The Clouds, Aerosol and Precipitation from Multiple Instruments using a Variational Technique (CAPTIVATE) algorithm has been developed for synergistic retrievals from EarthCARE’s cloud profiling radar, atmospheric lidar and multispectral imager. The properties of ice clouds and snow, liquid clouds, rain, and multiple aerosol species, and their associated uncertainties are retrieved simultaneously within a robust optimal estimation framework that is fully configurable for spaceborne, airborne and surface-based platforms.
The benefit of a unified retrieval that assimilates all available active, passive and integrated measurements is the capacity to represent the combined effect of multiple constituents within complex and layered cloud and precipitation scenes, while making consistent microphysical and radiative assumptions throughout the atmospheric profile. For example, radar path-integrated attenuation and solar radiances simultaneously constrain retrievals of rain and liquid cloud, and the precipitation mass flux can be required to be conserved across the melting layer to ensure physical consistency between snow and rain rates.

In preparation for EarthCARE we have applied CAPTIVATE to analogous instruments from the A-Train constellation of satellites. We present the results of a novel unified CloudSat-CALIPSO-MODIS retrieval of cloud and precipitation, and show that the retrievals perform well when compared against established synergistic and single-instrument retrievals of specific cloud or precipitation properties. To assess the radiative closure, a radiative transfer model is used to model longwave and shortwave fluxes along the A-Train track for comparison with top-of-atmosphere broadband fluxes from CERES. Finally, we select key cloud regimes over which to evaluate the sensitivity of CAPTIVATE retrievals and radiative closure to the representation of cloud and precipitation microphysics and optical properties.

Downwelling spectra in the 100-1600 cm-1 interval, measured by the Radiation Explorer in the Far Infrared-Prototype for Applications and Development (REFIR-PAD) spectroradiometer on the Antarctic plateau since year 2012, are ingested by an automatic machine learning algorithm named CIC, in order to detect and classify Antarctic clouds.
The CIC algorithm is a modified version of the one chained in the End-2-End Simulator (EES) of the Far-infrared Outgoing Radiation Understanding and Monitoring (FORUM) mission. The FORUM EES is a tool used to predict and assess the Level-1 and Level-2 performances of the next ESA 9th Earth Explorer. CIC sits at the centre of the decision tree of the FORUM EES for retrieving clear-sky atmospheric products or cloud properties.
Co-located lidar measurements are exploited to define training sets composed of radiance spectra in presence of clear sky, ice clouds or mixed-phase clouds. The CIC is initially tested on a controlled verification subset for optimization. It is demonstrated that the information content in the far infrared (FIR) part of the spectrum is critical for the improvement of the performances of the algorithm to identify thin clouds and for cloud phase classification. A test set of 1726 spectra is then used to estimate the classification error. Unprecedented cloud occurrence statistics concerning 4 years of data are provided for multiple time scales and related to meteorological parameters such as surface air temperature and wind direction.
The results indicate a clear sky mean annual occurrence of 72.3%, while ice and mixed-phase clouds are observed in 24.9% and 2.7% respectively, with an inter-annual variability of a few percent. The seasonal occurrence of clear sky shows a minimum in winter (66.8%) and maxima (75-76%) during intermediate seasons. In winter the mean surface temperature is about 9°C colder in clear conditions than when ice clouds are present. Mixed-phase clouds are observed only in the warm season (November-March). In summer they amount to more than one third of total observed clouds. Their occurrence is correlated with warmer surface temperatures. In the austral summer, the mean surface air temperature is about 5°C warmer when clouds are present than in clear sky conditions.
A comparison of the CIC classification with available satellite Level-2 (L2) and Level-3 (L3) products, is provided. Passive (Infrared Atmospheric Sounding Interferometer - IASI, and Moderate Resolution Imaging Spectroradiometer - MODIS), and active (Cloud-Aerosol LiDAR with Orthogonal Polarization - CALIOP, and the Cloud Profiling Radar - CPR) sensors are considered.
For selected case studies, a direct comparison between co-located L2 satellite products and parameters retrieved from ground-based observations (REFIR-PAD, Lidar, MMR) is performed. In case of cloudy scenes, retrieved cloud parameter are used to simulate FORUM like observations and to evaluate the impact of the additional FIR part of spectrum in satellite cloud retrievals from infrared passive measurements.
For L3 satellite products, the monthly gridded data are used for comparison. The differences observed among the considered products and the CIC results are analysed in terms of footprint sizes and sensors' sensitivities. The comparison highlights the ability of the CIC/REFIR-PAD to identify multiple cloud conditions from high spectral resolution radiances.

Keywords: Cloud occurrence, Antarctic plateau, FORUM mission, cloud retrievals

The contribution of Arctic mixed-phase clouds to the Arctic Amplification is still not clear as there are major deficits in their representation in regional and climate models. One of the main interests of the Transregional Collaborative Research Center (TR 172) "Arctic Amplification: Climate Relevant Atmospheric and Surface Processes, and Feedback Mechanisms (AC)3" is to increase the understanding of these clouds by observation and modeling activities. Within the framework of (AC)3, the polar research aircraft of the Alfred-Wegener-Institute for Polar and Marine Research, Germany, were operated from Longyearbyen at Spitsbergen equipped with remote sensing and in situ probes to investigate Arctic boundary layer mixed-phase clouds within three airborne field campaigns. Key instruments of the remote sensing setup are the MiRAC 94 GHz FMCW radar along with passive microwave observations between 22 and 340 GHz and the AMALi lidar. In total, 192 flight hours have been collected with the remote sensing instruments on board.

