Phytoplankton species predictability increases towards warmer regions
Limnol. Oceanogr., 57(4), 2012, 1126-1135 | DOI: 10.4319/lo.2012.57.4.1126
ABSTRACT: We explored systematic patterns in predictability of phytoplankton species from 83 lakes over a gradient ranging from subpolar to tropical regions in South America. We estimated the explained variance (proxy of predictability) of the presence and biomass (estimated as biovolume) of species using multiple regressions from commonly measured environmental variables such as nutrient levels, light, mixing depth, temperature, and zooplankton biomass. Both the presence and biomass of species occurring at least in 10 lakes were quite well predicted from the environmental variables, with average values of 35% and 58%, respectively. Predictability was not systematically related to phylogenetic affiliation or particular functional groups as defined by morphology. However, biomass predictability decreased with increasing occurrence, and improved with larger species size (maximum linear dimension). Species that were predictable in terms of biomass (R2 ≥ 0.5, p ≤ 0.05) had, on average, a larger volume, and were relatively more frequent in lakes from warmer regions, with high water temperature, low chlorophyll a, low nutrient concentrations, and low total zooplankton biomass. Although we cannot diagnose the mechanisms involved, our finding that the number of predictable species increases towards warmer regions resembles situations where competition for nutrients and grazing are likely to be less severe, and may imply that in a future warmer world phytoplankton will be easier to predict.