Implementing GNR analysis on the NI PXI7831R FPGA

1
Navys MURI Impact on UW Hyperspectral Activities

Allen Huang Cooperative Institute for
Meteorological Satellite Studies (CIMSS) Space
Science Engineering Center (SSEC) Univ. of
Wisconsin-Madison 5th Workshop on
Hyperspectral Science of UW-Madison MURI,
Airborne, LEO, and GEO Activities   The Pyle
Center University of Wisconsin?Madison 702
Langdon Street, Madison (608-262-1122)   7-9 June
2005
2
UWs road to the Hyperspectral (Next Generation)
Sounders
HES (1600 GEO O)
GIFTS (1600 GEO E)
CrIS (2215 LEO O)
UW has played a significant roles in the past,
current, and future Hyperspectral Sounders
(labeled in green)
IASI (8000 LEO O)
AIRS (2200 LEO E)
NAST-I (8220 Airborne)
IMG (18400 LEO E)
S-HIS (4840 Airborne)
GOES Sounder (18 GEO O)
( of spectral bands) O Operational E
Experimental
HIS (4492 Airborne)
VAS (12 GEO O)
VTPR, HIRS (18 LEO O)
IRIS (862 LEO E)
Time
1978
2012
3
UWS Hyperspectral End-to-End Simulation Effort
Mesoscale Modeling
Radiative Transfer Modeling
FTS Simulator
Interferograms
Profiles Clouds Surface temp Wind
Top of Atmosphere radiances
Compression
Trade Study
Compressed Data (Rad. Counts)
Profile Tracking
Instrument Design Compression Impacts
Wind
Validation
Off-Axis Normalization
Calibration
Retrieval
Normalized INFGs
Spectra
Profiles
Outputs
4
Navys MURI Impact on UW Hyperspectral Activities
5
Navys MURI Impact on UW Hyperspectral Activities
Current UW Direct Broadcast End-to-End Processing
Capability
6
Single-scattering Properties of Ice
Crystals--Database and parameterization

• Yang, P., H. Wei, H.-L. Huang, B. A. Baum, Y.
  X., Hu, G. W. Kattawar, M. I. Mishchenko, and Q.
  Fu, 2004 Scattering and absorption property
  database for nonspherical ice particles in the
  near- through far-infrared spectral region,
  Appl.Opt. (accepted).

7
Bulk Scattering Models Available for Multiple
Instruments
Provide bulk properties (mean and std. dev.)
evenly spaced in Deff from 10 to 180 ?m
for asymmetry factor phase function single-scat
tering albedo extinction efficiency cross
sections IWC Dm Models available at
http//www.ssec.wisc.edu/baum for IR Spectral
Models (100 to 3250 cm-1) MODIS AVHRR AATSR MI
SR VIRS MAS (MODIS Airborne Simulator) ABI
(Advanced Baseline Imager) POLDER
(Polarization) SEVIRI (Spinning Enhanced Visible
InfraRed Imager)
8
UW Hyperspectral Sounder Simulator Processor
(HSSP) Simulator - Radiance and Model Component
9
UW Hyperspectral Sounder Simulator Processor
(HSSP) Simulator - Radiance and Model Component

• Effect/Feature Included
  Notes
• Cloud Microphysics
  yes Measurements, NWP model output
• Single Scattering Parameterization
  Partial ongoing effort
• DISORT yes ongoing effort
• Cloud Layer Albedo Transmittance Par.
  Partial ongoing effort
• Fast Cloudy RT Model Partial
  under development
• Atmospheric profile data base yes
• LBLRTM yes
• Water Vapor Spectroscopy
  yes ongoing effort
• Fast Clear RT Model yes PLOD
• Adjoint operator yes MATLAB version
• Tangent Linear
  yes MATLAB version
• Ocean Surface Emissivity Model yes IRSSE Model
  (Van Delst)
• Land Surface Emissivity Model not yet under
  development
• Aerosol Parameterization not yet under
  development
• Solar Spectrum not yet
• RT Model validation partial ongoing effort
• RT Model consolidation no
  coordination PLOD RTTOV OPTRAN OSS
  Mesoscale NWP MODEL yes MM5 and WRF
• Improved Cloud Physics in NWP no
  cloud spectral bin modeling

10
Radiative transfer modeling of atmospheric gases
absorption
Surface Type

• LBLRTM based PLOD fast model
• LBLRTM runs
• HITRAN 96 JPL extended
• spectral line parameters
• CKD v2.4 H2O continuum
• Spectral Characteristics
• 586-2347 cm-1
• 0.8724 cm MOPD
• Kaisser Bessel 6 apodization
• Fast Model
• 32 profiles from
• NOAA database
• 6 view angles
• AIRS 100 layers
• Fixed, H2O, and O3
• AIRS PLOD predictors
• Run time
• 0.8 Sec on a 1 GHz CPU

