Plans for a new land surface model for NCEP operations and for use

Report
New Directions for WRF
Land Surface Modeling
Michael Barlage
Research Applications Laboratory (RAL)
National Center for Atmospheric Research
Polar WRF Workshop – 3 November 2011
1
Noah LSM in NCEP Eta, MM5 and WRF Models
(Pan and Mahrt 1987, Chen et al. 1996, Chen and Dudhia 2001,
Ek et al., 2003)
Canopy Water
Transpiration Evaporation
Turbulent Heat Flux to/from
Snowpack/Soil/Plant Canopy
Precipitation
Condensation
on
vegetation
Direct Soil
Evaporation
Runoff
on
bare
soil
Deposition/
Sublimation
to/from
snowpack
Evaporation
from Open Water
Snowmelt
D Z = 10 cm
Soil Moisture
Flux
Interflow
Soil Heat Flux
D Z = 30 cm
Internal Soil
Moisture Flux
D Z = 60 cm
Internal Soil
Heat Flux
D Z = 100 cm
Gravitational Flow
2
Noah LSM in NCEP Eta, MM5 and WRF Models
(Pan and Mahrt 1987, Chen et al. 1996, Chen and Dudhia 2001,
Ek et al., 2003)
Canopy Water
Transpiration Evaporation
Turbulent Heat Flux to/from
Snowpack/Soil/Plant Canopy
Precipitation
Condensation
on
vegetation
Direct Soil
Evaporation
Runoff
on
bare
soil
Deposition/
Sublimation
to/from
snowpack
Evaporation
from Open Water
Snowmelt
D Z = 10 cm
Soil Moisture
Flux
Interflow
Soil Heat Flux
D Z = 30 cm
Internal Soil
Moisture Flux
D Z = 60 cm
Internal Soil
Heat Flux
D Z = 100 cm
Gravitational Flow
3
Noah LSM Performance
• Noah does some things well
– Surface fluxes without snow
present
– Summertime simulation in general
– Noah is relatively simple, less
parameters
• Noah structure good for
satellite-derived surface
properties
– Albedo, observed from satellite, is
a bulk property (vegetation, snow,
soil)
– Vegetation properties like green
vegetation fraction are easily used
as prescribed vegetation condition
4
4
Noah LSM Deficiencies
• Related to Snow Physics
– Combined snow/vegetation/soil layer
– No explicit canopy or liquid water retention
– Currently one-layer snow
• Results in:
– Under-prediction of snow throughout season
– Snow melts too early in spring
– Surface skin temperature is limited to (near) freezing with
snow on ground (cannot produce a “warm” canopy)
– Limits 2m temperature in cases of warm air advection and
when significant energy absorbed by canopy
5
Noah LSM Deficiencies
Flagstaff WRF T2m
simulation compared
to METAR
observations
February
Courtesy Mike Leuthold, U. Arizona
6
Noah LSM Deficiencies
Flagstaff WRF T2m
simulation compared
to METAR
observations
• Cold bias during the day results from
capped surface temperature at freezing
February
• Bias recovers during the night
• When snow is gone, bias is low
7
Noah LSM Deficiencies
Flagstaff WRF T2m
simulation compared
to METAR
observations
• Cold bias during the day results from
capped surface temperature at freezing
February
• Bias recovers during the night
• When snow is gone, bias is low
8
Noah LSM Deficiencies
Flagstaff WRF v3.2
T2m simulation
compared to METAR
observations
• Cold bias during the day results from
capped surface temperature at freezing
February
• Bias recovers during the night
• When snow is gone, bias is low
9
Simulations compared to SNOTEL observations
Modified Noah
SWE, snow melt and
sublimation between the
control simulation and
simulation with all changes
Sublimation reduced
consistently throughout
simulation
Noah v3.0
Resulting pack increase melts
in spring
Legend legend
GS: GOES SW forcing
ML: model level forcing
LV: Livneh albedo
TA: terrain adjustment
CH: WRF MYJ stability
85: Max albedo = 0.85
10
ZE: Zo = f(exposed veg)
Simulations compared to Niwot Ridge observations
Diurnal average
sensible heat flux for
January 2007
Both Noah-MP and
Noah-UA do better
with fluxes at night
Noah-MP does very
well with daytime flux
Noah-UA improves
greatly upon both
version of current
Noah
