Skip to main content
ARS Home » Research » Publications at this Location » Publication #119096


item Bonta, James - Jim

Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 7/16/2001
Publication Date: N/A
Citation: N/A

Interpretive Summary: Methods to characterize and estimate dry times between storms are needed for providing input to storm simulation models where there are no data and for studies of crop stress, drought, etc. Therefore, a basic understanding of the seasonal and spatial variabilities of parameters that make up the inputs to the models is needed, and then methods for estimating the parameters must be developed. A parameter used to identify the minimum dry times between storms was mapped over a 225,000 square kilometer plains area east of the Rocky Mountains (called the "critical duration" or CD). The parameter varied greatly from month to month and location to location. Maps were made of the variation for each month. A method for describing how frequently dry times occur was found to work well. However, a method for estimating CD that worked well over a range of climates in another study did not work well for estimating monthly CD values in the study area. Another equation that has worked well elsewhere provided poor estimates. Average values of CD were found to be good estimates of monthly trends, but not of spatial variability. Monthly mapped CD values were the best source of data for model parameterization. The study has utility for guiding further development and parameterization of a storm generator model and for drought studies. Federal, state and university scientists and consultants will benefit from this study.

Technical Abstract: Methods to estimate a parameter were studied that are used to identify storms in a precipitation record for the purpose of simulating the occurrence of storms. The parameter (critical duration, CD) is fundamental to a storm-generator model. It is the minimum dry period between storms that separates bursts of rainfall into statistically independent storms. Data covered an area of about 225,000 km2 in the plains of Colorado, Nebraska, Kansas and Wyoming for May through September. A method developed in another study that yields an exponential frequency distribution of times between storms (TBS) gives good visual fits to the data. However, some precipitation data yield CD estimates that appear to be outliers. CD varies noticeably with season of year and location. Mapping CD is the best method to estimate CD on ungauged areas. An equation developed in another study for estimating CD over a wide range of climates is inadequate for monthly CD values. Computing CD by collapsing precipitation data into 2-, 3-, 4- and 5- onth periods did not improve the estimates of that equation. Computing separate regressions for monthly CD data against monthly precipitation yielded statistically insignificant results for all months and led to averages for each month. Averages do not give information on the spatial variability of the data. A multiplicative power equation that was successful in another study to estimate monthly CD did not work well in the present study. The study results are useful for guiding storm generation and drought studies.