Laserfiche WebLink
<br />0.... "l! ~. -,... <br />ll\ll:n <br /> <br />Monthly flows for into-basin imports were obtained from the USGS and <br />irrigation districts. The import term is the monthly sum of all <br />import flows into the basin. upstream of the station. <br /> <br />Incidental depletions or miscellaneous adjustments account for such <br />uses as stock pond evaporation, fish and wildlife uses, etc. <br />Generally, only annual totals were estimated with monthly values <br />being determined by a fixed percentage distribution. <br /> <br />Computer programs have been developed to handle the great amount of <br />data necessary to make the adjustments from historical flow to <br />natura 1 flow. <br /> <br />After all of the historical flow data had been converted to natural <br />flows, it was necessary to extend the data back to 1906 at those sta- <br />tions which had no record in the early years, as well as fill-in any <br />gaps which may have existed. Generally, this was done by multiple <br />regression techniques using data from nearby stations. Step-wise <br />procedures were used so that the best correlations were found using <br />from one to four other stations. In some cases, the Upper Colorado <br />River Commission's Engineering Advisory Committee had made previous <br />estimates of virgin flows for the early years. To check for <br />accuracy, the CRSS natural flow data was checked against this earlier <br />work. The coefficient of determination (r2) for most of the exten- <br />sions were above 0.90 with many in the 0.95 to 0.98 range. This <br />would indicate that reasonably accurate extensions of data were <br />obtai ned. <br /> <br />Natural Salinity <br /> <br />Natural salinity was derived using the following procedure. First, <br />the historical salt load vs. the historical flow was plotted on log- <br />log paper at each station for each month~ It was found that best-fit <br />curves with this data were straight lines on log-log paper, i.e., the <br />relationship had the form TDS=AQB. where TDS was tons of salt, Q was <br />monthly discharge in 1,000 acre-feet, B was the slope of the line, <br />and A was a constant. It was reasoned that irrigated acreage can be <br />used as an index of man-caused salinity and that most of the <br />deviation of the points from the best-fit line can be accounted for <br />by changes in irrigated acreage from year to year. The deviation of <br />each point from the best-fit line was then plotted on arithmetic <br />paper against the estimate of irrigated acreage for that particular <br />year. The best-fit straight-line curve was computed and the inter- <br />cept val ue or the "zero-acreage factor" was determi ned. The zero- <br />acreage factors were then plotted against the month of the year and <br />generally were found to fall into a rough sinusoidal pattern. By <br />trial, the best-fit sine curve was plotted and new "smoothed" values <br />for zero acreage factors were determined. Zero-acreage or curves of <br />natural salt vs. flow were then computed for each month from the <br />equation TDS=A'QB, where A' was the adjusted zero acreage factor. It <br />was assumed that the slope, B, would remain the same as before. <br /> <br />20 <br />