Improvement in the Forecasting of Low Visibility over Guizhou, China, Based on a Multi-Variable Deep Learning Model

Dongpo He,Yuetong Wang,Yuanzhi Tang,Dexuan Kong,Jing Yang,Wenyu Zhou,Haishan Li,Fen Wang

DOI: https://doi.org/10.3390/atmos15070752

IF: 3.11

2024-06-25

Atmosphere

Abstract:High-quality visibility forecasting benefits traffic transportation safety, public services, and tourism. For a more accurate forecast of the visibility in the Guizhou region of China, we constructed several visibility forecasting models via progressive refinements in different compositions of input observational variables and the adoption of the Unet architecture to perform hourly visibility forecasts with lead times ranging from 0 to 72 h over Guizhou, China. Three Unet-based visibility forecasting models were constructed according to different inputs of meteorological variables. The model training via multiple observational variables and visibility forecasts of a high-spatiotemporal-resolution numerical weather prediction model (China Meteorological Administration, Guangdong, CMA-GD) produced a higher threat score (TS), which led to substantial improvements for different thresholds of visibility compared to CMA-GD. However, the Unet-based models had a larger bias score (BS) than the CMA-GD model. By introducing the U2net architecture, there was a further improvement in the TS of the model by approximately a factor of two compared to the Unet model, along with a significant reduction in the BS, which enhanced the stability of the model forecast. In particular, the U2net-based model performed the best in terms of the TS below the visibility threshold of 200 m, with a more than eightfold increase over the CMA-GD model. Furthermore, the U2net-based model had some improvements in the TS, BS, and RMSE (root-mean-square error) compared to the LSTM_Attention model. The spatial distribution of the TS showed that the U2net-based model performed better at the model grid scale of 3 km than at the scale of individual weather stations. In summary, the visibility forecasting model based on the U2net algorithm, multiple observational variables, and visibility data from the CMA-GD model performed the best. The compositions of input observational variables were the key factor in improving the deep learning model's forecasting capability, and these improvements could improve the value of forecasts and support the socioeconomic needs of sectors reliant on visibility forecasting.

environmental sciences,meteorology & atmospheric sciences

What problem does this paper attempt to address?

This paper mainly focuses on improving the low visibility forecast in Guizhou, China. In the study, the authors constructed several prediction models based on deep learning, especially utilizing the Unet and U2net architectures for hourly visibility forecast with a lead time ranging from 0 to 72 hours. Through the optimization of different meteorological input variables, the model they established showed an improved Threat Score (TS) compared to the China Meteorological Administration Guangdong Model (CMA-GD) at multiple thresholds, especially with more than an eight-fold enhancement when visibility was less than 200 meters. However, the Bias Score (BS) of the Unet model was relatively high, while the introduction of the U2net architecture almost doubled the TS of the model, significantly reducing the BS and enhancing the stability of the model's predictions.

The study also found that multivariable observational data is crucial for improving the predictive capability of deep learning models, which can increase the value of forecasts and support economic and social sectors that rely on visibility predictions. Although previous studies have applied deep learning methods to visibility forecasting, these models (e.g., LSTM, MLP, and CNN) have limitations in handling spatiotemporal variations. In contrast, the U2net, with its deeper hierarchical structure and parameters, can improve feature representation capability and application accuracy.

Finally, the paper points out that Guizhou province has complex terrain and frequent low visibility weather, and the existing meteorological stations are insufficient to comprehensively reflect the visibility conditions in the entire province. Therefore, using the gridded observational data from the China Meteorological Administration and the CMA-GD model output, as well as the U2net-based model, can provide more detailed and accurate visibility forecasts.