Enhancing carbon emission reduction strategies using OCO and ICOS data

Area of Science:

• Environmental Science
• Atmospheric Science
• Data Science

Background:

• Accurate local carbon dioxide (CO2) monitoring is crucial for effective climate change mitigation.
• Traditional methods often lack the spatial resolution needed to capture localized emission patterns.
• Integrating diverse data sources offers a promising avenue for improving CO2 monitoring granularity.

Purpose of the Study:

• To develop a novel methodology for high-resolution local CO2 estimation.
• To integrate satellite data (OCO-2, OCO-3), ground observations (ICOS), and weather data (ERA5).
• To validate the model's performance in capturing local emission patterns.

Main Methods:

• Utilized multimodal data fusion techniques.
• Employed weighted K-nearest neighbor (KNN) interpolation combined with machine learning models.
• Leveraged satellite measurements, ground-level observations, and meteorological reanalysis data.

Main Results:

• Achieved a Root Mean Squared Error (RMSE) of 3.58 ppm for ground-level CO2 predictions.
• Demonstrated the effectiveness of data integration in capturing local emission dynamics.
• Validated the model's capability for high-resolution CO2 estimations.

Conclusions:

• The proposed methodology significantly enhances the granularity of CO2 monitoring.
• This approach provides precise insights valuable for targeted carbon mitigation efforts.
• The model represents a novel and adaptable application of machine learning and KNN in environmental monitoring.