WMOD00063 - Laserfiche WebLink

Home Browse Search

each seeded water year. The diflerences bctween the actually observed streamllow and the predicted streamflO\.... at the targct station were assumed to be the result of seeding during each water year. It was found that the average difference between actual and predicted FNF at KGF was +77,964 AF, an increase in streamflow of 4.79%. A t-test of the null hypothesis that the average differcnce between actual and prcdicted KGF streamllow was equal to zero yielded a 2- sided P-value of 0,048; thus, there was sufficient evidence to reject the null hypothesis, In a further analysis of thesc results, the series of water year ditTercnccs, represented as percentage changes due to seeding. \Vas plottcd as a function of water year (Figure I), It can be secn from Figure I that the dilTerences during the operational period were negative in about 35% of the water ycars and positive in about 65% of the water years with the percentage changc due to seeding ranging from -21.8% to +102.1%. It is interesting that Henderson (2003b) rcportcd, but didn't secm to be concerned by, percentage changes due to seeding ranging from -9.5% to +25.6% in his cvaluation of47 years of operations with negative seeding eflects occurring in 12 of those ycars, Apparcntly, the efTects of seeding are superimposcd on residual natural variability during the operational period that is not explained through the use of the historical regrcssion equation. This natural variability is illustrated in Figurc 2 in which the streamflow at Merccd Rivcr near Pohono Bridge (MDP), the control site, is plotted as a function of water year. The year-to-year variability as well as a long-ternl trend of increasing strcamllow with time is quite evident. The dillercnces betwecn observed and predicted streamllow arc confounded with and dominated by effects due to natural variability (year-to.year variability, short-tcrm cycles and long-term trend), It is possible that the obscrvcd minus predicted streamllow differences are also confounded by etTccts caused by changcs in the hydrologic system (changes in the streamflow vs precipitation relationship), Historical Regression Evaluation 120 ~ 105 -g 90 ~ 75 ,g 60 . " "" . co o ~ <: U 30 15 o -15 -30 45 I ;,-",J;'III,I'I(IIIJLI~lIl.rl.' ~':':fll " o ~ ~ ~ - " ~ - ~ ~ ~ - ~ ~ ~ ~ ~ ~ - ~ ~ ~ - - ~ ~ - ~ ~ ~ - ~ o o N Water Year Fi~ure I. \Vater year difTerences between the actual streamflow at the Kings Rivcr target site, KGF. and that predictcd by thc historical rcgression method. represented as percentage changcs 37