Hello, A yearly allocation for the LCRC cluster has been requested with the following updated information: Submitter/PI: Rao Kotamarthi Project Name: Atmos_Chem Division: EVS Project title: Process Scale Modeling of Atmospheric Chemistry and Transport Associated funding: DOE,DoD Other Systems: NERSC Science: A primary objective is to perform dynamic downscaling of climate model simulation results for regional scale applications. In particular, we will focus on the uncertainties in downscaled projections resulting from boundary conditions, the utility of spectral nudging and grid scale nudging for improving the energy balance of regional scale climate models over long integration times. In particular, we are exploring the development of new regional scale climate models as there is a need to develop a code base that bridges the scales from mesoscale to microscale using continuous mesh refinement that can allow for local refinement of grids from kilometers and scales with well established physics parameterization to meters, where the flow is completely resolved without the need of nesting. Project description: There are several existing mesh geometries and dynamical cores that allow continuously refined meshes, for example, the Model for Prediction Across Scales (MPAS) develop by Los Alamos National Laboratory and the National Center for Atmospheric Research (NCAR); and the High-Order Method Modeling Environment (HOMME) contributed by NCAR, DOE laboratories and universities. These models are built using various resolution meshes which have smoothly-varying mesh transitions. This capability ameliorates many issues associated with the traditional mesh refinement strategy of one-way and two-way grid nesting where the transitions are abrupt. However, currently, both of these dynamic cores are only used in global scale models. It is well known that running global scale models at very high resolutions is computationally expensive. For example, running a uniform 3-km mesh over globe contains 65.5 millions of horizontal grid cells. Tests on NCAR’s Yellowstone found that the uniform 3-km mesh with 41 layers requires around 190000 core-hours per simulated day; while BG/Q hardware at Argonne tends to require around five times the number of core-hours as Yellowstone (personal conversation with Michael Duda, NCAR). However, if one is only interested in resolving flow and atmospheric dynamics over a limited area of the earth system, then resolving flow in remote regions from the boundary of the limited area of interest is not necessary or computationally relevant. Thus, there is an urgent need to develop an exascale capable code for atmospheric simulations over limited regions based on the same or similar dynamical core as MPAS and/or HOMME. Doing so will only focus on the area of interest and allow continuous refinement from kilometers to meters with much cheaper computational resources. For this proposal the primary activity is developing a process for prescribing the boundary conditions to the limited area model (LAM) with unstructured grids. Three types of lateral boundary conditions will be tested: (a) Smooth flow boundary conditions, which fixes the velocity (U,V) at the boundary for the model simulation period; (b) Cyclic boundary conditions, which cycles a time varying boundary condition over a certain period of the model simulation (for example a day) over the entire simulation period (for example a month); and (c) variable (observational) boundary conditions, which provide temporal varying boundary conditions for the model. The terrain following coordinate or the hybrid coordinate used in the MPAS model will be used as the surface boundary in the LAM. The lateral boundaries will be modeled using a similar approach used in the fixed mesh mesoscale meteorological models, such as the WRF model, by defining a buffer region extending over several grid po ints and using interpolation of the flow fields to the buffer zone. Industry partnership: None Project URL: Current FY Hours Used: undetermined amount New FY Requested allocation: 200000 Q1: 50000 Q2: 50000 Q3: 50000 Q4: 50000 Justification: Storage requirements: 1TB Thank You, The LCRC Accounts System