Increasing pressure on available water resources in semi-arid region will affect the availability of water for irrigated agriculture. In this context, adoption of innovative and cost-effective tools for water management and analysis of water use patterns in irrigated areas is required for an efficient and sustainable use of water resources.
This study aims to evaluate a remote sensing-based approach which allows estimation of the temporal and spatial distribution of crop evapotranspiration and irrigation water requirements over large irrigated areas. The method consists of an adaptation of the daily step FAO-56 Soil Water Balance model combined with time series of basal crop coefficient (Kcb) and the fractional vegetation cover (fc) derived from high resolution satellite NDVI imagery (Normalized Difference Vegetation Index).
The model was first calibrated and validated at plot scale using evapotranspiration measured by eddy-covariance systems over wheat fields and olive orchards which represents the main crops grown in the study area of Haouz plain located around Marrakesh city. Then, the model was used to compare remotely sensed estimates of irrigation water requirements and observations of irrigation water use at plot scale over an irrigation district in Haouz plain along three agricultural seasons.
At plot scale, the results showed that the model provides good estimates of evapotranspiration for wheat and olive trees, given specific calibration of crop and soil related parameters that control transpiration and evaporation processes. At the irrigation district scale, the comparison of spatialized irrigation water requirements and irrigation water use showed great discrepancies indicating a temporal and spatial varying demand and supply of irrigation water over the seasons. In addition to the model and the observed data uncertainties, the variability observed could be influenced by different biophysical factors and the farmer’s behaviour and management practices. Also, some differences between observations and estimates were spatially correlated at some extent with the distribution of wells in the area, which shows a potential use of the method for monitoring groundwater withdrawals. These results suggested that the water supplied within the studied irrigation district needs to be improved for a better performance of irrigation. The findings demonstrate the potential interests for irrigation managers of using remote sensing-based models to assess irrigation water requirements and monitor irrigation water use for efficient and sustainable use of water resources.