Cloud macrophysical properties: The role of subgrid-scale heterogeneity, and attributing radiation biases
Maike Ahlgrimm (ECMWF)
Abstract
Forecast models are becoming increasingly skilled at predicting clouds in the right place, at the right time. Yet errors in cloud occurrence and their macrophysical properties remain a primary source of error for the calculation of radiative fluxes.
One such aspect is the treatment of subgrid-scale condensate heterogeneity in global models that do not explicitly assume a distribution of cloud properties within each grid box. Observations show that heterogeneity varies with cloud regime. We develop a new regime-dependent parameterization for subgrid-scale heterogeneity which is then tested with two options for the assumed condensate distribution shape within the gridbox (log-normal and Gamma). Given a realistic cloud state, the new parameterization helps to address long-standing bias patterns in the IFS radiation.
We also show how the contribution to the typical shortwave radiation bias found in marine boundary layer clouds can be attributed to specific assumptions and cloud properties through a combination of observational analysis and offline radiation calculations.
By incrementally adjusting the model's cloud properties towards more realistic (observed) values in offline radiation experiments, the contribution to the radiation bias from errors in cloud cover, water path, effective radius, subgrid heterogeneity and vertical overlap can be quantified. Understanding the relative contribution of individual model components or parameterizations can then help to prioritise development efforts.
The quest for consistent ice optical properties across the spectrum
Anthony Baran (Met Office)
Abstract
With the advent of space-based sensors that sample cirrus and ice clouds from across the spectrum, it is important to construct ice crystal scattering models that are representative of the observed irradiance and radiance properties of these cloud types. In this talk, an example application of an ensemble model of cirrus ice crystals is presented that has been used to parametrise ice optical properties within the Met Office’s suite of models. The parametrisation of ice optical properties is based on a couple between the model prediction of the ice mass mixing ratio and temperature, where the particle size distribution (PSD) between the cirrus microphysics and radiation schemes is the same within the suite of models, thus ensuring consistency. This couple circumvents the need to parametrise ice optical properties in terms of an effective dimension, but rather couples the ice optical properties directly to the model prediction of ice mass and environmental temperature. Thus, the irradiance and radiance simulations within the model depend more directly on how well the model predicts the ice mass mixing ratios and temperature profiles rather than on inconsistencies between model assumptions in the cloud and radiation schemes. Examples shall be presented of various comparisons of how well the microphysics-radiation coupled parametrisation performs in the spaces of irradiance and radiance, both from model simulations obtained from the Met Office suite of model’s and from aircraft campaigns.
Moreover, examples shall be given as to why the concept of an effective dimension fails to replicate the radiative properties of cirrus across the spectrum of importance to the energetics of the Earth’s atmosphere. Furthermore, it shall also be argued that the assumed shape of the PSD is as important as the choice of ice crystal scattering models, both in terms of the small and large ice modes of the PSD.
Application of Gauss-Legendre Quadrature for Computation of Atmospheric Radiative Flux Profiles in Atmospheric Models
Howard W. Barker (Environment and Climate Change Canada)
Abstract
We introduce two methods for computing radiative fluxes in atmospheric models. Both employ Gauss-Legendre Quadrature (GLQ). The first applies to models with large horizontal grid-spacings ∆x (e.g., GCMs) and is referred to as GLQ Independent Column Approximation (GLQ-ICA). GLQ-ICA is contrasted against its progenitor McICA. GLQ-ICA uses N stochastic subgrid-scale cloudy columns, sorts them according to cloud water path, and performs full SW and IR spectral integrations on nG<<N sub-columns. The nG sub-columns are identified by rules governing GLQ. Resulting flux profiles are weighted and summed giving domain-averages. While significant bias errors occur for nG = 1, they are almost eliminated by nG = 2, with corresponding random errors, for net fluxes at surface and top-of-atmosphere, typically smaller than McICA’s; because SW and LW solvers operate on common sub-columns. GLQ-ICA’s random errors for heating rates (HR) are, however, comparable to McICA's, even for the expensive nG = 3. An amendment to the initial GLQ-ICA reduces HR noise, but introduces biases.
