Sensitivities of Simulated Satellite Views of Clouds to Subgrid‐Scale Overlap and Condensate Heterogeneity. Issue 14 (21st July 2018)

• Record Type: Journal Article
• Authors: Hillman, B. R.; Marchand, R. T.; Ackerman, T. P.
• Is Part Of: Journal of geophysical research. Volume 123:Issue 14(2018)
• Page Start: 7506
• Page End: 7529
• Publication Date: 2018-07-21
• Subjects: global climate model -- satellite simulator -- COSP -- cloud overlap -- cloud condensate heterogeneity -- multiscale modeling framework; Atmospheric physics -- Periodicals; Geophysics -- Periodicals; 551.5
• DOI: 10.1029/2017JD027680
• Languages: English
• ISSNs: 2169-897X

Abstract

Abstract: Satellite simulators are often used to account for limitations in satellite retrievals of cloud properties in comparisons between models and satellite observations. The purpose of this framework is to enable more robust evaluation of model cloud properties, so that differences between models and observations can more confidently be attributed to model errors. A critical step in this process is accounting for the difference between the spatial scales at which cloud properties are retrieved with those at which clouds are simulated in global models. In this study, we create a series of sensitivity tests using 4‐km global model output from the multiscale modeling framework to evaluate the sensitivity of simulated satellite retrievals to common assumptions about cloud and precipitation overlap and condensate variability used in climate models whose grid spacing is many tens to hundreds of kilometers. We find the simulated retrievals are sensitive to these assumptions. Using maximum‐random overlap with homogeneous cloud and precipitation condensate leads to errors in Multiangle Imaging Spectroradiometer and International Satellite Cloud Climatology Project‐simulated cloud cover and in CloudSat‐simulated radar reflectivity that are significant compared to typical differences between the model simulations and observations. A more realistic treatment of unresolved clouds and precipitation is shown to substantially reduce these errors. The sensitivity to these assumptions underscores the need for the adoption of more realistic subcolumn treatments in models and the need for consistency among subcolumn assumptions between models and simulators to ensure that simulator‐diagnosed errors are consistent with the model formulation. Key Points: Simulated satellite retrievals from global model outputs are sensitive to subgrid assumptions Using maximum‐random overlap and horizontally homogeneous condensate leads to errors in simulated retrievals Errors in simulated satellite retrievals are reduced using more realistic overlap and horizontal variability