CISAR (Combined Inversion of Surface and AeRosol) is a generic algorithm for the joint retrieval of surface reflectance and aerosol properties with continuous variations of the state variables in the solution space. CISAR is based on the inversion of a fast Radiative Transfer Model (RTM) in an optimal estimation framework, including prior information on the state variable magnitude and regularization constraints on their spectral and temporal variability. The observed scenario is described through two atmospheric layers and an underlying surface, radiatively coupled with atmospheric scattering. CISAR retrieves the surface reflectance model parameters, the aerosol optical thickness and single scattering properties in each processed wavelength.

This study analyses CISAR algorithm performance when applied on satellite data acquired from radiometers flying at different orbits. For this purpose, SEVIRI (geostationary) and PROBA-V (polar orbiting) observations acquired over 20 AERONET stations observed by both instruments during year 2015 have been processed. SEVIRI data have been processed in the framework of the aerosol_cci2 project. The processing of PROBA-V data was part of the ESA PV-LAC project. The selected stations span different geometries and land cover type (vegetation, urban, desert, water bodies). Satellite observations are accumulated in time and the inversion is performed over one accumulation period. The length of this accumulation window is determined by the availability of cloud-free observations and thus is strongly dependent on the satellite orbit.

The information content of the SEVIRI and PROBA-V observations is quantified through the analysis of the entropy (the uncertainty reduction from the prior to the solution) and the Jacobians. The seasonal variation of the sensitivity of the TOA BRF to the state variable is also analysed through the Jacobians.

The concept of observation system will be first defined, including the characterization of all associated uncertainties. CISAR’s retrieved surface and aerosol properties will be evaluated against MODIS land product and AERONET data respectively. The performance differences resulting from the two types of orbit are discussed, analysing and comparing the information content of SEVIRI and PROBA-V observations.

Area C - Board 277

A ground-based bistatic lidar system located at an altitude of 3397 m above sea level at Mauna Loa Observatory (MLO) on the Big Island of Hawaii in the Pacific Ocean was used to detect aerosols entrained in air above the observatory. While Mauna Loa Observatory typically experiences clean tropospheric air, during certain seasons, dust from Asia can be lofted by strong storms and transported over the ocean to Mauna Loa and beyond. The lidar system was used to detect candidate aerosol layers above MLO. Radiosonde and photometer data were applied to better characterize the particles. Back-trajectory modeling was then used to investigate the origins of the aerosols.

The lidar system used in this study employed a 20 watt NdYAG laser at 532 nm as its transmitter. Laser light sent vertically up into the atmosphere scattered off clouds, air molecules and aerosols. Unlike in traditional elastic lidar where the backscattered light is collected and the time of flight of the laser pulse is measured to derive the altitude of the scattering, this system employed a CCD camera detector with wide angle optics to image the entire laser beam from ground to zenith from the side at a distance of 139 m from the laser. In this bistatic configuration where the laser and detector are spatially separated, the altitude of the scatterer can be determined from geometry without time of flight measurements of the laser pulse. The raw data image from the lidar detector consists of pixel intensities and locations. The scattered signal at each altitude is determined by fitting a Gaussian plus a constant across the width of the beam in the image. The area under the Gaussian is taken as the side-scattered laser signal. The constant represents the sky background. Molecular scattering is removed from the signal by fitting a molecular model to the scattering in an aerosol-free region and subtracting the molecular portion from the side scattered signal. The remainder is taken to represent cloud/aerosol scatter. Since the scattering angle changes with altitude for this system, an aerosol phase function (which describes the angular dependence of the aerosol scatter) is needed. The lidar system operates only at night. During the day a Cimel photometer from the Aerosol Robotic Network (AERONET) conducts measurements at Mauna Loa, and AERONET derived aerosol scattering phase functions from these data were used in the lidar analysis. The signals were corrected for transmission, converted to total scatter, and a single scattering albedo was assumed to derive aerosol extinction as a function of altitude.

In this study, on two different days during Asian dust season at MLO, aerosol layers were observed for a few hours above Mauna Loa. Balloon-borne radiosonde measurements of relative humidity were used to discriminate between aerosols and tenuous ice clouds. The radiosondes were launched from Hilo, Hawaii, 60 km from MLO. Layers where relative humidity was low were studied further. Figure 1 shows the lidar derived aerosol extinction vs. altitude above sea level for the two dates, with aerosol layers of interest noted by the arrows. Origins of the dry layers were studied using the National Oceanic and Atmospheric Administration’s (NOAA) Hybrid Single Particle Lagrangian Integrated Trajectory (HYSPLIT) model to conduct back-trajectories of the parcels above MLO at the altitudes of the identified aerosol layers. Figure 2 shows the back trajectories of the highlighted aerosol layers. HYSPLIT results showed origins in arid regions of East Asia. Study results suggest that the detected layers are dry aerosols originating from arid East Asian regions during dust season. The bistatic camera lidar system was shown to be useful in detection of entrained dust layers and aerosols and can be combined with modeling and other data to provide insights on transport phenomena with aerosols as tracers.

