High-resolution Mapping of Irrigated Areas & Changes at Planetary Scale
Sound water resources planning and management require adequate data with sufficient spatial and temporal resolution. This is especially true in the context of irrigated agriculture, which is one of the main consumptive users of the world's freshwater resources. While ground-based data on acreages of irrigated land are scarce and often unreliable, the adoption of remote sensing technology in the planning and management of irrigated agricultural systems has never been fully operational in the development context. The reason for this is high costs, the need for skilled workers and remote sensing methods that usually only work in one place and time. Recent developments in cloud computing with services available such as Google Earth Engine (GEE) can help to overcome some of these constraints.
hydrosolutions Ltd. has implemented a fully automatic irrigation mapping procedure in GEE that uses surface reflectance satellite imagery from different sensors. Ground-based data are not required. The method is based on unsupervised clustering of image pixels within training areas identified through object-based segmentation. The clustering technique distinguishes productive irrigated fields from non-productive and non-irrigated areas on a monthly basis. A number of automatized post-processing steps are then carried out, making use of multi-temporal data, to further improve the classification accuracy and to produce final monthly and annual irrigation maps (30 m resolution).
The novel method is applied to Inner Asia, including Afghanistan, Central Asia, and Western China where we have calculated the development of irrigated areas from 2000 to 2017 and assesses the classification results in terms of robustness and accuracy. Based on six available testing scenes (in total 700'000 pixels) the classification accuracy is 77 - 93 percent.
For the Central Asian Chu and Talas River Basins that are shared between upstream Kyrgyzstan and downstream Kazakhstan, we have deployed a monitoring tool that allows to map irrigated areas and, based on these data, compute irrigation scheme efficiencies. Results show that on the Kyrgyz side of the Talas basin, derived increasing trends over the years are highly significant (23 percent area increase between 2000 and 2017). In the Chu Basin, we identify significant correlations between in-season precipitation and annual irrigated areas, reflecting on how agricultural production is limited by access to scarce water resources. We show how the multi-temporal, high-resolution irrigation maps derived in this study can help local stakeholders to study the performance of vast and complex irrigation systems and help in monitoring irrigation activities from field scales up to transboundary levels.
The tool can also be used to monitor canal catchment level development of irrigated areas. This allows authorities to readily obtain important indicators of the performance of the vast irrigation systems in the region.
Jointly with the World Bank, hydrosolutions Ltd. is also developing a monitoring technique to monitor the development of irrigated areas in sub-humid Eastern Africa with a special focus on Mozambique.