In order to determine the nature of the high salinisation rate of the waters of Lake Sidi Boughaba, which has been listed in the Ramsar list since 1980, 23 samples that were taken during four sampling operations were subjected to physicochemical analyses. The obtained results were processed using a combination of bi-varied methods (correlation tests) and multivariate statistical methods (principal component analysis – PCA). The physicochemical analyses reveal that they are alkaline waters with a pH ranging between 8.38 and 9.03, an electrical conductivity ( EC) of the order of 12.4 to 17.4 mS∙cm ^–1, and high levels of Na ⁺ and Cl ^–, up to 3700 and 6630 mg∙dm ^–3 respectively, indicating a marine origin of these waters. In addition, the statistical treatment revealed that the mineralisation of the waters of this ecosystem is controlled by four main mechanisms of the salinisation; the main mechanism underlying this strong mineralisation is due to the impact of the marine spray. The second-order processes are about the phenomenon of the ion exchange, the dissolution/precipitation of evaporitic and carbonate formations, the oxidation–reduction processes, notably the reduction of sulphates as well as biochemical phenomena due to the selective absorption of certain ions by fauna and flora.

This article accounts for the development of a powerful artificial neural network (ANN) model, designed for the prediction of relative humidity levels, using other meteorological parameters such as the maximum temperature, minimum temperature, precipitation, wind speed, and intensity of solar radiation in the Rabat-Kenitra region (a coastal area where relative humidity is a real concern). The model was applied to a database containing a daily history of five meteorological parameters collected by nine stations covering this region from 1979 to mid-2014.
It has been demonstrated that the best performing three-layer (input, hidden, and output) ANN mathematical model for the prediction of relative humidity in this region is the multi-layer perceptron (MLP) model. This neural model using the Levenberg–Marquard algorithm, with an architecture of [5-11-1] and the transfer functions Tansig in the hidden layer and Purelin in the output layer, was able to estimate relative humidity values that were very close to those observed. This was affirmed by a low mean squared error ( MSE) and a high correlation coefficient ( R), compared to the statistical indicators relating to the other models developed as part of this study.