Agriculture land use recognition by multi-temporary middle resolution satellite data in Northern Kazakhstan

Spring crops are a monoculture in Northern Kazakhstan. Mid-resolution satellite data of red and near infrared spectral bands from the Russian satellite, Resource MSU-SK and Terra´s MODIS satellite can be used for assessment of regional landuse and land cover. The objective of this research is to improve the classification of spring crops, range and fallow land from mid-resolution satellites. The study was conducted for the Akmola Oblast for the 2000 and 2001 vegetation season. Three specific phenological periods were selected to maximize the spectral differences between landuse types. Images were acquired for clear sky conditions and composites developed for each for the three periods. Training sites were identified and supervised classification performed from the multi-temporal composite. Classification accuracy was assessed based on farmer interviews for information on landuse for more then 800 agricultural fields. The classification error analyses suggested 13% for spring wheat, 141% for fallows, 20% for multiyear grasses classes. The errors in classification are attributed primarily to the unusual spectral characteristics of certain agricultural fields and the lack of cloud free imagery during the critical phenological periods.