Improving PM_2.5 predictions during COVID-19 lockdown by assimilating multi-source observations and adjusting...

Area of Science:

• Atmospheric Chemistry and Physics
• Environmental Science
• Data Assimilation Techniques

Background:

• The COVID-19 pandemic led to reduced anthropogenic emissions, impacting air quality.
• Existing emission inventories for air quality models lack real-time updates, causing uncertainty in PM2.5 predictions.
• Accurate PM2.5 forecasting is crucial for public health and environmental monitoring.

Purpose of the Study:

• To improve the accuracy of PM2.5 predictions using the WRF-Chem model during the COVID-19 pandemic.
• To investigate the effectiveness of data assimilation (DA) and joint emission adjustment for enhancing air quality modeling.
• To assess the impact of real-time emission adjustments on forecast error accumulation.

Main Methods:

• Implemented a three-dimensional variational (3D-Var) data assimilation approach.
• Assimilated multi-source PM2.5 data from satellite and ground observations.
• Jointly adjusted emissions within the WRF-Chem model framework.
• Conducted experiments over Hubei Province, China, from January 21st to March 20th, 2020.

Main Results:

• PM2.5 prediction accuracy significantly improved across validation sites.
• Data assimilation benefits persisted for up to 48 hours, though diminishing with forecast length.
• Joint emission adjustment substantially slowed down error accumulation in forecasts.
• A notable 19.49% RMSE improvement was achieved at 48-hour forecasts through emission adjustments.

Conclusions:

• Data assimilation combined with emission adjustment enhances PM2.5 forecasting accuracy.
• The method provides more reliable air quality predictions, especially when emission patterns change rapidly.
• This approach offers a robust strategy for improving air quality models under dynamic emission scenarios.