Laserfiche WebLink
<br />, <br /> <br />and Arizona were surveyed I to 6. times between January I and May I <br />in order to forewarn water users of the flows to be expected during the <br />runoff season. Seasonal forecasts for 33 gaging stations w~re issued <br />on the first of February, March, April and May. The Weather Bureau <br />printed water year forecasts for 44 gaging stations as of the first of <br />each month January through May. The two agencies pooled informa- <br />tion .and coordinated their forecasts. <br /> <br />,. <br /> <br />Most interest is in the forecast for Colorado River near Grand <br />Canyon, measuring the inflow to Lake Mead. Representatives of the <br />Bureau of Reclamation, City of Los Angeles, Southern California Edison <br />Company, Soil Conservation Service, Weather Bureau, and Geological <br />Survey, all responsible for gathering the data or using the forecasts, <br />meet each April to compare data and techniques of forecasting the flow <br />at that station and to improve methods through interchange of ideas. <br />The average forecast error for this group decreased from 14 percent <br />in 1948 to 3 percent in 1953. <br /> <br />.. <br /> <br />The Bureau of Reclamation issues forecasts of the April-July <br />inflow to Lake Mead early each month from January through July. They <br />stress the precipitation at 13 stations having 30 years of record and a <br />high degree of individual correlation with runoff. All the forecasts <br />except April are based on accumulated precipitation since October 1 <br />correlated with the runoff. No snow survey data are available in <br />January, and the Bureau found no significant correlation between run- <br />off and water equivalent of snow on February 1 or March 1 at snow <br />courses having 18 years of record. The April forecast is obtained by <br />using the cumulative precipitation to March I, plus the March increase <br />in water content of snow at high snow courses. The Bureau feels that <br />forecasts based on snow surveys show promise, but the 18 years of <br />record are too short to give stability. The average error during the <br />past 5 years indicates Httle difference in the results by the two methods. <br />The official forecasts since 1948 have an average error of 11%. <br /> <br />i <br /> <br />.. <br /> <br />The Department of Water and Power of the City of Los Angeles <br />and the Southern California Edison Company, users of power from <br />Hoover Dam, forecast the flows into Lake Mead to aid in setting up <br />preliminary power system schedules. Their basic forecasts are made <br />from April I surveys at snow courses selected on the basis of location, <br />length of record, and correlation with streamflow. For each of seven <br />basins the mean water content of the snow on April 1 is expressed as <br />percent of average for the period of record. This percentage is multi- <br />plied by a weighting factor, which expresses the normal contribution of <br />the stream to the flow of Colorado River near Grand Canyon. The sum <br />of the 7 weighted percentages, adjusted to compensate for unmeasured <br />flow into the main stem, is applied to the mean April-July runoff near <br />Grand Canyon for the period of snow course records to forecast the run- <br />off. During the pas,t 6. years the average error was 11 % in both the City <br />and Edison Company forecasts. <br /> <br />- <br /> <br />The Soil Conservation Service has investigated methods of fore~ <br />casting from snow survey data. Two methods have come out of the most <br />recent analyses. One is based on the total of forecasts for 10 major <br />tributaries to the Upper Colorado in reference to the past average, with <br />a percentage added for the flow near Grand Canyon. The other is based <br /> <br />-29- <br />