Generating high-resolution concentration maps for atmospheric gases using geography-informed machine learning

An information processing system for generating one or more high-resolution atmospheric gas concentration maps using geography-informed machine learning, the information processing system comprising:

• a processor;
• memory communicatively coupled to the processor;
• and
• an atmospheric gas mapping unit communicatively coupled to the processor and the memory, wherein the atmospheric gas mapping unit:
• implements at least one geography-informed machine learning (GIML) model trained on a first remote sensing dataset corresponding to a first geographic area, the first remote sensing dataset including a first set of atmospheric gas concentration data for at least one atmospheric gas, a first set of multispectral data, and a first set of spatially autocorrelated land use classifications;
• provides input to the at least one GIML model, the input comprising a second remote sensing dataset corresponding to a second geographic area, the second remote sensing dataset including a second set of atmospheric gas concentration data for the at least one atmospheric gas, a second set of multispectral data, and a second set of spatially autocorrelated land use classifications;
• generates, using the at least one GIML model, a plurality of predicted atmospheric gas concentration values for the at least one atmospheric gas and associated with the second geographic area, and wherein the plurality of predicted atmospheric gas concentration values has a first spatial resolution that is greater than a second spatial resolution of the first and second sets of atmospheric gas concentration data;
• and
• processes the plurality of predicted atmospheric gas concentration values.

A method for generating one or more high-resolution atmospheric gas concentration maps using geography-informed machine learning, the method comprising:

• training a machine learning model with a training dataset comprising a first plurality of data points corresponding to a first geographical area, wherein each data point of the first plurality of data points is associated with a first set of atmospheric gas concentration data for at least one atmospheric gas, a first set of multispectral data, and a first set of spatially autocorrelated land use classifications;
• providing input to the trained machine learning model, the input comprising a second plurality of data points corresponding to a second geographical area, wherein each data point of the second plurality of data points is associated a second set of atmospheric gas concentration data for the at least one atmospheric gas, a second set of multispectral data, and a second spatially autocorrelated land use classification;
• generating, by the trained machine learning model, a third plurality of data points responsive to the input, wherein each data point of the third plurality of data points is associated with a predicted atmospheric gas concentration value for the at least one the at least one atmospheric gas, wherein the third set of data points has a spatial resolution that is greater than a spatial resolution of the first plurality of data points and the second plurality of data points;
• processing the third plurality of data points;
• and
• generating, based on the predicted atmospheric gas concentration value associated with each processed third plurality of data points, at least one atmospheric concentration map.