5th WRF LSM Workshop, NCAR, 9/13/05

Overview of the Unified Noah LSM in WRF Recent Enhancement, Test, and Future Plan

Fei Chen

Research Application Laboratory, NCAR

NCARKevin ManningMukul Tewari Jimy DudhiaDave Gochis

Peggy LeMone

NCEPKen MitchellMichael Ek

AFWAJohn EylanderJerry Wegiel

NRLTeddy Holt

UniversitiesDev Niyogi (Purdue)

Eric Small (CU)Venkat Laskhmi (USC)Alexis Lau (HKUST)

Other InstitutionsHiroyuki Kusaka (CREPI, Japan)

Bill Coirier (CFD Res. Corp)

5th WRF LSM Workshop, NCAR, 9/13/05

WRF R&D aims

Priority for 1-10 km grid applicationsAdvanced data assimilation and model physicsPortable and efficient on parallel computers

Well-suited for a broad range of applications

Community model with direct path to operations

5th WRF LSM Workshop, NCAR, 9/13/05

Development and Implementation of the Community Noah Land Surface Model

• Collaborative Noah LSM development effort: NCEP, NCAR, U.S. Air Force Weather Agency, university community

• Support WRF mission and community requirements for high-resolution forecast of:– Severe weather (flash flooding) – Air pollution, transport and dispersion of toxic chemicals– Conditions affecting surface and air transportation– Extreme-temperature in the energy/utility industry

loading– Agricultural applications (irrigation management, pest

control)

5th WRF LSM Workshop, NCAR, 9/13/05

Mission

• Integrate advance land models/modules, new data sets, and land data assimilation techniques to improve representation of land and boundary layer processes in WRF

5th WRF LSM Workshop, NCAR, 9/13/05

Major Milestones in WRF/LSM Efforts

• SI include background fields (FY 2001): – 30-second global USGS 24-category landuse map– 30-second global hybrid (30-sec for CONUS and 5-min elsewhere) top and

bottom soil texture– 1-deg annual mean air temperature as lower boundary temperature for soil– NESDIS 0.144-deg monthly 5-year climatology green vegetation fraction– NESDIS 0.144-deg monthly 5-year climatology albedo

• SI Initialize soil moisture, temperature, snow, and sea-ice from AVN and Eta/EDAS (FY 2001)

• Implementation of OSULSM (WRF release 1.2 Beta, FY 2001)

• Inclusion of OSULSM for idealized WRF cases (available for mass version, FY 2001)

5th WRF LSM Workshop, NCAR, 9/13/05

Major Milestones in WRF/LSM Efforts

• Since 1 Dec. 2001: WRF/OSULSM coupled model has been running in real-time at NCAR

• SI add capability to read ARGMET soil fields as initial land state conditions (FY 2002)

• Implementation of FSL RUC LSM (FY 2002)• NCEP group delivers the quasi-unified Noah LSM for

internal test (FY 2002)• SI add nearest neigh approach for interpolation of external

landuse and soil texture, and land state variables (FY 2002)

5th WRF LSM Workshop, NCAR, 9/13/05

Major Milestones in WRF/LSM Efforts• Implement new surface driver (FY 2003)

– Separated from PBL driver

• SI add Maximum snow albedo database (FY 2003)• Implement quasi-unified Noah LSM (FY 2003)

– Snow and frozen-ground physics– Soil thermal conductivity– Patchy snow cover– Snow density– Soil heat flux treatment under snow pack– Snow roughness length

5th WRF LSM Workshop, NCAR, 9/13/05

Major Milestones in WRF/LSM Efforts• Implement the unified Noah LSM in WRF 2.0 and set up

CVS for Noah support (FY 2004) • Enhance the unified Noah LSM (FY 2004)

– Seasonal varying surface emissivity– Simple urban landuse treatment

• Apply high-resolution land data assmilation (HRLDAS) to ARW realtime 5-km winter and 4-km summer experiments

• Land surface models currently in test (FY 2004)– CLM– Pleim-Xu LSM – Single layer urban canopy model (close to be released)

5th WRF LSM Workshop, NCAR, 9/13/05

Provide documentation for the unified Noah LSM for its offline and coupled application

The unified Noah LSM significantly improved the precipitation scorecompared to its predecessor OSULSM

Realtime 22-km CONUS 12Z Cycle initialized from 40-km EDAS

12 day 12-36 h forecasted rainfall from 15 to 31 May 2003 verified on #212 grid

Unified Noah LSM

OSULSM

Precip scores availableat NSSL website

WRF/Noah Snow Forecast CapabilitySnow Storm Case 18 March 2003

24-h snow water equivalent change valid at 06Z 19 March

Analysis: 24-h SWE change valid at 06Z 19 March

Snow meltedtoo quickly inthe OSULSM

accumulationmelt/sublimation

WRF/Noah more realistic snow forecast

5th WRF LSM Workshop, NCAR, 9/13/05

Landuse Based Verification for 15 June 2005

2m Temperature Wind Speed

RED: Dryland Cropland, Blue: Cropland, Black: GrasslandGreen: Shrubland, Yellow: Evergreen Needleleaf Forest

High-resolution Land surface and urban modeling and assimilation system

Vegetation type

Soil texture

Urban type

Terrain

snow

Vegetation cover

Leaf area index

High resolution land data assimilation system (HRLDAS)

Obs. PrecipitationRadiation, T, Q, U, V

Soil moisture, soil temperature, snow cover, canopy water, wall/roof/road temperature

Noah land surface model, Urban canopy model

Boundary layer parameterization Coupled mode

HRLDAS Spin-Up

Sensible heat flux Latent heat flux

≤ ≤5 5 < 10 10 <Surface heat flux RMS difference (Wm-2)

Coarsesoil

Medium soil

Fine soil

Hovmoller Diagrams of Rainfall on 4-km WRF GridTime: 12Z 9 June - 12Z 21 June 2002 Longitude: from 107 W to 86 W

Continuously cycled soil moisture Updated daily with HRLDAS soil Difference (HRLDAS-Cycled)

Both experiments able to simulate propagating convection. Soil conditions influence timing of location of storm origin.

