The log-Pearson Type III distribution has been widely used for hydrologic applications since its adoption by federal agencies. One of the most widely used applications of the log-Pearson Type III distribution is modeling annual flood series, using a method of moments estimator in log space to compute the input parameters of the distribution. However, the moments are constrained by available data, and can have temporal variations resulting from changes in land use, channel capacity, and recent hydrologic activity as was observed at many continuous flow monitoring stations in the states of Missouri and Illinois. A common approach to increase accuracy in flood flow frequency estimates is to combine the station skew with a regional skew value using the inverse of their mean square errors as weights. However, over the period of record of a continuous monitoring station, the station skew may vary and lead to increased uncertainty in flood flow frequency estimates. A database of unregulated continuous monitoring flow gages within the states of Missouri and Illinois with a period of record greater than 40 years was obtained. Using the flow database, the objective of this study was to quantify the spatiotemporal distribution of stations skews in Missouri and Illinois, and the resulting envelope of expected flood flow frequency estimates utilizing the station skew range.
Lisa Andes– Hydrologic Engineer, US Army Corps of Engineers, St. Louis District