Mapping crop types and the crop water requirements over small-sized irrigated fields in the Khost Province of Afghanistan

Abstract

  Khost province of Afghanistan faces scarcity of water because of immature agricultural water management. Neither detailed information on agricultural land cover for specific crop types are available, nor are crop water demands estimated. Satellite remote sensing is a powerful tool for monitoring cropping areas and estimating crop water requirements. However, small fields in the Khost province makes a field-level analysis from satellite challenging. In this study, a crop-type classification was conducted by NDVI time-series provided by Sentinel-2 satellite images with 10-m resolution. The accuracy assessment using 1,804 ground-truth data indicated the crop classification was successful, with 90% overall accuracy. The accurate crop information enables us to estimate the crop water requirements of the agricultural region, with the spatial distribution. Crop water requirements, which is identical to the field evapotranspiration requirements, were estimated for each crop type, by referencing evapotranspiration computed by ground-measured weather data and by determining crop coefficients, which were determined by interviewing the local farmers and literature. The resultant map of crop water requirements is the spatial information of field evapotranspiration requirement, and it is expected to contribute towards advanced irrigation and water resources planning and management in the Khost province in future.