Modeling Validation with NASTM and a CloudResolving Model at 50430 GHz

1
Modeling Validation with NAST-M and a
Cloud-Resolving Model at 50-430 GHz

• R.V. Leslie, L. J. Bickmeier, W. J. Blackwell,
  and F. W. Chen
• Contributions from C. Surussavadee, P.
  Rosenkranz, and D. Staelin (MIT RLE)
• Paul Bieringer and Jonathan Hurst (MIT LL)
• 3rd International Precipitation Working Group
  Workshop
• Melbourne, Australia
• October 25, 2006

This work was sponsored by the National Oceanic
and Atmospheric Administration under contract
FA8721-05-C-0002. Opinions, interpretations,
conclusions, and recommendations are those of
the author and are not necessarily endorsed by
the United States Government.
2
Outline

• Introduction Reconciling models measurements
• NAST-M instrument
• Numerical Weather Prediction (NWP) model
• Radiative transfer model tuning
• Comparison of simulated data observations
• Summary

3
Reconciliation of Models and Measurements

• Numerical Weather Prediction (NWP) Models
• Given atmospheric state at t0, predict
  atmospheric states at t1, t2,
• Microphysical Electromagnetic Models
• Given microphysical properties of precipitation
  (particle size/abundance of ice, liquid, hail,
  graupel, etc.) calculate radiative properties
  (scattering and absorption of microwave
  radiation)
• High-Resolution Microwave Radiance Measurements
• NAST-M aircraft instrument
• Goal Optimization of the radiative transfer
  model through aircraft validation

Observations
NWP Modeling
Radiative Transfer
4
NPOESS Aircraft Sounding Testbed - Microwave
(NAST-M)

• Cruising altitude 17-20 km
• Cross-track scanning
• Scan angle -65º to 65º
• Swath width of 100 km
• 7.5º antenna beam width (FWHM)
• 2.5 km nadir footprint diameter

Developed by MIT RLE
5
Atmospheric Opacity at Microwave and
MillimeterWave Frequencies

• Four Spectrometers
• 24 Oxygen Channels
• 6 Water Vapor Channels
• Millimeter-wave Propagation Model
• Standard Atmosphere

O2
O2
H2O
O2
6
CRYSTAL-FACE July 11 2002
7
Mesoscale and Cloud Models

• Why use mesoscale models?
• Explicit forecasts of cloud and precipitation
  hydrometeors
• Clouds
• Convective storms
• Detailed initial condition specification
• Terrain
• Land-use
• Meteorological observations
• Approach
• Detailed storm simulations
• Validate with surface radar observations
• Apply satellite radiative transfer algorithms

8
Mesoscale Model v5 (MM5)Parameterizations

• 1 km horizontal resolution
• 32 vertical levels (surface to 100 mb)
• 15 minute resolution output
• Lower/lateral boundary conditions from Rapid
  Update Cycle (RUC-20 km)
• Explicit microphysics (Reisner2 - six phases)
• Boundary layer physics (MRF)
• Radiation scheme (IR SWLW cloud interactions)
• Cold starts ( 2-5 hours before target time)

9
Radiative Transfer Models

• Atmospheric absorption and scattering TBSCAT
• P. Rosenkranz, Radiative Transfer Solution using
  Initial Values in a Scattering and Absorbing
  Atmosphere with Reflective Surface, IEEE
  Transactions on Geoscience and Remote Sensing,
  40(8)1889-1892, Aug. 2002
• Surface emissivity
• Water - fastem
• S. English T. Hewison, A fast generic
  millimetre-wave emissivity model, In Proceedings
  of SPIE, Vol. 3503, 1998
• Land - Used randomly chosen values based on
  measurements from
• F. Weng, et al., A microwave land emissivity
  model, J. of Geophysical Research, Vol. 106, No.
  D17, Sept. 2001

10
Radiative Transfer / NWP Interface Issues
Marshall-Palmer
Mass Density g/m3
US Standard 1976
Sekhon-Srivastava
100 mb
Radius mm
MM5
snow
graupel
Each level requires hydrometeor density per drop
radius
Pressure mb
rain
Mass Density g/m3
11
Histograms of Simulated and Observed Brightness
Temperatures (MIT RLE)
AMSU-A channel 5 (53.60 GHz)
AMSU-B channel 5 (183.317 GHz)
Histograms of MM5-simulated and AMSU-observed
brightness temperatures for twenty-four storms at
15-km resolution (3000 km square)
Surussavadee Staelin July 2005
AMSU Advanced Microwave Sounding Unit
(space-based)
12
Electromagnetic Modeling of Precipitation Example
Frequency Dependence of Particle Type
Sphere
Column
Plate
Fitting for snow F(l)
Fitting for graupel F(l)
Rosette
Ice habits studied (DDSCAT)
Surussavadee Staelin July 2005
F(?) is ice factor (normalized density)
13
Reflectivity Comparison for 11Jul02
Composite radar reflectivity over a GOES visible
image
Simulated reflectivity using MM5 output
14
50.3-GHz TB Image Comparison
Brightness Temperature Kelvin
Simulated (MM5)
Actual (NAST-M)

• Histograms of the images above are on the next
  slide

15
50.3-GHz Histogram Comparison
Simulated Data (MM5)
Actual Data (NAST-M)
Simulations (Precipitation only)
Brightness Temperature Kelvin
16
118.75 /- 3.5-GHz TB Image Comparison
Brightness Temperature Kelvin
Simulated (MM5)
Actual (NAST-M)

• Histograms of the images above are on the next
  slide

17
118.75 /- 3.5-GHz Histogram Comparison
Simulated Data (MM5)
Actual Data (NAST-M)
Simulations (Precipitation only)
Brightness Temperature Kelvin
18
183.31 /- 10-GHz TB Image Comparison
Brightness Temperature Kelvin
Simulated (MM5)
Actual (NAST-M)

• Histograms of the images above are on the next
  slide

19
183.31 /- 10-GHz Histogram Comparison
Simulated Data (MM5)
Actual Data (NAST-M)
Simulations (Precipitation only)
Brightness Temperature Kelvin
20
424.76 /- 2.15-GHz TB Image Comparison
Brightness Temperature Kelvin
Simulated (MM5)
Actual (NAST-M)

• Histograms of the images above are on the next
  slide

21
424.76 /- 2.15-GHz Histogram Comparison
Simulated Data (MM5)
Actual Data (NAST-M)
Simulations w/ precipitation
Brightness Temperature Kelvin
22
Final Thoughts

• Numerical atmospheric and radiative transfer
  modeling capabilities are rapidly expanding.
• Resources needed to further improve and validate
  these models are becoming available on a wide
  scale
• Computational capacity
• Global, high-resolution microwave and
  millimeter-wave observations
• New statistical characterizations of model
  performance can be used to calibrate
  model-generated ground-truth data for retrieval
  simulations, etc.
• Towards all-weather radiance assimilation
• Improved reconciliation of modeled and measured
  radiances is vital
• Study of spectral/spatial/temporal system
  requirements

23
Backup Slides
24
Simulations from 11Jul02
Simulated (MM5)
Actual (NAST-M)
50.3-GHz (50.21-50.39)
Brightness Temperature Kelvin
118-GHz (118.75 /- 3.5)
25
Simulations from 11Jul02
Simulated (MM5)
Actual (NAST-M)
183-GHz (183.31 /- 10)
Brightness Temperature Kelvin
425-GHz (424.76 /- 2.15)