The second method pertains to models with small 
such as high-resolution CSRM or LES models which often have >105 columns. Typically, 1D RT solvers are applied to each column - the ICA. To reduce execution time, RT calculations are not done at every time-step. As RT time-step increases, however, the rest of the model’s physics can become out of step with radiative HRs. While cloud conditions can be averaged over sub-domains and 1D RT models applied, like those discussed above, sub-domain average HRs must be down-scaled to full-resolution. A method employing GLQ is described which provides what might be sufficiently accurate radiative HRs at the inner-scale at each model time-step. It is demonstrated diagnostically using CSRM and NWP domains consisting of >1.4 x 106 columns. It is expected that in a good many cases, errors associated with this method will be smaller than those due to neglect of 3D radiative transfer.
Line-by-line absorption of atmospheric gases and uncorrelated k-distribution models
Jürgen Fischer (Free University Berlin)
Abstract
Beside line-by-line radiative transfer modelling, appropriate methods have to be introduced to simulate the impact of the high spectrally variation of the absorption by atmospheric gases. k-distribution methods are found as efficient tools for large spectral band radiation transfer
simulations, but a weakness of mainly used correlated k-distribution methods is the treatment of layered atmospheres with variable gas absorption. We have calculated spectrally high resolved absorption coefficients of atmospheric gases with a line-by-line model and used these to estimate k-coefficients by an uncorrelated k-distribution method. The impact of the vertical profiles of different atmospheric gases, but also just on the impact of pressure broadening of the absorption lines on the derived pseudo k-absorption coefficients, is discussed. Beside simulations of spectral radiance measurements, examples of broadband simulations, performed by the radiative transfer model MOMO, which rely on uncorrelated k-distribution estimates, were compared to RRTMG simulations, whereby RRTMG-based TOA (top-of-atmosphere) SW fluxes were up to 10 W/m² lower (or ~2.5%, at SZA~26°) in contrast to MOMO, while TOA LW fluxes deviated less than 1.5 W/m² (or ~0.5%) from one another.
Evaluation of IFS surface radiation from the ground and satellite
Thomas Haiden (ECMWF)
Abstract
Verification of downward radiation at the surface is of interest both from the modelling perspective as well as for forecast users, in particular for renewable energy applications. Operational evaluation of downward solar radiation at ECMWF is currently based on three independent observational datasets: Climate Monitoring Satellite Application Facility (CM SAF) products, Baseline Surface Radiation Network (BSRN) observations, and SYNOP observations. Verification against CM SAF reveals regional differences and the scale-dependence of cloud forecast skill and allows monitoring of IFS skill evolution over large areas. For the evaluation of systematic errors, independent in-situ data is required since the CM SAF product depends on NWP and surface albedo input. Verification against BSRN shows that biases in downward shortwave and longwave radiation are of the order of 10 W/m2 (positive for shortwave, negative for longwave), and have generally decreased in recent years. Recently, SYNOP downward solar radiation has been used as well, and the correspondence between the three datasets is analysed. A wintertime negative bias of about 1 K in 2m temperature in the IFS forecast can at least partly be traced back to cloud effects, i.e. a negative bias in downward longwave radiation due to a slight underestimation of low cloud cover and/or optical depth.
Aerosols-cloud-radiation interactions
Jim Haywood (Met Office)
Abstract
Aerosols have a potentially large effect on climate, particularly through their interactions with clouds, but the magnitude of this effect is highly uncertain. Large volcanic eruptions produce sulfur dioxide, which in turn produces aerosols; these eruptions thus represent a natural experiment through which to quantify aerosol–cloud interactions. Malavelle et al (2017) show that the massive 2014–2015 fissure eruption in Holuhraun, Iceland, reduced the size of liquid cloud droplets—consistent with expectations—but had no discernible effect on other cloud properties. The reduction in droplet size led to cloud brightening and global-mean radiative forcing of around −0.2 watts per square metre for September to October 2014. Changes in cloud amount or cloud liquid water path, however, were undetectable, indicating that these indirect effects, and cloud systems in general, are well buffered against aerosol changes. This result will reduce uncertainties in future climate projections, because we are now able to reject results from climate models with an excessive liquid-water-path response. Further impacts on precipitation processes and the role of aerosols in weather forecasting will be discussed.
Malavelle, F., et al., Strong constraints on aerosol-cloud interactions from volcanic eruptions, Nature, 546, 485–491,doi:10.1038/nature22974, 2017.
