Laserfiche WebLink
Table 2: Statistical characteristics of the Primary Climatic Gage Data <br />(Values represent June-August season totals in inches.) <br />TargetControlNorth Control South Control <br />(10 gages) (21 gages) (10 gages) (11 gages) <br />Median73.79125.0666.0962.39 <br />Mean74.405126.82965.27161.558 <br />(per gage) (7.44)(6.04)(6.53)(5.60) <br />Std. Deviation 19.4335.8318.5719.80 <br />Coefficient of 0.2610.2830.2840.322 <br />Variation <br />Correlation with Time 0.005-0.0870.069-0.221 <br />Correlation with ----0.7710.7610.683 <br />Target Area <br />Correlation with North ---------0.744 <br />Control <br />The correlations between areas are more or less typical of this kind of data. They <br />suggest that ordinary linear regression would account for about half of the variance in the <br />predictand data. It may, however, be worth noting that a randomized-crossover rain <br />enhancement experiment in Italy (List et al. 1999) with similar correlation between areas <br />did not identify a significant seeding effect. <br />Interestingly, the overall frequency distributions of the data (Figure 1) suggest <br />that they can be represented fairly well by normal distributions – except for the “outlier” <br />very wet year of 1993. This could be useful in simulations of seeding experiments or <br />other similar studies. <br />6 <br />