A Temporal and Spatial Analysis and Prediction of the Impact of the Fukushima Nuclear Accident on Nuclear Radiation in Zhejiang

The problem of nuclear radiation caused by nuclear accidents poses a serious challenge to human society and the natural environment, and after the Fukushima nuclear accident in 2011, global monitoring and analysis of nuclear radiation levels have become more important, especially to assess the potential impact of remote nuclear events on other regions. In this study, Zhejiang Province, China, was selected as the study object, focusing on analyzing the spatial and temporal distribution characteristics of the airborne absorbed dose rate in the region and assessing the possible impacts of external nuclear events (e.g., the Fukushima nuclear accident) on the region. To this end, the present study systematically collected air-absorbed dose rate data from different areas of Zhejiang Province, and combined with time-series analysis methods, explored whether the Fukushima nuclear accident had a significant impact on nuclear radiation levels in Zhejiang Province. The results of the analysis show that there is a statistically significant correlation between the air-absorbed dose rate in some areas of Zhejiang Province and the Fukushima nuclear accident after the Fukushima nuclear accident. In addition, two time series prediction models, Transformer and Long Short-Term Memory (LSTM) network, were used in this study to predict and analyze the air absorbed dose rate in Zhejiang Province. The experimental results show that the Transformer model is excellent in handling complex time series data and capturing nonlinear relationships, and its prediction accuracy is significantly better than that of the LSTM model. In addition, through the analysis of SHAP values, this study reveals that wind direction is the key internal meteorological factor affecting the air absorbed dose rate, which provides important support for the interpretability of the model.