Carbon dioxide column concentration space-time sequence adjustment method based on machine learning

Abstract

The invention discloses a carbon dioxide column concentration space-time sequence adjustment method based on machine learning. The method comprises the following steps: firstly, through an empirical Bayesian Kriging interpolation (EBK) theory, completing point-to-surface conversion of discrete data in spatial properties to obtain a spatial fitting value result; secondly, constructing a time parameter library, carrying out reverse statistics on an annual XCO2 rule of a satellite, fitting satellite data of which the effective value of a single pixel is located in October to 12 months by adopting a specific formula, putting obtained parameters into the parameter library, and labeling corresponding point location information in the time parameter library; and then matching the space fitting value result with the fitting value of the corresponding point location in the step S2 by adopting a transfer learning TCA technology, and distributing the obtained parameters in the time parameter library to each point location of the global research area; and finally, taking each point location as a basic unit, substituting the distributed parameters into a specific formula, fitting is carried out again, wherein a fitting result is a data product after space-time adjustment.

Description

technical field [0001] The invention relates to the field of atmospheric remote sensing, in particular to a method for adjusting the time-space series of carbon dioxide column concentration based on machine learning, thereby filling the data blank area in greenhouse gas satellite observation. Background technique [0002] Due to the rapid increase in global emissions of major greenhouse gases, the greenhouse effect is intensifying, affecting the health of the planet's ecosystems and economic prosperity. Since the industrial revolution in the last century, the global carbon dioxide concentration has risen from 278ppm before the industrial era to 410ppm in 2020. Increases in atmospheric carbon dioxide concentrations have caused global climate change, with a range of societal impacts. As a result, the scientific community has invested enormous efforts in understanding carbon cycle mechanisms through actual measurements and advanced modeling tools. Burning of fossil fuels and ...

AI Technical Summary

Problems solved by technology

However, clouds and aerosols will cause spectral interference to the carbon dioxide signal, and at the same time, the data gap between orbits will be caused by the satellite observation mechanism

Therefore, XCO 2 The number of retrieved products is small and the data availability is low, thus limiting the application of these valuable satellite data

Images