Area C - Board 278

Even though the aircraft equipment-technologies are developed, harsh atmospheric phenomena have high potential causing socioeconomic damages in the aviation. Such visible phenomena (e.g., Tropical cyclone and convective cloud) are easily avoidable, but the invisible phenomenon (i.e., Inner cloud icing) is not simple. Aviation icing is an ice-film forming phenomenon on the aircraft surface (e.g., Aircraft wings, engine and antenna) and causes aviation to miss-control, when the aircraft goes through the super-cooled droplet (SCD) rich cloud. Theoretical estimation of SCDs in the cloud contains lots of uncertainties, thus, aircrafts usually detouring thick and cold clouds even it spends unnecessary time and fuel consumption. Thus, Sim et al., (2018) suggested a potential icing detection method using machine learning approaches to avoid the uncertainties of the theoretical estimation of icing. Sim’s study is estimating the icing potential location based on current values of the geostationary satellite pixel, thus it may cause spatio-temporal discontinuity resulting in inconstant aviation path. Therefore, in this study, advanced icing potential cloud detection methods are proposed, which produce a stable product in spatial and temporal change. To consider spatio-temporal information in the icing potential cloud, statistical variable (e.g., standard deviation and mean) and convolutional filtered value (e.g., edge emphasizing filter, horizontal line filter and vertical line filter) were added to the input variable with their temporal changes.
Communication, Ocean and Meteorological Satellite (COMS) Meteorological Imager (MI) sensor, and Himawari-8 Advanced Himawari Imager (AHI) products are used as input data in this study. Decision trees (DTs) and random forest (RF) is used to produce icing potential cloud detection. Quality checked pilot reports (PIREPs) were used as in-situ data. Result using COMS data with the RF approach from the Sim’s study shows the best accuracy—detection accuracy of 87.1%, false alarm rate of 24.3%, and overall accuracy of 79.5%— compared with the other approaches. The advanced model will contain the spatio-temporal, changing the property of the icing potential cloud with high performance.

Area C - Board 279

Aerosols are an important regulator of the Earth’s climate. However, the magnitude of aerosol radiative effects remains the single largest uncertainty in current estimates of anthropogenic radiative forcing. One of the key quantities needed for accurate estimates of anthropogenic radiative forcing is an accurate estimate of the radiative effects from natural unperturbed aerosol. The dominant source of natural aerosols over Earth’s vast forested regions is biogenic volatile organic compounds (BVOC) which, following oxidation in the atmosphere, can participate in new particle formation or condense onto aerosol particles to form secondary organic aerosol (SOA). Consequently, BVOC emissions could introduce a regionally relevant cooling feedback in a warming climate. However, the evidence for a link between air temperature and the aerosol radiative effects due to increased BVOC emissions is not yet definite because in addition to BVOC emissions photochemistry and synoptic meteorology also depend on temperature and can affect SOA formation.

The main objective of this study is to provide a quantitative estimate of the regional aerosol direct radiative effect caused by the temperature-dependent biogenic emissions over the boreal forests in present day conditions and in a warmer future. The study is done using a combination of satellite data and climate modeling. Key remote sensing data used are the AATSR based aerosol optical depth (AOD) and land surface temperature (LST) products available from the Aerosol-CCI and GlobTemperature projects, together with ancillary data, such as column concentrations of CO and water vapour from AIRS, Fire Radiative Power from MODIS, and NO2 column densities from OMI. The aerosol-chemistry climate model used is ECHAM-HAMMOZ, which describes the relevant atmospheric aerosol processes. It includes all the main atmospheric aerosol compounds and BVOC emissions are computed online using the Model of Emissions of Gases and Aerosols from Nature (MEGAN), which enables the simulation of the effects of temperature changes on atmospheric aerosol load.

Our analysis shows that there could be a small temperature dependence in AOD over the boreal forests but it cannot be reliably detected from the observations or the simulations. The only subregion with a clear temperature dependence in AOD was found over western Russia. Anthropogenic emissions affect this subregion more than the other regions analyzed thus, it is likely that in addition to BVOC emissions hygroscopic sulfate aerosols affect the temperature dependence of AOD.

In a warmer future the clear-sky radiative forcing caused by biogenic aerosols will increase, following the increase of BVOC emissions, but if anthropogenic emissions will decrease at the same time the total clear-sky forcing will also decrease.