Ozone
Temp.
Dust/Aerosol
CO
Temp.
Water Vapor
11
Radiative transfer approximation of single cloud
layer model
12
Two layer cloud model from Texas AM coupled with
UW/CIMSS clear-sky model
3 ice cloud models, 1 water cloud model 100-3246
1/cm (3-100 um)
Tropical De 16-126 um
Mid-latitude De 8-145 um
Polar De 1.6-162 um
Water-spheres De 2-1100 um
13
A fast infrared radiative transfer model (FIRTM2)
for overlapping cloudy atmospheres

• Niu, J., P, Yang, H.-L. Huang, J. E. Davies, J.
  Li, B. A. Baum, and Y. Hu, 2005 A fast infrared
  radiative transfer model for overlapping cloudy
  atmospheres. J. Quant. Spectroscopy Radiative
  Transfer (to be submitted).

14
How to extract the cloud information?

• AIRS sub-pixel cloud detection and
  characterization using MODIS data (Li et al.
  2004a)
• Cloud property retrieval from AIRS radiances (Li
  et al. 2004b 2005) with the help of MODIS

15
An Aerosol Database
16
UW Hyperspectral Sounder Simulator Processor
(HSSP) Simulator - Sensor Component

• Effect/Feature Included
  Notes
• Instrument Emission yes
• Instrument Responsivity yes
• Numerical Filter yes filter function set to
  unity
• Instrument Phase yes varies linearly with n
• Phase variation across FPA not yet
• Off-axis OPD sampling yes
• ILS variations yes
• pixel-to-pixel offset variations
  yes 12(LW), 5(SMW) random variation
• pixel-to-pixel gain variations yes 8-40(LW),
  2-5(SMW) of full well depth
• pixel operability not yet
• FPA center not aligned with FTS axis yes 1-2
  pixels, non integer
• LW/SMW FPA misalignment
  no retrieval issue
• Detector non-linearity no small
• Detector noise yes
• Photon noise yes
• Quantization noise yes
• OPD scan mirror velocity variation no small
• OPD scan mirror tilt no small

17
UW Hyperspectral Sounder Simulator Processor
(HSSP) Processor - Measurement
Retrieval/Product Component

• Effect/Feature Included
  Notes
• Calibrated radiances yes generate
  sensor spectral measurements
• Geo-location yes based on nominal geo orbit
• Total sensor noise yes mainly random detector
  noise
• Diffraction blur partial
  simulated to demonstrated band to band reg. Error
  effect
• 4-km sampling yes
  MM5 meso-scale run
• 15 to 30 minutes sampling yes MM5
  meso-scale run
• Clear radiances
  yes Latest PLOD fast clear model run
• Cloudy radiances yes Water Ice Clouds
  (includes size effect)
• Aerosol/Dust radiances not yet Extinction
  modeling underdevelopment
• Ocean emissivity
  yes IRSSE model
• Land emissivity
  not yet underdevelopment (UH-UW)
• Clear regression retrieval yes demonstrated by
  simulation, air/space borne
• Clear physical retrieval yes developed under
  testing
• Cloudy retrieval down to cloud level
  partial demonstrated by simulation and airborne
• Cloudy retrieval transparent clouds not
  yet under design
• Altitude resolved water vapor wind yes demonstrate
  d by simulation and airborne
• 3D water vapor wind not
  yet under development
• Cloud detection partial
  under development

18
AIRS Std. Operational Product
CIMSS
19
AIRS/MODIS Synergistic C.C. can Supplement
AIRS/AMSU C.C. Especially over Desert Region
AIRS/AMSU C.C. (3 by 3 AIRS FOV) V4.0 -
Blue AIRS/MODIS C.C. (1 by 2 AIRS FOV) Multi-Ch.
- Black Single-Ch. Band 31 Green Band 22 - Red
South Africa Granule
20
(No Transcript)
21
AIRS Absolute Emissivity
Ozone Not Fit
Atm. Corr. Relative IR Emiss
Absolute IR Emiss

• Squares are using 281 Select AIRS channels only.
  It Works !!!

22
July 2003
12 ?m Emissivity
MODIS
AIRS
AIRS - MODIS
23
Altitude Resolved Water Vapor Wind Demonstration
GIFTS - IHOP simulation 1830z 12 June 02  
GOES-8 winds 1655z 12 June 02 
Simulated GIFTS winds (left) versus GOES current
oper winds (right)
24
(No Transcript)
25
Selecting Computing Hardware

• Cluster options were evaluated and found to
  require significant time investment.
• Purchased SGI Altix fall of 2004 after extensive
  test runs with WRF and MM5.
• 24 - Itanium2 processors running Linux
• 192GB of RAM
• 5TB of FC/SATA disk
• Recently upgraded to 32 CPUs, 10TB storage.

