Land use and land cover (LULC) projections are critical for climate models to predict the impacts of LULC change on the Earth system. Different assumptions and policies influence LULC changes, which are a key factor in the decisions of planners and conservationists. Therefore, we predicted and analyzed LULC changes in future scenarios (SSP1-26, SSP2-45, SSP5-85) in the middle reaches of the Yangtze River basin (MYRB). We obtain historical (i.e., 2005-2020) LULC data from the Google Earth Engine (GEE) platform using the random forest (RF) classification method. LULC data for different future scenarios are also obtained by the driving factors of LULC changes in future shared socioeconomic pathways (SSPs), representative concentration pathways (RCPs) (SSP-RCP) scenarios (i.e., 2035-2095) and the patch-generated land use simulation (PLUS) model. The major findings are as follows: (1) simulation using the PLUS model based on the acquired classification data and the selected drivers can obtain accurate land use data in MYRB and a Kappa coefficient of 89.6% and 0.82, respectively; (2) as for the LULC changes in the MYRB, forests increased by 3.9% and decreased by 1.2% in the SSP1-26 and SSP5-85 scenarios, respectively, while farmland decreased by 9.2% and increased by 13.4% in SSP 1-26 and SSP 2-45, respectively, during 2080-2095; and (3) the main conversions in LULC in the MYRB were farmland to forest, forests/water bodies to farmland, and forests/grasslands to farmland/buildings in SSP1-2.6, SSP2-4.5, and SSP 5-8.5, respectively. This can be mainly attributed to gross domestic product (GDP), population (POP), temperature, and precipitation. Overall, this study not only contributes to the understanding of the mechanisms of LULC changes in the MYRB but also provides a basis for ecological and climatic studies.
Keywords: Patch generated land use simulation model; Random forest; Remote sensing; SSP-RCP; Yangtze River basin.
Copyright © 2022 Elsevier B.V. All rights reserved.