Challenges for radiation in NWP models
Robin Hogan (ECMWF)
Abstract
Radiation is a fundamental process that drives atmospheric flows at all scales, and in addition to its importance in improving short-range surface temperature forecasts, it has a key role to play in pushing the boundaries of predictability at medium-range to seasonal timescales. In this talk I will discuss some of the challenges for radiation in NWP models, and how we are tackling them at ECMWF. I will first introduce the “ecRad” radiation scheme that became operational in July 2017, and whose improved efficiency and flexibility is underpinning these developments. We are working on the introduction of urban areas into the ECMWF model, and a key part is to represent radiative interactions within the urban canopy consistent with the atmosphere above. I will show how we can improve on the common but crude assumption in urban radiation models that the urban canopy can be treated by a single, infinitely long street in vacuum. I will then address the question of whether improving clear-sky radiative transfer has the potential to improve extended-range predictive skill, for example via the impact of aerosols on monsoon systems, and how a better stratosphere can improve monthly forecast skill. Finally I will discuss how we can improve the efficiency of radiation codes and choose the optimal trade-off between the spectral, spatial and temporal resolution of the radiation calls. A related issue is the cost of improvements to the accuracy to the solver: the introduction of longwave scattering increases the cost of ecRad by 32% overall, but I will show how we can represent longwave scattering by clouds, but not aerosols, with only a 3.3% cost increase.
Several Issues for Current Radiation Algorithms in Climate Models
Jiangnan Li (Canadian Center for Climate Modeling and Analysis)
Abstract
I will address a few problems of current radiation algorithm in weather and climate models
CKD cloud: A new scheme of water cloud optical properties is proposed for correlated k-distribution (CKD) models, in which the correlation in spectral distributions between the gaseous absorption coefficient and cloud optical properties is maintained. This is an extension of the CKD method from gas to cloud by dealing with the gas absorption coefficient and cloud optical properties in the same way. It is shown that the band-mean cloud optical property scheme can overestimate cloud solar heating rate, with a relative error over 30%. However, it is found that the error in the longwave caused by the band-mean cloud scheme is very small.
Gaseous overlap: How to handle gaseous overlap is a key issue in for any CKD model. In many spectral ranges, more than three different gases are important. I will discuss how to sort them in them efficiently in a CKD model.
Solar and infrared interaction: The solar spectrum extends into the infrared, with about 12 Wm-2 in the 4--1000 μm range. Usually, the solar spectrum comprises wavelengths shorter than 4 μm with all incoming solar energy deposited in that range. RRTM creates a special solar band over all the infrared range. Both methods can cause large errors. A simple method is proposed wherein the longwave radiative transfer equation with solar energy input is solved. In comparison with traditional methods, the new solution results in more solar energy absorbed in the atmosphere and less at the surface.
Modelling non-LTE effects
Manuel López-Puertas (IAA-CSIC)
Abstract
In this talk I will present a review of the major effects of non-LTE (non-local thermodynamic equilibrium) on the retrieval of temperature and species from satellite IR instruments, as well as on the radiative IR and near-IR heating/cooling rates in the middle/upper atmosphere. A brief introduction to non-LTE will be given followed by some examples (including pros and challenges) on their effects on retrievals and heating/cooling rates. Examples of the most recent Earth’s observations of temperature and key atmospheres species (O3, H2O, CO2, CO, NO2, NO), taken by the SABER/TIMED and MIPAS/Envisat instruments, and retrieved under non-LTE conditions, will also be presented. The second part of the talk will be devoted to non-LTE effects on cooling/heating (IR, NIR) rates. I will present a review of the major radiative heating and cooling rates in the middle and upper atmosphere, the region where non-LTE is most important. They will be illustrated with examples of the most recent observations taken by SABER and MIPAS. An overview of the current non-LTE parameterizations of those heating/cooling rates will also be given. Finally, the major problems, uncertainties and current lacks in their modelling in the context of future climate projections will be discussed.
Radiative transfer for extra-solar planets: bringing it down to Earth
James Manners (Met Office)
Abstract
An ongoing collaboration with the Exoplanets group at Exeter University has led to adaptations of the Met Office Unified Model to allow flexible simulation of general planet atmospheres. Here I will outline some of the latest developments to the radiation scheme that may also be of benefit to NWP and climate applications for Earth. This will touch on the flexible generation of gas optical properties, the ability to diagnose simulated observations from the GCM at high(ish) wavelength resolution, and in particular, the treatment of spherical geometry for the direct solar beam. Some initial results will show the effect of spherical geometry on the shortwave fluxes and heating rates for Earth. I will also discuss the possible future benefits of using spherical geometry on other aspects of the radiation parametrisation including the treatment of orography and cloud.
