Pixel-level land cover change detection in the Loess Plateau based on different data

The implementation of ecological engineering such as afforestation and reforestation in the Loess Plateau over the past two decades has brought significant changes to its land use and land cover (LULC). However, the differing spatial resolutions of land use data from various sources hinder an objective analysis of the effects of national ecological projects on vegetation growth in the Loess Plateau. Based on the commonly used MODIS LULC data and the latest CLCD LULC dataset from Wuhan University, the land use transition matrix calculation and change pixel detection were conducted for the period between 2001 and 2019 in the Loess Plateau. The analysis of the CLCD LULC data revealed that grass, cropland, and forest dominate the land use types in the Loess Plateau. Over the past two decades, approximately 15% (around 97000 km² ) of the Loess Plateau experienced land use changes. During this period, the forest and grass areas increased by 14025 km² and 7262 km² , respectively. These increases were mainly due to the conversion from cropland and shrub, which saw decreases in area by -18843 km² and -1671 km² , respectively. The patterns of land use changes were consistent across the seven provinces and regions within the Loess Plateau, with notable afforestation effects observed in Shanxi and Shaanxi provinces, where forest increased by 6303 km² and 4143 km² , respectively. Inner Mongolia primarily witnessed grassland restoration, resulting in a substantial growth of 5879 km² in grass. However, the results obtained from MODIS LULC data were contradictory. According to these results, the cropland area in the Loess Plateau increased by 20688 km² , while the grass area decreased by 26635 km² between 2001 and 2019, constituting 32.4% and 41.7% of the total changed area, respectively. The study outcomes provide valuable insights for scientifically analyzing the ecological benefits of afforestation and reforestation projects in the Loess Plateau, serving as a reference for improving the accuracy of carbon sink estimation in ecological engineering efforts.