PNNL

Abstract

Cloud models are applied extensively in many branches of the atmospheric science as a tool for generating synthetic datasets that can be used as proxies for real data. Model-generated cloud fields are often used to test retrieval algorithms for ground-based and satellite remote sensors. When embedded into a global climate model as, e.g., in a Multiscale Modeling Framework, cloud models are trusted to generate realistic statistics of cloud properties over a wide range of synoptic regimes. However, the credibility of all these important and promising applications will remain limited until the ability of these cloud models to reproduce correctly cloud structure and variability observed in nature is proven. We test such ability in a case study of low-level clouds from the Earth Observing System (EOS) validation program conducted in March 2000 in north central Oklahoma. This field project resulted in an observational dataset that is unique in its completeness. During that period there were five days for which, in addition to continuous ground-based observations from the Atmospheric Radiation Measurement Program, in-cloud aircraft observations are available for the time of Terra overpass. Furthermore, for two days, including the one considered in this study, the Terra overpasses were complemented with high spatial resolution measurements from airborne versions of the Moderate-Resolution Imaging Spectrometer (MODIS) and Multi-angle imaging spectro-radiometer (MISR), flown on the high altitude NASA ER-2 aircraft. Synthetic cloud properties are generated by a high-resolution three-dimensional cloud model with a sophisticated explicit microphysics scheme. The main advantage of the explicit microphysics scheme is that it rids all ad hoc assumptions on the shape of cloud particle spectrum present in bulk microphysics parameterizations. Instead, this scheme predicts evolution of the size distributions of cloud hydrometeors by explicitly solving the system of condensation and coagulation equations. By keeping track of the evolution of cloud particle spectrum in each grid point, particle number concentration, mean size, extinction coefficient, water content, and radar reflectivity factors, which are proportional to the 0th, 1st, 2nd, 3rd, and 6th moments of the cloud particle size distribution, respectively, can be calculated independently and with a high degree of fidelity, albeit at great computational expense. Unlike many previous modeling studies, in which comparisons with observations were limited to a few highly integrated cloud characteristics, such as a cloud fraction or averaged liquid water profile, we employ a concept of an instrument simulator. In this approach, we simulate how a specific sensor would see the synthetic cloud field. For example, the top of atmosphere radiances are obtained using a 3D radiative transfer model. One of the advantages of this approach is that the sensitivity and error propagation analysis are much easier to perform using forward calculations than inversion transformations (as in retrievals). Comparisons for selected ground-based, satellite, suborbital, and in-situ measurements will be presented. Benefits and limitations of the approach will be discussed, as will be some aspects of accuracy and numerical implementation of various simulator’s algorithms.