For each of the campaigns, various atmospheric models on different resolutions in space and time have been employed. A special focus has been put on the Large Eddy Simulation (LES) version of the ICON model which allows a resolution O(50 m) while using heterogeneous surface and lateral boundary conditions to capture the observed synoptic situations. The analyzed simulations are using a 2-moment microphysics parameterization.

Within this presentation we will juxtapose Arctic mixed-phase clouds as seen in measurements taken during the different airborne campaigns to their representation within the ICON-LEM model. This will be done in physical variable space as well as in observational space by applying the PAMTRA forward simulator for passive and active observations to ICON-LEM model fields. The setup also allow to i) improve the representation of clouds by modifying the cloud scheme and comparing to real measurements and ii) to explore the benefit of future (virtual) instrumentation, e.g. G-Band radar.

The variation in climate sensitivity estimated by leading climate models is substantial. For a large part this is attributable to our lack of understanding of cloud feedback processes and aerosol-cloud interactions and our inability to represent such processes in models. Improved global satellite observations of clouds are crucial to further study cloud processes and to better constrain models. Fundamental liquid cloud properties include the liquid water path and droplet number concentration, both of which can be derived from cloud top effective radius and cloud optical thickness under reasonable assumptions of the cloud vertical structure. While uncertainties in cloud top effective radius and cloud optical thickness both equally affect uncertainties in derived liquid water path, uncertainties in derived droplet number concentrations are mostly driven by the uncertainties in retrieved effective radius. A common method to infer cloud top effective radius and cloud optical thickness uses combinations of visible and shortwave infrared reflectances (hereafter referred to as the bi-spectral approach). Application of this approach to long term data records of, e.g., MODIS has provided invaluable information on cloud microphysics. However, it is known that the bi-spectral approach is sensitive to the neglect of, e.g., 3D radiative transfer, sub-pixel cloudiness and vertical inhomogeneity, which may lead to substantial biases, especially in situations of small, broken, thin, precipitating or mixed-phase clouds. Multi-angular polarimetry offers an alternative approach to infer cloud top droplet size distributions. Making use of the relative structures of the primary and supernumerary cloudbow feature observed in polarized reflectances, this approach is relatively insensitive to the aforementioned issues hampering the bi-spectral approach. Here, we evaluate both the bi-spectral and polarimetric approaches to infer cloud top effective radius, liquid water path and droplet number concentrations. The study uses multi-angle total and polarized reflectances observed by the Research Scanning Polarimeter (RSP) during NASA’s CAMP2Ex airborne campaign. CAMP2Ex was conducted in the maritime continent surrounding the Philippines, which is one of the most challenging regions for cloud remote sensing on the globe. The unique observations of RSP at 9 spectral bands and up to 152 viewing angles allow a pixel-level application of polarimetric retrievals, in addition to bi-spectral retrievals at multiple wavelengths and viewing angles. We show that cloud retrievals based on polarimetry agree well with in situ observations and theoretical cloud physics, while bi-spectral retrievals lead to substantial biases. Using the retrievals in combination with in situ observation, we explore possible reasons for these biases. Furthermore, we show that the full droplet size distribution can be retrieved using polarimetry and demonstrate that this allows detection of precipitation onset in the observed clouds. In addition to RSP observations, simulated measurements based on realistic cloud model simulations are used to evaluate bi-spectral and polarimetric retrievals of effective radius, number concentrations and liquid water path. We also discuss the application of the polarimetric approach applied to spaceborne polarimeters and its limitations, such as those related to the instrument’s angular resolution and scattering angle sampling. We show that upcoming space-borne polarimeters, such as HARP-2 on the NASA’s PACE satellite (NASA) and the MAP on CO2M (ESA), allow application of the approach on a pixel-level, which is unprecedented. These provide the exciting prospect of global observations of highly accurate cloud top effective radius, number concentrations and liquid water paths and their usage in studies on cloud physics and aerosol-cloud interactions.

Breakthroughs in radar technology (Schottky diode-based frequency multiplied sources and solid-state transmitter) have recently fostered the construction of two G-band (frequency between 110 and 300 GHz) ground-based systems: the NASA-JPL VIPR system and the UKSA/CEOI funded GRaCE (G-band Radar for Cloud Experiments). Measurements taken from the two systems in two field campaigns (at the NY StonyBrook observatory and at the UK Chilbolton observatory) in synergy with other remote sensing instrument (cloud radars and lidars) confirm the potential of these high frequency systems in characterizing ice and light rain microphysics. In this paper we will present the first measurements of reflectivity and Doppler spectra ever measured at G-band and we will discuss how such measurements can be used when collected in synergy with more conventional cloud radars. G-band signals in fact experience non-Rayleigh scattering in regions where Ka- and W-band signal do not, thus demonstrating the potential of G-band radars for sizing sub-millimeter ice crystals and droplets. In rain G-band Doppler spectra show significant reductions in the spectral power return as compared to theoretical Rayleigh scattering in regions of high Doppler velocities (i.e. large raindrops) with the presence of multiple peaks and Mie notches in correspondence to the maxima and minima of the raindrop backscattering cross sections. The first two G-band Mie troughs correspond to smaller velocities/sizes than the first W-band Mie notch, as expected from the reduced size parameter. These features offered by G-band radars pave the way towards applying, in rain, Mie notch vertical wind retrievals and multi-frequency drop size distribution microphysical retrievals to smaller rain rates and smaller characteristic sizes than ever before.
Finally, the presentation will discuss the possibility of the deployment of a G-band radar in a space mission in the framework of Earth weather and climate monitoring programs. Simulations of radar observables will be provided for a high latitude scene will be used to draw a first iteration of instrument requirements.