The single-band approach – a way to full cloud-radiation interaction (WITHDRAWN)
Ján Mašek (Czech Hydrometeorological Institute)
Abstract
Although heavily parameterized, radiative transfer schemes constitute a very expensive part of model physics. Their computational cost is usually further reduced by calling them intermittently or with reduced spatial resolution. However, such an approach deteriorates the cloud-radiation interaction for cloud fields strongly varying on the resolved spatio-temporal scales.
An alternative solution is offered, ensuring full cloud-radiation interaction at comparable cost. It solves the radiative transfer equation at every model grid-point and time-step, combining actual cloud optical properties with hourly updated gaseous transmissions. Memory storage required for carrying gaseous transmissions between model time-steps is made affordable by using single shortwave and single longwave spectral intervals. The accuracy of the single-band approach is significantly increased by parameterizing previously ignored spectrally unresolved phenomena. With clouds entering the scene, the heating rate error due to hourly calculated radiative transfer would be one order of magnitude larger than the error resulting from hourly update of gaseous transmissions.
Linear scalability of the longwave calculations with respect to the number of layers is achieved via decomposition of the net flux into cooling to space, exchange with the surface, and internal atmospheric exchanges. Expensive internal exchanges are approximated by a bracketing technique, interpolating the associated net flux between its cheap minimum and maximum estimates. A key finding is that interpolation weights determined in clear-sky conditions are applicable also in the presence of clouds. Therefore, they can be calculated using accurate and costly absorptivity-emissivity method applied 3-hourly, replacing a cheap but much less accurate statistical fit.
The developed radiation scheme reaches an optimal error balance between stand-alone radiative transfer calculations and selectively applied intermittency. It is primarily targeted for short range numerical weather prediction, which is only weakly sensitive to the accuracy limits reached by the single-band approach. At present, the scheme is used operationally in eight countries that use the ALADIN NWP system.
Infrared radiance modelling and assimilation
Marco Matricardi (ECMWF)
Abstract
The quality of the products retrieved from satellite infrared spectra hinges on the accuracy of the forward calculations carried out in the algorithms used in the retrieval processes. More specifically, the exploitation of infrared satellite radiance data requires the use of an accurate radiative transfer (RT) model to simulate radiances from an input atmospheric profile. There are two main types of RT models: computationally expensive line-by-line (LBL) models based on first principles and fast RT models. Fast RT models are generally based on LBL models and use efficient parameterisations that allow the simulation of radiances at a fraction of the cost required by a LBL model. Fast RT models are accurate and fast enough to fulfil the current NWP requirements of near real-time monitoring and assimilation of satellite radiance data. In our presentation, we will discuss the accuracy of LBL simulations and errors associated with the parameterisations used in fast RT models putting emphasis on the practical use of the RTTOV fast RT model in the ECMWF 4D-var assimilation system.
3D radiation in cloud resolving models
Bernhard Mayer (Ludwig-Maximilians-University (LMU), Munich)
Abstract
Atmospheric circulation is driven by solar and thermal radiation. Radiation is also a key driver for cloud development. Clouds, as soon as they exist, have a strong impact on radiation, not only by casting shadows on the ground, but also by absorption and emission at cloud edges, causing strong heating and cooling rates at cloud level. Absorption and emission of radiation has an effect on cloud microphysics in the first place. The heat is then dissipated to the surrounding air and in turn impacts cloud dynamics. Diabatic heat by absorption and emission of radiation is in the same order of magnitude as latent heat, when integrated over a cloud’s lifetime. In atmospheric models, radiation is generally treated in one-dimensional, plane-parallel approximation, neglecting heating and cooling at cloud sides and causing mis-placed shadows at the ground. In recent years we have developed fast three-dimensional radiation schemes which allow studying these effects interactively in cloud-resolving models. The “TenStream” solver and the “Neighboring Column Approximation” are accurate yet fast approximations for three-dimensional fluxes and heating rates. Using these solvers, a number of effects was shown. These effects include changes in local cloud dynamics as well as cloud field organisation. A pronounced result is the effect of cloud shadows at the ground; effects of cloud side heating and cooling have been clearly demonstrated as well. An open question is how to consider these effects in NWP models with much lower resolution.
