If you have any problems related to the accessibility of any content (or if you want to request that a specific publication be accessible), please contact (email@example.com). We will work to respond to each request in as timely a manner as possible.
Evaluating and Improving the Multi-Moment Cloud Microphysics Scheme
AdvisorWilcox, Eric M.
AltmetricsView Usage Statistics
Four different studies examining parameterizations of particle size distribution, sedimentation, autoconversion, and vapor condensation in a scheme for representing cloud microphysics in numerical weather and climate models are conducted with an emphasis on warm liquid-phase cloud systems. In the first study, the expression of rain droplet size distribution is investigated to determine the optimal approach to fit the particle spectra of rain droplets. The widely used size distribution functions (SDFs) that are investigated here are the Marshall-Palmer SDF and the gamma SDF where the parameters of the gamma SDF are determined from a variety of approaches, including diagnosis from empirical functions and algebraic solutions. The results of sensitivity experiments illustrate that a solution for the parameters of the gamma SDF based on mathematical deviation is an effective pathway to improve the ability of the triple-moment gamma SDF to fit the observed droplet SDF. Additionally, we also attempt to select the optimal moment order group to construct the equation group used to solve for the parameters of the gamma SDF and found that the group consisting of the 0th, 3rd, and 4th moments of the SDF (the 0-3-4 moment order group) provided better performance than the 0-3-6 moment order group. In the second study, the focus shifts to the prediction of precipitation improved by incorporating the new triple-moment treatment of rain droplet spectrum in the sedimentation parameterization. The implementation of triple-moment gamma SDF in the Morrison cloud microphysical scheme enabled this scheme to simulate the evolution of the rain droplet mass and number maxing ratios during a heavy rainfall event in better agreement with the observed evolution of these quantities. Additionally, the modeling results and analysis of parameterization algorithms suggest that two microphysical terms contribute to an overestimate of the rain droplet number mixing ratio should be removed from the source and sink terms in the parameterization scheme so as to yield the correct magnitude of rain droplet number. Furthermore, since the sedimentation strength and evaporation strength are continuous functions of cloud particle droplet size, the application of the triple-moment gamma SDF is effective at improving the parameterizations of these physical processes. In the third study, the investigation of precipitation particle generation showed that the original Morrison scheme underestimated rates at which rain droplets are generated and their sedimentation rate. The sensitivity experiments demonstrated that the Tripoli autoconversion scheme and triple-moment treatment of rain droplet spectrum are needed to correctly model rain droplet formation and settling rates. A remaining issue is the overestimated magnitude of rain droplet number mixing ratio induced by the monodisperse size distribution of rain droplets resulting from the autoconversion process. Additionally, it is noted that the accretion term in Morrison scheme cannot simulate the precipitation suppression related to air pollution since the accretion term is not sensitive to aerosol-induced cloud droplet number perturbations. A new subgrid scale vapor phase change scheme is presented and implemented in the Morrison cloud microphysics scheme which has been modified as in the studies described above. The simulated results from warm cloud cases observed at sites in the mountains and on plains are presented and contrasted with observations. The evaluations indicate that although the grid spacing was set as 4km, the original version of Morrison scheme is not capable of realistically forecasting the initialization of small-scale warm clouds. A potential source of error is the absence of subgrid-scale vapor condensation induced by the heterogeneous distribution of vapor mixing ratio within model grid cells. The biases associated with small cloud formation was decreased by the assumption that the subgrid-scale perturbations of vapor mass mixing ratio followed a normalized gamma distribution function. Further development of this subgrid scheme is needed to yield the precise reflection of vapor mass fluctuations and subsequently forecast the occurrence of subgrid clouds.