Studies of global cloud field using measurements of GOME, SCIAMACHY and GOME-2

Tropospheric clouds are main players in the Earth climate system. Characterization of long-term global and regional cloud properties aims to support trace-gases retrieval, radiative budget assessment, and analysis of interactions with particles in the atmosphere. The information needed for the determination of cloud properties can be optimally obtained with satellite remote sensing systems. This is because the amount of reflected solar light depends both on macro- and micro-physical characteristics of clouds. At the time of writing, the spaceborne nadir-viewing Global Ozone Monitoring Experiment (GOME), together with the Scanning Imaging Absorption Spectrometer for Atmospheric Chartography (SCIAMACHY) and GOME-2, make available a unique record of almost 17 years (June 1996 throughout May 2012) of global top-of-atmosphere (TOA) reflectances and form the observational basis of this work. They probe the atmosphere in the ultraviolet, visible and infrared regions of the electromagnetic spectrum. Specifically, in order to infer cloud properties such as optical thickness (COT), spherical albedo (CA), cloud base (CBH) and cloud top (CTH) height, TOA reflectances have been selected inside and around the strong absorption band of molecular oxygen in the wavelength range at 758-772 nm (the O2 A-band). The retrieval is accomplished using the Semi-Analytical CloUd Retrieval Algorithm (SACURA). The physical framework relies on the asymptotic parameterizations of radiative transfer. The generated record has been throughly verified against synthetic datasets as function of cloud and surface parameters, sensing geometries, and instrumental specifications and validated against ground-based retrievals. The error budget analysis shows that SACURA retrieves CTH with an average accuracy of ±400 m, COT within ±20% (given that COT > 5) and places CTH closer to ground-based radar-derived CTH, as compared to independent satellite-based retrievals. In the considered time period the global average CTH is 5.2±3.0 km, for a corresponding average COT of 20.5±16.1 and CA of 0.62±0.11. Using linear least-squares techniques, global trend in deseasonalized CTH has been found to be -1.78±2.14 m * year-1 in the latitude belt ±60°, with diverging tendency over land ( 0.27±3.2 m * year-1) and water (-2.51±2.8 m * year-1) masses. The El Nino-Southern Oscillation (ENSO), observed through CTH and cloud fraction (CF) values over the Pacific Ocean, pulls clouds to lower altitudes. It is argued that ENSO must be removed for trend analysis. The global ENSO-cleaned trend in CTH amounts to -0.49±2.22 m * year-1. At a global scale, no explicit patterns of statistically significant trends (at 95% confidence level, estimated with bootstrap resampling technique) have been found, which are representative of peculiar natural climate variability. One exception is the Sahara region, which exhibits the strongest upward trend in CTH, sustained by an increasing trend in water vapor. Indeed, the representativeness of every trend is affected by the record length under study. 17 years of cloud data still might not be enough to provide any decisive answer to current open questions involving clouds. The algorithm used in this work can be applied to measurements provided by future planned Earth's observation missions. In this way, the existing cloud record will be extended and attribution of cloud property changes to natural or human causes and assessment of cloud feedback sign within the climate system can be investigated