Lazaros Oreopoulos

THE I3RC: Bringing Together the Most Advanced Radiative Transfer Tools for Cloudy Atmospheres

The interaction of clouds with solar and terrestrial radiation is one of the most important topics of climate research. In recent years it has been recognized that only a full three-dimensional (3D) treatment of this interaction can provide answers to many climate and remote sensing problems, leading to the worldwide development of numerous 3D radiative transfer (RT) codes. The international Intercomparison of 3D Radiation Codes (I3RC), described in this paper, sprung from the natural need to compare the performance of these 3D RT codes used in a variety of current scientific work in the atmospheric sciences. I3RC supports intercomparison and development of both exact and approximate 3D methods in its effort to 1) understand and document the errors/limits of 3D algorithms and their sources; 2) provide “baseline” cases for future code development for 3D radiation; 3) promote sharing and production of 3D radiative tools; 4) derive guidelines for 3D radiative tool selection; and 5) improve atmospheric science education in 3D RT. Results from the two completed phases of I3RC have been presented in two workshops and are expected to guide improvements in both remote sensing and radiative energy budget calculations in cloudy atmospheres.

Overlap properties of clouds generated by a cloud‐resolving model

[1]xa0The overlap properties of ∼850 snapshots of convective cloud fields generated by a cloud-resolving model are studied and compared with previously published results based on cloud radar observations. Total cloud cover is overestimated by the random overlap assumption but underestimated by the maximum overlap assumption and two standard implementations of the combined maximum/random overlap assumption. When the overlap of two layers is examined as a function of vertical separation distance, the value of the parameter α measuring the relative weight of maximum (α = 1) and random (α = 0) overlap decreases in such a way that only layers less than 1 km apart can be considered maximally overlapped, while layers more than 5 km apart are essentially randomly overlapped. The decrease of α with separation distance Δz is best expressed by a power law, which may not, however, be suitable for parameterization purposes. The more physically appropriate exponential function has slightly smaller goodness of fit overall but still gives very good fits for Δz between 0 and 5 km, which is the range of separation distances that would be of most importance in any overlap parameterization for radiative transfer purposes.

Cloud three‐dimensional effects evidenced in Landsat spatial power spectra and autocorrelation functions

An analysis of nadir reflectivity spatial Fourier power spectra and autocorrelation functions for solar wavelengths and cloudy conditions is presented. The data come from Landsat thematic mapper (TM) observations, while Monte Carlo (MC) simulations are used to aid the interpretation of the observations and to examine sensitivity to various factors. We show that shortwave radiative processes produce consistent signatures in power spectra and autocorrelation functions. Power spectra take a variety of shapes not shown or explained in previous observational studies. We demonstrate that TM spectra can potentially be affected by radiative “roughening” at intermediate scales (∼1–5 km) and radiative “smoothing” at small scales (<1 km). These processes are wavelength-dependent, with systematic differences between conservative (for cloud droplets) TM band 4 (∼0.8 μm) and absorbing band 7 (∼2.2 μm). Band 7 exhibits more roughening and less smoothing than band 4 and faster decrease in autocorrelation. Roughening is more prevalent at large solar zenith angles due to optical and/or geometrical side illumination and shadowing. MC spectra illustrate that scale invariant optical depth fields can produce complex power spectra that take a variety of shapes under different conditions. Radiative roughening increases with decreasing single scattering albedo and increasing solar zenith angle (as in the observations). For low solar zenith angles, there is a clear shift in the radiative smoothing scale to smaller values as droplet absorption increases. Power spectra also show stronger decorrelations between optical depth and reflectivity when cloud top variations are more pronounced. Finally, it is shown that power spectral analysis is a useful tool for evaluating the skill of novel optical depth retrieval techniques in removing three-dimensional radiative effects. New techniques using inverse nonlocal independent pixel approximation and normalized difference of nadir reflectivity yield optical depth fields which better match the scale-by-scale variability of the true optical depth field.

Radiative susceptibility of cloudy atmospheres to droplet number perturbations: 2. Global analysis from MODIS

[1]xa0Global distributions of albedo susceptibility for areas covered by liquid clouds are presented for 4 months in 2005. The susceptibility estimates are based on expanded definitions presented in a companion paper and include relative cloud droplet number concentration (CDNC) changes, perturbations in cloud droplet asymmetry parameter and single-scattering albedo, atmospheric/surface effects, and incorporation of the full solar spectrum. The cloud properties (optical thickness and effective radius) used as input in the susceptibility calculations come from MODIS Terra and Aqua Collection 5 gridded data. Geographical distributions of susceptibility corresponding to absolute (“absolute cloud susceptibility”) and relative (“relative cloud susceptibility”) CDNC changes are markedly different indicating that the detailed nature of the cloud microphysical perturbation is important for determining the radiative forcing associated with the first indirect aerosol effect. However, both types of susceptibility exhibit common characteristics such as significant reductions when perturbations in single-scattering properties are omitted, significant increases when atmospheric absorption and surface albedo effects are ignored, and the tendency to decrease with latitude, to be higher over ocean than over land, and to be statistically similar between the morning and afternoon MODIS overpasses. The satellite-based susceptibility analysis helps elucidate the role of present-day cloud and land surface properties in indirect aerosol forcing responses. Our realistic yet moderate CDNC perturbations yield forcings on the order of 1–2 W m−2 for cloud optical property distributions and land surface spectral albedos observed by MODIS. Since susceptibilities can potentially be computed from model fields, these results have practical application in assessing the reasonableness of model-generated estimates of the aerosol indirect radiative forcing.