Area C - Board 280

The Arctic environment is a very sensitive area to changes of climate conditions. The yearly mean temperature did raise on a much higher level compared to other regions of the Earth. Aerosol parameters such as type, number concentration, distribution or the aerosol optical depth can have impacts on the energy fluxes. Secondary effects on clouds and cloud formation have possible impacts on number concentration and cloud condensation nuclei. The characterization of the status of the atmosphere and the knowledge of aerosol parameters help to examine and understand such effects and processes. The scattering function of particles strongly depend on type and size of the particles. During clean air conditions the pattern of the polarized atmospheric radiation of scattered Sun light differs strongly compared to Arctic haze, polluted or fog conditions. The derived scattering function can give information of type, particle number concentration and other aerosol related parameters. A full stokes polarimeter (URMS/AMSSP: Universal Radiation Measurement System/ Airborne Multispectral Sun- and Polarimeter) was deployed on a research vessel (Polarstern) during an Arctic ship cruise (PS106). The cruise went up 2 times over a period of time of 8 weeks in Summer 2017 North West and North East of Svalbard up to 82° and 84° North. In that time of year the Sun elevation changes from about 10° to 30° during a 24 hour period. The system did measure approximately every 0.5 hours the polarized radiance's partially in the upper and lower hemisphere to derive polarized radiance pattern of scattered Sun light and snow and water reflectances. The measurements in this special environment during night and day with low Sun elevation angles and different meteorological conditions will be presented. In addition the current status of laboratory calibration, data processing procedures and first radiative transfer simulations of such condition are part of the presentation.

Area C - Board 281

Both in-situ measurements and satellite observations unequivocally show that temperature inversions are common phenomena in the Arctic during polar winter. This invisible barrier controls the atmosphere and surface interactions, including aerosol vertical distribution. Having knowledge about the relationship between aerosols and inversions is important as the radiative and dynamic impacts of aerosols in the Arctic are strongly dependent on how they are distributed in the atmosphere in the presence of inversions. However, there exists a key knowledge gap about this relationship. Once transported into the Arctic, the final fate of the aerosols and their overall impact on the atmosphere and surface depends heavily on local meteorological conditions such as inversions. This is one of the research areas where comparatively little is known about the Arctic, due to the lack of suitable observations.

In this study, we investigate how the varying magnitude of the atmospheric stability can affect the dispersal of the aerosols. Ten years of collocated aerosol retrievals from CALIOP-CALIPSO and temperature and humidity profiles from AIRS-Aqua, both sensors being part of the A-Train constellation are used for this study. It is shown that, as the degree of atmospheric stability increases, the fraction of aerosol vertical distribution trapped below inversion layer during winter also increases by up to 21%, 17% and 14% in the Atlantic, Barents/Kara Sea and Labrador Sea sectors respectively. The average aerosol optical depths also increase consistently with increased frequency. In contrast, aerosol layers are often accumulated above inversion layers during spring months additionally by up to 15% in the Eurasian and Pacific sectors as the stability increases. These contrasting sectorial differences can be explained by the nature of long-range pollution transport into the Arctic. For the first time, collocated satellite observations are used to quantify co-variability of aerosols and temperature inversions in the Arctic and to infer the role of long-range transport. The results have implications for the upcoming ESA EarthCARE mission.

Area C - Board 282

Since their successful launch on 16/02/2016 and 25/04/2018, Sentinel 3A and B satellites provide several products dedicated to land and maritime objectives, but also on aerosol characteristics. Both optical instruments, Ocean and Land Color Instrument (OLCI) and Sea and Land Surface Temperature Radiometer instrument (SLSTR) have their own land L2 products including land surface temperature, global vegetation index, terrestrial chlorophyll index or global integrated water vapor. In addition, by combining the radiometric and angular information from OLCI and SLSTR instruments, the SYNERGY Level 2 processing chain provides Surface Directional Reflectances for considered channels and Aerosol Optical thickness data on a 300m resolution.

To strengthen the Sentinel 3 catalog regarding land and aerosol products, three new processors have been developed and specified to be operationally deployed for systematic production in 2019: one dedicated to the detection and characterization of fires and two further ones dedicated to aerosol parameter retrievals. These advanced algorithms are not only focusing on land surfaces but also include aerosol retrieval over ocean by day, and detection of oceanic gas flares by day and night.

Global aerosol processors are based on a similar algorithm developed by P. North and A. Heckel from SWANSEA University. The differences between these two processors consist of the considered inputs files and timeliness: SLSTR L2 AOD products will be operated by EUMETSAT in NRT timeliness and derived from SLSTR L1 data whereas SYN L2 AOD products will be generated by ESA in NTC timeliness and derived from SYN L1 data. Both products will be provided at the resolution of 4.5 km but will include extended aerosol and surface datasets (including single scattering albedo or fine model fraction data for instance) compared to the 300m SYNERGY L2 SDR products.

