Inference of an Optimal Ice Particle Model through Latitudinal Analysis of MISR and MODIS Data

Abstract

The inference of ice cloud properties from remote sensing data depends on the assumed forward ice particle model, as they are used in the radiative transfer simulations that are part of the retrieval process. The Moderate Resolution Imaging Spectroradiometer (MODIS) Collection 6 (MC6) ice cloud property retrievals are produced in conjunction with a single-habit ice particle model with a fixed degree of ice particle surface roughness (the MC6 model). In this study, we examine the MC6 model and five other ice models with either smoother or rougher surface textures to determine an optimal model to reproduce the angular variation of the radiation field sampled by the Multi-angle Imaging Spectroradiometer (MISR) as a function of latitude. The spherical albedo difference (SAD) method is used to infer an optimal ice particle model. The method is applied to collocated MISR and MODIS data over ocean for clouds with temperatures ≤233 K during December solstice from 2012-2015. The range of solar zenith angles covered by the MISR cameras is broader at the solstices than at other times of the year, with fewer scattering angles associated with sun glint during the December solstice than the June solstice. The results suggest a latitudinal dependence in an optimal ice particle model, and an additional dependence on the solar zenith angle (SZA) at the time of the observations. The MC6 model is one of the most optimal models on the global scale. In further analysis, the results are filtered by a cloud heterogeneity index to investigate cloudy scenarios that are less susceptible to potential 3D effects. Compared to results for global data, the consistency between measurements and a given model can be distinguished in both the tropics and extra-tropics. The SAD analysis suggests that the optimal model for thick homogeneous clouds corresponds to more roughened ice particles in the tropics than in the extra-tropics. While the MC6 model is one of the models most consistent with the global data, it may not be the most optimal model for the tropics.

Ice Cloud Optical Thickness, Effective Radius, And Ice Water Path Inferred From Fused MISR and MODIS Measurements Based on a Pixel‐Level Optimal Ice Particle Roughness Model

“…In this study, one full year (2013) of MISR and MODIS data is fused based on cloud‐top properties detailed in Liang et al () and Liang and Di Girolamo (), with updates in Wang et al (). There was no strong signal from El Nino nor La Nina in 2013, based on the ENSO index.…”

“…The filtered pixel radiances are analyzed using the spherical albedo difference (SAD) method developed by Doutriaux‐Boucher et al (), which was applied to a MISR dataset in our previous study (Wang et al, ). To use the SAD methodology, an adding‐doubling radiative transfer model (Huang et al, ) is used to compute the spherical albedos in conjunction with the use of the 12 roughness values.…”

“…The pixel locations are restricted to 60°N–60°S to avoid the effects of snow and sea ice on the measured reflectance. Compared to Wang et al (), we add an additional quality‐control step to remove potential multilayer clouds (McFarlane & Marchand, ; Naud et al, ). Specifically, we identify the potential presence of multilayer clouds as occurring when the MISR stereo height values and MODIS cloud top height values differ by more than 2000 m (Mueller et al, ).…”

“…To use the SAD methodology, an adding‐doubling radiative transfer model (Huang et al, ) is used to compute the spherical albedos in conjunction with the use of the 12 roughness values. The spherical albedo is a variable that is independent on satellite viewing geometry and is a nonlinear function of τ. τ is not fixed in the computation of spherical albedo; more details are in Wang et al (). The SAD values for a roughness model are computed from the differences of the computed spherical albedo values and the average spherical albedo of all applicable angles (i.e., the scattering angle measured among AA, AN, and AF).…”

“…A recent study examined the agreement between Multi‐angle Imaging SpectroRadiometer (MISR) sensor measurements and radiative transfer simulations based on an ice habit model in conjunction with a varying degree of surface roughness. The surface roughness was found to be an important parameter in inferring ice cloud properties due to a latitudinal dependence that varied seasonally with the thermal equator annual cycle (Wang et al, ). The MISR cameras view a wide range of scattering angles for a given pixel containing an ice cloud and provide valuable data for determining a variable ice habit model that specifies an optimal value of the degree of ice particle surface roughness in fitting measurements at the pixel level.…”