Laserfiche WebLink
<br />conducted during a gem.'f1llly dry period which <br />received only 69% of the long-tenn average SWE for <br />Dt.'Cembe:r 2003 through March 2000las lTK.mitored by <br />a nearby SNOTELsnow pillow, A total of9~ EUs <br /> <br />wert ohtained. of which 9~ wert paired in 2 hour <br />blocks. The remaining"' had their paired EU rcjtth."t.! <br />be:cause of very light winds, resulting in unreliable <br />estimation of plume transit times. <br /> <br />- <br /> <br /> <br />L <br />I <br /> <br />'I' <br /> <br />IA. <br />it' <br />~. <br /> <br />Figure 4. TIle ~lral plllJ'IIne di~penscr and IOWeT at lhe High Allllooe Sc:eJing (HAS) luciltiun wllh the view IU"llrd the ~lllth. <br />soulh<:asl where lhe Wa\lltch f>lall:;tu lop is in cloud. The propane conlroller and dala logger IU(' at the ba~e orthe tower. <br /> <br />Stall\11c1l1 h,\tlng <br /> <br />Three statistical methods, all specified prior to data <br />collection, ......ere applied 10 the precipitalion gauge <br />SWE observations for the population of EUs. All <br />three uscd the crosswind control gauge to reduce <br />natural variance, an approach specified Ixfore the <br />e'l.pt.'1"iment was comiuctro. The methods were <br />sc!c:cted 10 c'l.plorc different means of objectively <br />measunng potential seeding clTccts. They arc <br />rcfcm:d to as pairs testing, linear ranking of residual <br />predpitation amounts (Wilcoxon test), and mean <br />double ratio (~[)R) tcstmg. All three tests provided <br />impressive and ollen similar results as discussed by <br />Super and IIcimbach (2005a. 2005b) with the <br />Wilcoxon being somewhat less sensitive as expeeted. <br /> <br />lIigh sensiti,'ity and inclusion of contidenee <br />intCf'Vals gi....e-s the MDR approlleh advantages, in the <br />interesls of brevity and simplicily of explanation <br />most results presentoo herein will Ix from ~mR <br />testing. Ilowever, referctle-e will be made to ~mean <br />resiJuals~ ealculatoo by the pairs method becau~e <br />they pro\'id... estirn.1leS of the mean difference in <br /> <br />SWE be:twccn givet! groupings of scl"deJ and <br />nonsceded EUs. <br /> <br />The traditional but wrnewhat arbitrary (I-value used <br />to define "signifieanceM in wealher modification <br />e\'alualion is 0.05, Because: it is assumed that <br />scl"ding cannot decrease precipitation. the tests were <br />I-Iaile-d. Slatistical tests giving probability ,'alues Jp. <br />\'alues) '$ a are termed significant. A P-valuc ~ual <br />to a - 0.05 means Ihere are 5 changes out of 100 fur <br />the apparenl seeding increase to be due 10 chance <br />rather than a real effect. The lower the (I-value- the- <br />more significant the result, indiclIting greater <br />likelihood of having detl'Cled an actual seeding <br />etTce!. <br /> <br />v"hen applied to all 98 EUs, ~1DR testing provided <br />the re-sults shO\" n in Table I where targe-t gauges are <br />ordcrl"t.! by increasing distance hum the liAS sl"\:dlllg <br />site. Eneh tar~et gauge is well correlated with the <br />GNO control gauge observations so much of the <br />natural variance was eliminated by using Gl'O as a <br />covariate. Significant P-values between 0.03 to O,Of! <br />are seen for the three larget glluges OSC. locllled <br /> <br />. <br />