Identification of suitable habitats and priority conservation areas under climate change scenarios for the Chinese alligator (Alligator sinensis).

Amphibians and reptiles, especially the critically endangered Chinese alligators, are vulnerable to climate change. Historically, the decline in suitable habitats and fragmentation has restricted the distribution of Chinese alligators to a small area in southeast Anhui Province in China. However, the effects of climate change on range-restricted Chinese alligator habitats are largely unknown. We aimed to predict current and future (2050s and 2070s) Chinese alligator distribution and identify priority conservation areas under climate change. We employed species distribution models, barycenter migration analyses, and the Marxian model to assess current and future Chinese alligator distribution and identify priority conservation areas under climate change. The results showed that the lowest temperature and rainfall seasonality in the coldest month were the two most important factors affecting the distribution of Chinese alligators. Future predictions indicate a reduction (3.39%-98.41%) in suitable habitats and a westward shift in their distribution. Further, the study emphasizes that suitable habitats for Chinese alligators are threatened by climate change. Despite the impact of the Anhui Chinese Alligator National Nature Reserve, protection gaps persist, with 78.27% of the area lacking priority protected area. Our study provides crucial data for Chinese alligator adaptation to climate change and underscores the need for improved conservation strategies. Future research should refine conservation efforts, consider individual plasticity, and address identified limitations to enhance the resilience of Chinese alligator populations in the face of ongoing climate change.

Hierarchical attention network for multivariate time series long-term forecasting.

Multivariate time series long-term forecasting has always been the subject of research in various fields such as economics, finance, and traffic. In recent years, attention-based recurrent neural networks (RNNs) have received attention due to their ability of reducing error accumulation. However, the existing attention-based RNNs fail to eliminate the negative influence of irrelevant factors on prediction, and ignore the conflict between exogenous factors and target factor. To tackle these problems, we propose a novel Hierarchical Attention Network (HANet) for multivariate time series long-term forecasting. At first, HANet designs a factor-aware attention network (FAN) and uses it as the first component of the encoder. FAN weakens the negative impact of irrelevant exogenous factors on predictions by assigning small weights to them. Then HANet proposes a multi-modal fusion network (MFN) as the second component of the encoder. MFN employs a specially designed multi-modal fusion gate to adaptively select how much information about the expression of current time come from target and exogenous factors. Experiments on two real-world datasets reveal that HANet not only outperforms state-of-the-art methods, but also provides interpretability for prediction.