| description abstract | Data from two dense networks of raingages in Illinois were used to obtain estimates of sampling errors in the measurement of areal mean precipitation on areas of 50?550 mi2 with investigations being made of storm, monthly, and seasonal precipitation. Storm data were grouped also according to season, precipitation type, and synoptic storm type to evaluate the relation between these factors and the sampling error. Storm sampling error was then related to areal mean precipitation, storm duration, area, and gage density within each data grouping. Several transformations of the above parameters were tested and results indicated a slight superiority for logarithms, followed by cube roots, square roots, and a logarithm-square root combination of the parameters. For a given sampling error, the gage density needed in warm season storms was 2?3 times greater than that required in the colder part of the year. Air mass storms required the greatest sampling density among synoptic storm types. Unstable types of rainfall in the warm season were found to require twice as many gages as steady rain. Considerable difference in the magnitude of storm sampling errors was found between consecutive 5-year periods on the same network, as well as between storms with apparently similar characteristics. Sampling requirements for monthly and seasonal precipitation were substantially lower than for storms. | |
