Nature Notes

Volume XXX - 1999

The Filling of Crater Lake
By F. Owen Hoffman

It was 30 years ago last summer since I last worked at Crater Lake. From 1966 to 1968 I was a seasonal park naturalist and during that time I conducted research on the vertical migrations of zooplankton in the lake. This research led to a masters degree in limnology and ecology from Oregon State University in 1969. Last August I volunteered for one week as a way to reacquaint myself with the park. I soon realized that I had forgotten many facts and quite a few had been revised since my last time in uniform. This dictated a visit to the park library.

One of the intriguing questions I came across in my brief survey of park-specific literature was: how quickly did the lake form, or how many years did it take to reach the present water level? I looked in the library, but found little or no information about what might seem to be a worthy focus of scientific investigation. The best thing I found was a paper published in 1994, where the authors claim that Crater Lake reached its current level in only 300 years (Nelson, et al. 1994). Being curious, and traveling with my laptop computer, I began to develop a simple mathematical model of the lake aimed at obtaining the answer (if only in an approximate way) to this question.


My mathematical model was initially constructed on the basic assumption that the hydrological processes of the past are not dramatically dissimilar to those of the present. For starters, I assumed:

(a) annual precipitation of 30.4 billion gallons (Phillips and VanDenburgh 1968);

(b) annual average evaporation of 15.2 billion gallons, based on an estimated evaporation rate of 120 cm per year (Redmond 1990);

(c) subsurface seepage that depends on the amount of water in the lake at any time.

For my initial calculations, I set seepage at 0.33 percent of the lake's volume per year (as derived from a net annual input of 15.2 billion gallons divided by the lake volume of 4.6 trillion gallons). Since evaporation will change somewhat with lake depth, I added an adjustment factor to account for the increase in surface area occurring as the lake filled to its present depth. I included an additional adjustment to the seepage rate to simulate more seepage with increasing depth as the lake filled. This adjustment accounts for both increased pressure from a rising lake level and increased bottom surface area. In the beginning, when there were only 230 billion gallons of water (assumed to have accumulated from ground water and hydrothermal sources), seepage is therefore estimated at 0.03 percent of lake volume. Conversely, when the lake became half full, seepage was about 0.2 percent of lake volume.

diagram estimating water volume in Crater Lake
Illustration used by J.S. Brode, estimating water volume in Crater Lake, September 1934.

The model is based on values for present conditions, but this cannot be assumed over the entire period of lake formation. Analysis of the lake sediment and samples of pollen indicate that, at least initially, the climate was not unlike today (Nelson et al., 1994). It is thought that Wizard Island formed not too long after the collapse of Mount Mazama and that the lake was within 250 feet of its current level at that time, Consequently, I assume that conditions of precipitation and temperature similar to the present day, existed for roughly 400 to 500 years after Mazama's collapse. Soon afterwards, however, a warmer and drier period commenced and lasted as long as 1,000 years. This drier period affected the entire region of the Pacific Northwest. Exactly how dry was it? No one really knows for sure, but to simulate this effect I reduced precipitation during this time period to between 30 and 50 percent of present day averages while also increasing evaporation by 10 to 20 percent.

To simulate the filling of Crater Lake, I used the software package STELLA II which has been specifically designed for the development of models for time dependent phenomena. This software is ideal for examining the effect of different assumptions about climate on the formation of Crater Lake.


Crater Lake was formed in the collapsed caldera of Mount Mazama. Its formation began soon after the collapse, almost as quickly as the caldera floor cooled to the point where water could accumulate on the bottom, The evidence to date suggests that the young lake probably resulted from ground water draining back into the caldera from the surrounding slopes and hydrothermal springs. Since annual precipitation did not vary substantially from that of the present climatic regime, the lake rapidly (in geological terms) increased in depth--gaining much more than the present net input of 15.2 billion gallons of water per year. The actual gain would depend upon the amount of evaporation, relative to the increase in surface of the lake, with early seepage having a negligible contribution to the annual loss.

chart show four model simulations of the filling of Crater Lake
Fig. 1: Four model simulations of the filling of Crater Lake with time. The first simulation (1:Crater Lake) assumes that present day annual precipitation has been constant over time. The second through fourth simulations (2-4:Crater Lake) assume for a 1,000 year altithermal period that prevailed between 500 and 1,500 years post collapse a reduced precipitation rate that is 0.7, 0.6 and 0.5 times that of the prsent, and an evaporation rate that is increased respectively 10%, 15%, and 20% above the present.

