Laserfiche WebLink
r <br />Crystal Reservoir Temperature Study <br />Page 5 <br />Septe~aber 23, 1988 <br />sets 503 confidence limits for the predicted water temperatures. <br />Thus, 503 of the actual water temperatures fall within ~ ± S of the <br />model predictions. The explicit equations for ~ and S are: <br /> <br />S = .6745 ST.x <br />where <br />T; ith predicted temperature from heat transport model <br />Tj jth predicted temperature from regression model <br />N number of predicted temperatures in sample set <br />i j <br />ST,x =_ modified standard difference of estimate <br />Calibration of the Temperature Model was accomplished by adjusting <br />the internally-defined model coefficient for air temperature until <br />the probable difference was reduced as much as possible while <br />increasing the coefficient of determination. Special emphasis was <br />placed on the Stateline Gage, although attention was given to all <br />validation nodes during the calibration process. Adjusting the air <br />temperature coefficient accounts for transposition of air tempera- <br />ture from the Grand Junction weather station to each node in the <br />network. An assumption is made that the validation gage regression <br />models are accurate. <br />The probable and mean differences for the Gunnison River Gage below <br />Gunnison Tunnel on an annual basis were -4.47°C and 1.25°C, respec- <br />tively. Sporadic water temperature data at the Gunnison River Gage <br />below the Gunnison Tunnel were obtained from the USGS database. The <br />large discrepancy between regressed and model temperatures (Table 5) <br />may be attributed to one or more of several factors. First, the <br />Gunnison diversion dike is immediately above this gage; water is <br />backed up and additional heating of the water occurs. This anomaly <br />is not accounted for in the temperature model. Second, Water <br />temperatures below Crystal Dam were obtained from the Colorado <br />State Fish and Game Commission; they recorded temperatures using <br />continuous thermographs; if placement of these thermographs were in <br />abnormally deep water, water temperature data may be lower than what <br />actually occurred Nonetheless, these anomalies do not explain all <br />of the error, so further analysis at this gage is needed. <br />The mean difference and probable errors for the Gunnison Gage Near <br />Lazear are in Table 5. The model over-predicted water temperatures <br />at this point for all time periods. The annual mean difference (all <br />time periods) was -1.67°C; the annual probable difference was <br />1.76°C. <br />