Spatiotemporal trend analysis of runoff and actual evapotranspiration in Northern Algeria between 1901 and 2020

Abstract

In water resources management, the estimation of the water balance component is mainly used to control dams, agriculture, and irrigation developments and also to provide water supply for drinking and industry. The spatial trend analysis applied to the Northern Algerian region to determine the significant changes in rainfall–runoff (IARR) and actual evapotranspiration (IAEa) required a sufficient period. In this study, we carried out this analysis in four periods between 1901 and 2020, according to the data availability of climate variables (IAR, and IAT), since the hydrological measurement posed a big problem during this period. In this regard, a set of nonparametric and empiric water balance models were applied and compared on five bioclimatic floors of Northern Algeria between 1991 and 2020 to estimate IARR and IAEa between 1901 and 1990. The results obtained from several tests, such as regression analysis, scattergram graphs, R2, R2Adj, MSE, RMSE, and DW, proved that Yang's model gives the best performance and good results in very humid regions when replacing n by 2. On the other hand, Schreiber and Pike’s models are more efficient in the humid and semi-humid regions, respectively. Furthermore, the model proposed by Sharif is well performed in the Mediterranean and semi-dry areas. The variability analysis given by the t test and the regression curves is applied on IAEa and IARR in each bioclimatic floor during intervals measurement of [1901, 1930], [1931, 1960], [1961, 1990], and [1991, 2020]. It showed a significant change in both components between 1961 and 1990 in very humid, semi-humid, and Mediterranean regions when compared with observations done between 1931 and 1960. Moreover, a large variation and significant trend in IARR were also observed in humid and semi-humid areas between 1991 and 2020 compared to the previous periods.