Based on these assumptions, Crater Lake probably reached about half of its present size within 150 years (Fig. 1), rising at a rate of some three to four feet per year. After roughly 300 years, the lake would have reached the level that was present at the time of Wizard Island's formation. This also means that at 400 years since the collapse of Mount Mazama, Crater Lake reached about 90 percent of its present volume.

Approximately 500 years after the collapse, a 1,000 year period of drier climate ensued. This caused the lake surface to fall steadily and by the time this "altithermal" or "xeric" period gave way to cooler and more humid conditions (not unlike the present time), the lake could have lost between 40 and 80 percent of the peak volume reached at the onset of the dry period. As precipitation increased, the lake once again started to rise and reached 95 percent of present day volume about 2,000 years after Mount Mazama collapsed. The final stage of filling Crater Lake took another 500 years and required conditions that produced a relative equilibrium among precipitation, seepage, and evaporation. Because of the changes in climate, complete equilibrium among precipitation, evaporation, and seepage did not occur for another 1,500 to 2,000 years thereafter. In other words, it took about 2,200 to perhaps more than 3,000 years for the lake to reach the present state of complete equilibrium with an average depth of 1,066 feet (325 m) and a maximum depth of 1932 feet (589 in). If it were not for the 1,000 year dry period, Crater Lake would have reached its present level between 1,100 and 1,500 years after Mount Mazama's collapse 7,700 years ago. Due to the 1,000 year dry period, evaporation and seepage exceeded precipitation and the lake fell to more than half its present level (Fig. 1). During this dry millennium it is likely that the lake water would also have been richer in minerals, more biologically productive, and thus less transparent than it is at present.

Crater Lake in the Future

Is the lake always going to stay like it is today? Given enough time, most certainly not. Perhaps the most imminent change, however, will be that of climate and its effect on precipitation. Any change in annual precipitation will have a direct effect on the level of Crater Lake. If precipitation begins to decline as climate changes to a drier condition, evaporation and seepage will again exceed precipitation so the lake level will drop. If precipitation increases, the lake level should rise and perhaps find a new equilibrium. There is no evidence to date that Crater Lake ever reached levels substantially above present. Nevertheless, such changes may occur over the next century or so. Over the next few thousand years, however, pronounced climate changes are certainly anticipated and with these changes will come fluctuations in lake levels, If we project ahead even further in time to, say, one million years or more, then it is likely that renewed volcanic activity or some other process of mountain building will occur, These major processes will dramatically change the appearance and structure of Crater Lake as we now know it. Few lakes on Earth have had a life-span that transcends a million years.


C.H. Nelson, et al., "The volcanic, sedimentologic, and paleolimnologic history of the Crater Lake caldera floor, Oregon: Evidence for small caldera evolution," Geological Society of America Bulletin 106 (May 1994), pp. 684-704.

K.N. Phillips and A.S. VanDenburgh, Hydrology of Crater, East, and Davis Lakes, Oregon. USGS Water Supply Paper 1859-E. Washington, D.C.: Government Printing Office, 1968.

K.T. Redmond, "Crater Lake climate and lake level variability," pp. 127-141 in E.T. Drake, et al (eds.), Crater Lake: An Ecosystem Study. San Francisco: Pacific Division, American Association for the Advancement of Science, 1990.

F. Owen Hoffman spent the summers of 1966 through 1968 as a seasonal naturalist at Crater Lake. He volunteered at the park for one week in 1998.

<<< Previous
> Cover <
Next >>>