WMOD00063 - Laserfiche WebLink

Home Browse Search

very robust to departures from its inhcrent assumptions and approximate results from re- randomization analyscs extremely well, The bias-adjusted RR values are dcsignated as RRA. The results of the evaluation for the water year streamtlow for all the targets listed in Table 1 are presented in Table 3. In addition to showing the valuc of RRA , the 90% confidcnce interval and the level of confidence that the secding elTect was positive (;?: 0%) are shown, The Icvel of confidence that thc annual averagc seeding effect is 2: I % is also shown, It has been conscrvatively assumed that a 1 % increase is nceded to onset the annual cost of the seeding opcrations so the level of confidence that the annual average sceding effect is 2: 1% is taken as an indicator of the cost-cffectiveness of the seeding operations. lIendcrson (2003a) has shown that the benelit-to-cost ratio of increasing streamflow by I % through the snowpack augmentation programs under consideration in this study is ab()Ut 6,8: I, so a thrcshold of cost efTectivcness of 1% is very conscrvative indeed. It is emphasized that the calculations could be done for any specified threshold of cost effectiveness. Tahle 3, Results of the evaluation of the operational seeding programs ovcr their respcctivc periods of record, Results arc given for thc proportional elTect of seeding, 0(%) = 100*(RRA.I), where RRA is the bias-adjusted Regression Ratio, Po is the probability (%) that 0 2: 0%, and PI is the probability (%) that 0 2: 1% (the hypothetical threshold of cost efTectiveness). Operational Seeding Program Lake Uppcr Mokelumne Eastern Kings Almanor American River Sierra Rivcr FPR SCU MMF OWL KGF 39 24 50 26 49 1954-92 1969-92 1954~03 1978-03 1955-03 +9,5 + 10,1 + 10,1 -0.5 +3.0 Statistical Resuh No Ycars Seeded Evaluation Period 0(%) 90% Conf Interval LB(%) UB(%) 0%$0$ UB (%) Confidence (%) UB(%) Po P, Kawcah River KWT 28 1976-03 +0.9 Kern River KRI 26 1978-03 -1.4 +1.5 +{J.6 +1.7 +18,2 +20.4 +19.2 95.0 92.3 95.4 +19.9 +21.1 +21.2 97.5 96.1 97.7 95.9 94.1 96.3 -6.1 +5,5 -2.7 -7.3 +9.0 +9.7 59.9 14.6 +6,] +1.7 80.0 57.3 71.6 49,1 -12.2 +10.3 44.2 H 41.0 36.1 It can be seen from Table 3 that the Lake Almanor, Uppcr American and Mokelumne River operational seeding programs experienccd the most positive seeding effects \vith the strongest statistical support. The 90% contidcnce interval lor all the other operational seeding programs included the null efTect even though the Kaweah River and Kings River programs indicated a small positive seeding elTect. What is striking about this result is that the Lake Almanor and Upper American evaluations, due to the limitations in available FNF data for thosc targets, were conducted after water year 1992 while thc others wcrc conducted al1er \'...ater year 2003. To investigate the significance of this finding, a variation of the ratio statistics evaluation method was used to rcveal the time evolution of the seeding cfTect. A progressive statistical evaluation using ratio statistics. called thc cumulati....e year statistical evaluation. was conductcd whereby the secding is evaluated as a function of the cumulative number of years of seeding operations (initially the first 5 operational years. then the first 6 opcrational years, then the first 7 operational years. ... . and finally all operational years). A signilicant change in trend in the plot 40