Fast inversion method for atmospheric methane on basis of short wave infrared hyperspectral data under condition of cirrus

Abstract

The invention belongs to the technical field of atmospheric remote sensing and discloses a fast inversion method for atmospheric methane on the basis of short wave infrared hyperspectral data under the condition of cirrus. The method comprises the following steps: obtaining world-wide representative atmospheric temperature, humidity and methane profile, and obtaining reflection spectral characteristic information of a typical ground object; obtaining a simulation spectral data set of a methane absorption band; performing PCA (principal component analysis) on the basis of simulated and calculated short wave infrared methane absorption spectral data set under the condition of global cirrus; training an established neural network by using first eight principal components of an observation spectrum, observation pixel height, surface air pressure, sun angle and satellite observation angle as input of the neural network and using methane input by a forward radiative transfer model as outputof the neural network; establishing an short wave infrared atmospheric methane fast inversion model based on PCA and the neural network under the condition of cirrus. Fast and precise inversion of atmospheric methane on the basis of the short wave infrared hyperspectral data under the condition of cirrus can be realized with the method.

Description

technical field [0001] The invention belongs to the technical field of atmospheric remote sensing, and in particular relates to a fast inversion method for atmospheric methane of short-wave infrared hyperspectral data under cirrus cloud conditions. Background technique [0002] At present, the existing technologies commonly used in the industry are as follows: Methane (CH 4 ) is an important greenhouse gas in the atmosphere, and its greenhouse effect is second only to atmospheric carbon dioxide. The increase of methane concentration has an important impact on the atmospheric energy balance and global climate change. At present, the measurement methods of methane mainly include: direct sampling measurement, laser radar, ground-based spectrometer and satellite observation. Among them, satellite observation methods can obtain the distribution of methane concentration on a global scale, mainly including short-wave infrared and thermal infrared hyperspectral observation method...

AI Technical Summary

Problems solved by technology

[0003] In summary, the problems in the prior art are: At present, the research on shortwave infrared hyperspectral data in the case of cirrus clouds mainly uses the physical method of optimal estimation theory to invert methane information. The algorithm is easily affected by prior information such as atmospheric profiles, cirrus clouds, and surface parameters. During the evolution process, a large amount of radiative transfer calculation time is required due to the scattering process involving clouds

[0005] Since the current numerical prediction model and assimilation model cannot provide accurate atmospheric temperature and humidity profiles, and conventional observation methods cannot provide reliable cirrus cloud height, thickness, optical characteristics and other macro-micro parameters, so the optimal estimation theory based It is difficult for the physical inversion algorithm to obtain reliable methane information under cirrus cloud weather based on shortwave infrared hyperspectral data

Images