WSP10847 - Laserfiche WebLink

Home Browse Search

A statistical test (Kendall Tau) to identify possible time trends in the discharge data was computed for each streamflow-gaging station, using its annual mean discharge record. The analysis indicated that at the I-percent level of significance, an upward trend statistically different from zero occurred at station 07086000 (Arkansas River at Granite). This increase in discharge (see fig. 58) can be explained by a general increase through the years in transmountain diversions that enter the river upstream from this station. The analysis also indicated that a downward trend in discharge sta- tistically different from zero at the I-percent level of significance occurred at stations 07119500 (Apishapa River at Fowler), 07123000 (Arkansas River at La Junta), 07128500 (Purgatoire River near Las Animas), 07130500 (Arkansas River below John Martin Reservoir), and 07133000 (Arkansas River at Lamar). Joining this group of stations, the analysis indicated that a downward trend in discharge probably different from zero at the 5-percent significance level occurred at stations 07089000 (Cottonwood Creek below Hot Springs), 07106000 (Fountain Creek near Fountain), 07116000 (Huerfano River below Huerfano Valley Dam), 07124500 (Purgatoire River at Trinidad), 07126500 (Purgatoire River at Ninemile Dam), 07134100 (Big Sandy Creek near Lamar), and 07137500 (Arkansas River near Coolidge). Stations 07106000 and 07134100 have rather short per- iods of record, and the trends may result from an inopportune period of col- lecting data (both stations included prolonged drought periods near the end of their record). No explanation is available as to why station 07089000 should show a trend. The rest of these stations are at the downstream end of the basin, where much interstate concern occurs over the apparent decrease in discharge. The construction of John Martin Reservoir has been mentioned as an explanation for downward trends in discharge in the main stem of the Arkansas River. However, the fact the stations upstream from the reservoir indicate downward trends and that the periods of record for stations 07130500 and 07137500 are more recent than the dam construction would discount the reservoir's having much effect. The pumping of numerous irrigation wells often has been suspected of causing decreased main-stem discharge, but this consideration does not account for the downward trend in the discharge of the tributaries. To consider the effects of the apparent downward trends in the tributary discharge on the main-stem discharge, simple linear regression was used to compute the flow change with time at each of the stations discussed before. It must be recognized that the following discussion is based on regression analysis with rather low correlation coefficients, but it does provide some insight to the situation. Based on the regression analysis (see table 8), the annual mean discharges for tributary stations 07116000 and 07119500 have decreased at average rates of about 2.2 (ft'/s)/yr and 0.6 (ft'/s)/yr respec- tively. This obviously could account for the computed average decrease of 2.3 (ft'/s)/yr in the annual mean discharge at station 07123000. The addi- tional decrease of 2.4 (ft'/s)/yr on the average for its period of record at station 07128500 easily accounts for the computed decreases of 4.5 (ft'/s)/yr at station 07133000 and 4.4 (ft'/s)/yr at station 07137500. Removing an extremely high value recorded early in the record for station 07130500 from the analysis results in a computed decrease in the annual mean discharge of 3.7 (ft'/s)/yr on the average for its period of record. Regressions were computed for the two anomalous stations along this reach of the Arkansas River that did not indicate trends using the Kendall Tau analysis discussed previously. The regression analysis indicates that the annual mean discharge 66