Interannual fluctuations in primary production: Direct physical effects and the trohic cascade at Castle Lake, California

Abstract

Direct physical effects and cascading trophic interactions operate together to determine interannual variability in the seasonal pattern of primary productivity at Castle Lake, California. Principal component analysis was used to investigate the depth‐time distributions of productivity for summers 1961–1986. Two characteristic patterns were found, together accounting for 60–70% of the year‐to‐year variability. The first pattern corresponds to the deep productivity maximum that forms in the hypolimnion in June and July. Variability in this first pattern arises from the direct effects on phytoplankton populations of year‐to‐year changes in timing of ice breakup and hydraulic flushing in spring. The first pattern is unusually strong or weak, compared to the long‐term average, during the phenomenon of El Niño‐Southern Oscillation. A second characteristic pattern corresponds to the mixed‐layer productivity maximum that develops in August and September. Variability in this second pattern arises from trophic interactions at higher levels: rainbow trout feed on late‐summer Daphnia rosea populations, which in turn graze on the mixed‐layer algal community. In addition, some of the variability in the rainbow trout population arises from year‐to‐year differences in angling pressure from humans. These results demonstrate how interannual variability in an ecosystem property (primary productivity) can be controlled by forces acting simultaneously through the top and the base of the food web.