Keeping an up-to-date and accurate global database of the volume and source of atmospheric pollutants is key in achieving air quality and climate goals set out in the Paris Agreement (and hopefully any goals set at COP26).
One way to help achieve this is through satellite observations. Dr Douglas Finch and his team at the University of Edinburgh have developed and trained a machine learning model to identify plumes of nitrogen dioxide, a tracer for combustion, from observations gathered by the TROPOMI instrument on board the Sentinel 5-P satellite. This approach allows them to efficiently exploit the growing volume of satellite data available to characterise the Earth's climate.
Dr Finch trained his model using six thousand 28×28-pixel images of TROPOMI data and found that it can be used to identify nitrogen dioxide plumes with a success rate of more than 90%. Using this model, they have found more than 310,000 individual plumes that can be attributed to large urban centres, oil and gas production, and major power plants as well as biomass burning.
During his talk at the School of Earth and Environmental Sciences, he will cover the development of the model, results and possible further uses for this technique.
More information on this event