The availability of multi-spectral remote sensing imagery (from multiple sources including Landsat, MODIS, Sentinel etc) with a long term archive and machine learning techniques allows for the possibility of estimating with great degree of spatial resolution the area sown under different crop types.
If this could be successfully achieved it would open up the possibility of calculating with a high degree of certainty river basin or watershed level agricultural water demand. Agriculture being by far the largest sectoral user of water, knowing the scale of its water footprint accurately is vital to water resource planning. Scan this page for updates on our progress in this effort.