Abstract:

The aim of this study is to identify an optimal management rule of the water storage of the Nebhana on the basis of historical available data. This dam is located in the central part of Tunisia and was been constructed in 1967 for irrigation purposes mainly.

The time series of water inflow (water volume runoff to the dam) and outflow (evaporation, infiltration, seepage, spilled way and crops water needs) are constructed and analysed firstly. In a second step, probability distribution function and forecasting models were been calibrated respectively for the water inflow and outflow series.

A dynamic stochastic model was been applied to carry out the optimal management rule of the water storage in the Nebhana dam. This rule will permit to estimate the optimal water volume that could be used for irrigation on the basis of actual storage water volume in this dam. In the dynamic stochastic model, the variability of the water inflow is expressed by its occurrence probability, the water outflow losses by theirs forecasting models and the management goals by weighting coefficients.

In the case of the Nebhana dam, the inflow water series is characterized by its comparative high variability, a non-normal statistical distribution and a decrease tendency. The water volume losses by seepage and infiltration throughout the dam are highly correlated to the stored water volume, this correlation is of quadratic type with a very high significant level. Compared to the stored volume, these losses are not negligible. Also, the lost water volume by evaporation are important and are related to storage water volume by linear model that is specific of the considered month.

The management objectives considered in this study are conflict and concern both the agricultural water demand satisfaction and a minimal water storage safeguard in the dam. By considering different weighting values affected to these objectives, different optimal management rules were been obtained. The identification of the best optimal management rule is accomplished by taking into account descriptor indicators deduced from simulation analysis. These indicators characterising the risk, the vulnerability and the resilience of the phenomena, such as water storage failure, spilled way volume or unsatisfactory of the water demand, are useful to analyse the system performance.

The predicted optimal spill way water volumes by means of the best optimal management rule identified in this study vary from one month to another and depend upon the stored water volume; in all the cases, they are rather lower than the maximal agricultural water needs of the hall irrigated area.

1. Ecole Supérieure d’Agriculture de Mograne, 1121 Mograne, Tunisia
2. Institut National Agronomique de Tunisie, 43 Avenue de Charles Nicolle, 1082 Tunis-Mahrajène, Tunisia