The fire processor will use the medium infrared, thermal infrared and specifically dedicated fire channels acquired by the SLSTR instrument to detect land and oceanic hotspots and provide their Fire radiative power characteristics at 1 km resolution. All detected fires will be gathered in a single file providing fire characteristics but also background conditions. This algorithm has been defined by M. Wooster and W. Xu from KCL and developed to be operated in Near Real Time (by EUMETSAT) and Non-Time Critical (by ESA).

For each of these advanced processors, a development and validation status will be provided. Their product specificities and contents as long as their limitations and possible evolutions will be then described.

Area C - Board 283

The Sentinel-3A and -3B satellites carry the Sea and Land Surface Temperature Radiometer (SLSTR), the latest evolution of the (A)ATSR series of instruments carried onboard the ERS and ENVISAT series. SLSTR incorporates the functionality of AATSR alongside many new features, including two dedicated 'fire' channels (F1 and F2) specifically designed to provide unsaturated spectral radiance observations over high temperature targets in the midwave and longwave infrared spectral regions. The Sentinel-3 SLSTR Active Fire Product is now running in test mode and will soon become operational, generating active fire, gas flare and volcanic information from targets worldwide. The algorithm uses data from both the ‘F1’ fire channel and the SLSTR ‘standard’ channels spanning the visible to longwave infrared regions to first detect sub-pixel high temperature targets, and then to characterise their fire radiative power (FRP) output. Landscape fires are the main target, and the fire radiative power (FRP) measures produced by the algorithm will be used to estimate fire emissions of trace gases and aerosols, for example within the Copernicus Atmospheric Monitoring Service (CAMS). The Sentinel-3 overpass times in the morning and evening mean that this Sentinel-3 level 2 product generates data at a similar time to the very widely used MODIS Terra Active Fire and Thermal Anomaly products, thus being able to fill this gap when Terra finally reaches its end of life. Here we present details of the Sentinel-3 Active Fire Product algorithm, which is as significant advance on the basic prototype algorithm due to specific characteristics of the SLSTR spectral bands and data, and we illustrate some of the first outputs, indicating their scientific validity and degree of similarly and difference with the MODIS fire products.

Area C - Board 284

The Combined Inversion of Surface and Aerosol (CISAR) algorithm developed at Rayference allows the simultaneous retrieval of surface reflectance and aerosol properties. It embodies a Radiative Transfer Model (RTM) that accounts for the surface reflectance anisotropy and its coupling with aerosol scattering. The RTM explicitly solves the radiative transfer equation during the inversion process, without relying on pre-calculated integrals and allowing for a continuous variation of the state variables in the solution space. An Optimal Estimation (OE) approach is used, seeking the best balance between the observation and the a priori information. The latter is provided by any additional knowledge on the observed scene, and it can either concern the magnitude of the state variable or constraints on temporal and spectral variability.

Within the aerosol_cci Phase 2, the SEVIRI (Spinning Enhanced Visible and InfraRed Imager) hourly Aerosol Optical Depth (AOD) product has been generated using the CISAR algorithm. SEVIRI observations in the spectral bands VIS0.6, VIS0.8 and NIR1.6 are accumulated during several days to document the surface anisotropy and minimise the impact of clouds. While surface radiative properties are supposed constant during this accumulation period, aerosol properties are derived on an hourly basis. Hence, the AOD value at 0.55μm, the associated uncertainty and the fine mode fraction are retrieved every hour for clear-sky pixels over both land and sea surfaces. This product has been compared with the Copernicus Atmospheric Mxx Service (CAMS) aerosol product which has a 3-hour temporal resolution.

Additionally, this dataset has been combined with cloud data from the Royal Netherlands Meteorological Institute (cloud type, cloud droplet number density, cloud optical thickness and cloud droplet effective radius) generated in the framework of the Climate SAF to produce an MSG/SEVIRI hourly level-3 cloud/aerosol product. The potential of this combined level-3 product to provide information on the direct and indirect effects of aerosol on clouds is shown with practical examples.

