Ensemble machine learning methods have been widely used for modeling landslide susceptibility, but there has been no uniform ensemble method for this problem. The main objective of this study is to compare popular ensemble machine learning-based models and apply them to landslides susceptibility mapping. The selected models include the random forest (RF), which is a typical bagging ensemble model, and three advanced boosting models, namely, adaptive boosting (AB), gradient boosting decision trees (GBDT), and extreme gradient boosting (XGBoost). This study considers 94 landslide points and 12 affecting factors. The data are divided into a training dataset consisting of 70% of the overall data, and a validation dataset, containing the remaining 30% of the data. The models are evaluated using the area under the receiver operating characteristic curve (AUC) and three common performance metrics: sensitivity, specificity, and accuracy. The results indicate that the four ensemble models have an AUC of more than 0.8, suggesting that they can appropriately and accurately predict landslide susceptibility maps. In particular, the XGBoost model achieves the best performance among all models, having a sensitivity of 92.86, specificity of 90.00, and accuracy of 91.38. Furthermore, the bagging model has a sensitivity of 89.29, specificity of 86.67, and accuracy of 87.93, and it is superior to the GBDT, which achieves a sensitivity of 86.21, specificity of 86.21, and accuracy of 86.21, and the AB, reaching a sensitivity of 82.14, specificity of 80.00, and accuracy of 81.03. The results presented in this study indicate that the advanced ensemble model, the XGBoost model, could be a promising tool for the selection of ensemble models for predicting landslide susceptibility mapping.

In recent years, most of domestic and foreign researches about heavy metal pollutions of metal mine mainly focus on water, soil and plants on the surface. There is lack of researches about heavy metal pollution in groundwater of metal mine. In this research, a certain antimony mine area is selected as a typical study area. Also, the study about statistical characteristics of heavy metals in groundwater has been carried out. Furthermore, the interrelationships have been preliminarily discussed through related analysis, such as relevant analysis, cluster analysis and principle component analysis. The results show that: the excessive elements in groundwater of study area are Sb, As, Pb, Se, and Ni. The average mass concentration of Sb, As, and Pb is higher than that of drinking water standards (GB5749-2006). The concentration of most heavy metals in dry season is lower than or equal to that in wet season for groundwater. Zn is the only metal in groundwater showing a different pattern, the concentration of which in dry season is higher than that in wet season. Under the impacts of stratum leaching and absorption effect, the concentration of heavy metals (except Pb and Ba) in groundwater are lower than or equal to that in surface water. As and Se, the two heavy metals have a significant positive correlation, which shows the two elements might have gone through similar environmental geochemical effect. Also, the connection among Zn, Hg, Pb, and Mn is not obvious; therefore, the sources of those elements are quite different. In addition, the elements of Se and As have obvious positive interrelationship with elements of CO₃^2- and F^-. Also, the Pb has significant positive correlation with PO₄^3-, H₂SiO₃ and oxygen consumption. The results of cluster analysis show that 9 different heavy metals in the study area can be divided into 3 categories: Zn, Cd, Mn, Hg, Cu, and Cr belong to the first category, Se and As belong to the second one, and the last category is Pb. Also, the principle component analysis divides 6 heavy metals (Zn, As, Hg, Pb, Mn, and Se) into 4 different principle components, which can be utilized to assess heavy metals pollution situations in groundwater. The reliability of this method is higher than 91%. Moreover, the research provides theory basis and models for establishing evaluation index system and exploring the evaluation method of heavy mental pollution in groundwater.

The dry area in northwest China (ARNC), with its tough climate, serious soil erosion, and poor soil quality, is one of China’s most fragile ecosystems. Studying changes in plant growth cycles here is very important for improving environmental monitoring and making plans to adapt to climate change. While vegetation growing season parameters (Start/End of Season: SOS/EOS) serve as vital indicators of ecosystem dynamics, comprehensive understanding has been constrained by limited long-term phenological datasets and insufficient exploration of multi-factor interactions. This study used PLS-SEM to analyze 27-year (1990–2016) vegetation index data, systematically quantifying spatiotemporal variations in growing season phenology and disentangling climate–terrain driving mechanisms. The results revealed the following key findings. (1) Spatial heterogeneity in phenological patterns, with the annual average Start of Season (SOS) and End of Season (EOS) being 114.7 Day and 301.7 Day, respectively, exhibiting a northwest–high to southeast–low gradient. The findings indicate a prolongation of the vegetation growing season, with significant spatial variability. (2) Interannual fluctuations showed the SOS and EOS coefficient of variation (CV) values of 0.230 and 0.234, respectively, with southeastern regions displaying higher instability than northwestern counterparts. (3) The spatial variation in SOS/EOS is primarily influenced by meteorological and geographical factors, with an explanatory power exceeding 30%. This research advances mechanistic understandings of arid ecosystem responses to environmental stressors, providing a scientific foundation for targeted ecological restoration, desertification mitigation, and sustainable land management in climate-sensitive drylands.