This website's content is no longer actively maintained, but the material has been kept on-line for historical purposes.
The page may contain broken links or outdated information, and parts may not function in current web browsers.

Please note that ISCCP data processing is now performed at NOAA/NCEI.
Please visit the NOAA/NCEI website for access to ISCCP H data products and other up-to-date information.

Cloud Analysis - Part 6

Mesoscale Optical Inhomogeneity of Clouds

Clouds vary on all spatial scales from planetary down to about 30m, but practical considerations limit representation of cloud variability in global climate and weather models to spatial scales larger than about 100-300 km. Since the relationship between cloud properties and radiative fluxes is not linear, the presence of cloud variability at smaller scales (we call scales < 300 km, mesoscale) creates biases in the modeled radiative fluxes if it correctly predicts the cloud property averaged over the smaller scales.
The figures below illustrate a climatology of cloud property variations for spatial scales < 280 km from the ISCCP cloud dataset; the full climatology can also be viewed in more detail or downloaded.


Mesoscale Variability of Cloud Optical Thickness

The effects of mesoscale cloud optical thickness variations on solar radiative transfer can be accounted for approximately by re-scaling the area-mean optical parameters (e.g., optical thickness, single scattering albedo, asymmetry parameter) using a simple parameter, ε (Cairns et al. 2000), that is given by the expression (Rossow et al. 2001):

ε = 1 - τ̂ / τ̅

where τ̅ is the linear average of the varying optical thickness over the area and τ̂ is the "radiative-average" that gives the correct cloud albedo. Thus, the optical thickness value that gives the correct albedo for a spatially inhomogeneous cloud is τ̂, given by

τ̂ = (1 - ε) × τ̅

The annual mean values of epsilon are shown in the Figure 1 below for all clouds and for low-level clouds (top pressure > 680 mb), middle-level clouds (680 mb > top pressure > 440 mb) and high-level clouds (440 mb > top pressure). This parameter can be scaled to represent the variations for other sized regions by assuming that epsilon varies approximately as the region area. Figure 2 shows the annual average albedo bias, caused by neglecting the mesoscale cloud variability, where the bias is weighted by the cloud fraction to give a scene albedo bias.

Figure 1: Annual Mean Cloud Optical Thickness Inhomogeneity Parameter

All Low-Level Middle-Level High-Level
Annual Mean Cloud Optical Thickness Inhomogeneity Parameter: All Annual Mean Cloud Optical Thickness Inhomogeneity Parameter: Low-Level Cloud Annual Mean Cloud Optical Thickness Inhomogeneity Parameter: Middle-Level Cloud Annual Mean Cloud Optical Thickness Inhomogeneity Parameter: High-Level Cloud

Figure 2: Annual Mean Scene Albedo Bias

All Low-Level Middle-Level High-Level
Annual Mean Scene Albedo Bias: All Annual Mean Scene Albedo Bias: Low-Level Cloud Annual Mean Scene Albedo Bias: Middle-Level Cloud Annual Mean Scene Albedo Bias: High-Level Cloud

The mesoscale variations of cloud optical thickness also affect the thermal infrared radative fluxes by changing the effective emissivity of the clouds. A similar ε - ir parameter can be defined to describe this effect (Rossow et al. 2001). The annual mean value of ε - ir is shown in Figure 3. Figure 4 shows the scene (cloud-fraction-weighted) emissivity bias that is produced by the mesoscale cloud variability.

Figure 3: Annual Mean Cloud Emissivity Inhomogeneity Parameter

All Low-Level Middle-Level High-Level
Annual Mean Cloud Emissivity Inhomogeneity Parameter: All Annual Mean Cloud Emissivity Inhomogeneity Parameter: Low-Level Cloud Annual Mean Cloud Emissivity Inhomogeneity Parameter: Middle-Level Cloud Annual Mean Cloud Emissivity Inhomogeneity Parameter: High-Level Cloud

Figure 4: Annual Mean Scene Emissivity Bias

All Low-Level Middle-Level High-Level
Annual Mean Scene Emissivity Bias: All Annual Mean Scene Emissivity Bias: Low-Level Cloud Annual Mean Scene Emissivity Bias: Middle-Level Cloud Annual Mean Scene Emissivity Bias: High-Level Cloud

Mesoscale Variability of Cloud Top Temperature

The thermal infrared radiation is also affected by mesoscale variations of cloud top temperature because the radiative flux is a non-linear function of temperature. A larger negative bias in emitted flux is produced by increasing variability. If clouds are categorized by the cloud top height, as in the figures above, the cloud top temperature variability is restricted. Figure 5 shows the annual mean standard deviation of cloud top temperature variability only for all clouds together. The variability is caused by mixtures of different cloud types; the occurrence of upper-level clouds produces the larger effects.

Figure 5: Annual Mean Spatial Standard Deviation of Cloud Top Temperature

Annual Mean Spatial Standard Deviation of Cloud Top Temperature


Cairns, B., A.A. Lacis, and B.E. Carlson, 2000: Absorption within inhomogeneous clouds and its parameterization in general circulation models. J. Atmos. Sci., 57, 700-714. (Read abstract.)

Rossow, W.B., C. Delo, and B. Cairns, 2002: Implications of the Observed Mesoscale Variations of Clouds For Earth's Radiation Budget. J. Clim., 15, 557-585. (Read abstract.)


Part 1 | Part 2 | Part 3 | Part 4 | Part 5 | Part 6 | Part 7 | Part 8

Data Analysis