Potentials of RapidEye time series for improved classification of crop rotations in heterogeneous agricultural landscapes: experiences from irrigation systems in Central Asia

Abstract

In Central Asia, more than eight Million ha of agricultural land areunder irrigation. But severe degradation problems and unreliable waterdistribution have caused declining yields during the past decades.Reliable and area-wide information about crops can be seen as importantstep to elaborate options for sustainable land and water management.Experiences from RapidEye classifications of crop in Central Asia areexemplarily shown during a classification of eight crop classesincluding three rotations with winter wheat, cotton, rice, and fallowland in the Khorezm region of Uzbekistan covering 230,000 ha ofirrigated land. A random forest generated by using 1215 field sampleswas applied to multitemporal RapidEye data acquired during thevegetation period 2010. But RapidEye coverage varied and did not allowfor generating temporally consistent mosaics covering the entire region.To classify all 55,188 agricultural parcels in the region threeclassification zones were classified separately. The zoning allowed forincluding at least three observation periods into classification.Overall accuracy exceeded 85 % for all classification zones. Highestaccuracies of 87.4 % were achieved by including five spatiotemporalcomposites of RapidEye. Class-wise accuracy assessments showed theusefulness of selecting time steps which represent relevant phenologicalphases of the vegetation period. The presented approach can supportregional crop inventory. Accurate classification results in early stagesof the cropping season permit recalculation of crop water demands andreallocation of irrigation water. The high temporal and spatialresolution of RapidEye can be concluded highly beneficial foragricultural land use classifications in entire Central Asia.