-, and 100-yr return periods are between 14 and 33% lower than the values reported by Osborn and Renard (1988). The analyses presented in this work are statistically accurate, but any user of these data should be aware that early records on Walnut Gulch suggest greater intensities.