progress on mpas land ice model development · 2014. 2. 6. · operated by los alamos national...

32
Operated by Los Alamos National Security, LLC for the U.S. Department of Energys NNSA U N C L A S S I F I E D Progress on MPAS Land Ice Model Development Matt Hoffman, Steve Price, LANL Mauro Perego, Andy Salinger, Irina Kalashnikova, SNL Lili Ju, Univ. South Carolina Max Gunzberger, Florida State Univ. Wei Leng, Chinese Academy of Sciences, Beijing With help from: Bill Lipscomb, Doug Jacobsen, Todd Ringler (LANL)

Upload: others

Post on 25-Jan-2021

2 views

Category:

Documents


0 download

TRANSCRIPT

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA U N C L A S S I F I E D

    Progress on MPAS Land Ice Model Development

    Matt Hoffman, Steve Price, LANL Mauro Perego, Andy Salinger, Irina Kalashnikova, SNL Lili Ju, Univ. South Carolina Max Gunzberger, Florida State Univ. Wei Leng, Chinese Academy of Sciences, Beijing With help from: Bill Lipscomb, Doug Jacobsen, Todd Ringler (LANL)

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Outline

    MPAS Overview

    Status of Velocity Solvers

    Coupling & Science Applications

    Slide 2

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    MPAS-Land Ice •MPAS - Model for Prediction Across Scales: A climate modeling framework that supports dynamical cores on unstructured Voronoi (SCVT) meshes (MPAS Atmosphere, Land Ice, Ocean, Sea Ice cores)

    1Leng et al., JGR, 117 (2012) 2Perego et al., J. Glac. 58 (2012)

    -Funded under PISCEES project: -LANL, ORNL, SNL, NCAR, FSU, USC, UT, MIT -Goal: Hierarchical suite of FEM-based ice sheet dynamical cores (Stokes1, 1st-order2, LIL22, etc.) based on MPAS CVT mesh generation and modeling framework

    Allows high resolution in regions of interest, reducing necessary number of grid cells by ~10x

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    FO-FELIX (Using Trilinos); SNL Stokes-FELIX (Using PHG); USC, FSU

    MPAS-LI; LANL

    Current Breakdown of Responsibility

    Additional physics (e.g., basal processes) Coupling to Earth Systems Models

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 5

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 6

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 7

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Status of MPAS-Land Ice Initial Public Release, November 2013: http://mpas-dev.github.io/ Native SIA solver in MPAS (convenient for testing) Interfaces to LiveV (SSA, L1L2, FO, Stokes solvers), Albany-FELIX (FO solver), PHG

    (Stokes solver) Forward Euler Time Integration scheme

    • Optional adaptive time stepping Thickness evolution

    • FO Upwind • Margin advance & retreat • Surface Mass Balance • Higher-order Flux Corrected Transport

    Tracer (temperature) transport: First-Order & Flux Corrected Transport (FCT) scheme Tools external to MPAS (written mostly in python)

    • Setup land ice grids on regular planar hex mesh or variable resolution spherical mesh • Ability to setup test cases: Halfar dome, EISMINT, ISMIP-HOM (in progress) • Copy CISM datasets to MPAS grids • Visualization tools (not many off-the-shelf options) • Automated testing using 'lettuce' • 'loose' coupler to MPAS-Ocean

    Slide 8

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Status of Velocity Solvers

    Slide 9

    Native Fortran SIA (LANL)

    Albany-FELIX First Order (SNL)

    PHG-FELIX Stokes (USC, FSU)

    Modeled thickness (m) Analytic thickness (m) Model error (m) Halfar analytic test case for SIA after 1000 years on 2 km mesh

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    MPAS SIA solver running on a sphere (FO, Stokes can too) Simplistic "Laurentide-like" simulation: inception, 10's ka simulation time, multiple ice masses

    Slide 10

    Surface Topography N-S Ice Speed

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Albany-FELIX, First-Order Solver (see talk by Irina Kalashnikova)

