WMOD00277 - Laserfiche WebLink

Home Browse Search

are either inadequate or missing. Most of these stations are near the Continental Divide in Colorado or on streams from the Uinta Mountains in Utah and ViTyoming. For purposes of snow runoff relationships the,se diversions tend to be relatively constant and of lesser effect on larger streams. The lack of records on diversions account for low correlation on smaller stC'eams to some degree. Snow accumulation to April 1 (or 1113.ximum date) is the major factor affecting flow in snow melt streams. There are several other factors. These include precipitation after April 1 through the snow melt season, soil moisture conditions, and ground water levels on the watershed. The snow melt or temperature sequence also affects total runoff and especially peak stage. In most years these other effects tend to balance and approach an average over a season. They are difficult to evaluate unless there is an extreme and lengthy deviation of climatic condi- tions from normal. In the past 20 years, the 1957 water year was noted for excessive snow accumulation after April 1 and a delayed snow melt in Colorado mountains. Flows of the previous year or more have a substantial effect on many watersheds. It is noted that flows for the years 1953, 1958, 1963 and 1965 tended to exceed that which should have been expected from the snow cover. These followed heavy runoff years. On the other hand flows in 1956 and 1962 were generally less than expected because these years followed two or three years of low runoff. The period April - August, inclusive, is generally the best period to represent snow melt runoff in determining snowpack-streamflow relation- ships. At lower elevations the inclusion of March flows would be slightly preferable. Any period from March through September that includes May, June and July are highly correlated and represents a fair index of snow melt runoff in the Upper Colorado B8sin. Direct snow melt runoff ranges from 75 to 85% of total annual runoff. If the flow that is delayed through groundwater storage is added, the snow melt runoff would probably account for near 90% of total annual flow ~ Precipitation records during the winter and spring months at high elevation stations are also us ed to forecast streamflow. The practice is to adjust monthly total precipitation to give greater weight to mid-winter months. Where good precipita- tion records are available these records can be used as a satisfactory parameter to relate to subsequent runefL No studies of these relations have been made at this time. Runoff data, elevation of gaging stations and their drainage area were taken from appropriate tables in Sur f ace W ate r Sup ply Pap e r s published by the U. S. Geological Survey. Snow course deLta was obtained from data summary publications of the Soil Conservation Service. Detailed descriptions of stations and snow courses are shown in the respec- ti'\Te reports and are consequently not included here. Table X:;(II presents a summary description of some oi -{ne rn.ore in'lportant geographic and hydrologic characteristics of selected strearris within the Colorado River Basin and for dowrl.wind basins that would also be affected by Weather bodification within the Colorado River Basin. I The mean elevation of the waLersheds I was estimated roughly from 500-f<j>ot contour-scale 1:500,000 maps of states. The es~imated elevation is believed to be correct within 300 feet. The minimum snow'line elevation is based on five inches of snow water equivalent on April 1 of an average year. The elevation of this line varies with aspect and exposure. Estimates were made from data on low eleva~ion snow courses, and personal observation of snow pack conditions for several seasons. An average of f~ve inches water content is near the minimum snoW: pack that can produce runoff after satisfying so~l moisture deficits, and transpiration and evaporation losses during the snow melt season. , --! The acre inches per acre is figured by dividing mean April-August runoff by the total area of the \vatershed. The runoff per 'unit area follo\X/s a pattern related to the size of the snow pack area of the watershed. This may not be r:elated to total watershed area. The estimate of acre-feet runoff per inch of snow water equivalent is based on the slope of the relationships points (estimated least square line). It is adjusted somewhat to reflect any deviation of the average of the elevations of, the snow courses used from the average snow pack ~levation of the watershed. The percentage column represents the percent that one inch of snow water equivalent will produce of the total April-August runOff. The correlation coefficient was calculated by the rank-difference (Spearman) method. This is a rough estimate. Correlation coefficients using the actual snow course and runoff data may vary up to .05 from that calculate;d by the rank- difference method. Typically the 'variation will be about.02. Data is available to c~1culate the least square line, the standard deviation of runoff, and a more accurate correlation coefficient. In the Colorado Rh;er at Lee's Ferry, Arizona, one inch of snow water equivalent at the 10, OOO-foot level over the basin '*ill produce about three quarter million acre feet during the snow melt season or one million acre feet annually. This excludes some two and one half to three million acre feet diverted or used in the Upper Basin. Figures 51 through 56 show the relationship between April 1 snow, cover and runoff for a few selected stations in the Colorado River Basin. These are .considered representative of relationships that can be established for other areas. A summary of the distribution of correlation coefficients obtained f9r the 70 streams studied is shown in Figure 57. It can be noted that 8, or 110/0, of the, streams have corfelations with snowpack of . 9 or better and 50, or 71%, are correlated at .8 or better. 64