Field campaign data analysis in support of the future FORUM and CAIRT ESA missions

This work aims at analysing the measurements obtained from two instruments, FIRMOS and GLORIA, during their employment in the HEMERA ballon cam- paign in Timmins (Canada) in the period 23-24 August 2022. The first part of this work deals with the geolocation of the GLORIA data. After that the focus is posed on the scene classification which is obtained by the application of the CIC (Cloud Identification and Classification) algorithm. CIC is a supervised machine learning code based on PCA Principal Component Analyses that is able to perform cloud identification and classification from high spectral resolution data at infrared wavelengths. The CIC code has been recently tested on the identification of cloudy scenes. The classification method is based on a distributional approach of similarity index computed from the element of the dataset with respect to the training set made available to the algorithm. Two different versions of the classificator are tested. For the first time, the CIC algorithm is applied to upwelling radiance fields in this configuration. Another novelty is the CIC application to high spatial resolution data from GLORIA++ to perform a soil classification. The exercise is an important test within the studies aiming at improving the initial guess of the geophysical retrieval of future satellite sounders. The last section of this work describes the application of FARM algorithm, an optimal estimation based retrieval code relying on an innovative forward model σ − F ORU M . σ − F ORU M is a modify version of the σ − IASI code which is a monochromatic fast code for calculating synthetic radiances in the 10-2760 cm−1 spectral range. The goal of this analyses is to retrieve atmospheric and surface parameters such as: surface temperature and emissivity, and temperature profiles and H2O vertical profiles.