As an important production base for livestock and a unique ecological zone in China, the northeast Tibetan Plateau has experienced dramatic land use/land cover (LULC) changes with increasing human activities and continuous climate change. However, extensive cloud cover limits the ability of optical remote sensing satellites to monitor accurately LULC changes in this area. To overcome this problem in LULC mapping in the Ganan Prefecture, 2000–2018, we used the dense time stacking of multi-temporal Landsat images and random forest algorithm based on the Google Earth Engine (GEE) platform. The dynamic trends of LULC changes were analyzed, and geographical detectors quantitatively evaluated the key driving factors of these changes. The results showed that (1) the overall classification accuracy varied between 89.14% and 91.41%, and the kappa values were greater than 86.55%, indicating that the classification results were reliably accurate. (2) The major LULC types in the study area were grassland and forest, and their area accounted for 50% and 25%, respectively. During the study period, the grassland area decreased, while the area of forest land and construction land increased to varying degrees. The land-use intensity presents multi-level intensity, and it was higher in the northeast than that in the southwest. (3) Elevation and population density were the major driving factors of LULC changes, and economic development has also significantly affected LULC. These findings revealed the main factors driving LULC changes in Gannan Prefecture and provided a reference for assisting in the development of sustainable land management and ecological protection policy decisions.

Grazing is one of the main human disturbance factors in alpine grassland on the Qinghai-Tibet Plateau (QTP), which can directly or indirectly influence the community structures and ecological functions of grassland ecosystems. However, despite extensive field grazing experiments, there is currently no consensus on how different grazing management approaches affect alpine grassland diversity, soil carbon (C), and nitrogen (N). Here, we conducted a meta-analysis of 70 peer-reviewed publications to evaluate the general response of 11 variables related to alpine grassland ecosystems plant diversity and ecological functions to grazing. Overall, the results showed that grazing significantly increased the species richness, Shannon–Wiener index, and Pielou evenness index values by 9.89% (95% CI: 2.75–17.09%), 7.28% (95% CI: 1.68–13.62%), and 3.74% (95% CI: 1.40–6.52%), respectively. Aboveground biomass (AGB) and belowground biomass (BGB) decreased, respectively, by 41.91% (95% CI: −50.91 to −32.88%) and 17.68% (95% CI: −26.94 to −8.52%). Soil organic carbon (SOC), soil total nitrogen (TN), soil C:N ratio, and soil moisture decreased by 13.06% (95% CI: −15.88 to −10.15%), 12.62% (95% CI: −13.35 to −8.61%), 3.27% (95% CI: −4.25 to −2.09%), and 20.75% (95% CI: −27.89 to −13.61%), respectively, whereas, soil bulk density and soil pH increased by 17.46% (95% CI: 11.88–24.53%) and 2.24% (95% CI: 1.01–3.64%), respectively. Specifically, moderate grazing, long-durations (>5 years), and winter grazing contributed to increases in the species richness, Shannon–Wiener index, and Pielou evenness index. However, AGB, BGB, SOC, TN, and soil C:N ratios showed a decrease with enhanced grazing intensity. The response ratio of SOC was positively associated with AGB and BGB but was negatively related to the Shannon–Wiener index and Pielou evenness index. Furthermore, the effects of grazing on plant diversity, AGB, BGB, SOC, and TN in alpine grassland varied with grazing duration, grazing season, livestock type, and grassland type. The findings suggest that grazing should synthesize other appropriate grazing patterns, such as seasonal and rotation grazing, and, furthermore, additional research on grazing management of alpine grassland on the QTP is needed in the future.

Above-ground biomass (AGB) is a key indicator for studying grassland productivity and evaluating carbon sequestration capacity; it is also a key area of interest in hyperspectral ecological remote sensing. In this study, we use data from a typical alpine meadow in the Qinghai–Tibet Plateau during the main growing season (July–September), compare the results of various feature selection algorithms to extract an optimal subset of spectral variables, and use machine learning methods and data mining techniques to build an AGB prediction model and realize the optimal inversion of above-ground grassland biomass. The results show that the Lasso and RFE_SVM band filtering machine learning models can effectively select the global optimal feature and improve the prediction effect of the model. The analysis also compares the support vector machine (SVM), least squares regression boosting (LSB), and Gaussian process regression (GPR) AGB inversion models; our findings show that the results of the three models are similar, with the GPR machine learning model achieving the best outcomes. In addition, through the analysis of different data combinations, it is found that the accuracy of AGB inversion can be significantly improved by combining the spectral characteristics with the growing season. Finally, by constructing a machine learning interpretable model to analyze the specific role of features, it was found that the same band plays different roles in different records, and the related results can provide a scientific basis for the research of grassland resource monitoring and estimation.

