Laserfiche WebLink
2.3. Use of SNOTEL Precipitation Data to Estimate There was a CREST-related analysis (Super and <br /> Snowpack Ablation McPartland 1993) of snowpack-runoff relationships <br /> for fourteen watersheds in Colorado, Wyoming and <br /> To estimate snowpack ablation, we compared Utah. The selected watersheds were not significantly <br /> November 1 to April 1 accumulated gauge snowfall affected by upstream trans-mountain or trans-basin <br /> at 16 SNOTEL sites against April 1 SNODAS SWE diversions and not regulated by upstream reservoirs. <br /> there.November 1 is a nominal date after which most This analysis performed a long-term linear regression <br /> precipitation falls as snow rather than rain. The 16 of snow course/snow pillow SWE and stream gauge <br /> sites were from Colorado,Utah, Wyoming, and Ari- data and assumed 10% SWE increases from seeding. <br /> zona, at dispersed geographic locations and eleva- Correlation coefficients between the two datasets was <br /> tions. Interestingly, the SNOTEL precipitation-to- low for some watersheds, usually because the snow <br /> SNODAS SWE ratios were 1.22 in 2004 and 0.89 in courses/snow pillows were relatively low in elevation <br /> 2005, respectively. The latter ratio, indicative of and didn't reflect higher altitude snowpack (this <br /> greater snowpack water than precipitation, might shortcoming could be alleviated by the spatially con- <br /> reflect a problem with gauge precipitation measure- tinuous SWE fields of SNODAS, if one were to do a <br /> ment or the arbitrary November 1 start date of the new regression analysis with that system). Given the <br /> snowfall season. There is significant gauge under assumed 10% SWE increase, April to July seasonal <br /> catch of snowfall in wind-exposed locations that runoff increases varied from 6%to 21%. This varia- <br /> might explain much of the problem. The bottom line tion was attributed either to poor representation of the <br /> is that we cannot trust the SNOTEL gauge-measured snow course/snow pillow SWE data or to differing <br /> seasonal precipitation to estimate seasonal melt of the basin hydrologic or meteorological characteristics, as <br /> snowpack.There is one other option to estimate melt. related in the preceding paragraph. Porous geology <br /> The SNODAS model outputs snow melt at the base such as sinkholes may divert meltwater away from <br /> of the pack.We post daily melt products on our Colo- stream gauges, leading to decreased runoff measure <br /> rado web site (Hunter 2004). To generate seasonal ments, whereas impermeable soils such as clay may <br /> melt,we would have to sum the daily values over an increase runoff percentages. Again, these complex <br /> entire winter. This is beyond the scope of the current factors will affect any additional runoff produced by <br /> study but may pursued later. seeding-induced precipitation increases and should be <br /> weighed when selecting target areas. It is logical to <br /> 3. RESULTS AND INTERPRETATION assume that the farther the target area is from the <br /> The reader is cautioned that water volumes re- mainstem of the Colorado River,the greater the run- <br /> sulting from increasing existing April 1 snowpacks off losses at the river. Examples of such areas are the <br /> via cloud seeding do not necessarily equal runoff Wyoming potential targets at the northern extremity <br /> increases. The latter increases may be changed by a of the basin (see Fig. 1). On average, however, 10% <br /> given basin's hydrologic processes such as soil infil- runoff increases might be expected to result from <br /> tration, antecedent soil moisture, slope and aspect, 10% snowpack increases (Arlin Super, personal <br /> and vegetative cover. Other factors affecting a ba- communication). <br /> sin's precipitation-runoff relationship are spatial dis- <br /> tribution of the snowpack, amount and timing of any Table 4 lists the water volumes produced by 10% <br /> rainfall on the pack,temperature,and evapotranspira- increases of the snowpack SWE on April 1 for exist <br /> lion of snowmelt water. ing target areas and for the potential target areas. <br /> Table 4. Areas and water yields for 10%snowpack SWE increases from seeding,for existing(opera- <br /> tional)seeding targets and potential new targets. <br /> Area(km) April 1,2004 April 1,2005 Mean Yield <br /> (Dry)Yield (Wet)Yield 04-05 <br /> ac-ft ac-ft ac-ft <br /> Existing Areas <br /> Utah 12,992 1289902 294,527 2119715 <br /> Colorado 17,767 240,852 499,190 370,021 <br /> Total 30,759 369,754 793,717 5815736 <br /> Potential Areas All States Total 13,611 217,890 352,978 285,434 <br /> Existing+Potential Areas Total 449370 1 5879644 1 1,146,695 8679170 <br />