Area C - Board 285

The physical and optical properties of aerosol particles determine their radiative effects, and information on these properties can be used to predict and assess global and regional changes in the Earth’s climate.
Detailed information on aerosol optical depth and types (e.g. size and absorption properties) is needed to understand the effects of aerosols on the Earth’s climate and the environment. In particular, the aerosol direct radiative forcing is sensitive to the aerosol radiation absorptivity. Therefore, it is important to understand the connections between the aerosol types and absorption properties because that information can be used for comparisons and validation of chemical transport models and satellites.
In this study, aerosol measurements from AERONET at several sites across Europe and Mediterranean are used over a period of several years to better characterize the climatology of the aerosol microphysical and optical properties and examine direct aerosol radiative forcing. As a result of legislations regarding atmospheric pollution, negative trends in aerosol amounts have been observed during the last years in many areas of Europe, the so-called “brightening effect”.
Due to the difference of the sources, aerosols exhibit noticeable differences in physical and optical properties with respect to the location and season. Fine Mode Fraction (FMF) refers to the fraction of Aerosol Optical Depth (AOD) due to fine-mode aerosol particles with sizes smaller than 1 µm. The Ångström exponent (AE) is the exponent for the power law describing the wavelength dependence of AOD. Using FMF and AE, the dominant size mode of aerosol can be determined. Moreover, the absorbing from non-absorbing aerosols, based on measurement of Single Scattering Albedo (SSA) can be distinguished.
We used the above-mentioned parameters to classify the fine-mode and coarse-mode particles from each site into eight groups (from highly absorbing fine-mode particles to weakly absorbing coarse-mode particles). The results are expected not only contribute to our understanding of regional climate and local air quality impacts, but they also will be useful for validation of satellite data and improving the performance of models and remote sensing algorithms in the future.

Area C - Board 286

Solar irradiance is considered as the key factor for assessing the radiative forcing of the climate system and the efficient planning of solar energy applications. The transfer of solar irradiance through the atmosphere depends on scattering and absorption by gases, aerosols and clouds. Time series of solar radiation components at high spatial and temporal resolution are necessary; however, the existence and the availability of high-quality assured long-time series of solar products is limited worldwide either in temporal and spatial basis due to the sparseness and the inhomogeneous distribution of the solar monitoring locations globally, especially in desert and sub-desert locations.
To overcome this potential issue, the use of satellite or modeled data for the derivation or the computation of SSI provides as the alternate. Even it sounds as the best and very optimistic solution, solar products or solar-related parameters from satellites include numerous sources of uncertainties (not accurate representation of solar-resource conditions at finer spatial and sub-pixel scale, misrepresentation at cloud overcast conditions etc.). For sound solar resource assessment, all satellite/modeled products must be extensively compared and validated against locally available ground-based measurements leading to the application of site-adaptation techniques. Site-adaptation methodologies aim to compare, validate and correct satellite-derived solar products against real pin-point measurements in order to reduce the systematic biases-uncertainties and to further create substantially improved satellite solar resource. Among other methods, the Least Squares Linear technique and the Empirical Quantile Mapping have been widely applied for the statistical bias adjustment of modeled data.
In this study, SSI measurements at high temporal resolution for nine (9) stations located in the Mediterranean region are used to validate the outcomes of Radiative Transfer – Satellite synergistic model. Firstly, modeled SSIs are generated using a combined- backward multi linear procedure based on Radiative Transfer Model outputs (LibRadTran, module SBDART) and atmospheric parameters from MODIS (v6.1). Goodness of fit statistics (ME, MAE, RMSE and Pearson’s R) are calculated to estimate and understand the various uncertainty sources between the ground truth and the satellite-derived SSI. Finally, the Empirical Quantile Mapping approach is evaluated. The bias correction algorithm leads to wide reduction of error metrics for all stations eliminating also the presence of systematic biases (ME=0). For Global Horizontal Irradiance (GHI), MAE and RMSE are reduced by >50% while the correlation coefficient slightly changes. On the other side, improvements in Direct Normal Irradiance (DNI) derived in all error metrics with lower reduction percentages compared to GHI results due to the more complicated relationship of DNI with aerosols and water vapor.

Area C - Board 287

Cirrus clouds play a key role in the Earth’s radiation budget and they represent an important variable in climate changes. Unfortunately these clouds are still uncertain and not completely understood due to the lack of measurements. The uncertain in the estimate of the radiative impact is mainly due to the high variability in the size/shape distribution of ice particles . The radiative effect produced by these clouds is strong in the atmospheric window between 820–960 cm−1 (8–12 μm) but the contribution in the far infrared (FIR) of the spectrum, below 667 cm−1 (> 15 μm), is important as well. In fact we know that more than 40% of the entire thermal flux is emitted in the FIR spectral range where the spectrum is very sensitive to the effective sizes as well as the ice crystals habit.
So far, most of the meaurements of the spectral radiance emitted by the Earth are limited to the mid infrared part of the spectrum, up to the CO2 band at 667 cm-1, hence only few measurements covering the FIR region are available. However the sounding of the FIR is very important to validate the current clouds models and improving the characterisation of the cirrus radiative effect.
A sinergy of different instruments is in progress at the Concordia Station, Dome-C (3233 m a. s.l.), over the Antarctic Plateau, since December 2011. The FIRCLOUDS (Far Infrared Radiative Clouds) project, selected by PNRA (Programma Nazionale per la Ricerca in Antartide), has the aim to achieve a radiative closure experiments through the comparison of the retrieved clouds properties and the measurements performed by means of the Fourier transform spectroradiometer REFIR-PAD (Radiation Explorer Far Infrared – Prototype for Applications and Development) operating in the FIR spectral ragion, the backscattering/depolarization lidar, an optical Ice-Camera and, starting from this year, a Micro Rain Radar at 24 GHz, which will be installed at the site, in order to measure the Ice Water Content (IWC) of the Antarctic clouds. Furthermore some other ancillary measurements, such as daily radiousoundings, are available for an indipendent characterisation of the atmospheric state.
This work introduces the retrieval techniques to invert remote sensing data of the ice clouds to obtain their optical and micro-physical properties, such as the crystals effective diameter, the ice water path, the optical depth and the ice fraction, in case of mixed phase clouds. The retrieval has been performed by using the Simultaneous Atmospheric and Clouds Retrieval (SACR) code, which is based on the optimal estimation approach. The algorithm allows the simultaneous retrieval of the atmospheric variables, such as the vertical profiles of water vapour and temperature, and the cirrus clouds properties by using the single scattering coefficients of different ice crystal habits tabulated in specific databases.
In this work the results of the analysis of the first 3 years of measurements will be shown together with the preliminary results coming from the last field campaign.

