1 / 41

Recent Development of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS TM )

Recent Development of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS TM ). J. Doyle, S. Chen, R. Hodur, T. Holt, M. Liu, K. Sashegyi, J. Schmidt, S. Wang, D. Westphal, J. Cummings, X. Hong, and J. Pullen Naval Research Laboratory Introduction Recent Development

ona
Download Presentation

Recent Development of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS TM )

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Recent Development of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPSTM) • J. Doyle, S. Chen, R. Hodur, T. Holt, M. Liu, K. Sashegyi, J. Schmidt, S. Wang, D. Westphal, J. Cummings, X. Hong, and J. Pullen • Naval Research Laboratory • Introduction • Recent Development • Atmospheric model • Ocean model • Future Plans

  2. Navy Mesoscale Modeling Strategy Telescoping Systems Strategy for Mission Success • NOGAPS: (Fleet Numerical) • Global coverage • 1–10d forecaster guidance • COAMPS: (Fleet Numerical) • High resolution, nested regional coverage • 0-72h forecaster guidance Observations BC, IC Local Model Output • DAMPS: (Regional Centers) • On-scene tactical-scale weather • 0-48h forecaster guidance Data Fusion AI Nowcast On-Scene Obs • COAMPS-OS: (Shipboard-NITES) • Battlegroup data assimilation system • 6-12h data assimilation cycle On-Scene Obs Battlespace Awareness Cube • Nowcast: (Shipboard-NITES) • Real-time, automatic, 4D data fusion • Warfighter time & space requirements • Common situational awareness TDAs

  3. Navy Strategy COAMPS: Flexibility in System Design Single CM System: Open-MP / MPI Shared or Distributed Memory Architectures Ocean Data Assimilation System Atmospheric Data Assimilation System

  4. COAMPS Coupled Ocean/Atmosphere Mesoscale Prediction System: Atmospheric Components • Complex Data Quality Control • Analysis: • Multivariate Optimum Interpolation Analysis (MVOI) (Near Future: 3D Var) • Initialization: • Hydrostatic Constraint on Analysis Increments; Digital Filter • Atmospheric Model: • Numerics: Nonhydrostatic, Scheme C, Nested Grids, Sigma-z, Flexible Lateral BCs • Parameterizations: PBL, Convection, Explicit Moist Physics, Radiation, Sfc Layer • Features: • Globally Relocatable (5 Map Projections) • User-Defined Grid Resolutions, Dimensions, and Number of Nested Grids • 6 or 12 Hour Incremental Data Assimilation Cycle • Can be Used for Idealized or Real-Time Applications • Single Configuration Managed System for All Applications • Operational at FNMOC: • 9 Areas, Twice Daily, using 81/27/9 km or 81/27 km grids • Forecasts to 72 hours • Operational at all Navy Regional Centers (w/ GUI Interface)

  5. NRL Background in Community Modeling • Began development of COAMPS in 1988 • Began distribution to limited research community in 1995: • Lawrence Livermore National Laboratory • University of Oklahoma • North Carolina State University • Jackson State University • Desert Research Institute • U. S. Army Research Laboratory • Began operations in 1997: • FNMOC • Navy Regional Centers • COAMPS Process Action Team (PAT) recommended general release of COAMPS in 2000: • Release via Web; open to all • Support for release now funded • Naval Postgraduate School • San Diego Supercomputer Center • Oregon State University • NOAA Forecast Systems Laboratory • Goddard Space Flight Center • Tulane University

  6. COAMPS Recent Development • Atmospheric Model • NAVDAS • Moist physics • Aerosols • Moving nests • Land surface processes • Upper Boundary Condition • Ocean Model • Analysis • Forecast • WAM

  7. z z x x NAVDAS Navy Atmospheric Variational Data Assimilation System • cast in observation space • vertical profiles are transformed into coefficients • of background error correlation eigenvectors (10-25 fold speedup) • non-separable error correlation functions providing scale-length • variations with height and location • correlation functions can be defined on isentropic surfaces • (along dry air flow) • error correlations have scale dependence • operates on all grids for NOGAPS and COAMPS global regional • uses Message Passing Interface for Massively Parallel Computers • directly assimilates measured quantities SSM/I wind speed, TOVS • radiances, SSM/I precipitable water