Representing spectroscopy and the solar spectrum in atmospheric models
Eli Mlawer (Atmospheric and Environmental Research (AER))
Abstract
Accurate weather and climate simulations depend on high quality values for a variety of spectrally varying values, as well as their proper treatment in radiative transfer calculations. Examples of these dependences include spectroscopic parameters, solar irradiance, cloud optical properties, etc., and their spectral correlations with each other. This talk will illustrate issues related to the treatment of spectrally varying quantities in radiation codes. Particular areas of focus will be the far-IR and UV.
Balancing accuracy, efficiency, and flexibility in a radiation code for atmospheric models
Robert Pincus (CIRES University of Colorado)
Abstract
We describe the design of RTE+RRTMGP, a new radiation code for computing fluxes intended primarily for use in models of the atmosphere. RTE+RRTMGP is based on state-of-the-art spectroscopy but is, in all other respects, functionally similar to codes employed in global models for decades, using a k-distribution to describe the optical properties of atmospheric gases, the Monte Carlo Independent Column Approximation to treat the vertical and horizontal variability of clouds, and a combination of two-stream and adding methods to compute the radiative transfer. The code diverges from its predecessors, most notable RRTMG, in being designed from scratch to balance accuracy, efficiency, and flexibility defined quite broadly.
The code is strictly modular in that gas optics, condensate optics including sampling of sub-grid variability, radiation transport, and reduction (the summarizing of spectrally- and vertically-detailed calculations for use in the host model) are all independent. Data is provided externally allowing for the development of k-distributions optimized to the task at hand. The core is a set of small, high-efficiency computational kernels; the user-controlled inner dimension allows the code to operate on optimally-sized problems for the computing architecture at hand, while language-interoperable interfaces and supporting unit tests mean that the kernels can be re-implemented if desired. The kernels are wrapped in a Fortran 2003 class structure that organizes computation, minimizes data movement between classes, and provides user support e.g. in the combining of optical properties. The class structure also supports efficiency by allowing user customization of reduction.
Because it is intended for use in a wide range of atmospheric models RTE+RRTMGP is less focused on providing a comprehensive radiation package than on providing a flexible facility. We have begun efforts to implement the code in several such models; time will tell whether the attempts to balance efficiency, accuracy, and flexibility will be embraced by the community.
Surface-radiation interactions in Polar Regions: challenges and future perspectives
Roberta Pirazzini (Finnish Meteorological Institute)
Abstract
The representation of snow–radiation interaction in numerical weather prediction models is oversimplified, causing large errors particularly in the near-surface temperature during spring. Advancements in the simulation of the surface-radiation interaction can be accomplished through improvements of the parameterizations of snow albedo and of the effect of snow on the canopy albedo.
The explicit representation of the inherent scattering-absorption properties of the snow crystals and the prognostic equations of the snow microstructure evolution will allow the coupling between the penetration of solar radiation in the snow, the snowpack bulk properties (density, thermal profile), its microstructure characteristics (grain size and shape), and the surface albedo. At the same time, the simulation of the transfer of solar radiation in the snowpack will open the door for direct assimilation of the satellite reflectance data for constraining the forecast snowpack properties.
The multiple scattering between the canopy and snow on the forest floor contributes markedly to the diurnal variation of direct-beam albedo. The effect is largest at high latitudes, where the midday sun elevation is low and the canopy small and sparse. A multiple scattering scheme that depends on leaf area index and leaf single scattering albedo would improve the simulation of the albedo of canopy over snow-covered ground.
Why do we need improved vegetation-radiation interactions in earth system models?
Tristan Quaife (University of Reading)
Abstract
It is 30 years since the publication of Piers Sellers' seminal paper describing a two-stream radiation scheme for vegetation canopies. Owing to its computational efficiency and full treatment of the scattering problem it remains the de facto standard for calculating absorbed and reflected radiation in many of the big land surface models.
An assumption in its derivation is that scattering elements (i.e. leaves) are infinitesimally small and randomly distributed in space in a semi-infinite plane-parallel medium. In effect this does not allow for the impact of structure i.e. the clumping of vegetation into discrete tree crowns, on the radiative transfer problem. The Earth, according to these assumptions, looks like it is covered in a uniform layer of green gas.
This talk explores the possibility of adding more complex treatments of 3D vegetation structure without significantly increasing the computational burden on the model. As well as improving estimates of albedo and radiation absorbed in the canopy this should also allow for better integration of remote sensing products into the models. We discuss potential modifications to the Seller’s model and the application of the SPATACUS 3D cloud model to vegetation.