26
Model Configuration

• 42 hr simulation initialized at 1200 UTC 23 June
  2003
• 290 x 290 grid point domain with 4 km horizontal
  spacing and 50 vertical levels

MM5
WRF

• Goddard microphysics
• MRF PBL
• RRTM/Dudhia radiation
• Explicit cumulus convection
• OSU land surface model

• WSM6 microphysics
• YSU PBL
• RRTM/Dudhia radiation
• Explicit cumulus convection
• NOAH land surface model

27
Global training database for hyperspectral and
multi-spectral atmospheric retrievals Suzanne
Wetzel Seemann, Eva Borbas Allen Huang, Jun Li,
Paul Menzel
28
Non-dimensional Tb Sensitivity to Atmospheric
Temperature (Thermal Source only)
Clear sky
Cloudy
29
Data Compression Demonstration
30
Ground Segment Processing Demonstration

• GIPS Design Elements
• Monitoring, Control, and Data Channels
• Parallel Processing Pipeline Architecture
• Modular Software Component Design

31
Navys MURI Impact on UW Hyperspectral Activities

• Itemized Impacts
• Physical Modeling
• Clear Sky RTE Forward Model Enhancement/Improvemen
  t
• Cloud/Aerosol Microphysical Property Database
  Development
• Cloudy Sky RTE Forward Model Development
• Surface Property
• High-spatial Resolution NWP Model Simulation
• Sensor Measurements Simulation
• Level 0 to Level 1 and Level 1 to Level 2
  Processing Algorithm Development Demonstration
• Hyperspectral/Multispectral Synergy
• Hyperspectral/Multispectral Applications
• Hyperspectral Science Education Training

32
Navys MURI Impact on UW Hyperspectral Activities
Overall Impact Every Element of a Truly
End-To-End Infrastructure Under Construction at
SSEC/CIMSS of UW-Madison in Support of NPP/NPOESS
GOES-R Activities Through Three-Pillar
Partnership
33
Monday-Thursday 1-4 August 2005 Atmospheric and
Environmental Remote Sensing Data Processing and
Utilization Numerical Atmospheric Prediction and
Environmental Monitoring   330 to 530 pm Monday
1 August 2005
Panel on Three-Pillar Partnership in Remote
Sensing the Roles of Government, Industry, and
Academia Moderator James F. W. Purdom, Colorado
State Univ. Paneilists Philip E. Ardanuy,
Raytheon Technical Services Co. LLC Michael J.
Crison, Colleen Hartman, National Oceanic and
Atmospheric Administration Henry E. Revercomb,
Univ. of Wisconsin/Madison Steven W. Running,
Univ. of Montana Merit Shoucri, Northrop Grumman
Space Technology Tentative commitments at time
of publication, subject to change.   This panel,
organized by the track and conference chairs of
the Remote and In Situ Sensing program track,
offers the opportunity to discuss the roles of
government, industry, and academia in the era of
NPOESS and GOES-R, these being our nations
preeminent environmental satellite programs in
the coming decades. The revolution in the last 40
years to date in remote sensing that has taken
place in the United States could not have
occurred without the closest cooperation between
these three pillars.   The unrelenting growth in
processing complexity and measurement data
volume, arising from maturing environmental
satellite systems, triggered NOAA and NASA to
jointly task the National Academy of Sciences to
conduct an end-to-end review of current
practices, including characterization of process
weaknesses, assessment of resources and needs,
and identification of critical factors that limit
the optimal management of data including the
strategic analysis for maximum environmental
satellite data utilization. The Committee on
Environmental Satellite Data Utilization (CESDU)
was formed in early 2003 to respond to this
charge.   CESDU recommended a partnership
strategy between the government, industry, and
academia (the CESDU report is available from
http//www.nap.edu/openbook/0309092353/html/1.html
). This three-pillar partnership strategy was
identified as a significant factor in the success
of ozone retrievals in a CESDU case study. The
strategy for future system acquisitions will be
discussed in light of these recommendations.   Sho
rt Presentations on Government
Perspective Industry Perspective Academia
Perspective National Academy of Sciences CESDU
report   Key Discussion Issues Contention Only
a fully integrated team--a joint three-pillar
partnership--working together in a seamless
manner with a relentless determination to excel,
will achieve total user satisfaction and
comprehensive data utilization. Examples from
the past NPOESS partnerships
GOES-R partnerships.