Possibilities of IPDA spaceborne lidar and neural networks for measuring methane concentration

The possibilities of retrieving the mean relative concentration of methane using pseudo-inverse matrices, fully connected and convolutional neural networks based on signals from the Earth received by a 450 km orbit space-based lidar are considered. It is shown that the random error of methane concentration retrieving for lidar with laser pulse energy of 9 mJ and repetition rate 20 Hz, receiving system size of 68 cm, resolution of 50-60 km is not higher than 30 ppb.