Radiation and orography in weather models
Laura Rontu (Finnish Meteorological Institute)
Abstract
Radiation fluxes are modified by the underlying topography in different spatial and temporal scales. Slopes get different amount of direct solar radiation, neighbouring hills create shadows and restricted sky view influences the longwave and diffuse shortwave radiation. Modified radiation fluxes influence the surface energy balance directly and may change the cloud-radiation interactions in the models.
Integrated orographic radiation parametrizations are applied in all main European mesoscale NWP model systems. For AROME-SURFEX framework, these parametrizations are being developed using the “ororad” approach, which has been applied in HIRLAM NWP model since 2010. Alternatively, it is possible to take the orographic radiation effects into account when postprocessing the NWP grid-scale results for downstream applications like the road weather models.
Handling the orographic effects is basically a classical trigonometric exercise. For the NWP models, the usage and treatment of the subgrid-scale orography features is the main challenge. Fine-resolution surface elevation data up to a few metre horizontal resolution are available – how to use them optimally in the scales of weather? Application of a few underlying principles will be discussed, based on the HIRLAM – ALADIN experience:
Average the fluxes, not the orography.
Mind the physics related to the different spatial scales.
KISS: Keep It Simple, Stupid
Validation of the effects due to the orographic radiation parametrizations is challenging. Over the complex orography observations are sparse and may represent an environment different from that of the forecast grid-scale variables. When introducing new parametrizations into a NWP model, we create new interactions , dependencies and sources of uncertainty in the model system. Simple sensitivity experiments, using single-column approach may give valuable information about these interactions. Such experiments are especially applicable in the radiation studies, where the diurnal cycle of solar radiation and topography properties represent well defined and dominating forcing.
A fast radiative transfer method for the generation of synthetic visible satellite images
Leonhard Scheck (Ludwig-Maximilians-University (LMU), Munich)
Abstract
High-resolution satellites images provide a wealth of information about atmospheric conditions and are therefore seen as an important type of observation for data assimilation (DA) and model evaluation. In particular the visible and near-infrared channels provide information on the distribution of clouds and aerosols, microphysical properties and cloud structure with high temporal and spatial resolution. However, in operational DA systems currently only thermal infrared and microwave radiance observations are used, which provide mainly temperature and humidity and only limited cloud information. Also for model evaluation mainly infrared radiances have been used. The main reason for this is that sufficiently fast and accurate forward operators for solar channels were up to now not available, which is related to the fact that multiple scattering makes radiative transfer at solar wavelengths complicated and computationally expensive. To address this problem, we developed MFASIS, a 1D radiative transfer (RT) method based on compressed loop-up tables, which is orders of magnitude faster than conventional radiative transfer solvers for the visible spectrum (like the discrete ordinate method) and similarly accurate. Here we discuss advantages and limitations of the new method and present fast parameterizations for 3D RT effects that increase the accuracy and consistency of the operator without strongly increasing the computational effort. In particular, the impact of inclined cloud tops and the effect of the slant viewing angle on the subgrid cloud overlap are taken into account. Examples for using MFASIS in convective-scale data assimilation and for high-resolution model evaluation are presented and we discuss the possibility to apply the method also for aerosols.
Evaluation of Radiation Schemes using Aircraft Observations
Manfred Wendisch (University of Leipzig)
Abstract
Solar and terrestrial irradiance data collected by airborne measurements during two observational campaigns are compared with the output of the IFS radiation model (ECRAD). The measurements were performed using broadband and spectral sensors installed on the High Altitude and Long Range Research Aircraft (HALO). The data stem from the North Atlantic Waveguide and Downstream Impact Experiment (NAWDEX) and the Arctic CLoud Observations Using airborne measurements during polar the Day (ACLOUD) campaign.
After a short introduction and an overview of the campaigns and measurements, the general simulation set-up will be introduced. This includes input parameter extraction, preparation for the radiative simulations, and adjustment of the measurements to the simulation with respect to temporal and spatial differences. A short test of the representativeness of the radiation measurements on the 2D radiative field is shown by using true and artificial flight tracks.
Simulated and measured broadband and spectral values of upward and downward irradiances and derived quantities, such as cloud top albedo and cloud radiative forcing, will be presented. Both, time series and probability density functions of these measured and calculated quantities are compared for different cloud conditions (e.g., liquid / ice water cloud). Further on, the spectral resolution of the measurements will be used to find potential spectral deficiencies in the radiative parameterization of the model.