Accurate runoff simulation and prediction are crucial for water resources management, especially under the impact of climate change. In this study, a multi-physics ensemble Bayesian model averaging (MPE-BMA) model is developed to improve runoff prediction accuracy by integrating a soil and water assessment tool (SWAT), hydrologiska byråns vattenbalansavdelning (HBV) model, and Bayesian model averaging (BMA) into a general framework. The MPE-BMA model integrates the strengths of the SWAT and HBV models. This approach enhances the robustness of simulation outputs and reduces uncertainties from single-model methods. MPE-BMA is subsequently employed to simulate and predict runoff for the upper reaches of Xiangxi River Basin (XXRB) in China, where four general circulation models (GCMs) and three shared socioeconomic pathways (SSP126, SSP245, and SSP585) are considered. Multiple statistical metrics (R², NSE, and RMSE) prove that the MPE-BMA model outperforms the single models of SWAT and HBV. Results reveal that higher-emission scenarios generally lead to significant decreases in runoff, particularly by the 2080s. Specifically, under SSP585, runoff is projected to decrease by approximately 4.61–12.68% by the 2040s and 5.96–11.28% by the 2080s compared to the historical period. From the perspective of monthly and seasonal runoff changes, the peak runoff is projected to shift from June to May by the 2080s. Additionally, under SSP585, spring and summer runoffs tend to significantly increase, while winter runoff decreases sharply, leading to wetter summers and drier winters. These findings underscore the importance of enhancing water use efficiency, upgrading hydropower stations, and implementing watershed management practices to ensure sustainable water resources management in the XXRB amidst climate change.

In this study, a fuzzy bi-level chance constraint programming (FBCP) model is developed for urban ecological management in Xiamen, China. FBCP has advantages in balancing trade-offs between multiple decision makers and can address fuzzy and stochastic uncertainty in ecosystem management. It also can reflect the impact of different violation risk levels and emission reduction measures on system benefit, ecosystem service value, and land resource allocation. Then, the conversion of land use and its effects at small regional extent (CLUE-S) model is employed to provide the spatial allocation of future land resources under different scenarios. Results reveal that (i) carbon fixation and climate regulation are the major contributors to the ecosystem service value, with a proportion of [15.4, 15.6]% and [18.5, 18.8]%, respectively; (ii) the main environmental problem in Xiamen is the water pollution caused by the excessive discharge of commercial and residential land, with COD and NH₃-N account for [68.81, 69.33]% and [67.65, 68.20]% of the total discharge of the city, respectively; (iii) the violation risk p level is the most impact factor, and the schemes with high system benefit would face greater default risk and lower ecological quality; (iv) FBCP model considers the trade-off between economic benefit and ecological quality, while the fuzzy chance constraint programming (FCP) model achieves a high system benefit at the expense of the environment. These findings help decision makers to understand the impact of parameter uncertainty and pollutant discharge policies on system benefit, and adjust land-use patterns to weigh ecological environment protection with urban economic development.

The rise in global temperatures and increased extreme weather events, such as heatwaves, underscore the need for accurate regional projections of daily maximum temperature (T_max) to inform effective adaptation strategies. This study develops the CNN-BMA-QDM model, which integrates convolutional neural networks (CNNs), Bayesian model averaging (BMA), and quantile delta mapping (QDM) to downscale and project T_max under future climate scenarios. The CNN-BMA-QDM model stands out for its ability to capture nonlinear relationships between T_max and atmospheric circulation factors, reduce model uncertainty, and correct bias, thus improving simulation accuracy. The CNN-BMA-QDM model is applied to Fujian Province, China, using three CMIP6 GCMs and four shared socioeconomic pathways (SSPs) to project T_max from 2015 to 2100. The results show that CNN-BMA-QDM outperforms CNN-BMA, CNNs, and other downscaling methods (e.g., RF, BPNN, SVM, LS-SVM, and SDSM), particularly in simulating extreme value at the 99% and 95% percentiles. Projections of T_max indicate consistent warming trends across all SSP scenarios, with spatially averaged warming rates of 0.0077 °C/year for SSP126, 0.0269 °C/year for SSP245, 0.0412 °C/year for SSP370, and 0.0526 °C/year for SSP585. Coastal areas experience the most significant warming, with an increase of 4.62–5.73 °C under SSP585 by 2071–2100, while inland regions show a smaller rise of 3.64–3.67 °C. Monthly projections indicate that December sees the largest increase (5.30 °C under SSP585 by 2071–2100), while July experiences the smallest (2.40 °C). On a seasonal scale, winter experiences the highest warming, reaching 4.88 °C under SSP585, whereas summer shows a more modest rise of 3.10 °C. Notably, the greatest discrepancy in T_max rise between the south and north occurs during the summer. These findings emphasize the importance of developing tailored adaptation strategies based on spatial and seasonal variations. The results provide valuable insights for policymakers and contribute to the advancement of regional climate projection research.