Prediction of the Dynamics of Total Chlorophyll-a in the Waters of the Ebrie Lagoon, Abidjan (Ivory Coast)

Abstract

The aquatic environments which represent a vital resource for man are threatened by various pollution of anthropogenic origin. Chlorophyll-a being the consequence of the nutrient enrichment of water and linked to primary production, participates in the degradation of water quality by the proliferation of phytoplankton. Algal development depends on physical, chemical and biological factors in the aquatic environment. The objective of the study is to determine a better model for predicting chlorophyll-a concentrations from the environmental variables of the waters of the Ebrié lagoon. The predictive model by multilinear regression of total chlorophyll-a allowed to generate excellent statistical indicators and final descriptors of influence (R² = 0.9076; RMCE = 0.1296; F = 216.009). The external validation tests of Tropsha et al and Roy et al are all verified and the model obtained has a good predictive power. Thus, suspended solids (SS) (33%) and pH (23%) were the most influential descriptors of the water model of the Ebrié lagoon. At the level of the applicability domain of the model the value of the threshold lever is h*=0.875. The observations of the training set are all within the domain, so the prediction of chlorophyll-a concentrations is reliable.