Laserfiche WebLink
<br />The sampling variance numerical experiment generated a total of 530 data <br />points - five gage densities for' each of 100 isohyetal patterns and two gage <br />densities for each of 15 isohyetal patterns. Before relating the sampling <br />variance to the isohyetal pattern characteristics, the input parameters were <br />generalized by removing the scale factors. Thus, the pattern shape, E, which <br />implicitly defines the raincell area in the simulations, was combined with <br />the appropri ate ly scal ed gage n,etwork densiti es to form the more general <br />variable called GPR, the number of gages per raincell. Thus, for example, a <br />raincell that has an area of 100 km2 (E approximately equal to 2) and is <br />sampl ed by a gage network with a density of 10 km2 per gage (approximately <br />4 mi2 per gage) will have an average GPR equal to 10. The range of variation <br />of GPR corresponding to the range of variation of E and gage density used in <br />this study is 0.157 to 714.0. In addition, the maximum precipitat'ion point <br />offset from the geometri cal center, ori g; nally expressed as X offsE~t and Y <br />offset, was vectorially combined into a single variable which was called F. <br /> <br />Using the coefficient of variation, s/i, as the dependent variable, a <br />stepwise linear regression model was fitted to the independent varirables <br />since s/x~O as GPR~DO, regl',ession through the origin was fOrCE!d <br />and the independent variables uSI~d in the regression were GPR, powers and <br />roots of GPR, and combinations of the products of GPR and its power' functions <br />with G and F and their power functions. The following regression equation <br />was obtained as the best fit over the range of variation of the input data: <br /> <br />s/i = 0.7105/GPR + 0.5079 G/GPR - 0.1381 G/GPR2 <br />+ 0.0121 G/GPR3 - 0.0531/GPR2 <br /> <br />(2) <br /> <br />11 <br />