  8. Adjustments to COAMPS Bulk Microphysics • Original RH83 scheme lacked secondary ice nucleation, CCN, drizzle, aggregates, and liquid to ice conversions (homogeneous freezing, graupel and/or hail production) • Modified Adjustment to Saturation Scheme: Implicit solution for T,q (Soong and Ogura, 1974) Normalized microphysical rates • Modified ice nucleation (Meyers et al. 1992; Hallet and Mossop, 1974) • Allow nonzero fall speed for pristine ice • Implemented various autoconversion schemes • Implemented the RH84 graupel scheme and homogeneous freezing • Adapted the two-moment Khairoutdinov and Kogan (2000) drizzle parameterization (originally implemented by Dave Mechem, OU) • Modified the turbulence closure for mixed-phase clouds • Implemented a Hybrid time scheme (Clark 1979; Smolarkiewicz and Clark, 1986; Tripoli 1992; Wicker and Wilhelmson 1995) • Implemented a forward positive definite advection scheme (Bott, 1989) • Developing a full two-moment mixed-phase microphysics scheme (Reisner et al., 1998; Meyers et al., 1997; KK 2000 ) • Coupling cloud microphysics with aerosol model(s)

  9. Original RH83 Scheme RFC Analysis Modified Scheme Moist Physics Comparisons Winter-time Precipitation (mm) 24 Hour COAMPS Grid 3 Forecast Valid 01/25/02

  10. COAMPS CONUS One-Week (20020125-20020201) Precipitation Scores

  11. Addition of Aerosol Microphysics to COAMPS • Objective: COAMPS with interactive clouds, radiation and aerosols • Goal: COAMPS predictions of dust storms as weather events for strategic, tactical, • surveillance, and operational uses • Approach: • Include source, dry deposition, and wet removal terms for dust (COAMPS already has accurate tracer transport code) • Use remote sensing to improve specification of dust source areas • Validate model forecasts of occurrence and intensity of dust events in Southwest Asia region for spring 2002

  12. Simulated Orographic Rainfall Structure Rain water (g/kg) 21600 sec Rain water (g/kg) 21600 sec Aerosol Concentration / cm3 Aerosol Concentration / cm3 Marine Environment Continental Environment Aerosol Void Aerosol Void

  13. Mt. Etna Ash Plume COAMPS Transport Simulation 24-h forecast Valid 12 UTC 24 July 2001 Color: Mass Load (kg m-2) SeaWiFS Satellite Image 24 July 2001

  14. COAMPS MPI Moving Nests Hurricane Gordon 00Z September 17- 00Z September 19, 2000 Nest 2: 36 h forecast valid 12 UTC 18 September 2000 Sea level pressure (hPa) Fixed Nest Option Moving Nest Option 10-m wind speed (m/s) (m/s) 0 4 8 12 16 20

  15. J E A F I B C G H R Q D K P L O N M Proposed COAMPS Land Surface Model (LSM) System • DATABASES • Vegetation type (USGS 1-km global) • Soil texture (1-km USDA STATSGO; 1o GED) Seasonal variation (satellite-derived NDVI) LSM Meteorological model input T,q, ps, u,v, precip. LW, SW radiation LSM processes A = precipitation B = condensation C = on vegetation D = on bare soil E = transpiration F = canopy water evaporation G = direct soil evaporation H = evaporation from open water I = deposition/sublimation to/from snow pack J = turbulent heat flux to/from snow pack/soil/plant canopy K = soil heat flux L = interflow M = internal soil flux N = gravitational flow O = internal moisture flux P = soil moisture flux Q = runoff R = dust processes S = urban effects S

  16. Urban Canopy Parameterization • Originally developed by Brown and Williams (1998), modified • by Chin et al. (2000) (2) Turbulence production (1) Momentum loss (3) Radiation absorption • Roof albedo (a) = 0.22 • Roof emissivity (e) = 0.91 • Roof heat capacity (Croof) =9.681 e4 • Roof drag coefficient (CDroof) = 7.1e-3 • Urban drag coefficient (Cd) = 1.2e-2 • Extinction coefficient (k) = 0.1 • Bowen ratio (Br) = 1.5 • Canopy area density (a(z))= linear in z (4) Surface energy budget (1) (2) (3) (4)

