WSP00601 - Laserfiche WebLink

Home Browse Search

N N ..t:o. ..t:o. projects would have become more substantial after the mid-1980's. The LOWESS smooth curves for the flow-adjusted periodic concentration (figs. 9 and 10) and monthly load data (fig. 6) for 1970--93 and 1980--93 for station 09163500 indicated the beginning of downward trends after about 1985 or 1986, which is consistent with the trend test results. The LOWESS smooth curves for 1970--93 also indicated decreasing concentrations and loads in at least part of the pre-1980 record at station 09163500. The trend results and the LOWESS smooth curves indicated fairly definitive, downward trends in dissolved-solids and major-ion data at station 09163500 during 1986-93. However, a relation between the salinity- control projects and trends at station 09163500 is not straightforward because of the trend results for station 09095500, the inflow site on the Colorado River that is upstream from the salinity-control projects. Significance levels and direction of change (downward) in the trend slopes for station 09095500 generally were the same as for station 09163500 for water years 1986-93 for many of the same variables, and LOWESS curves for the two stations show similar trends. For station 09095500 for 1986-93, all trends were highly significant and indicated decreasing flow- adjusted constituent concentrations and dissolved- solids loads. Although none of the flow-adjusted trends in salinity variables at station 09152500 on the Gunnison River were statistically significant (alpha 0.05) for 1986-93, all of the trend slopes were downward for flow-adjusted data, except for monthly dissolved-solids loads. Trends at station 09152500 could have cumulative effects with trends in the Colorado River at station 09095500 that could induce significant trends at station 09163500. The downward trends at station 09163500 during 1986-93 could have been related to other factors that affected dissolved solids in the Colorado River upstream from Cameo and in the Gunnison River Basin that were not related to the salinity-control projects. In attempting to directly relate the trend results to salinity-control projects, it might seem logical to com- pare the magnitude of the trend slopes among the three stations to determine if upstream changes in concentra- tions and loads could cause the trends at the State line station on the Colorado River. If upstream changes were not sufficient to cause all of the downward trend at the State line station, then it could be inferred that the salinity-control projects were causing some of the decrease in dissolved solids in the Colorado River. However, the trend slopes computed for the nonpara- metric seasonal Kendall test are estimates of the overall monotonic rate of change for the period. The slopes are medians of all possible pairwise comparisons, and such values are not that informative for making direct com- parisons between stations, which is especially true for the slope estimators for log-transformed data that were done for the trend tests on periodic data. Therefore, comparison of trend slopes among stations for vari- ables that were determined by the nonparametric method, which includes dissolved-solids and major-ion concentrations and monthly dissolved-solids loads, was not done. If trend slopes can be compared among the three stations for the trend tests that were done using para- metric methods, such comparisons could indicate if the trends at station 09163500 on the Colorado River were related to salinity-control projects or if the trends were solely the result of trends at the inflow stations. The trend slopes for the annual and seasonal dissolved- solids loads were determined by linear regression, which is a parametric method. Trend slopes deter- mined by linear regression represent a linear rate of change, and comparison of slopes among stations is more justified than it is for non parametric trend slopes. However, many of the trends in the annual and seasonal loads (after flow adjustment) were not statistically significant (tables 4 and 5). The only annual or sea- sonalload that had significant trends at both Colorado River stations for the same time period were the trends for flow-adjusted annual dissolved-solids loads for water years 1986-93. Both stations on the Colorado River have highly significant downward trends in flow- adjusted annual loads for 1986-93. The trend slope in annual loads for the station near the State line (09163500) is 21,300 tons greater than the trend slope for the station near Cameo (09095500) (table 4). That annual rate of decrease represents a total decrease in dissolved-solids load of about 170,000 tons over the 8-year period. The 170,000-ton decrease in load at station 09163500 would be caused by decreases in dissolved-solids loads from the Grand Valley and from the Gunnison River. Compared to the projected decreases in dissolved-solids loads through 1993 for the salinity-control projects in the Grand Valley and Lower Gunnison Basin Units, the decrease in annual dissolved-solids load at station 09163500 seems plau- sible. If the trend slope (not significant) for the Gunnison River (station 09152500) also is subtracted from the slope for station 09163500, the net decrease in annual dissolved-solids load at station 091635000 30 Trend Analysla of Salected Water-Quality Data Aasoclated With Salinity-Control proJecta In the Grand Valley, In the Lower Gunnison River Basin, and at Meeker Dome, Western Colorado