Minimize Web Content Display
Minimize Water vapor, clouds, and aerosols

Monitoring of the global distribution of water vapor, clouds and aerosols is essential for the understanding of the Earth's energy budget. All three quantities are interconnected in several ways: for example clouds can form when relative humidity exceeds saturation levels, the properties and amount of aerosols determine cloud properties (acting as cloud condensation nuclei), and precipitation is an effective removal process for aerosols. Because water vapor, clouds and aerosols are also involved in several climate feedback mechanisms they are important for the understanding and prediction of climate change. Water vapor, clouds and aerosols are highly variable making global observations from satellite highly desirable. Although instruments like GOME, SCIAMACHY and OMI (GOME-type instruemnst) were not primarily designed for the observation of these quantities, they can provide very valuable information on the global distribution of water vapor, clouds and aerosols.


Chapter Editor

Thomas Wagner (MPIC)


Kelly Chance (SAO), Steffen Dörner (MPIC), Michael Grzegorsky (EUMETSAT), Margherita Grossi (DLR), Joanna Joiner (NASA/Goddard), Alexander Kokhanovsky (EUMETSAT), Luca Lelli (IUP Bremen), Sebastian Mieruch (KIT), Stefan Noël (IUP Bremen), Marloes Penning de Vries (MPIC), Holger Sihler (MPIC), Gijsbert Tilstra (KNMI), Omar Torres (NASA/Goddard), Huiqun Wang (SAO)



Global water vapour total columns have been derived from measurements of the Global Ozone Monitoring Experiment (GOME) on ERS-2 and the Scanning Imaging Absorption Spectrometer (SCIAMACHY) on ENVISAT in the spectral region around 700 nm using the Air Mass Corrected Differential Optical Absorption Spectroscopy (AMC-DOAS) method. From these data, gridded daily means with a spatial resolution of 0.5° x 0.5° have been determined. They are further averaged to get monthly and annual means. For each spatial grid point, linear trends were then derived based on the monthly means from GOME (1996-2002) and SCIAMACHY (2003-2007). Because GOME and SCIAMACHY measurements had different spatial and temporal sampling and resolution, the trend fit considered a potential offset / jump in the time series when switching from GOME to SCIAMACHY data (see Mieruch et al., 2008). Courtesy: Stefan Noël , Sebastian Mieruch (all U. Bremen).


This map was based on 7 years of absorbing aerosol index information measured by SCIAMACHY. Only clear aerosol events (AAI > 0.5) were averaged, meaning that the map also contains one-time events over the 7-year period. The detected aerosols originate mainly from desert dust storms and seasonally recurring biomass burning events, but forest fires, such as those in Siberia and Canada, and occasional volcanic eruptions, also contribute to the global aerosol distribution (Tilstra et al., 2012). Courtesy: Gijsbert Tilstra, Martin de Graaf, Bas Mijling, Piet Stammes (all KNMI).


Aerosol source type climatology (2007-2011) derived using the Global Aerosol Classification Algorithm (GACA). Combining aerosol data, e.g. from MODIS, with trace gas data from GOME-2 (NO2, HCHO, and SO2) and MOPITT (CO) is an innovative approach to the difficult issue of aerosol classification from space. GACA exploits the co-existence and correlation between aerosols and trace gases (due to e.g. co-emission, as in the case of HCHO, CO, and small absorbing particles from biomass burning) to derive the dominant aerosol source type. The sources discriminated are (see legend): biomass burning (BB), desert dust (DD), secondary aerosols of biogenic (BIO) or urban/industrial (URB) origin, aged/transported aerosols (AGED), volcanic sulfate (VOG), and sea salt (SS); unknown aerosol mixtures are designated XX; regions with very little aerosol or trace gas data were not assessed (na). (Penning de Vries et al., 2015). Courtesy: Marloes Penning de Vries (MPIC).


