Analysis of cropping patterns in Sudan's Gash Spate Irrigation System using Landsat 8 images

Abstract

Spate irrigation has been conducted in eastern Sudan's Gash Delta; however, availability of data necessary to understand actual conditions is limited. Therefore, we identified annual land cover types (irrigated crop land, non-crop land, and shrub land) in the Gash spate irrigation system using images obtained from Landsat 8. Further, we analyzed spatial cropping patterns and crop-fallow rotation characteristics for six irrigation blocks (Kassala, Mekali, Degain, Tendelai, Metateib, and Hadaliya) based on land-cover classification. To classify land cover types, we compared three vegetation indices, including the normalized difference vegetation index (NDVI), and examined six water-related indices, including the normalized difference water index (NDWI). Comparing the accuracy of these indices shows that NDVI and NDWIGS1 calculated from green and short-wave infrared performs best for land cover classification. We identified irrigated areas with a decision tree model using NDVI and NDWIGS1 time series data. The results showed that irrigated areas vary annually and are determined by available water resources from the Gash River. We found that the coefficient of variation of the irrigated areas is lower in the upper irrigation blocks (Kassala and Mekali) and higher in the lower irrigation blocks. We investigated crop-fallow rotation patterns and show that a two-year rotation pattern, in which irrigation and fallowing are conducted alternately, is dominant in the two upper irrigation blocks (Kassala and Mekali), while a three-year rotation pattern in which land lies fallow for two years is the main type adopted in the two lower irrigation blocks (Metateib and Hadaliya).