Finance Committee Meeting 2009 - Laserfiche WebLink

Home Browse Search

12 Journal of Weather Modification Volume 41 • The combined flows from GUP,OBL,GBH,OPT, don that are most favorable to seeding, statists- GBO, and MID, hereafter called Gore Creek catty significant increases in precOltation from (OCT),flows kite the Eagle River about 12 miles +28% to +45% were apparent (Hjerrnstad and south of the target area at a point 3 miles down- M elte,1978). stream of Mintum,Colorado.An evaluation of the seeding effect fior OCT indicates a s istically Lest the rather large seeding effect estimate for significant increase in OCT etreamtow of+9.5% GBH be an anomalous function d the control that • with 90%confidence that the true effect of seed- was used (FRR), the evaluation of GBH was re- ing lies between+2.9%and+16x4%. pealed using several differont controls. The re- sults of the evaluations of GBH with aid the it can also be seen from Table 3 that the target- controls are shown in Table 4.It can be seen that control correlation coefficients range from 0.775 the results of the evaluation using the other 7 to 0.918,accounting for about 589E to 85%of the controls are statistically comparable to the results variance of the target strearrdlows, respectively. obtained using FRR as the control. However, ,i There are two noteworthy aspects to this finding: using FRR as the control yields the most precise i (1)despite the fact that the targets are very close result i.e., the lowest standard error of the est- to one another, there is a big difference in their mate of RRn. Therefore, it is reasonable to con- correiaton coefficients with the control, and (2) dude that the estknates of the seeding effect us- , , the target-control correlation coefficients for the ing FRR as the control for all of the targets, as r:. Val targets we substantially smaller than those given in Table 3,are statistically credible. found for the evaluation of the operational seed- . ing programs in the watersheds of the Sierra Ne- Table 4. Same as Table 2 except for the 1.1 vada Mountains which ranged from 0.93 to 0.98. evaluation of GBH using different controls as It appears that the spatial l variability of annual indicated. !, runoff among watersheds in the Rocky Moun- tains is greater than that for watersheds in the Con Vol 8 CI90L CI9OU p Sierra Nevada Mountains. in view of the pr FRR +28.8 +1e 6 +422 0.776 • creased standard error of the estimate for tics Vail targets. It is Impressive that there is such WDC +23.6 +10.9 +37.9 0.703 i. strong statistical support for the estimated its creases NF +26.8 +13.9 +41.2 0.718 ases in atrearnftow due to seeding. ' WSF +33.4 +19.2 +49.3 0.679 The+28.8%increase in water year streamflow at GBH is rather large compared to the results from FRR,WSF +28.7 +16.6 +42.0 0.768 Sirniar seeding programs, especially when one FRR.WNF +28.4 +16.4 +41.8 0.771 takes into account that the water year streamffow is being affected by seeding during only 3 WDC.WSF +27.9 +14.9 +42A 0.716 months of the year. Silverman (2007, 2008) wDC,VYNF +25.2 +12 8 +39.1 0.732 evaluated the Kings River,Kern and San Joaquin operational cloud seeding programs and found that they produced a positive, statistically signifi- 5. ME EVOLUTION OF THE SEEDING cant increase in water year streamiow ranging EFFECT • from +4.8% to +12.2%. Ciiirnax I and Climax II resulted in increases in precipitation of+9%and The time evolution of the seeding effect for the :: +13%, respectively, that could reasonably have Vail primary targets is shown in Fig. 3. In Fig. 3 occurred by chance (Grant and Kahan, 1974); the seeding effect calculated for each seeded ! however, for data stratifications of temperature water year Is the value that would have been ob- and wind that are most favorable to seeding, sta- tahied if the evaluation were done for ail seeded tistk&ly significant increases in precipitation as years up to and inckaing that water year. It can high as +55.2% were Indic at d (Misice it aL, be seen that the seeding effect appears to be 1981).The evaluation of the Colorado River Ba- consistent over time. For clarity of presentation, ' sin Plot Project resulted in no statistically signiti• the 90 percent confidence limits are not shown. cant Increase in precipitation(Eilott at ad., 1978); For each of the targets the 90 percent confidence however, a statistical reanalysis of the data sug- limit lines follow the pattern of their point value r gassed that,for data stratifications of wind di c- plot and narrow with time as the standard error of 1 -Scientific Papers-