Mechanisms driving streamflow alterations in algerian catchments along a hydroclimatic gradient

Water security is a major issue in semi-arid and arid regions in which the hydrological cycle is adversely impacted by extensive anthropogenic and climatic changes. Consequently, understanding the mechanisms that underline the hydrologic alteration of rivers in these regions is a key to sustainable water resources management. Algeria is a country in these regions that can be considered a hot spot for climate change. Due to its particular position in North Africa, its climate differs significantly from east to west. Also, other geographical properties such as topography and land use differ along the climatic gradient. As a result, the streamflow regime of Algerian rivers is characterized by high spatial and temporal variability and subjected to long-term alterations. Yet the contribution of the different climatic and geographic variables to these alterations still needs to be quantified. The major aim of this thesis is to investigate the long-term changes of hydroclimatic variables in three catchments in Northern Algeria (the Medjerda, the Isser, and the Macta catchment) and to explore the potential driving factors of streamflow alteration in these catchments. This investigation is expected to identify to what extent the hydrological changes in the Algerian basins are driven by climate or human activities. In the first part, we analysed the temporal trends of the different hydrometeorological variables using non-parametric Mann-Kendall and Sen's slope estimator tests respectively. In the second part, we used an empirical method, Convergent Cross mapping (CCM), to identify causal drivers of streamflow change in hydrological systems. This latter method is based on nonlinear state space reconstruction and has been considered recently as being a powerful method for causality analysis in complex dynamic systems. The trend analysis showed that the streamflow of the Medjerda and Macta basins have predominantly decreasing trends, while the trend is increasing for the Isser catchment. The observed trends of streamflow are not necessarily coherent with the trends of climate and crop vigor factors, such as temperature, evapotranspiration, and NDVI, which all are increasing within the 3 basins. This incoherency between trends of streamflow and possible driving factors justifies that more advanced causal analysis methods are used to unravel the determining factors of streamflow alteration. The results of such advanced analysis using CCM analysis indicate that climate-related factors, particularly precipitation variability, temperature, and evapotranspiration, are the main drivers of streamflow changes in the selected basins. In a second order, human activities, such as changes in land use, increased agriculture area, water withdrawal for irrigation, and excessive groundwater extraction, have also played a role in altering streamflow. Yet, the alteration in streamflow is largely attributed to the interaction between climate change and anthropogenic activities. The specific mechanisms by which human activities alter streamflow and the extent to which they contribute to these changes on top of climate change require further analysis. In summary, this part of the research shows that the CCM method can provide valuable causal links from common hydroclimatic and environmental time series, which is crucial to a wide range of applications, but it should be used and interpreted with caution.