5th WRF LSM Workshop, NCAR, 9/13/05

Some solutions: Dynamic modeling of land-Some solutions: Dynamic modeling of land-surface hydrology with ‘surface hydrology with ‘Noah-RouterNoah-Router’’

(NCAR Tech Note: Gochis and Chen, 2003)(NCAR Tech Note: Gochis and Chen, 2003)

2-Dimensional2-DimensionalDiffusive WaveDiffusive Wave

Overland Flow RoutingOverland Flow RoutingOgden, 1997Ogden, 1997

Saturated Subsurface RoutingSaturated Subsurface RoutingWigmosta et. al, 1994Wigmosta et. al, 1994

Ponded Water EvaporationPonded Water Evaporationand Re-infiltrationand Re-infiltration

Direct EvaporationDirect Evaporation

Re-infiltrationRe-infiltration

Surface Exfiltration fromSurface Exfiltration fromSaturated Soil ColumnsSaturated Soil Columns

Lateral Flow fromLateral Flow fromSaturated Soil LayersSaturated Soil Layers

5th WRF LSM Workshop, NCAR, 9/13/05

Terrain and Land-use of the Model Domain with 10-km grid spacing

1200km

Sea of Japan

Pacific Ocean

5th WRF LSM Workshop, NCAR, 9/13/05

Terrain and Land-use around Tokyo

Tokyo Metropolitan area

5th WRF LSM Workshop, NCAR, 9/13/05

2-m Temperature on 1500 LT: Obs and WRF (Case UCM)

5th WRF LSM Workshop, NCAR, 9/13/05

2-m Temperature on 0300 LT: Obs and WRF (Case UCM)

High-Resolution WRF/Noah/Urban Modeling Capability

Complex terrain on WRF nested D-5 (0.5 km grid spacing) over the Slat Lake City area Complex Urban land use

distribution over Salt Lake City

Single layer Urban Canopy Model

Domains: 40.5,13.5,4.5,1.5,0.5 km

High-Resolution WRF/Noah/Urban Modeling Capability: Coupled to CFD-Urban

WRF-Noah/UCM coupled model forecast

Down-Scale

Up- Scale

Coupling

•

CFD-Urban:Hi-Res Urban Model

Diurnal Wind Direction at North Downtown

NIGHT DAYRed: Obs, Green: WRF/Noah, Blue: WRF/Noah/UCM

5th WRF LSM Workshop, NCAR, 9/13/05

• Entire IOP 10

• 3 Releases/Pauses

• WRF Data for BC

•Quasi-steady approach:

• Wind/Turbulence fields at 15 minute intervals

• Unsteady T&D using Unified Frozen Hydro Solver

• Flow turning is replicated, which causes plume to travel NNW

WRF/UCM - CFD Transport and Dispersion Preliminary Results: Urban IOP 10 Urban 2000

5th WRF LSM Workshop, NCAR, 9/13/05

WRF/UCM - CFD Transport and Dispersion Preliminary Results: Urban IOP 10 Urban 2000

• Urban 2000: Field Test conducted in Salt Lake City

• SF6 released in Central Business District

• Samplers located in CBD and on “arcs” located downstream

•Statistical Comparison of Predicted to Measured Concentration Data

Acceptable values:

• FAC2 > 0.5

• -0.3 < FB <0.3 (0.7 <MG < 1.3)

• NMSE < 4 (VG <1.6)

5th WRF LSM Workshop, NCAR, 9/13/05

Preliminary Results: IOP 10 Urban 2000• Three sets of calculations:

• Raging Waters Input: Use sounding data (single sounding) at all boundary faces

• WRF Boundary Conditions: Unsteady Flow, Turbulence and Contaminant

• WRF Boundary Conditions: Steady Flow (“Wind Library”)Near Source R2 R3 R4 All

FAC2: RW 0.12 0.17 0.36 0.38 0.18FAC2: Unsteady 0.08 0.17 0.36 0.38 0.16FAC2: Quasi-Unsteady 0.57 0.42 0.36 0.5 0.51MG: RW 25.42 14.11 4.58 5.06 15.83MG: Unsteady 15.89 11.64 4.77 5.679 11.69MG: Quasi-Unsteady 0.74 1.59 1.96 2.05 1.04

• Quasi-Steady approach appears to be best mode of operation:

• Steady-state wind/turbulence fields at set intervals in time using WRF data as boundary conditions

• Significantly improve FAC2 (FAC>0.5 acceptable) and MG (0.7<MG<1.3 acceptable)

•Unsteady flow/turbulence/transport: Time step restrictions

•Too costly for accuracy or inaccurate because timestep too big

5th WRF LSM Workshop, NCAR, 9/13/05

Summary• We have achieved main tasks we set four years ago• Unified Noah LSM modifications currently tested at NCAR

– Vary LAI as landuse type and scale it by vegetation fraction– Define all vegetation parameters in VEGPRM.TBL (part of them are

now defined in LANDUSE/TBL) – Adjust parameters (emissivity, roughness length, albedo)

• Need to incorporate new satellite data • Further changes/improvements brought up at the workshop• Need more systematic evaluation of coupled model (surface

heat fluxes, radiation, for instance) • Implement the unified Noah in NMM, AGRMET, COAMPS