Search

An innovative methodology for Cloud Identification and Classification (CIC) is presented. The CIC algorithm is a machine-learning technique, based on Principal Component Analysis, performing binary classification for clear/cloudy scene detection and classification (i.e. liquid/ice cloud phase). Firstly, CIC is tested on a wide synthetic dataset of high spectral resolution radiances in the far and mid infrared (FIR and MIR) part of the spectrum simulating measurements from the ESA Earth Explorer Fast Track 9 competing mission FORUM (Far Infrared Outgoing Radiation Understanding and Monitoring) that is currently (2018/19) undergoing the industrial and scientific Phase-A studies. Simulated spectra represent observations covering many diverse climatic areas, ranging from the tropical to polar regions. Secondly, CIC is applied to a small dataset of high spectral resolution radiance observations taken from aircraft on the 13th March 2015 off the north-east coast of Scotland, UK, during the Cirrus Coupled Cloud-Radiation Experiment (CIRCCREX). Measurements in the FIR are made by the Tropospheric Airborne Fourier Transform Spectrometer, TAFTS, (Canas et al. 1997) and in the MIR by the Airborne Research Interferometer Evaluation System, ARIES, (Wilson et al. 1999). The main result demonstrates the high information content of spectral radiance in the FIR region of the electromagnetic spectrum which is exploited here to improve the detection performance in identifying cloudy scenes, specifically thin cirrus clouds.

In November 2017 the European Space Agency selected FORUM (Far-infrared Outgoing Radiation Understanding and Monitoring) as one of the two instrument concepts to be developed further and to compete to be the ninth Earth Explorer mission. FORUM will be a Fourier Transform Spectrometer measuring the upwelling spectral radiance emitted by the Earth across the most relevant infrared part of the electromagnetic spectrum from satellite. In particular, the instrument will cover the spectral range from 100 to 1600 cm -1 (from 100 to 6.25 microns in wavelength), covering the Far InfraRed (FIR), between 100 and 667 cm -1 , which has never been done from space before. Idealized model simulations suggest that between one quarter and one third of the total clear-sky long-wave cooling of the Earth to space occurs within the FIR spectral region. An even larger fraction of this cooling is predicted to occur in the FIR under all-sky conditions, because the presence of clouds causes lower emitting temperatures, and hence a shift to longer wavelengths of the peak of the black-body curve. For this reason, a detailed study of the Earth's radiation budget requires accurate, frequent, and global-coverage measurements of the total upwelling irradiance. In this paper we show that FORUM measurements will permit to obtain a very accurate estimate of the upwelling FIR irradiance. The achieved accuracy will be much better than that achievable with state-of-the-art forward model calculations of the FIR spectrum, based on the water vapor distribution determined from operational Mid- Infra-Red measurements, such as those from the Infrared Atmospheric Sounding Interferometer (IASI) on Met-Op satellites. For simplicity reasons the presented study is limited to clear-sky. In cloudy conditions, due to the large uncertainties in cloud composition, forward model simulations are expected to be even less accurate.

A new cloud identification and classification algorithm named CIC is presented. CIC is a machine learning algorithm, based on principal component analysis, able to perform a cloud detection and scene classification using a univariate distribution of a similarity index that defines the level of closeness between the analysed spectra and the elements of each training dataset. CIC is tested on a widespread synthetic dataset of high spectral resolution radiances in the far- and mid-infrared part of the spectrum, simulating measurements from the Fast Track 9 mission FORUM (Far-Infrared Outgoing Radiation Understanding and Monitoring), competing for the ESA Earth Explorer programme, which is currently (2018 and 2019) undergoing industrial and scientific Phase A studies. Simulated spectra are representatives of many diverse climatic areas, ranging from the tropical to polar regions. Application of the algorithm to the synthetic dataset provides high scores for clear or cloud identification, especially when optimisation processes are performed. One of the main results consists of pointing out the high information content of spectral radiance in the far-infrared region of the electromagnetic spectrum to identify cloudy scenes, specifically thin cirrus clouds. In particular, it is shown that hit scores for clear and cloudy spectra increase from about 70 % to 90 % when far-infrared channels are accounted for in the classification of the synthetic dataset for tropical regions.