New directions in the radiative transfer of cloudy atmospheres

Atmospheric radiative transfer plays a central role in understanding global climate change and anthropogenic climate forcing, and in the remote sensing of surface and atmospheric properties. Because of their opacity and highly scattering nature, clouds (covering more than half the planet at any time) pose unique challenges in atmospheric radiative transfer calculations. n nSome widely-used assumptions regarding clouds—such as having a flat top and base, horizontal uniformity, and infinite extent—are amenable to simple one-dimensional (1-D) radiative transfer and are therefore attractive from a computational point of view. However, these assumptions are completely unrealistic and yield errors. The ever-increasing need to realistically simulate cloud radiative processes in remote sensing and energy budget applications has contributed to the recent rapid growth of the three-dimensional (3-D) radiative transfer (RT) community [e.g., Marshak and Davis, 2005].

Radiative susceptibility of cloudy atmospheres to droplet number perturbations: 1. Theoretical analysis and examples from MODIS

[1] Theoretical and satellite-based assessments of the sensitivity of broadband shortwave radiative fluxes in cloudy atmospheres to small perturbations in the cloud droplet number concentration (N) of liquid water clouds under constant water conditions are performed. Two approaches to study this sensitivity are adopted: absolute increases in N, for which the radiative response is referred to as absolute cloud susceptibility, and relative increases in N or relative cloud susceptibility. Estimating the former is more challenging as it requires an assumed value for either cloud liquid water content or geometrical thickness; both susceptibilities require an assumed relationship between the droplet volume and effective radius. Expanding upon previous susceptibility studies, present radiative calculations include the effect of AN perturbations on droplet asymmetry parameter and single-scattering albedo, in addition to extinction. Absolute cloud susceptibility has a strong nonlinear dependence on the droplet effective radius as expected, while relative cloud susceptibility is primarily dependent on optical thickness. Molecular absorption and reflecting surfaces both reduce the relative contribution of the cloud to the top-of-atmosphere (TOA) flux and therefore also reduce the TOA albedo susceptibility. Transmittance susceptibilities are negative with absolute values similar to albedo susceptibility, while atmospheric absorptance susceptibilities are about an order of magnitude smaller than albedo susceptibilities and can be either positive or negative. Observation-based susceptibility calculations are derived from MODIS pixel-level retrievals of liquid water cloud optical thickness, effective radius, and cloud top temperature; two data granule examples are shown. Susceptibility quantifies the aerosol indirect effect sensitivity in a way that can be easily computed from model fields. As such, susceptibilities derived from MODIS observations provide a higher-order test of model cloud properties used for indirect effect studies. MODIS-derived global distributions of cloud susceptibility and radiative forcing calculations are presented in a companion paper.

Horizontal radiative fluxes in clouds and accuracy of the independent pixel approximation at absorbing wavelengths

In order to correctly interpret shortwave cloud radiation measured by satellites and ground-based radiometers, or by two aircraft flying above and below clouds, we need to better understand interactions between inhomogeneous clouds and solar radiation. The discrepancies between shortwave absorption inferred from measurements and predicted by models, between cloud optical depths estimated from satellites and ground measurements, between single scattering albedo retrieved from in situ radiation measurements and computed from measured droplet size distribution, among others, are strongly affected by cloud horizontal inhomogeneity. Net horizontal photon transport (i. e., horizontal fluxes) are a direct consequence of the inhomoqeneity in cloud structure. Horizontal fluxes and their effect on the accuracy of the pixel-by-pixel one-dimensional (1 D) radiative transfer calculations has recently undergone close scrutiny for conservative scattering. However, the properties and magnitude of horizontal fluxes in absorbing wavelengths are still poorly understood. As far as we are aware, only Ackerman and Cox and Titov discussed correlations between horizontal fluxes at absorbing wavelengths, though these were far from comprehensive. This paper partly fills this gap. We discuss here of whether the accuracy of the Independent Pixel Approximation (IPA), a 1 D radiative transfer approximation for each pixel, is a better model for multiple scattering at conservative or at absorbing wavelengths. Issues addressed here are: (1) dependence of net horizontal fluxes on single scattering albedo; (2) connection between pixel-by-pixel accuracy of the IPA and horizontal fluxes and (3) radiative smoothing and horizontal fluxes at absorbing wavelengths. In contrast to the traditional understanding of IPA, we study IPA accuracies not only for reflectance but also for transmittance and absorptance at both conservative and absorbing wavelengths. In spite of the apparent similarity between the three processes, dependence of IPA accuracies on single-scattering albedo is completely different. As a result, cloud optical properties retrieved from high resolution satellite images and ground-based measurements using IPA at absorbing channels will have different accuracies.