Keep snow at the expense of energy
11
11
Addressing with Two Approaches
Noah-UA
• Wang et al. 2010
– Canopy shading effect
– Reduce exchange
coefficient under canopy
– Adjust roughness length for
snow and vegetation
fraction
– Additional snow cover
fractions
• Advantages
– Easy to implement
– Maintains Noah structure
(added as namelist option)
• Disadvantages
– Skin temperature still limited
Noah-MP
• Liang/Niu et al. 2011
–
–
–
–
Explicit canopy
Multiple snow layers
Snow liquid water retention
Two-stream canopy
radiation
– Multiple temperatures
• Advantages
– More physical surface
representation
– Surface exchange
consistent with LSM
• Disadvantages
– Complexity/cost
– More parameters
12
Noah-UA: Canopy Shading
Noah
SWdn
Noah-UA
SH
SWdn
SH + Δcan
Δcan
(1-α)SWdn
Δcan = solar radiation
intercepted
by canopy
= f(LAI, canopy
reflectance,
snow
albedo)
(1-α)SWdn- Δcan
13
Noah-MP: Canopy Fluxes
Canopy Fraction
Bare Fraction
• Separate exchange coefficients
–
–
–
–
Bare ground to atmosphere
Under-canopy ground to canopy
Canopy to atmosphere
Leaf to canopy
• Flux balance
– Iterate leaf and canopy temperatures so that heat flux to atmosphere is
balanced with flux from canopy to leaf and canopy to ground
14
Simulations compared to SNOTEL observations
Modified Noah
Noah-MP
Noah-MP improves both
peak SWE simulation and
spring melt timing
Noah v3.1+
Noah
Noah v3.1
15
Simulations compared to Niwot Ridge observations
Diurnal average
sensible heat flux for
January 2007
Both Noah-MP and
Noah-UA do better
with fluxes at night
Noah-MP does very
well with daytime flux
Noah-UA improves
greatly upon both
version of current
Noah
16
16
Simulations compared to Niwot Ridge observations
Diurnal average
sensible heat flux for
January 2007
Both Noah-MP and
Noah-UA do better
with fluxes at night
Noah-MP does very
well with daytime flux
Noah-UA improves
greatly upon both
version of current
Noah
17
17
Coupling Noah-MP to WRF
•
•
•
•
•
•
Noah-MP is coupled to WRF and currently
going through testing
12 Km horizontal resolution with
NARR data is used as initial condition
WRF Runs starts 1 March 2008, 12Z
– Using WRFV3.3/Noah
– Using WRFV3.3/Noah-MP
Models are integrated for 15 days.
Results are compared
–
Noah vs Noah-MP
Sensible Heat Flux at Niwot Ridge, CO
Noah-MP
Noah
Obs
Snow Model Intercomparison
• Coordinated effort by NCAR to compare surface processes within snow
components of land models
• Volunteer participation by several universities
• Phase-1a: Control experiment at SNOTEL sites. All forcing comes from WRF
simulation except GOES observed solar radiation
• Phase-1b: Same as Phase-1a except daily precipitation from SNOTEL
observations
• Phase-1c: Same as Phase-1b except diurnal hourly precipitation
distribution is based on WRF monthly-averaged diurnal distribution
• Phase-1d: Same as Phase-1a except that SWE is reset to SNOTEL observed
SWE on the date of maximum
• Phase-2a: 2004-2008 simulations for AmeriFlux sites (Niwot Ridge and
GLEES). Forcing comes from NARR except precipitation(NLDAS) and solar
radiation
–
–
–
–
Phase-2a1: Replacing the 2m Temperature forcing data with the 21m forcing.
Phase-2a2: Ameriflux SW/LW replacing GOES/NARR SW/LW (no obs 2004-2005)
Phase-2a3: 2a1+2a2
Sensitivity with forcing height (ZLVL)
20
Snow Model Intercomparison
LEAF
VIC
SAST
CLM
Noah
NoahMP
Snow Model Intercomparison
LEAF
VIC
SAST
CLM
Noah
NoahMP
Summary
• Other Noah-MP features
– Dynamic vegetation
– Groundwater treatment
– Photosynthesis-based canopy resistance
• A new model (Noah-MP) and new processes within
the existing Noah (Noah-UA) are planned to be
released in the next WRF release
– Both models attempt to address Noah deficiencies in snow
treatment
– Noah-MP contains several options for physical
parameterizations within the LSM
23

similar documents