Laserfiche WebLink
<br />Thc result of the cumulative year statistical evaluation of the operational seeding <br />programs in Group 2 is shown in Figs, 3b. The decrease in seeding effectiveness for all of these <br />seeding programs starting in about 1990, givc or take a year or two, is quite evident. Whereas <br />these secding programs indicated weak seeding effccts when evaluated aOer seeding year 2003 <br />(see Table 3), they indicatcd very strong seeding effects whcn evaluated after seL'ding year 1990 <br />(see Table 4). It can bc seen from Figure 3b that the decrease in seeding effectiveness was quite <br />sharp for the Kern River, Kaweah River and Eastern Sierra seeding programs. The decrease was <br />less sharp but no Icss signilicant than the decrcase found for the Kings River seeding program. <br />Noting the gencral decrease in seeding effcctiveness that started in about 1990, give or <br />take a year or two, an evaluation was donc on the Tuolumnc River operational seeding program <br />using the only potential target with suitable FNF data, TLG, the Tuolumne River-La Grange <br />Dam. This was done because seeding in that program first began in 1991. The evaluation of the <br />Tuolumne River seeding program at TLG showed no definitive evidence ofa seeding efTect. The <br />cumulativc year secding elTect varied between -2 and +2 percent throughout thcir pcriod of <br />operations with very weak statistical support for the results. Without additional information, it is <br />not known whethcr the lack of a seeding efTect is due to the fact that TLG is not a good seeding <br />target or that the efl'cct of seeding in the Tuolumne River watershed was severely limited by <br />whatever caused the general decrease in seeding elTectiveness in all the other operational sceding <br />programs. <br />The decay in sceding etTectiveness that started about 1990 needs extensivc study to <br />explain. Apparent in Figs, 3a and 3b are othcr shorter-term trend changcs bcfore and aftcr 1990 <br />that are interesting and worthy of study; however, this paper will focus on the decay in seeding <br />effectiveness starting in 1990 in the discussions to come and then in only a preliminary way. <br /> <br />6. Pooled estimates ofseedill~ <br />To undcrscore the significance of the decrease in secding efTectiveness that started in <br />about 1990, the estimates of seeding efTectiveness for the 7 openuional seeding programs <br />evaluated aftcr water year f990 (see Table 4) will be pooled and compared with the pooled <br />estimate of seeding for the 5 targct watersheds evaluated aftcr watcr year 2003 (see Table 3). <br />Pooling of the estimates was donc according to the method of Gabriel (2002), The weight <br />assigned to the statistical result for each of target \...'atershed was dircctly proportional to the <br />number of years that target was secded and indirectly proportional to that target's standard error <br />of estimate. It was found (see Tablc 5) that the common etTect of seeding on all 7 target <br />watersheds after water year 1990 could be said \...,ith 90% conlidence to be between +6.0% and <br />+ 13.0%. The point valuc pooled estimate of seeding is +9.4% and thc probability that the <br />seeding elTcct is grcatcr than 0% is 100%. For the cvaluation of the 5 larget watcrsheds aftcr <br />water year 2003. it was lQlInd that the common effect of secding on all 5 target watershcds could <br />be said with 90% confidence to be between -D. 1% and +6.8%, The point value pooled cstimate <br />of seeding is +3,]% and the probability that the sceding elTect is greater than 0% is 94.4%. <br />Although a positive elTect of sccding is suggcsted either way. the motivation to explain and <br />compcnsate. if possible, for the decay in seeding cffectiveness is ob\'ious and compelling, <br /> <br />TaMe 5. Pooled opcrational program results for the 7 operational programs aOer water year 1990 <br />and the 5 operational programs aOer water year 2003. <br /> <br />43 <br />