ever, below those observed in the past 10 years without risk of adverse occurrences that are not described in the detailed monitoring record (1990-present; analyses complete through 1998). A negative association between the welfare of the species and the lake level could emerge if lake levels are reduced below those of recent historical experience. The absence of any empirical connection between the observed lake levels and the welfare of the endangered suckers cannot be taken as justification for continuous or frequent operation of the lake at the lowest possible levels, given that the effects of operating the lake at lower levels are undocumented. Thus, while the observational record contradicts important underlying assumptions of the RPA, it does not provide an endorsement for the lake levels proposed in the USBR biological assessment, which, if implemented, could take interannual mean lake levels well below those of recent historical observation.
The potential benefits of higher lake levels in Clear Lake, Gerber Reservoir, and Tule Lake sump are more difficult to evaluate, because the record of analysis and observation for these water bodies is not as extensive as that for Upper Klamath Lake. These lakes have not suffered notable mass mortality in association with low lake levels, but Clear Lake populations showed poor body condition following severe drawdown in the early 1990s. The USFWS provides reasonable support for lake levels in Clear Lake no lower than the recent drought-related minimum (1992–1993:4,519 feet). The RPA reasonably adds a margin of two feet (4,521) to allow for water loss in the absence of withdrawals under drought conditions.