Area C - Board 288

EMI (Environmental Measuring Instrument) is an imaging spectrometer with a convex grating design, acquiring high accuracy Earth’s radiance and solar irradiance in a nadir push-broom mode. The objective of EMI is to monitor atmospheric composition: ozone (O3), nitrogen dioxide (NO2) and sulfur dioxide (SO2). EMI on board GF-5 satellite was launched on 9 May 2018. EMI has a wavelength range from 240 to 710nm with 4 channels and the spectral resolution is 0.3-0.5 nm. EMI cloud product is needed for cloud correction of O3, NO2 and other trace gas retrievals. In this study, we choose DOAS (differential optical absorption spectroscopy) fit of the O2-O2 absorption feature around 477nm (the fitting window wavelength range from 460 nm to 490 nm) to acquire cloud pressure and effect cloud fraction. Compared to the O2 A band at 760 nm, the O2-O2 fitting window is closer to the NO2 an UV fitting windows and has lower surface albedo over vegetated land. Solar spectrum measured by EMI was chosen as reference spectrum in the DOAS fit. The slant column of O2-O2 and the continuum reflectance at 477 nm are the retrieved parameters. Then we use look-up tables to acquire O2-O2 VCD (vertical column density) in which multiple Rayleigh scattering is taken into account. The parameters for this look-up table include solar zenith angle, viewing zenith angle, relative azimuth angle, surface albedo, cloud pressure, and cloud fraction. Since the vertical density profile of O2-O2 is known, namely quadratic with pressure, the column density of O2-O2 can be derived from the measured O2-O2 band, which leads to the cloud pressure.
In this presentation, we will describe the EMI O2-O2 cloud retrieval algorithm and compare the cloud pressures retrieved from EMI with TROPOMI (Tropospheric Monitoring Instrument) cloud products. Additional, we should account for the difference between them and analyze it.

Area C - Board 289

Knowledge of the spectral reflectivity of the Earth's surface is important for the interpretation of satellite measurements and for understanding the Earth's radiation budget. Having accurate surface reflectivity maps is essential for the retrieval of many trace gases, clouds, and aerosols.

In the framework of the AC SAF (Satellite Application Facility on Atmospheric Composition Monitoring) of EUMETSAT we therefore produce a global surface reflectivity climatology based on observations performed by the 2nd Global Ozone Monitoring Experiment (GOME-2) instruments onboard the MetOp-A, -B, and -C satellites.

The directional surface reflectivity product is the directionally dependent Lambertian-equivalent reflectivity of the Earth's surface (DLER). Because of its directional dependence, it provides surface BRDF for GOME-2 and for other instruments put in a polar orbit with an equator crossing time close to 09:30 LT. The DLER is the improved follow-up of earlier surface reflectivity databases based on GOME-1 (on the ERS-2 satellite), SCIAMACHY (on the Envisat satellite) and OMI (on the Aura satellite) which only provided non-directional Lambertian-equivalent reflectivity (LER).

The GOME-2 surface DLER is retrieved from the main science channels for 26 pre-defined 1-nm wide wavelength bands ranging from the UV to the near-infrared. The wavelength range that is covered this way is 328-772 nm and in this interval the DLER offers support for most retrieval methods and windows, including those that use the oxygen-A and oxygen-B band windows. The surface DLER is retrieved for land and sea surfaces, including snow and ice surfaces. The spatial resolution is 0.25 by 0.25 degrees.

