It is increasingly acknowledged that land-use and land-cover change has become a key subject that urgently needs to be addressed in the study of global environmental change. In the present study, supported by the long-time-series of land-use and land-cover data from 1990, 2000, and 2017, we used the land-use transition matrix, Markov chain model and Moran’s I to derive detailed information of the spatial patterns and temporal variation of the land-use and land-cover change; additionally, we highlight the deforestation/afforestation conversion process during the period of 1990–2017. The results show that a total of 4708 km² (i.e., 2.0% of the total area) changed in Guangxi from 1990 to 2017, while 418 km² of woodland has been lost in this region. The woodland lost (deforestation) and woodland gained (afforestation) were collocated with intensive forest practices in the past 27 years. The conversions from woodland to cropland and from woodland to grassland were the dominant processes of deforestation and afforestation, respectively. Steep slope cropland was one of the major conversion patterns of afforestation after 2000. This result is mainly explained by the implementation of the “Grain for Green Program” policy and the large-scale development of eucalyptus plantations. Further efforts should be made to control deforestation in this area. These findings can also be used as a reference in the formulation and implementation of sustainable woodland management policies.

Machine learning (ML) and data-driven approaches are increasingly used in many research areas. Extreme gradient boosting (XGBoost) is a tree boosting method that has evolved into a state-of-the-art approach for many ML challenges. However, it has rarely been used in simulations of land use change so far. Xilingol, a typical region for research on serious grassland degradation and its drivers, was selected as a case study to test whether XGBoost can provide alternative insights that conventional land-use models are unable to generate. A set of 20 drivers was analysed using XGBoost, involving four alternative sampling strategies, and SHAP (Shapley additive explanations) to interpret the results of the purely data-driven approach. The results indicated that, with three of the sampling strategies (over-balanced, balanced, and imbalanced), XGBoost achieved similar and robust simulation results. SHAP values were useful for analysing the complex relationship between the different drivers of grassland degradation. Four drivers accounted for 99 % of the grassland degradation dynamics in Xilingol. These four drivers were spatially allocated, and a risk map of further degradation was produced. The limitations of using XGBoost to predict future land-use change are discussed.

Land-use and land-cover changes have important effects on ecology, human systems, the environment, and policy at both global and regional scales. Thus, they are closely related to human activities. The extraction of more details about land-use change and grassland degradation is necessary to achieve future sustainable development in Inner Mongolia. The current study presents the patterns and processes of land-use changes over space and time, while also analyzing grassland degradation that is based on an analysis of land-use changes using a transition matrix, the Markov chain model and Moran’s I index, and a combination of long-time-scale remote sensing data as the data source. The major results indicate the following. (1) In 1990–2015, 13% (123,445 km²) of the total study area, including eight land-use types, changed. Woodland increased the most and moderate grassland decreased the most. (2) Grassland degradation, which occupied 2.8% of the total area of Inner Mongolia, was the major land-use conversion process before 2000, while, after 2000, 8.7% of the total area was restored; however, grassland degradation may still be the major ecological issue in Inner Mongolia. (3) Environmental protection policies show a close relationship with land-use conversion.

Land-use and land-cover change (LUCC) are currently contested topics in the research of global environment change and sustainable change. Identifying the historic land-use change process is important for the new economic development belt (the Zhujiang–Xijiang Economic Belt, ZXEB). During this research, based on long-time-series land-use and land-cover data, while using a combination of a transition matrix method and Markov chain model, the authors derive the patterns, processes, and spatial autocorrelations of land-use and land-cover changes in the ZXEB for the periods 1990–2000 and 2000–2017. Additionally, the authors discuss the spatial autocorrelation of land-use in the ZXEB and the major drivers of urbanization. The results indicate the following: (1) The area of cropland reduces during the two periods, and woodland decreases after the year 2000. The woodland is the most stable land-use type in both periods. (2) Built-up land expansion is the most important land-use conversion process; the major drivers of built-up land expansion are policy intervention, GDP (gross domestic product), population growth, and rural population migration. (3) Transition possibilities indicate that after 2000, most land-use activities become stronger, the global and local Moran’s I of all land-use types show that the spatial autocorrelations have become more closely related, and the spatial autocorrelation of built-up land has become stronger. Policies focus on migration from rural to urban, and peri-urban development is crucial for future sustainable urbanization.