Average aerosol optical depth from GOME-2 (METOP-A) for June/July 2013. The data is retrieved by EUMETSAT's Polar Multi-sensor Aerosol product (PMAp), combining GOME-2 and AVHRR/3. AVHRR is used for cloud correction, volcanic ash and dust detection. The interpolation of the AOD and the selection of the microphysical properties are based on GOME-2. As expected, hotspots are found for dust transport from the Sahara, biomass burning in central Africa, pollution in China and dust between the Arabian Peninsula and India. The time period contains a few exceptional events, such as very strong wild-fires in Canada and western USA. This results in higher values than usual in the North Atlantic, the Hudson Bay and at the Pacific coast of the USA. An inter-comparison between chemical transport models and PMAp for the years 2013 and 2010 also shows that the dust plume from the Sahara is weaker in June/July 2013 than in previous years. Note that some pixels on earth have only a few cloud-free overpasses from METOP-A. Consequently some of the hotspots are dominated by one or two measurements only. AOD retrievals over ocean are available operationally in near-real time. AOD retrievals over land, including additional information on the aerosol type based the GOME UV index and dust/ash detection from IASI, are available on a prototype level and are expected to become operational in Q1/2016. Courtesy: Michael Grzegorsky (MPIC).


Eight-year global record of aerosol absorption properties derived from OMI near-UV observations. Global seasonal averages of 388 nm aerosol single scattering albedo are shown on the left panel. Time-series of monthly, zonal averages of single scattering albedo (top) and aerosol absorption optical depth (bottom) are depicted on the right panel (Torres et al., 2013). Courtesy: Omar Torres (NASA).


Radiance measurements in limb geometry are affected by scattering events in the volume of sensitivity (VOS). Therefore, strong gradients, for example in volcanic plumes lead to underestimations of plume altitude and aerosol optical thickness due to assuming horizontal homogeneity while retrieving aerosol extinction from the measured radiance. This issue can be solved by using proxy data which locate a plume horizontally. The method is applied to a SCIAMACHY measurement on 13th June 2011, right after the eruption of the Nabro volcano. Here, SO2 VCD retrieved from SCIAMACHY measurements in nadir geometry are used to define the location of the volcanic plume (Penning de Vries et al., 2014). Courtesy: Steffen Dörner.


Comparison between the Total Column Water Vapour (TCWV) retrieved with the GOME-2 instrument on board of the MetOp-A satellite and the co-located model data from the ECMWF ERA-Interim reanalysis in February 2013 (top panels) and August 2013 (bottom panels). The TCWV distribution follows the seasonal cycle of the near surface temperature: the H2O total column has a maximum during the northern hemisphere summer, and a minimum in winter. The GOME-2 retrievals capture the overall spatial variability in the H2O total column values quite well both over ocean and land surfaces. The mean global bias between the two data sets is rather small: 0.036 g/cm2 in February and 0.066 g/cm2 in August 2013. While in February the bias is overall low, in August larger relative differences between GOME-2A and ECMWF ERA-Interim data can be seen. Looking at the bottom right panel of the figure, we can see that an underestimation of the TCWV (blue regions) is located in land areas with a very high humidity. Dry bias is also observed in regions with high surface albedo values, like the Northern Africa, the Arabian Peninsula, India and part of the East Asia and Central America (Grossi et al., 2015). Courtesy: Margherita Grossi (DLR).


Relative anomalies of water vapor, cloud shielding and cloud fraction during the strong ENSO event of 1997/98 derived from GOME-1 observations (Wagner et al., 2005). Strong enhancements for all quantities occur over the tropical Pacific, where the sea surface temperature is also enhanced. The O2 cloud shielding indicates the reduction of the O2 absorption at 650 nm caused by clouds. Note that significant El-Niño induced anomalies are found not only for the tropics but also for mid and high latitudes (Wagner et al., 2005). Courtesy: Thomas Wagner (MPIC).