  17. COAMPS Upper Boundary Conditon Linear Hydrostatic Gravity Wave Test Analytic Solution Rigid Lid (wtop=0) w (105 m s-1) w (105 m s-1) Radiation Condition (Klemp&Durran) MM5 Local Radiation Condition w (105 m s-1) w (105 m s-1)

  18. COAMPS Coupled Ocean/Atmosphere Mesoscale Prediction System: Ocean Components • Data Quality Control • Analysis: • 2D MVOI of Sea Surface Temp on All Grids • 3D MVOI Analysis of Temperature, Salinity, Surface Height, Sea Ice, Currents • Ocean Model: Navy Coastal Ocean Model (NCOM) • Numerics: Hydrostatic, Scheme C, Nested Grids, Hybrid Sigma/z • Parameterizations: Mellor-Yamada 2.5 • Features: • Globally Relocatable (5 Map Projections) • User-Defined Grid Resolutions, Dimensions • Can be Used for Idealized or Real-Time Applications • Single Configuration Managed System for All Applications • Loosely coupled to COAMPS atmospheric model • Strategy for testing coupled system: • construct a Mesoscale Atm-Ocean Data Assimilation System for the Med. Sea • quantify the skill of system • examine the Adriatic Sea at high resolution as a test-bed for coupling strategies

  19. NCOM Forecasts Surface Velocity (cm/s) 5 Oct 1999 3 Oct 1999 1 Oct 1999 Sea Surface Temperature (C) 5 Oct 1999 3 Oct 1999 1 Oct 1999

  20. Coupled COAMPS/WAM Simulation of TC Bonnie Significant Wave Height (m) 24-h Forecast Valid at 1200 UTC 24 August 1998 (Dx=6 km) NASA Scanning Radar Altimetry Uncoupled Coupled 12.8 10.9 • • Wright et al. (2000)

  21. Future Plans • Development and Implementation of: • 3D variational analysis (NAVDAS) • aerosol model • air-ocean coupled system • land-surface model (LSM) (NOAH) • improved microphysical parameterization • improved Mellor-Yamada Level 2.5 BL Param. • incorporation of WRF KF scheme • LES Option • mesoscale verification • efficiency improvements

  22. Future Plans • COAMPS and WRF Comparisons: • Idealized Simulations • Real Data Simulations • Interchange of Key COAMPS/WRF Modules • Collaboration With WRF Community • Land Surface Model (NOAH) • Physical Parameterization & Numerical Techniques • Ocean/Atmosphere Coupling Methods • Tropical Cyclones • Efficiency/MPI Issues • Proposed Next Generation Micro-a Scale Model

  23. COAMPS FNMOC Operational Areas As of: February 25, 2002 27 km 27 km Europe W_Atl 81 km 81 km 27 km W_Pac 81 km 27 km E_Pac 81 km 9 km 27 km 27 km Southwest_Asia Conus 81 km 81 km 9 km 27 km 27 km 27 km Cent_Am 81 km Area 1 Arabian Sea 81 km 81 km

  24. NAAPS/NOGAPS Simulates Large-scale Dust Storms TOMS Aerosol Index NAAPS Aerosol Optical Depth red – sulfate green – dust blue - smoke SeaWiFS True Color image 13 February, 2001

  25. Dust Storms: A Recurring, Worldwide ProblemMobilization and Transport Controlled by Mesoscale Dynamics Korea, 31 March, 2001 Africa Coast, 21 April, 2001 Arabian Sea, 7 December, 1999 Mediterranean, 18 April, 2001 SeaWiFS images from Kuring/GSFC

  26. COAMPS MPI Moving Nest Software Development • Software developed using MPI • Makes use of existing COAMPS nesting software • Advantages: • Allows for smaller nests (less resources required) • Flexibility in movement of nests: • Namelist specified options: • Battle group option (“target” times/locations) • User specified grid point movement • Nests automatically move together • Automated tropical cyclone movement option (under development) Recent Developments of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS), Ninth Conference on Mesoscale Processes