Due to climate change and human activities, the eco-environment quality (EEQ) of eco-fragile regions has undergone massive change, especially in the Tibet Plateau. The Qilian Mountains (QLM) region is an essential ecological function zone in the northeastern Tibet Plateau, which plays a vital role in northwestern China’s eco-environmental balance. However, EEQ changes in the QLM during the 21st century remain poorly understood. In this study, the spatiotemporal variations of the EEQ in the QLM were analyzed from 2000 to 2020 using a remote sensing ecological index (RSEI). The EEQ driving factors are identified by the geographic detector, and the spatial influence of critical factors is represented by a geographically weighted regression model. The results show low EEQ in the QLM. From 2000 to 2020, the EEQ initially slightly improved, then deteriorated, and finally gradually recovered. Spatially, the EEQ shows an increasing trend from northwest to southeast. Moran’s I of EEQ remains at around 0.95, representing high spatial aggregation. “High–High” and “Low–Low” clustering features dominate in the local spatial autocorrelation, indicating the EEQ of the QLM is polarized. Precipitation is the dominant positive factor in the EEQ, with a q statistics exceeding 0.644. Furthermore, the key factors (precipitation, distance to towns, distance to roads) affecting EEQ in different periods vary significantly in space. From results we can draw the conclusion that the natural factors mainly control the spatial patterns of EEQ, while the human factors mainly impact the temporal trend of EEQ, the EEQ in the QLM has been significantly improved since 2015. Our findings can provide theoretical support for future eco-environmental protection and restoration in the QLM.

Rubber plantation is an important strategic material related to the national economy and people's livelihoods. Up-to-date and accurate rubber plantation maps are critical for monitoring the area and spatial distribution of rubber plantations and assessing their impacts on society, the economy, and the environment. However, existing optical images are greatly limited by frequent cloud cover, which seriously affects the accuracy of rubber plantation area extraction. To overcome this issue, we used dense Landsat time series stacks based on Google Earth Engine, combined phenological features, and applied random forest algorithms to monitor rubber plantations in Xishuangbanna from 1987 to 2020. The results showed that 1 ) the phenological characteristics of rubber plantations in Xishuangbanna indicated that the leaf-off period lasts from late December to mid-February of the following year, while the leaf-on of rubber plantations occurred in other months; 2 ) the overall accuracy and kappa coefficient values ranged from 0.82 to 0.96 and 0.76 to 0.95, respectively, showing that the extraction accuracy of rubber plantation information can meet the accuracy requirement; 3 ) the rubber plantation area in Xishuangbanna increased between 1987 and 2020 from 7.05×10 ⁴ to 47.78×10 ⁴ hm ² . The peak rubber plantation area occurred in 2015 (49.60×10 ⁴ hm ² ) followed by a downward trend; 4 ) spatially, the rubber plantation is mainly distributed in Jinghong City and Mengla County, while less abundant in Menghai County. Overall, this article is expected to contribute to the rapid and accurate mapping of rubber plantations in large-scale applications and analysis.

The timely and accurate estimation of grassland aboveground biomass (AGB) is important. Machine learning (ML) has been widely used in the past few decades to deal with complex relationships. In this study, based on an 11-year period (2005–2015) of AGB data (1620 valid AGB measurements) on the Three-River Headwaters Region (TRHR), combined with remote sensing data, weather data, terrain data, and soil data, we compared the predictive performance of a linear statistical method, machine learning (ML) methods, and evaluated their temporal and spatial scalability. The results show that machine learning can predict grassland biomass well, and the existence of an independent validation set can help us better understand the prediction performance of the model. Our findings show the following: (1) The random forest (RF) based on variables obtained through stepwise regression analysis (SRA) was the best model (R²_vad = 0.60, RMSE_vad = 1245.85 kg DW (dry matter weight)/ha, AIC = 5583.51, and BIC = 5631.10). It also had the best predictive capability of years with unknown areas (R²_indep = 0.50, RMSE_indep = 1332.59 kg DW/ha). (2) Variable screening improved the accuracy of all of the models. (3) All models’ predictive accuracy varied between 0.45 and 0.60, and the RMSE values were lower than 1457.26 kg DW/ha, indicating that the results were reliably accurate.