Search

Abstract Natural world heritage sites face increasing risk under rapid climate change, especially considering the impacts of climate extremes. However, there is not yet enough understanding of the future extreme climate challenges at global natural world heritage sites. Here we identify the exposure of 250 natural heritage sites to extreme climate events under 4 different future scenarios. We found that by 2100 under the highest emission scenario, 248 out of 250 sites were exposed to extreme climate events. Forest natural world heritage sites may face increasing pressure to complex extreme phenomena under emission rise. In tropical regions, where high temperatures may magnify the vulnerability of biodiversity, we identify 14 natural world heritage sites to be prioritised that are poor in biodiversity and expected to face high temperatures. Hence, there is an urgent need for enhanced climate change adaptation at heritage sites to minimise loss of irreplaceable values.

Abstract Following a series of actions in China, the annual mean fine particle (PM2.5) concentrations have considerably reduced since 2013. However, in some cities, the PM2.5 concentrations exceed the national ambient air quality standard of 35 μg/m3. Herein, we evaluated the historical trends of PM2.5 from 2014 to 2022 in mainland China and found continuously declining trends. However, the annual decline rates decreased despite increasing investments on pollution control, showing considerably diminishing marginal effects of these investments. Accumulative reduction in PM2.5 concentrations is linearly dependent on the log‐transformed accumulative investment in air pollution control. The standard of 35 μg/m3 can be achieved by 2035 if the current efforts continue at the same pace. However, the standard of 35 μg/m3 can weaken the momentum of future actions in cities where the standard has already been implemented. We proposed and tested two tiers with four or six levels of specific regional standards based on the regression model predictions from the historical data and demonstrated that more rigorous standards for regions where the current standard is already attained could provide more momentum for strengthening abatement efforts. The results show that both tiers are feasible if the level of abatement endeavor is maintained even with marginal effects. We also demonstrate that additional 135,000 or 384,000 deaths can be avoided by applying the regional standards. Health benefits are monetized and compared with the control costs, suggesting that compared to Tier‐2, Tier‐1 is more cost‐effective and balanced across the mainland China.

The reduction in methane concentration is crucial for achieving the goals of the Paris agreement. However, its annual growth rate is unstable, and understanding the reasons for changes in methane growth is essential for climate policy-making. Currently, there is considerable uncertainty regarding its attribution. Here, we utilize multi-source data and optimal fingerprinting methods to detect the contributions of several key drivers to the methane trend and interannual variability. We find that the methane growth trend is primarily influenced by anthropogenic emissions, while interannual variability is predominantly determined by wetland and biomass burning emissions. This result underscores the central role of anthropogenic emissions in methane dynamics, providing confidence in the effectiveness of human efforts to control methane atmospheric concentrations through emission reductions. It also helps alleviate concerns about the recent surge in atmospheric methane concentration, as it may be a short-term peak caused by increased wetland emissions rather than a long-term change.

Tool wear state recognition is a challenging problem because the various types of signals corresponding to different wear states have similar features, making classification difficult. Convolutional neural networks are widely used in the field of tool wear state recognition. To address this problem, this study proposes a milling cutter wear state monitoring model that combines KANS and deep residual networks (ResNet). The traditional ResNet-34’s top linear classifier was replaced with a nonlinear convolutional classifier including _top_kan, and the data were preprocessed using continuous wavelet transform (CWT) to enhance the model’s immunity to interference and feature characterization. The experimental results based on the PHM dataset show that the improved KANS-ResNet-34 model improves accuracy by 1.07% compared to ResNet-34, making it comparable to ResNet-50, while its computation time is only 1/33.68 of the latter. This significantly improves computational efficiency, reduces the pressure on hardware resources, and provides an effective tool wear state recognition solution.