LES ARM Symbiotic Simulation and Observation (LASSO) Workflow: Ensemble Forcings and LES Sensitivity
Poster PDF
Category
ARM next generation – Megasite and LES activities
Authors
| Satoshi Endo — Brookhaven National Laboratory | Andrew M. Vogelmann — Brookhaven National Laboratory |
| Zhijin Li — University of California | Tami Fairless — Brookhaven National Laboratory |
| Xiaoping Cheng — National University of Defense Technology | Maike Ahlgrimm — Deutscher Wetterdienst |
| Heng Xiao — Pacific Northwest National Laboratory | Shaocheng Xie — Lawrence Livermore National Laboratory |
| William I. Gustafson — Pacific Northwest National Laboratory | Shuaiqi Tang — Pacific Northwest National Laboratory |
Description
The Atmospheric Radiation Measurement (ARM) Climate Research Facility is developing a routine large-eddy simulation (LES) modeling framework at its permanent sites, called the LES ARM Symbiotic Simulation and Observation (LASSO) Workflow, to supplement its extensive observations. An LES ensemble will be used based on multiple forcing data sets, as uncertainty in the forcing will be the biggest driver of simulation spread. We are exploring three forcing dataset methodologies and their variations: 1) the ARM continuous forcing data set based on a constrained variational-analysis approach that combines National Weather Service RAP analysis with ARM observations; 2) forcing data derived from diagnosed tendencies and grid point values from short-term ECMWF forecasts that incorporate ARM radiosonde data; and 3) forcing data derived by a WRF-3DVar-based multi-scale data assimilation (MS-DA) system. MS-DA efficiently assimilates high-resolution data using the community-based Gridpoint Statistical Interpolation (GSI) system in conjunction with a scale separation algorithm to combine observations representing coarse and fine scales. The MS-DA system assimilates ARM profiling and ground-level observations along with operational data assimilation input fields. We will explore initial/boundary conditions from different analysis products. Additionally, several choices must be made for the forcing calculations that impact the final results, e.g., domain size and averaging method. The forcing derivation methodologies and test LES simulations will be presented.Adjacent Poster
William I. Gustafson Jr., Andrew M. Vogelmann