We present the retrieval algorithm and explain the approach that is followed to handle the directionality. The DLER is compared to MODIS BRDF, and good agreement is found between the two albedo sources. Using the new DLER product as input for the FRESCO cloud retrieval we find clear improvements for both the retrieved cloud fractions and the cloud pressures.

This new directional surface reflectivity climatology is not only relevant for retrievals from the GOME-2 instruments. It can also be used for other instruments, such as UVNS on the future Sentinel-5 mission. The method that we used to generate the climatology can equally well be applied to other instruments, such as the TROPOMI instrument.

Area C - Board 290

We present a new, simple and fast method to retrieve cloud altitude from measurements of light absorption in the strongest molecular band of oxygen (the A-band), centred at 761 nm, by the POLarization and Directionality of the Earth’s Reflectances (POLDER) sensor. First, we assess the validity of the method against synthetic spectra as function of realistic cloud scenarios and satellite observation geometries. Second, the hurricane Ileana, overpassed the 30 August 2012 off the coasts west of Mexico and Southern California, is selected as test bed for the comparison of our retrievals against independent and nearly coincident cloud data inferred with the Moderate Resolution Imaging Spectroradiometer (MODIS) onboard Aqua and the Global Ozone Monitoring Experiment 2 (GOME-2) onboard MetOp-A. The average retrieval error amounts to less than 500 m. The absolute (relative) bias for low, mid and high-level clouds respectively amounts to −0.08 km (−3.36 %), −0.15 km (−2.94 %) and −0.27 km (−2.92 %). Additionally, we find that our approach can correctly reproduce the cloud height patterns, as long as clouds are not thin cirrus, and that the error dependence on other parameters, such as cloud optical thickness and geometrical extent, can be ignored in the first approximation. The reported accuracy is sufficient for applications including climate studies and ozone vertical column determination (above clouds). To this end, we stress that to achieve highest accuracy in total column ozone, retrieved from the UV, consistency of the cloud model implemented in the forward operator (either scattering or Lambertian) should be warranted across the spectrum. Our technique can also be used as a first guess approximation for more elaborated approaches for cloud top height retrievals, e.g. hyper-spectral measurements in the oxygen A-band, which in turn can serve as input for the retrievals of greenhouse gases from complementary spectral windows.

Area C - Board 291

Cloud optical properties are important climate variables for determining the Earth’s radiative energy balance. Particularly the cloud optical thickness (COD) and effective particle radius (Re) are the key parameters determining radiative properties of clouds such as reflection, transmission, and absorption of the solar radiation. Retrievals from geostationary and polar satellites, measuring the solar radiation reflected by clouds in the visible and near-infrared wavelengths are already provided by different sensors, however cloud optical properties can also be inferred complementarily from ground-based measurements of the transmitted solar radiation by a multispectral radiometers. In this work we present a new method developed within the SKYNET network (Khatri et al., 2017,2018) and applied to the PREDE-POM 02 sun-sky radiometers. The method consists in observing non-precipitating clouds through zenith measurements of radiance at 870, 1020 and 1640 nm, and retrieving the cloud optical properties through an iterative procedure that makes use of some input variables (columnar precipitable water vapour and surface reflectance) and the SBDART (Ricchiazzi et., 1998) forward model for reconstructing the measured radiance.
The new methodology will be applied to two sites of ESR/SKYNET network (www.euroskyrad.net), Lindenberg (Germany) and Valencia (Spain), and to Tohoku University site (Japan). A sensitivity study of the methodology to input parameters and wavelengths selection will be performed, and a comparison of the products against other available methodologies will be shown.

Khatri, P., T. Hayasaka, H. Iwabuchi, T. Takamura, H. Irie, T. Y. Nakajima, H. Letu, and Q. Kai, 2017: Cloud parameters from zenith transmittances measured by sky radiometer at the surface: Method development and satellite product validation. Proceedings of 19th EGU General Assembly, Vienna, Austria, EGU2017–6727.
Khatri, P., T. Hayasaka, H. Iwabuchi, T. Takamura, H. Irie, T. Y. Nakajima (2018), Validation of MODIS and AHI observed water cloud properties using surface radiation data, J. Meteor. Soc. Japan, 96B, pp. 151-172, doi:10.2151/jmsj.2018-036.
Ricchiazzi, Paul & Yang, Shiren, & Gautier, Catherine & Sowle, David. (1998) SBDART: A research and Teaching Software Tool for Plane-Parallel Radiative Transfer in the Earth’s Atmosphere. Bulletin of the American Meteorological Society. 79. 2101-2114.

Area C - Board 292

Clouds play a key role in the Earth’s radiation budget by reflecting shortwave radiation and reducing emissions of longwave radiation. A shift in a diurnal cycle of cloudiness may cause a change in the radiation balance, that potentially leads to significant feedbacks in the climate system. The CM SAF ClOud Fractional Cover (CFC) dataset from METeosat First and Second Generation - Edition 1 (COMET) provides high-temporal 30-minute cloud fraction estimates for 1991-2015. This allows to study spatial variability and trends in the diurnal cycle of cloudiness.

