Abstract
The impact of rainwater harvesting systems on biodiversity in hyperarid North African areas has not been assessed to date despite its demonstrated usefulness. The present study investigated this impact by considering the richness of wintering birds (RWB) in the Tataouine region (pre-Saharan Tunisia). We used data from three sets of variables-rainwater harvesting system type, microhabitat conditions, and topography-to determine the best predictors causing variation in the RWB by means of generalized linear mixed models. Our results show that the jessour system is the most attractive for the wintering bird species, followed by the tabia system, and lastly by the control areas. In the jessour system, RWB is influenced positively by slope and shrub cover and quadratically by tree cover, whereas this richness is affected positively by the extent of the herbaceous layer in the tabia system. In the control areas, RWB is negatively affected by elevation and quadratically affected by tree cover. Variation partitioning (VP) analysis reveals that (i) space is the most robust factor explaining the RWB in control areas (adj. R-2 = 0.15, p = 0.001), (ii) microhabitat is central in the tabia system (adj. R-2 = 0.10, p = 0.001), and (iii) the shared fraction between microhabitat and space is relevant in the jessour systems (adj. R-2 = 0.20). Specific management actions, especially preserving, maintaining, and promoting these traditional systems, are suggested to improve the attraction of wintering bird species to the Tataouine region. The implementation of a scientific watch system is recommended to understand how changes take place in such an arid environment.