Comparison of the Level 3 0.5  0.5 degree monthly mean SAO OMI Total Column Water Vapor (TCWV) product and GlobVapour combined SSM/I(over ocean)+MERIS(over land) TCWV product for (left) January and (right) July 2006. The top row shows the results derived from OMI using SAO's two-step retrieval algorithm where the vertical column is derived from the ratio of the slant column and the Air Mass Factor (AMF). The slant column retrieval uses the blue spectral range (430 – 480 nm), and considers water vapor, ozone, nitrogen dioxide, oxygen collision complex, liquid water, glyoxal, Ring, water Ring, wavelength shift, third order polynomial, spectral undersampling and common mode. The AMF is derived using the radiative transfer model VLIDORT with a priori information on surface albedo, water vapor vertical profile, cloud fraction and cloud top pressure. The middle row shows the GlobVapour SSM/I+MERIS product downloaded from SSM/I measures TCWV using microwave over the ocean. MERIS measures TCWV in the near IR over land. The third row shows the difference of OMI – GlobVapour. On the global scale, the mean difference is -0.40 cm in January and -0.30 cm in July. Over the land, the mean difference is 0.02 cm in January and -0.05 cm in July. Over the ocean, the mean difference is -0.58 cm in January and -0.41 cm in July (Wang et al., 2014). Courtesy: Huiqun Wang and Kelly Chance (SAO).


Trends in top altitude of clouds (in meter per year) derived from measurements of the spectrometers GOME, SCIAMACHY and GOME-2 in the period between June 1996 and May 2012. From top, clockwise: (a) map of the absolute changes in cloud top altitude; (b) map of the natural variability (standard deviation) of the changes in cloud top altitude; (c) map of those changes in cloud top altitude exceeding natural variability at the 95% confidence; (d) time series of cloud top height (CTH, top plot) and cloud horizontal extent (also termed cloud fraction, CF, bottom plot) and their correlation with ocean mean temperatures at the surface (El Niño 3.4 index) over the Central East Pacific (170°W - 120°W and 5°N - 5°S). Over this area (red outlined box), the El Niño-Southern Oscillation (ENSO) typically reaches its maturity, strongly modulating cloud properties. Globally, the ENSO is thought to pull clouds to lower altitudes (Lelli et al., 2014). Courtesy: Luca Lelli and Alexander Kokhanovsky (U. Bremen).


Detection of multi-layered clouds is important for obtaining accurate trace gas retrievals, evaluating model cloud parameterizations, and calculating radiative forcing. Multi-layer clouds can also present difficulties for trace-gas remote sensing, particularly when the gases are not well mixed (as is the case for NO2, SO2, HCHO, etc.). We need to be able to detect such problematic cases. We can do this using complementary information from OMI and MODIS (or with MODIS alone, using its shortwave H2O channel). The figure presents the fraction of OMI cloudy pixels containing distinct multi-layer clouds, July 2007. MODIS cloud-top pressure (sensitive to the physical cloud top) and OMI optical centroid cloud pressure (sensitive to bright lower cloud decks) are used to detect multi-layer clouds with good spatial coverage. The approach is tuned and validated using CloudSat radar data. (Adapted from Joiner et al., 2010). Courtesy: Joanna Joiner (NASA).


3D radiative properties become increasingly important as the resolution of satellite measurements increases: (a) The radiative transfer (RT) at cloud edges may have a brightening or darkening effect on the measured radiance, e.g. radiation is channeled into cloud shadow or cloud sides brighten the otherwise clear-sky scene. (b) At a wavelength of 440nm, the 2D box air-mass factor (boxAMF) of the clear-sky case illustrates the sensitivity distribution of DOAS measurements within and outside the geometric light path between satellite, surface and sun. (c) The sensitivity to near-surface trace-gas concentrations is decreased by the shadowing effect of a cloud with cloud optical thickness (COT) of 10, even though the geometric cloud fraction is zero (i.e. no cloud in the line-of-sight of the satellite). Furthermore, the sensitivity above the cloud is increased by a higher density of horizontal light-paths. Courtesy: Holger Sihler (MPIC).