Modeling of the Dynamics and Prediction of Transparency of a Mesotropic Tropical Lake: Bakré Lake (Abidjan, Côte d'Ivoire)

In this study, thirty samples were collected bimonthly from August 2015 to December 2016 at six locations of Bakre Lake and analyzed. Lake Bakre, located in the sub-equatorial Attiean climate, between the Atlantic Ocean in the South and the Ebrie Lagoon in the North, is an endorheic lake with a surface outfall, Abidjan (Cote d'Ivoire). A modeling study was conducted in order to determine a quantitative and qualitative relationship between transparency and five (5) physico-chemical descriptors (temperature, turbidity, nitrite ((NO 2 - ), nitrates (NO 3 - ) and ammonium (NH 4 + )). These descriptors were used as the explanatory and predictive parameters of the transparency of the selected samples taken at from Lake M'koa in a previous study. This study was conducted by using the Principal Component Analysis (PCA), the Ascending Hierarchical Classification (AHC). And then multilinear regression (MLR) and partial least squares (PLS) quantitative and qualitative models were proposed. These models were accredited to have good statistical indicators and so have been validated according to the principles established by the Organization for Economic Co-operation and Development (OECD). The statistical indicators of the MLR reveal more efficient predictions with the coefficient of determination R 2 = 0,920, the standard error RMCE = 0.169, correlation coefficient of cross-validation Q 2 cv = 0.920, and Fisher test F = 285.896. This model is acceptable with The values of the Transp theo / Transp exp report of the validation set tend (is closed) to unity, and the obtained results suggest that the combination of these 5 descriptors could be useful for predicting the property of Transparency. In addition, temperature is the priority (useful) descriptor for the prediction of transparency on the Bakre Lake stations.