Accurate assessment of the cooling effect from urban green space (UGS) is largely hindered by insufficient field samples or consideration of the internal and surrounding three-dimensional (3D) structure. This study developed a transferable modeling-optimization framework that integrated a multi-temporal sampling strategy, multimodal 3D environmental reconstruction, and Bayesian-based optimization. First, the potential influencing factors of the cooling effect were quantified from three aspects of inner 2D/3D structure, surrounding building ventilation, and background meteorology through fusing field measurements, multi-spectral UAV images, and Sentinel-2 images. Then, a generalized additive mixed-effects model was used to explore cooling-related patterns of UGS, and a Bayesian network was further applied to identify potential optimized configurations. The results suggest the following: (1) The adopted multi-temporal sampling strategy enhances the stability of detected cooling signals and minimizes spatial interference among neighboring UGS patches and water bodies. (2) Temporal changes in the cooling effect are mainly driven by average air temperature and maximum wind speed, while the spatial variation by the UGS inner characteristics of area and shape index and surrounding ventilation. (3) The “win–win” situation of cooling intensity and range occurred in UGSs with larger areas, higher shape regularity, and medium ventilation. This approach is useful for model-based planning of climate-responsive green infrastructure and city-scale ventilation systems in heat-vulnerable environments.

It is very important for the protection and utilization of water resources in the Yellow River Basin to explore the influencing factors and improvement paths of green water use efficiency based on different regions. The dynamic evolution characteristics, regional differences and internal inefficiencies of green water use efficiency for 48 cities in the Yellow River Basin from 2008 to 2018 are analyzed, based on their ecological geographical divisions, with the use of a DEA-SBM model, GML index decomposition and kernel density estimation. We further use the panel Tobit model to analyze the external influencing factors of green water use efficiency and propose ways to improve the utilization of water resources in different regions from both internal and external perspectives. The results are as follows: (1) During the study period, the green water use efficiency fluctuated between 0.58–0.67 and showed a trend of improving in arid areas and falling in ones. (2) Exploring the sources of inefficiency from the inside reveals that labor, capital, and wastewater redundancy in the semihumid area is higher, the energy redundancy in the semiarid area is higher, and the economic output in the arid area is insufficient. (3) From the GML point of view, the absolute difference in the green water efficiency of the cities in the Yellow River Basin is expanding. In terms of EC(technical efficiency index), the technical efficiency of the semiarid area has a convergence effect. In terms of TC(technical progress index), the gap in the arid area has been widening, and technology in the semihumid and semiarid areas is converging backward. (4) There are significant differences in the external factors affecting green water use efficiency in different ecogeographical regions. This study helps the government to consider eco-geographical factors when formulating water resources-related policies, and provides a scientific reference for how to better utilize water resources in different regions of the Yellow River Basin from both internal and external aspects.

Mountain disturbances, such as insect infestation, wildfire, timber harvesting and building construction, often cumulatively and interactively affect ecosystem services (ESs). There is a lack of studies exploring the impacts of natural and human disturbances on ESs and the trade-offs/synergies among ESs, particularly in mountain ecosystem with diverse landscapes. Using the Qilian Nature Reserve as a case study area, this study quantified mountain disturbances through Landsat time-series data and BEAST algorithm, evaluated three critical ESs of carbon sequestration, water yield and habitat quality through CASA model and InVEST model, using Bayesian Networks (BN) model to simulate ES, and quantified the impacts of disturbances on ESs and their trade-offs under different environment conditions through multi-scenario analysis. The results suggest that BEAST could effectively capture the patterns and dynamics of small-scale disturbances, which were difficult to be reflected by normal land use/cover products. The established BN model could simulate the spatio-temporal dynamics of carbon sequestration, water yield and habitat quality, with average classification error of 17.81%, 12.72% and 4.49% for each ES, respectively. Significant synergy exists between carbon sequestration and habitat quality at regional scale, while trade-off existed between water yield and the other two ESs. Specifically, these trade-offs/synergies among ESs tend to be weak at medium state of ESs, but stronger at higher and lower states. Thus, significant differences existed in the “win-lose” solutions between water yield and the other two ESs, further resulting the limited space to improve three ESs simultaneously. Disturbances at medium frequency and low-medium intensity are beneficial for the maintenance and improvement of three ESs. BN model has been demonstrated to be a promising decision support tool to integrate small-scale disturbances into ES evaluation and identify the most suitable management solutions for mountain ecosystems, which could provide critical information for decision-makers and guiding sustainable development.