The Data Assimilation Research Testbed: DART

Transcription

1 The Data Assimilation Research Testbed: DART Data Assimilation Research Section & Many Collaborators pg 1

2 DART provides data assimilation glue to build ensemble forecast systems for the atmosphere, ocean, land, Prediction Model Observing System DART Analysis Diagnostics pg 2

3 DART is: A state-of-the-art Data Assimilation System for Geoscience Works with many models (low order through state-of-the-art GCMs). Works with any observations: Real, synthetic, novel. Flexible, portable, well-tested, extensible, free! Fast and efficient, laptops to supercomputers. A Data Assimilation Research System Theory based, widely applicable general techniques. Localization, Sampling Error Correction, Adaptive Inflation, A Teaching Tool Extensive tutorial materials with examples, exercises, explanations. People: The DAReS Team pg 3

4 DART is used at: NCAR 48 UCAR member universities, More than 100 other sites, (More than 1500 registered users). pg 4

5 Global Atmosphere models: CAM Community Atmosphere Model; NCAR (all 3 dynamical cores) CAM/CHEM CAM with Chemistry NCAR WACCM Whole Atmosphere Community NCAR Climate Model AM2 Atmosphere Model 2 NOAA/GFDL NOGAPS Navy Operational Global US Navy Atmospheric Prediction System ECHAM4,6 European Centre Hamburg Model Hamburg Planet WRF Global version of WRF JPL MPAS Model for Prediction Across NCAR/DOE Scales LMDZ5 Laboratoire de Météorologie Indian Inst. Dynamique Zoom of Technology pg 5

6 Regional Atmosphere models: WRF/ARW Weather Research and NCAR Forecast Model WRF/CHEM WRF with Chemistry NCAR NCOMMAS Collaborative Model for NOAA/NSSL Multiscale Atmospheric Simulation COAMPS Coupled Ocean/Atmosphere US Navy Mesoscale Prediction System CMAQ Community Multi-scale Air Quality EPA COSMO Consortium for Small-Scale DWD Modeling pg 6

7 Ocean models: POP Parallel Ocean Program DOE/NCAR MIT OGCM Ocean General Circulation MIT Model ROMS Regional Ocean Modeling Rutgers System (under development) MPAS Model for Prediction Across DOE/LANL Scales (Under development) pg 7

8 Upper Atmosphere/Space Weather models: ROSE NCAR TIEGCM Thermosphere Ionosphere NCAR/HAO Electrodynamic GCM GITM Global Ionosphere Michigan Thermosphere Model Solar Dynamo Dynamo/sunspot model NCAR/HAO pg 8

9 Land Surface models: CLM Community Land Model NCAR NOAH Relatively simple land model Community CABLE Community Atmosphere CAWCR Biosphere Land Exchange (Australia) Hydrologic Model: WRF Hydro NCAR pg 9

10 U,V,T,Q U,V U,V (ocean surface) T,Q,refractivity of the atmosphere T,Q,Tsurface NCEP: Radiosonde, AIRCRAFT (commercial), ACARS NCEP: Cloud Drift Winds from satellite QUIKSCAT, including L2B (JPL) COSMIC Global Positioning Satellite radio occultation AIRS from Aqua/A-train satellite U,V,T,Q,Tsurface, pressure,altimeter MADIS: ACARS, Marine and MESONET surface, METAR, radiosonde, satellite wind Radar reflectivity, radial velocity NCEP pg 10

11 U,V U,V,T,Q,altimeter Cloud Liquid Water Path, Cloud Top and Base Pressures U,V CO (carbon monoxide) U,V MADIS; Wind Profilers, Atmospheric Motion Vectors (AMVs) OK mesonet (U. OK) GOES satellite, CIMSS SSEC (U Wisconsin): Cloud Drift Winds from satellite MOPITT GOES CIMSS (U. WI); rapid-scan AMVs (Atmospheric Motion Vectors), satellite cloud winds pg 11

12 T,Q,Total Precipitable Water Total Precipitable Water U,V U,V (at wind turbine hub height) Electron density U,V,T Chemical concentrations Aerosol optical depth (AOD) GOES CIMSS hyperspectral AIRS IR AMSR, MODIS Microwave Operational typhoon bogus winds, Taiwan Central Weather Bureau Seimens(?) COSMIC/FORMOSAT-3 GTS IASI on EUMETSAT Polar System MetOp satellite TERA and AQUA pg 12

13 Radiances, Occultation on Mars TES, limb sounder on Mars Density, ion concentrations Thermospheric Mass Densities Electron densities Total Electron Density Orbital element information Solar Magnetic Fields Rotational, Meridional Circulation CHAMP CHAMP, GRACE COSMIC Garner GPS Archive NORAD Wilcox, Mt Wilson, National Solar Observatories Mt Wilson, SoHO, SDO, HMI pg 13

14 T, Salinity Surface U, V currents World Ocean Database: Argo floats, CTD(ships), XBT, moored thermistors, drifting buoys(gt-spp) CODAR pg 14

15 Snow cover Leaf area index Total water storage MODIS MODIS GRACE Brightness AMSR-E temperature Heat Flux, Net Carbon Ameriflux tower network Soil Moisture COSMOS (neutron counter) pg 15

16 Provide State-of-the-Art Data Assimilation capability to: Prediction research scientists, Model developers, Observation system developers, Who may not have any assimilation expertise. pg 16

17 Models small to huge. Few or many observations. Tiny to huge computational resources. Entry cost must be low. Competitive with existing methods for weather prediction: Scientific quality of results, Total computational effort. pg 17

18 DART is public domain, mostly NCAR base funded. We provide consulting, support and implementation help. Train our customers who are model or observation experts to use DART; keep core group small. Accept community contributions for models, observations, diagnostics, and algorithms. pg 18

19 Sequential ensemble Kalman filters. Independent observation errors. Parallelization by MPI, no threads. Highly-scalable for computation and memory. First order scaling limit is metadata. pg 19

20 Anderson, J., Hoar, T., Raeder, K., Liu, H., Collins, N., Torn, R., Arellano, A., 2009: The Data Assimilation Research Testbed: A community facility. BAMS, 90, , doi: /2009BAMS pg 20