  27. (1,n) (1,n) (1,n) (m,n) (m,n) (m,n) Shifted region Interpolated region Time = t1 MPI communications needed for shifted and interpolated areas Dropped region (1,1) (1,1) (1,1) (m,1) (m,1) (m,1) COAMPS MPI Moving Nest Software Development (1,n) (1,n) (1,n) (m,n) (m,n) (m,n) Fixed Nest 1: (m x n) points 3 x 3 domain decomposition 2 Halo Points Moveable Nest 2: Time = t0 (1,1) (1,1) (1,1) (m,1) (m,1) (m,1) Recent Developments of the Coupled Ocean/Atmosphere Mesoscale Prediction System (COAMPS), Ninth Conference on Mesoscale Processes

  28. COAMPS CONUS Model Grid Setup Pacific Northwest (PNW) Mississippi River Basin (MRB)

  29. COAMPS CONUS 2001 Cold Season (20011113-20020430) Precipitation Scores

  30. COAMPS CONUS Regional One-Week (20020125-20020201) Precipitation Scores

  31. COAMPS CONUS One-Week (20020125-20020201) Precipitation Scores

  32. COAMPS Availability • Download COAMPS • http://www.nrlmry.navy.mil/projects/coamps/index.html

  33. NAVDAS: Nested COAMPS Wind speed analysis 300 mb 12 km 36 km • Innovations (ob-bckgnd) computed using highest resolution COAMPS background available for each observation • Single merged analysis grid vectors formed by combination of all the nested grid points • Analysis corrections added back to each nested grid on vertical sigma_z model levels • Display on pressure levels 108 km

  34. Moving Nest Option Fixed Nest Option Nest 2 at tau = 48 h 48 h Forecast position: 0-48 h (every 6 h) 0 h 27 km (46x46) Nest 2 at tau = 0 h 27 km (121x85) 81 km (61x61) 81 km (61x61) COAMPS MPI Moving Nests Hurricane Gordon 00Z September 17- 00Z September 19, 2000 Moving Nest Option is 2.7x Faster on O2K

  35. COAMPS Dipole Jewel 5 Simulations Impact of Increased Horizontal Resolution 10-m dosage (arbitrary units) and model terrain (m) 26-h forecast (valid 02 UTC 29 Oct) Nest 3 (3 km) Nest 4 (1 km) Nest 2 (9 km) log -3 -2 -1 0 1 2 Control Simulations

  36. COAMPSTM Urban 2000 Simulations Urban Control Obs Salt Lake City Nest 4 1-km resolution (49 x 49 km): 12-h fcsts 20 4 10-m air temp 10-m wind speed bias = 0.865, 0.099 rms = 1.712, 1.391 bias = 0.103, 0.966 rms = 0.394, 1.197 3 15 BG 2 10 1 5 Urban too warm at night 0 20 5 Reduction and improvement in day and nighttime winds bias =-0.292, 0.309 rms = 0.856, 0.991 bias = 0.470, -0.441 rms = 2.972, 2.849 15 RW 3 m/s deg C 10 2 1 5 Little daytime difference 0 5 4 15 SLC 3 10 2 1 5 0 12 00 12 00 12 UTC 12 00 12 00 12 UTC 16 Oct 17 Oct 18 Oct 16 Oct 17 Oct 18 Oct

  37. Coupled COAMPS/WAM Two-Way Coupling - Important in Boundary Layer(Janssen et al. 1989) - Improved Model Climate(Viterbo and Janssen 1996) - Extratropical Cyclones(Doyle 1995; Lalbebarry et al. 2000; Lionello et al. 1998) - ECMWF/WAM Coupled System (Janssen et al. 2001) WAM (Cycle 4) (WAMDI Group 1988) Wave Spectrum Predicted From Energy Balance Equation F(w,q): 2d Wave Variance Spectrum Coupling Methodology(Janssen 1989; Janssen 1991) COAMPS (zo, Fluxes, e, Kh, m, t, U10) WAM (a) zo=aU*2/g, where a=b(1-tw/t)-0.5

  38. COAMPS Atmospheric Forcing Surface Wind Stress (dynes/cm2) 5 Oct 1999 1 Oct 1999 3 Oct 1999 Surface Heat Flux (W/m2) 1 Oct 1999 3 Oct 1999 5 Oct 1999

  39. COAMPS Dipole Jewel 5 Simulations Nest4: Transport Observed: Cloud Track to T+105 min. 1-km Land-use Simulation DTRA Asterisks every 5 min from 2145-2330 UTC 28 Oct 98 24-h forecast valid 0000 UTC 29 Oct 98

More Related