Regional aerosol forecasts based on deep learning and numerical weather prediction
Regional aerosol forecasts based on deep learning and numerical weather prediction
复制标题
DOI：
10.1038/s41612-023-00397-0
复制
发表时间：
2023-06
影响因子：
9
通讯作者：
Yulu Qiu;Jining Feng;Ziyin Zhang;Xiujuan Zhao;Zi-ming Li;Zhiqiang Ma;Ruijin Liu;Jia Zhu
中科院分区：
文献类型：
--
作者：
Yulu Qiu;Jining Feng;Ziyin Zhang;Xiujuan Zhao;Zi-ming Li;Zhiqiang Ma;Ruijin Liu;Jia Zhu
近几十年来，大气化学传输模式在气溶胶预报中得到了广泛的应用，但由于排放率、气象数据的不确定性以及过于简化的化学参数化，这些模式面临着挑战。在这里，我们开发了一个时空深度学习框架，名为PPN（PM2.5污染预测网络），以准确有效地预测区域PM2.5浓度。它具有编码器-解码器架构，并结合了先前的PM2.5观测和数值天气预报。此外，该模型提出了加权损失函数，以提高预测性能的极端事件。应用该模型对京津冀地区3天PM2. 5浓度进行了3小时× 3小时的预报。总体而言，该模型表现出良好的性能，R2和RMSE值分别为0.7和17.7 μg m−3。它可以捕获南部高浓度的PM2. 5和北部相对低浓度的PM2. 5，并在未来24 h内表现出更好的性能。加权损失函数的使用降低了“高值低估，低值高估”的水平，而将先前的PM2. 5观测值纳入编码器相位提高了24 h内的预测精度。我们还比较了模式的结果，从一个国家的最先进的数值模式（WRF-Chem与污染物数据同化）。WRF-Chem的时间R2和RMSE分别为0.30−0.77和19−45 μg m− 3，而PPN模型的时间R2和RMSE分别为0.42−0.84和15−42 μg m−3。该模式具有较强的气溶胶预报能力，为区域污染事件的预警和管理提供了一种高效、准确的工具。
Atmospheric chemistry transport models have been extensively applied in aerosol forecasts over recent decades, whereas they are facing challenges from uncertainties in emission rates, meteorological data, and over-simplified chemical parameterizations. Here, we developed a spatial-temporal deep learning framework, named PPN (Pollution-Predicting Net for PM2.5), to accurately and efficiently predict regional PM2.5concentrations. It has an encoder-decoder architecture and combines the preceding PM2.5observations and numerical weather prediction. Besides, the model proposes a weighted loss function to promote the forecasting performance in extreme events. We applied the proposed model to forecast 3-day PM2.5concentrations over the Beijing-Tianjin-Hebei region in China on a three-hour-by-three-hour basis. Overall, the model showed good performance withR2and RMSE values of 0.7 and 17.7 μg m−3, respectively. It could capture the high PM2.5concentration in the south and relatively low concentration in the north and exhibit better performance within the next 24 h. The use of the weighted loss function decreased the level of “high values underestimation, low values overestimation”, while incorporating the preceding PM2.5observations into the encoder phase improved the predictive accuracy within 24 h. We also compared the model result with that from a state-of-the-art numerical model (WRF-Chem with pollutant data assimilation). The temporalR2and RMSE from the WRF-Chem were 0.30−0.77 and 19−45 μg m−3while those from the PPN model were 0.42−0.84 and 15−42 μg m−3. The proposed model shows powerful capacity in aerosol forecasts and provides an efficient and accurate tool for early warning and management of regional pollution events.