Impact of flooding in oligotrophic water measured by AGP test: A statistical approach

Abstract

The aim of this study was to present a methodology using principal component analysis and non-parametric correlations to investigate the usefulness of the Algal Growth Potential (AGP) test with respect to the impact of the process of flooding on an oligotrophic system. The bioassay tests were carried out on samples collected over a 2 1 2 - year period, representing pre- and post-flooding conditions, from sites situated within the basin of the Desaulniers river (Bay James region, Quebec). The sites were chosen so that some were affected by the flooding whereas others were not. Statistical analysis of the data indicates that significant correlations existed between all the sites before inundation. The correlations continued to exist in the case of sites which were little-or not at all-affected by the flooding, whereas no significant relationships were observed when comparing each of the former sites with those affected by the inundation. Principal component analysis of the non-parametric coefficients of correlation related to the AGP data obtained from all sites of the reservoir, after formation of the reservoir, have shown: 1. (1) A general effect, characterised by the first axis, which could probably be explained by seasonal effects on the sites. 2. (2) A special effect, as shown on the second axis, very probably due to the impact of inundation. 3. (3) Residual effects, shown on the third axis, which were possibly the outcome of the regional hydrological and geographical characteristics of the study area.