Heavy tail distribution and Deuterium excess for drought assessment case of Djelfa- watershed (Algeria)

Global warming has had significant effects on the hydrological cycle. In North Africa, these effects have resulted in a continued decrease in the amount of annual precipitation, greatly influencing the availability of surface and groundwater resources. This paper aims to evaluate drought episodes in a long dataset from Djelfa (Algeria) using a statistical approach based on a decision support system enabled by the Hyfran-plus software, to fit the best statistical distribution model and predict non-exceedance probability of precipitation. The standardized precipitation index is used to highlight sequences of meteorological drought and changes in weather patterns. Shift in precipitation has been identified by Wilcoxon Test for homogeneity; Wold-Wolfowitz test of independence, Kendall's test of trend and new cumulative mean difference curve (CMD) developed to visualize the trend of the precipitation and runoff dataset, Autocorrelation function (ACF) and Fourier spectrum and Deuterium excess isotopes has been used to visualize trends. Log–log plot; the mean excess function (MEF), AIC and BIC criteria are used to select the best distribution fitting model. The statistical results obtained indicate that the dataset follows a heavy-tailed distribution (Pearson type III), Values of excess Deuterium indicate that groundwater recharge has occurred in cold weather different from the actual period. SPI analysis shows that dry years cover more than 51% compared to 49% of wet years. The CMD curve shows a change and trend in the precipitation and runoff time series since 1974. Forecasts show drought recurrences to be between 3 and 200 years.