Cropping intensity map of China with 10 m spatial resolution from analyses of time-series Landsat-7/8 and Sentinel-2 images

Abstract

Cropping intensity maps at high spatial resolution play a crucial role in guiding agricultural policies and ensuring food security. So far, most of nationwide cropping intensity maps have been developed from satellite images at moderate or coarse resolutions. In this study, we first assembled and integrated time-series dataset with high spatial resolution, specifically Landsat-7, Landsat-8 and Sentinel-2 imagery in 2017. We then used an object- and phenology-based algorithm and integrated images to create a 10-m resolution cropping intensity map over China. The map evaluation results revealed an overall accuracy of 96.68 ± 0.01 % and a Kappa coefficient of 0.90. In 2017, single cropping dominated the agricultural practices in China, with an approximate area of 1.189 × 106 km2 ± 7.90 × 103 km2, constituted 79.26 % of the entire cropland area. Simultaneously, double and triple cropping covered approximately 0.306 × 106 km2 ± 8.03 × 103 km2 and 5.00 × 103 ± 1.75 × 103 km2, corresponding to 20.41 % and 0.33 % of the entire cropland area, respectively. On average, the national multiple cropping index (MCI) was 1.21. The results in the study prove the reliability of the generated mapping products and high potential of the developed mapping framework (the algorithm and integrated datasets), which can be readily applied to quantify the interannual changes of cropping pattern on a nationwide level with a high spatial resolution.

Keywords

• Cropping intensity
• Google Earth Engine
• Crop phenology
• Remote sensing
• Crop growth cycle