The Qilian Mountains (QLM) are an essential ecological security barrier in northwest China. Identifying the driven pattern of vegetation change is crucial for ecological protection and restoration in the QLM. Based on high-resolution vegetation coverage (VC) data in the QLM from 1990 to 2018, linear trend analysis was employed to examine the spatiotemporal dynamics of VC in the QLM, while correlation analysis was utilized to establish relationships between VC change and environmental factors. Multiple correlation analysis and residual analysis were adopted to recognize the climatically and anthropogenically driven pattern of VC change. The results showed that VC in the QLM presented a remarkable upward trend in volatility from 1990 to 2018. The significant increase areas accounted for 59.32% of the total, mainly distributed in the central and western QLM, and the significant decrease areas accounted for 9.18%, mostly located in the middle and eastern QLM. VC change showed a significant positive correlation with precipitation change and annual average temperature, while it exhibited a significant negative correlation with annual average precipitation, current VC status, livestock density, and slope. Climate change played a leading role in the increase of VC, and the impact of precipitation was significantly higher than that of temperature. Affected by climate change, the VC of alpine steppes and temperate steppes increased the most. Under the human interference, VC decreased significantly in 9.2% of the region, of which shrubs fell the most, followed by alpine meadows and forests. This study can provide certain guidance for local ecological protection and restoration efforts.

Urban spatial patterns regulate microclimates, yet those nonlinear interactive actions of 2D landscape patterns and 3D building morphology on land surface temperature (LST) among diverse urban functional zones (UFZs) remain insufficiently quantified. Using Shenyang as the case study, this study coupled Bayesian-optimized LightGBM and GeoShapley to disentangle these driving mechanisms. The models demonstrated high predictive accuracy (R² > 0.84 across all UFZs). Results show high-temperature zone proportions ranking as: commercial (17.93%) > public service (11.40%) > residential (8.70%) > industrial (3.97%). Further factor analysis reveals that while geographic location is the predominant overall driver, after removing this baseline effect, residential, public service, and commercial zones are mainly regulated by MNDWI and NDVI, while industrial zones are dominated by Spatial Congestion Degree (SCD) and Floor Area Ratio (FAR). Morphological metrics exhibit clear ecological thresholds: residential zones show turning points at FAR 1.19 and NDVI 0.13; commercial zones see SCD’s cooling effect reverse to warming beyond 0.27; industrial zones experience sharp deterioration when SCD and FAR exceed 0.09 and 0.51, respectively. These findings demonstrate the necessity of UFZ-based differentiated spatial regulation and provide actionable thresholds for climate-adaptive urban planning.

Land-use change substantially contributes to carbon emissions, yet systematic research on complex human–environment interactions in old industrial bases remains scarce. Here, we integrated multi-temporal land-use data and socio-economic statistics from 1990 to 2023 to analyze the spatiotemporal dynamics and drivers of land-use carbon emissions in Northeast China. The land-use transfer matrix, the carbon emission coefficient, exploratory spatiotemporal data analysis (ESTDA), standard deviational ellipses, and modified Kaya–LMDI models were applied. Construction land area expanded by 120%, with its share of total emissions increasing from 87% to 95%. Meanwhile, forest and grassland shrank, reducing their carbon sink capacity and increasing their net carbon emissions 1.9-fold. Spatially, emissions showed a weak global correlation but strong local lock-in (i.e., persistent stability of local spatial patterns over time), with the emission center of gravity shifting southwestward. Economic development was the dominant positive driver (provincial contribution rates: 275–529%), whereas energy intensity was the main mitigating factor (up to −409%). Population loss exerted a slight negative contribution, while energy structure showed only a weak inhibitory effect (−9.1%), reflecting the region’s path-dependent lock-in to fossil fuels. This study provides a scientific basis for differentiated carbon management strategies in Northeast China and analogous regions.