| During the last decades, new instruments have been designed and built to improve observations of atmospheric temperature, water vapor, and winds. In the area of infrared remote sensing, new technologies will enable the next generation of instruments, like the Geostationary Imaging Fourier Transform Spectrometer (GIFTS), to collect high spectral and spatial resolution data with very high data rates. If not properly compressed those data rates will exceed the capacity of the current operational downlink technology and will require expensive data systems to process the data on the ground. This dissertation focuses on establishing the compression and inversion procedures to reduce the volume of data with minimal information loss and with beneficial effects on the accuracy of the retrieved atmospheric variables.; To take full advantage of the large number of channels available and the high correlation between them, Principal Component Analysis (PCA) has been chosen as the basis for the compression procedure. By separating the atmospheric signal from the random component of the instrument noise, a PCA-based Compression algorithm (PCC) leads to high values of the compression ratio with an overall improvement of the signal-to-noise ratio.; The results obtained by applying PCA to both simulated and real data, show that it represents a key component in processing high spectral resolution data. PCA can be used to reduce the volume of data to be inverted for the retrieval of the atmospheric variables and to improve the signal to noise ratio. Both the data volume and the noise reduction have been demonstrated to be beneficial for the retrieval process, increasing the accuracy and the efficiency of the clear and cloudy sky inversion algorithms. Moreover, PCA has been proved useful in lowering the post-processing costs and improving the quality of the retrieved variables, the ultimate products of interest for the scientific community. |
