Patternising phytoplankton dynamics of two shallow lakes in response to restoration measures by applying non-supervised artificial neural networks

Author： Talib A. Recknagel F. Molen D.

Publisher： Springer Publishing Company

ISSN： 0251-1088

Source： The Environmentalist, Vol.27, Iss.1, 2007-03, pp. : 195-205

Disclaimer: Any content in publications that violate the sovereignty, the constitution or regulations of the PRC is not accepted or approved by CNPIEC.

Previous Menu Next

Abstract

Long-term time-series data sets of two shallow Dutch lakes, Lake Veluwemeer and Lake Wolderwijd were subjected to ordination and clustering by means of non-supervised artificial neural networks (ANN). Splitting of the data sets into sub-series corresponding with three different management periods have allowed a comparative analysis of both the short-term seasonal and long-term phytoplankton dynamics in relation to the restoration measures. The lakes were considered as hyper-eutrophic and have been managed both with bottom-up and top-down management approaches. Results of the study have demonstrated that non-supervised ANN allow to elucidate causal relationships of complex ecological processes (1) within the specific genus, Oscillatoria and Scenedesmus and (2) the combination of external nutrient control and in-lake food web manipulation of the two lakes achieved to control eutrophication.