Hydrology of Crater, East and Davis Lakes, Oregon by Kenneth N. Phillips
WATER LOSS FROM THE LAKE
SEEPAGE RATE IN RELATION TO LAKE LEVEL
The seepage rate varies with the stage of the lake, and its magnitude may be approximated, or its probable limits defined, by two different approaches.
Consider first the relationship between average lake stage and seasonal inflow with adjustment for changes in the volume of water stored in the lake. The lake level varies with annual runoff and is higher after periods of above-normal inflow. For example, in the 4 dry water years, 1940, 1941, 1942, and 1945, the average lake level was about 4,380 feet and the lake was lowered in the 4years by only a small volume, less than 1,000 acre-feet. The total runoff at the gaging station near Crescent in those 4 years was 159,000 acre-feet. The average annual seepage loss in those years, therefore, must have been about 40,000 acre-feet (!55 cfs) plus the unmeasured flow of streams draining 60 square miles below the station on Odell Creek. By contrast, in the 3 wet years, 1951, 1952, and 1956, the average lake level was about 4,391 feet; its volume increased by 30,000 acre-feet; and the measured runoff at the station near Crescent was 255,000 acre-feet. The annual seepage loss at that lake level must have been about 75,000 acre-feet (104 cfs) plus the flow of the ungaged tributaries. Therefore, the seepage rate must be more than 54 cfs when the lake is near altitude 4,380 feet and more than 104 cfs when it is near altitude 4,391 feet.
The other approach is to consider the relationship between the momentary lake level and the unsubmerged flow of the group of springs feeding Davis Creek downstream from Davis Lake and its lava dam. A well-defined and nearly straight-line relationship has existed for many years between the flow of those springs and the stage of Davis Lake (fig. 11), and that fact warrants the belief that those springs derive a large part of their flow from the seepage of Davis Lake, even though efforts to trace the lake outflow to the springs have not been successful (Oregon State Game Comm., written commun., 1961). The points ‘plotted in figure 11 cover a time period of 24 years, and the discharges for all plotted points lie within 10 percent of the discharge given by the average curve. Data are not available to plot antecedent lake levels against flows with various periods of delay for the probable time of underground travel. Both the stage of the lake and the flow of the springs no doubt respond to general runoff conditions, but the correlation shown in figure 11 seems so well defined that spring flow is believed to be a function of lake level. The lava that fills the creek channel does not seem to provide much storage to delay the passage of the water that leaks into it.