Prospective analysis of spatial heterogeneity influence on the concordance of remote sensing drought indices: a case of semi-arid agrosystems in Morocco (Moulouya and Tensift watersheds)

Several drought indices are commonly used to quantify the severity and magnitude of agricultural drought. A comprehensive assessment of temporal and spatial convergences is important. This study evaluates the spatio-temporal concordances of eight drought indices, between October and April, over the period 2003-2021 under the semi-arid conditions of two watersheds in Morocco. These include three monovariate, two bivariate and three composite indices. Taylor diagram, linear regression and Random Forest (RF) clustering were used to evaluate and compare the overall performance of the indices. The results indicate that spatio-temporal variability and spatial heterogeneity of the different agrosystems have little influence on the concordances of composite indices while TCI and VCI have a variable spatio-temporal concordance depending on the spatial extent considered. Nevertheless, depending on the spatial extent considered and the agroclimatic characteristics of each agrosystem, the seasonal correlations between the composite indices and precipitation anomalies, net primary productivity and cereal yield anomalies vary from 0.52 to 0.79, 0.23 to 0.78 and from 0.33 to 0.73, respectively. The correlations between the simple indices and precipitation anomalies, net primary productivity, and cereal yield anomalies range, respectively, from 0.33 to 68 0.13 to 0.72, and from 0 to 0.84. In general, the composite indices showed the best agreement between them and reference indicators but the VCI seems better suited for the assessment of the seasonal profitability of studied agrosystems.

Automated summary

This study evaluates the spatio-temporal concordances of eight drought indices in two semi-arid watersheds in Morocco. The results show that the composite indices have the best agreement with reference indicators, and the VCI is better suited for assessing the seasonal profitability of the studied agrosystems.