Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: November 15, 2002
Publication Date: February 1, 2003
Citation: BONTA, J.V. ESTIMATION OF PARAMETERS CHARACTERIZING FREQUENCY DISTRIBUTIONS OF TIMES BETWEEN STORMS. TRANS. OF THE ASAE. 2003. V. 46(2). P. 331-343.
Interpretive Summary: Precipitation data measured with minutely timed intervals are generally not available across the US. Computer simulation of such data is an alternative to measured data for many needs. A fundamental data need for computer simulation is a mathematical description of times between storms which is described by 2 variables - a minimum dry time between storms which separates storms between bursts of rainfall, and the average time between storms. A study was conducted using 34 rain gages in a 4-state area in parts of Colorado, Wyoming, Nebraska, and Kansas (approximately 225,000 kilometers squared) for May-Sep to investigate ways to estimate these 2 variables. Maps of average time between storms for each month were made. Mathematical relationships between minimum dry time between storms, average monthly rainfall, and average time between storms were developed. Of these, two equations were developed that are promising for estimation of minimum dry time between bursts of rainfall and average time between storms, allowing easy estimation of the variables for computer simulation purposes. Maps to use the simplified equations were developed. A comparison between measured average monthly rainfall and that estimated by PRISM, a model that was used to map average monthly rainfall on a 4 kilometers squared grid across the US, showed good agreement. The PRISM data provide a readily available database for average rainfall that can be used with the equations developed in this study to estimate minimum time between storms and average time between storms. The results are useful to university, foreign, and government scientists, consultants, and the Natural Resources Conservation Services (NRCS), for storm simulation modeling, drought studies, engineering design, etc.
Precipitation data of the order of minutes are often needed, but few such records exist over widespread areas. 24-hr precipitation totals are often used in research and practice, with poorly estimated storm intensity distributions. Storm simulation is a promising approach to meet data needs, however, simple characterization of storm occurrence is needed for storm simulation. Methods for estimating parameters describing the minimu time between independent storms (critical duration - CD), and the average time between independent storms (TBS) are investigated. Mapping of CD and TBS yields reliable monthly maps of the spatial variation of these two parameters. All data and regressions based monthly grouping of CD vs average monthly precipitation (Pmo), TBS and Pmo, and CD vs TBS yielded poor correlations. Monthly data at individual stations yielded good log-log correlations between TBS and Pmo, and fair correlations between CD and TBS. Intercept and slope parameters for the station analysis for the two regressions yielded correlated map surfaces for individual regressions. Regressions developed between the parameters allow simple estimation of CD and TBS using only one of the mapped parameter surfaces for each regression, reducing the size of the data base required for estimating CD and TBS. PRISM and measured Pmo data show good agreement, providing a source of readily available data for estimating Pmo. The methods developed provide simple parameterization techniques for describing independent storms and frequency distributions of times between storms for stochastic storm generation, drought studies, etc.