    Slide 11

    Sandia National Lab: Andy Salinger Irina Kalashnikova Mauro Perego

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 12

    Wei Leng, Lili Ju

    Univ. South Carolina: Lili Ju Chinese Academy of Sciences: Wei Leng Florida State Univ.: Max Gunzburger

    Stokes-FELIX

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Improved mass conservation of Enhanced Taylor-Hood Finite Elements

    Slide 13

    P1/P1 P2/P1 (Taylor-Hood)

    P2/P1+P0 (Enhanced Taylor-Hood)

    Velocity DOF Pressure DOF

    Global mass conservation only Local mass conservation

    Volume of ice in manufactured solution test case over 1000 yr:

    Wei Leng Lili Ju Max Gunzburger

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Application to Greenland

    Slide 14

    Velocity Difference between Taylor-Hood and

    Enhanced Taylor-Hood elements Flux Divergence

    Enhanced Taylor-Hood Taylor-Hood

    Wei Leng Lili Ju Max Gunzburger

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    5km Greenland: FO vs. Stokes

    Slide 15

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Elevation feedback SMB = f(elevation)

    Meltwater induced speedup Sliding = f(runoff)

    ice2sea Model Intercomparison: Greenland Sea Level Rise Predictions

    Results using MPAS model

    Edwards, et al. TC, 2014 Shannon, et al. PNAS, 2013

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    MPAS-Ocean: Support for sub-shelf circulation

    Slide 17

    Mark Petersen, LANL

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 18

    Mark Petersen, LANL

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Questions?

    Slide 19

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Slide 20

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 21

    MPAS Framework

    I/O MPI Grid management Timekeeping Shared operators …

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 22

    FELIX FELIX

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    What’s missing (short to medium term goals)

    Variable resolution planar hex meshes (talking with MPAS-A)

    Finish temperature implementation (vertical diffusion)

    Higher order thickness advection (using modified MPAS framework methods)

    Coupling to CESM

    More physics… (e.g. basal processes)

    Slide 23

    Ringler et al., Ocean Dyn. (2008)

    Velocity Cell Size

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Status of Velocity Solvers

    Slide 24

    Native Fortran SIA (LANL)

    Albany-FELIX First Order (SNL)

    PHG-FELIX Stokes (USC, FSU)

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    Dome: FO Velocity Solver – ‘ice stream’

    Slide 25

    Beta sliding coefficient

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 26

    Dome: FO Velocity Solver – ‘ice stream’

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA U N C L A S S I F I E D MPAS

    (FO) CISM

    (FO)

    5km Greenland, 2048 processors

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    CISM = solid lines

    MPAS = dashed lines

    Greenland Ice Sheet sea level rise ice2sea basal lubrication experiments

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 29

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 30

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D

    10 km Antarctica, no slip, with ice shelves, LifeV FO

    Slide 31

  • Operated by Los Alamos National Security, LLC for the U.S. Department of Energy’s NNSA

    U N C L A S S I F I E D Slide 32

    Wei Leng, Lili Ju

    Slide Number 1OutlineSlide Number 3Slide Number 4Slide Number 5Slide Number 6Slide Number 7Status of MPAS-Land IceStatus of Velocity SolversMPAS SIA solver running on a sphere (FO, Stokes can too)� Simplistic "Laurentide-like" simulation: �inception, 10's ka simulation time, multiple ice massesAlbany-FELIX, First-Order Solver�(see talk by Irina Kalashnikova)Stokes-FELIX�Improved mass conservation of �Enhanced Taylor-Hood Finite ElementsApplication to Greenland�5km Greenland: �FO vs. StokesSlide Number 16MPAS-Ocean: Support for sub-shelf circulationSlide Number 18Questions?Slide Number 20Slide Number 21Slide Number 22What’s missing (short to medium term goals)Status of Velocity SolversDome: FO Velocity Solver – ‘ice stream’Dome: FO Velocity Solver – ‘ice stream’Slide Number 27Slide Number 28Slide Number 29Slide Number 3010 km Antarctica, no slip, with ice shelves, LifeV FOSlide Number 32