A Synergy Cropland of China by Fusing Multiple Existing Maps and Statistics
Miao Lu,
Wenbin Wu,
Liangzhi You,
Di Chen,
Li Zhang,
Peng Yang,
Huajun Tang
Affiliations
Miao Lu
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Wenbin Wu
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Liangzhi You
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Di Chen
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Li Zhang
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Peng Yang
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Huajun Tang
Key Laboratory of Agricultural Remote Sensing, Ministry of Agriculture/Institute of Agricultural Resources and Regional Planning, Chinese Academy of Agricultural Sciences, Beijing 100081, China
Accurate information on cropland extent is critical for scientific research and resource management. Several cropland products from remotely sensed datasets are available. Nevertheless, significant inconsistency exists among these products and the cropland areas estimated from these products differ considerably from statistics. In this study, we propose a hierarchical optimization synergy approach (HOSA) to develop a hybrid cropland map of China, circa 2010, by fusing five existing cropland products, i.e., GlobeLand30, Climate Change Initiative Land Cover (CCI-LC), GlobCover 2009, MODIS Collection 5 (MODIS C5), and MODIS Cropland, and sub-national statistics of cropland area. HOSA simplifies the widely used method of score assignment into two steps, including determination of optimal agreement level and identification of the best product combination. The accuracy assessment indicates that the synergy map has higher accuracy of spatial locations and better consistency with statistics than the five existing datasets individually. This suggests that the synergy approach can improve the accuracy of cropland mapping and enhance consistency with statistics.