The study aims at evaluation of COMET dataset in terms of ability to reconstruct phase and amplitude of the CFC diurnal cycle. To this end, we derive CFC diurnal cycle from COMET 30-minute data and validate it using 111 SYNOP sites that provide 25-year seamless visual sky observations at a 3-hour step. Averaged over all sites, the results reveal the mean COMET bias in amplitude of 3% (0-100% CFC) varying from summer months (1%) to winter months (5%). Local time of the maximum CFC (i.e. phase of COMET CFC diurnal cycle) is on average approx. 15 minutes earlier than in SYNOP observations. Temporal stability of phase and amplitude has been assessed by means of Standard Normal Homogeneity Test (SNHT) that reveals there is no significant inhomogeneity in COMET dataset. The preliminary climate analysis of COMET dataset revealed a significant positive trend in cloud cycle phase exceeding 30 minutes per decade, as well as a negative trend in cloud diurnal cycle amplitude of 3% CFC per decade. Finally, we perform a grid-level inter-comparison of COMET-derived CFC diurnal cycles with other satellite data records (i.e. CM SAF CLAAS, ISCCP) as well as with ECMWF reanalyses (ERA-Interim and ERA-5). Our study builds a background for studying cloud diurnal cycle response to climate change, and cloud diurnal cycle feedback in the climate system.

Area C - Board 293

Polar-orbiting satellite sensors uniquely provide almost 4-decade-long global climate data records (CDR) of cloud cover. This requires combining data from consecutive satellite platforms, for instance NOAA and MetOp. One of the recently released CDRs generated in a frame of the Cloud_cci project consists of cloud physical properties derived from Advanced Very High Resolution Radiometer (1982-2014, Stengel et al., 2017). Generation of such long CDRs is challenging due to technical limitations related to: degradation of satellite sensors, different number of simultaneously operating satellites, radiometric and geometric distortions, as well as satellite orbital drift. As a result of the drift, the cycle of diurnal cloudiness is sampled at different local times, which in turn leads to ambiguous trends revealed during climatological analyses. The aim of the study is to quantify the uncertainty and ensuing spurious trends in level-3 AVHRR-based cloud climate data records caused by the satellite orbital drift.

The analysis is based on a high temporal 30-minute reference time series of cloud fractional cover (CFC) derived from the CM SAF Cloud Fractional Cover dataset from Meteosat First and Second Generation (COMET, Stöckli et al., 2018, Bojanowski et al., 2018). Initially the COMET dataset has been evaluated against synoptic observations in terms of ability to represent the CFC diurnal cycle. Further, the COMET time series has been sampled at times of NOAA and MetOp satellites overpasses. That formed the artificial CFC time series “as seen” by AVHRR sensors. Further, this time series has been aggregated to monthly means and compared with the reference COMET data. The comparison reveals the spatial and temporal variability of theoretical errors and spurious trends in AVHRR-based CFC time series related to different diurnal cloud regimes. Moreover, the study quantifies these errors separately for NOAA morning and afternoon satellites, and their ascending and descending nodes. Results of our study can serve as a background information for interpretation of global AVHRR cloud CDRs, as well as reveal a potential improvement of the level-3 CFC data accuracy after accounting for the orbital drift effect.

References:
Bojanowski, J.S., Stöckli, R., Duguay-Tetzlaff, A., Finkensieper, S., Hollmann, R., 2018, Performance Assessment of the COMET Cloud Fractional Cover Climatology across Meteosat Generations, Remote Sensing 2018 10(5) 804. doi:10.3390/rs10050804

Stengel, M., Stapelberg, S., Sus, O., Schlundt, C., Poulsen, C., Thomas, G., Christensen, M., Carbajal Henken, C., Preusker, R., Fischer, J., Devasthale, A., Willén, U., Karlsson, K.-G., McGarragh, G. R., Proud, S., Povey, A. C., Grainger, R. G., Meirink, J. F., Feofilov, A., Bennartz, R., Bojanowski, J. S., and Hollmann, R.: Cloud property datasets retrieved from AVHRR, MODIS, AATSR and MERIS in the framework of the Cloud_cci project, Earth Syst. Sci. Data, 9, 881-904, https://doi.org/10.5194/essd-9-881-2017, 2017.

Stöckli, R., Bojanowski J.S., John, V., Duguay-Tetzlaff, A., Bourgeois, Q., Schulz, J., and Hollmann, R.: Climatological Cloud Detection with Heritage Geostationary Satellite Sensors. J. Geophys. Res. 2018. (submitted)

Area C - Board 294