Laserfiche WebLink
<br />by use of Fig. 3 can be converted to G which, as we shall show 1 ater, is <br /> <br />related to the sampling variance of raingage networks. <br /> <br />A cursory check was made on the suitability of simulating convective raincells <br /> <br />by Eq. (1), which for the purposes of the following comparison is referred to <br />as Modell, by fitting an exemplary Montana raincell by a nonlinear least <br />squares regression method and comparing the model-derived rainfall volume <br />with that calculated from the point raingage data using the polygon method of <br />Thiesen (1911). The same procedure was followed for other proposed isohyet <br /> <br />models based on logarithmic and exponential functions of area, i.e., Model 2: <br />P(x) = Pm exp (-kAn) and Model 3: P(x) = Pm (l-kAn). The results of <br /> <br />this comparison are shown in Table 1. <br /> <br />Table 1. - Comparison of some measures of the goodness-of-fit <br />of the three isohyet models for a representative <br />Montana convective raincell <br /> <br /> Variance of Ratio of calculated Dm Ratio of model to <br />Mode 1 residualsa to theoretical Dm (a=.05)b computed ra i nfa 11 <br />1 1.11 .20 .95 <br />2 1.05 .20 .98 <br />3 1. 26 .61 .73 <br /> <br />a cr2 = ~ (e. _ e)2 / (N-l) where <br />e t , <br />b Based on the Komolgorov-Smirnov <br />lFo(x) - Sn(x)f and for which the <br />ratio is greater than one. <br /> <br />e. = Y. - Y. and e = ~ e./N <br />, " ; , <br /> <br />one-sampl e goodness-of-fit test where Dm = <br />null hypothesis is rejected if the indicated <br /> <br />7 <br />