Seamless Multi-scale Forecasting: Hybridizable Unstructured-mesh Modeling and Conservative Two-Way Nesting
Abstract
Project Summary / Abstract One of our research goals is to derive and apply advanced techniques for multiscale modeling of tidal-to-mesoscale processes over regional domains (nearshore-coastal-basin) with complex geometries including shallow seas with strong tides, steep shelfbreaks with fronts, and deep ocean interactions. On the one hand, our conservative implicit two-way nesting for realistic multi-resolution modeling has enabled such high-fidelity coupled multiscale dynamics studies. On the other hand, a high-order multi-dynamics modeling capability based on novel hybridizable discontinuous Galerkin (HDG) numerical schemes is also promising for seamless conservative multi-resolution forecasting. Our goal for this NOPP project is to improve, utilize and verify: (i) HYCOM downscaling schemes and conservative two-way nesting schemes for seamless multiscale forecasting and dynamical analyses of realistic coupled physics at abrupt topography; and, (ii) high-order non-hydrostatic HDG schemes for high-fidelity, conservative, and efficient multi-dynamics modeling. We plan to apply and further develop our downscaling and two-way nesting schemes for domains with abrupt topography in the West Pacific Ridge ocean region. For the downscaling and merging of varied modeling and observation inputs at initialization, we will utilize our new semi-analytical optimized initialization and downscaling methodology. Our implicit two-way nesting will aim for seamless and numerical-bias-free nesting of multiscale dynamics, from internal tides to larger-scale dynamics. We plan to evaluate the sensitivity to the HYCOM inputs and to the parameters of the two-way nesting. We will refine nested-grid boundary conditions and conservative multi-grid exchanges such that upscale and downscale effects of multiple dynamics, from nonlinear internal tides to basin-scale currents, are transferred accurately across the multi-resolution domains. We intend to continue our implementation of generalized adaptive vertical coordinate systems. We plan to contribute to real-time forecasting and data-assimilative simulations for the FLEAT observation region and for inputs to acoustics modeling and coupled simulations. We also plan to perform multi-resolution re-analyses, process studies and multiscale dynamics analyses focusing on interactions of current systems and circulation features with topographic waves, internal tides or bottom gravity currents. We plan to improve and utilize our HDG finite-element code to complete process modeling studies relevant to flows encountering abrupt topography. To do so, we will further develop the HDG schemes and their efficient implementation. The schemes combine the HDG method with a projection method and selective slope limiting to obtain high-order accurate schemes for nonhydrostatic ocean flows with a nonlinear free surface. After downscaling from HYCOM, we plan to research and implement our HDG finite-element schemes such that they can run with multiple dynamics within the same computational domain. With parallel computing, our HDG scheme then solve these sets of elements with different dynamics on different compute nodes. The dynamical goal for the high-order HDG simulations is to allow process modeling studies of flows encountering abrupt topography, especially multiscale phenomena that involve shelf physics, steep shelfbreak dynamics and non-hydrostatics physics in shallow and deep waters.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Aug 12, 2016
- Source ID
- N000141512597
Entities
People
- Pierre Felix Lermusiaux
Organizations
- Massachusetts Institute of Technology
- Office of Naval Research
- United States Navy