Performance of Goddard earth observing system GCM column radiation models under heterogeneous cloud conditions

We test the performance of the shortwave (SW) and longwave (LW) Column Radiation Models (CORAMs) of Chou and collaborators with heterogeneous cloud fields from a single-day global dataset produced by NCARs Community Atmospheric Model (CAM) with a 2-D Cloud Resolving Model (CRM) installed in each column. The original SW version of the CORAM performs quite well compared to reference Independent Column Approximation (ICA) calculations for boundary fluxes (global error - 4 W m -2 for reflected flux), largely due to the success of a combined overlap and cloud scaling parameterization scheme. The absolute magnitude of errors relative to ICA are even smaller (global error ∼ 2 W m -2 for outgoing flux) for the LW CORAM which applies similar overlap. The vertical distribution of heating and cooling within the atmosphere is also simulated quite well with daily averaged zonal errors always less than 0.3 K/day for SW and 0.6 K/day for LW heating (cooling) rates. The SW CORAMs performance improves by introducing a scheme that accounts for cloud inhomogeneity based on the Gamma Weighted Two Stream Approximation (GWTSA). These results suggest that previous studies demonstrating the inaccuracy of plane-parallel models may have unfairly focused on worst case scenarios, and that current radiative transfer algorithms in General Circulation Models (GCMs) may be more capable than previously thought in estimating realistic spatial and temporal averages of radiative fluxes, as long as they are provided with correct mean cloud profiles. However, even if the errors of our particular CORAMs are small, they seem to be systematic, and their impact can be fully assessed only with GCM climate simulations.

The Shortwave Radiative Forcing Bias of Homogeneous Liquid and Ice Clouds Observed by MODIS

We analyze the plane‐parallel bias of the shortwave cloud radiative forcing (SWCRF) of liquid and ice clouds at 1 degree scales using global MODIS (Terra and Aqua) cloud optical property retrievals for four months of 2005 representative of the meteorological seasons. The (negative) bias is estimated as the difference of the SWCRF calculated using the Plane‐Parallel Homogeneous (PPH) method and the Independent Column Approximation (ICA). These calculations require MODIS‐derived means (for PPH calculations) and distributions (for ICA calculations) of cloud optical thickness and effective radius as well as ancillary surface albedo and atmospheric information, that are inserted into a broadband solar radiative transfer code. The absolute value of global SWCRF bias of liquid clouds at the top of the atmosphere is ∼6u2009Wm−2 for MODIS overpass times while the SWCRF bias for ice clouds is smaller in absolute terms by ∼0.7u2009Wm−2, but with stronger spatial variability. Marine clouds of both phases are characterized by...

Evaluation of GCM Column Radiation Models Under Cloudy Conditions with The Arm BBHRP Value Added Product

The overarching goal of the project was to improve the transfer of solar and thermal radiation in the most sophisticated computer tools that are currently available for climate studies, namely Global Climate Models (GCMs). This transfer can be conceptually separated into propagation of radiation under cloudy and under cloudless conditions. For cloudless conditions, the factors that affect radiation propagation are gaseous absorption and scattering, aerosol particle absorption and scattering and surface albedo and emissivity. For cloudy atmospheres the factors are the various cloud properties such as cloud fraction, amount of cloud condensate, the size of the cloud particles, and morphological cloud features such as cloud vertical location, cloud horizontal and vertical inhomogeneity and cloud shape and size. The project addressed various aspects of the influence of the above contributors to atmospheric radiative transfer variability. In particular, it examined: (a) the quality of radiative transfer for cloudless and non-complex cloudy conditions for a substantial number of radiation algorithms used in current GCMs; (b) the errors in radiative fluxes from neglecting the horizontal variabiity of cloud extinction; (c) the statistical properties of cloud horizontal and vertical cloud inhomogeneity that can be incorporated into radiative transfer codes; (d) the potential albedo effects ofmorexa0» changes in the particle size of liquid clouds; (e) the gaseous radiative forcing in the presence of clouds; and (f) the relative contribution of clouds of different sizes to the reflectance of a cloud field. To conduct the research in the various facets of the project, data from both the DOE ARM project and other sources were used. The outcomes of the project will have tangible effects on how the calculation of radiative energy will be approached in future editions of GCMs. With better calculations of radiative energy in GCMs more reliable predictions of future climate states will be attainable, thus affecting public policy decisions with great impact to public life.«xa0less