Long-Term Patterns in the Population Dynamics of Daphnia longispina, Leptodora kindtii and Cyanobacteria in a Shallow Reservoir: A Self-Organising Map (SOM) Approach.

Abstract

The recognition of long-term patterns in the seasonal dynamics of Daphnia longispina, Leptodora kindtii and cyanobacteria is dependent upon their interactions, the water temperature and the hydrological conditions, which were all investigated between 1999 and 2008 in the lowland Sulejow Reservoir. The biomass of cyanobacteria, densities of D. longispina and L. kindtii, concentration of chlorophyll a and water temperature were assessed weekly from April to October at three sampling stations along the longitudinal reservoir axis. The retention time was calculated using data on the actual water inflow and reservoir volume. A self-organising map (SOM) was used due to high interannual variability in the studied parameters and their often non-linear relationships. Classification of the SOM output neurons into three clusters that grouped the sampling terms with similar biotic states allowed identification of the crucial abiotic factors responsible for the seasonal sequence of events: cluster CL-ExSp (extreme/spring) corresponded to hydrologically unstable cold periods (mostly spring) with extreme values and highly variable abiotic factors, which made abiotic control of the biota dominant; cluster CL-StSm (stable/summer) was associated with ordinary late spring and summer and was characterised by stable non-extreme abiotic conditions, which made biotic interactions more important; and the cluster CL-ExSm (extreme/summer), was associated with late spring/summer and characterised by thermal or hydrological extremes, which weakened the role of biotic factors. The significance of the differences between the SOM sub-clusters was verified by Kruskal-Wallis and post-hoc Dunn tests. The importance of the temperature and hydrological regimes as the key plankton-regulating factors in the dam reservoir, as shown by the SOM, was confirmed by the results of canonical correlation analyses (CCA) of each cluster. The demonstrated significance of hydrology in seasonal plankton dynamics complements the widely accepted pattern proposed by the plankton succession model for lakes, the PEG (Plankton Ecology Group), and may be useful for the formulation of management decisions in dam reservoirs.