ATMOSPHERIC INFRARED SOUNDER AIRS AND REGIONAL DATA ASSIMILATION

1
ATMOSPHERIC INFRARED SOUNDER (AIRS) AND REGIONAL
DATA ASSIMILATION Brad Zavodsky University of
Alabama Huntsville Huntsville,
Alabama Shih-hung Chou, Gary Jedlovec, Bill
Lapenta NASA / Marshall Space Flight
Center Huntsville, Alabama Joint NASA/NOAA
SOO/SPoRT Workshop July 11-13, 2006
2
Motivation

• Advanced Infrared Sounder (AIRS) is a prototype
  instrument for the CrIS (NPOESS) and HES
  (GOES-R)methodologies for AIRS data will be
  relevant to these upcoming instruments
• The hyperspectral nature of AIRS results in the
  highest quality remotely sensed vertical
  soundings
• AIRS provides complimentary data at locations
  and times at which traditional upper air data are
  not available
• Important for coastal offices where knowledge
  of upper air conditions over ocean is generally
  confined to large-scale model fields

0700 UTC 20 November 2005
3
AIRS Specifications

• Aboard Aqua polar orbiter
• early afternoon equator crossing
• data twice daily over SEUS at 0700 and 1800 UTC
  ( 1 hr)
• radiances 15 km spatial resolution (2378
  spectral channels 3.7 15.4 µm)
• profiles 50 km spatial resolution (3x3
  footprints 53-55 levels)
• coupling with AMSU allows for retrievals in
  clear and cloudy scenes
• verification studies have shown accuracies of
  1K in 1km layers and 10-15 RH in 2km layers in
  troposphere

4
Possible Uses for AIRS Data

• AIRS profiles could be used in a similar
  fashion to rawinsondes over oceans with the
  caveat that temperature and dew point represent
  layer averages rather than point measurements

• AIRS radiances can be assimilated in regional
  modelsJCSDA currently uses AIRS radiances in
  global modeling
• assimilating AIRS profiles is a more
  straightforward and less computationally rigorous
  method than the use of radiances
• SPoRT can provide to WFOs
• - an AIRS profile product
• - improved ICs (w/ AIRS) for local model runs
• - a version of WRF including AIRS

AIRS provides better vertical resolution than
model sounding
0700 UTC 20 November 2005 (30.51oN, 72.34oW)
5
AIRS Data Distribution and Quality

• Horizontal Distribution
• relatively narrow swath of data
• maximum 2 daily overpasses
• Vertical Distribution
• clouds
• land emissivity (surface/low level)
• Use maximum quality data
• quality indicators (QIs) allow for the removal
  of questionable data near surface or at/below
  cloud level (preprocessed with missing values)
• intelligent use of the QIs allows for the
  maximum amount of quality data to be assimilated
  thereby producing the best possible initial
  conditions for the WRF

0700 UTC 20 November 2005 AIRS swath
6
Case Study November 20-22, 2005

• rapidly intensifying storm in region relevant to
  SPoRT interests
• under forecasted by operational models at storms
  strongest pointpotential impact
• ample verification data available over the
  Eastern US (minimal terrain issues)

WRF Domain for November 2005 Case Study
L
L
L
7
ADAS Analysis with AIRS Profiles
700 hPa temperature (oC) at 0700 UTC 20 November
2005
analysis - background
WRF 1h forecast
analysis of AIRS data
ADAS analysis

• general pattern of AIRS analysis similar to
  backgroundadds validity to AIRS obs
• analysis 1 2oC warmer than background for
  much of Atlantic at 700 hPa
• similar magnitude at most other significant
  levels (some show cooling of 1 2oC)
• analysis of mixing ratio also impacted by AIRS
  data (not shown)

8
53h Forecast ImpactMean Sea Level Pressure
Valid at 1200 UTC 22 November 2005

• low pressure area deepens by more than 4 hPa
  with inclusion of AIRS data
• storm center shifts slightly to the south and
  west
• both alterations are improvements over the
  control when compared to surface analysis

9
41h Forecast Impact6h Cumulative Precipitation
Valid at 0000 UTC 22 November 2005

• subtle precipitation coverage difference
  between control and AIRS
• AIRS seems to better forecast intensity
  (especially over MD/DE/VA)

10
Summary

• What can SPoRT provide to the NWS in terms of
  AIRS data?
• AIRS profile product
• available in data void regions
• more vertical structure than a model field
• Improved initial conditions for local model
  runs
• intelligent use of QIs allows for assimilation
  of maximum high quality data
• assimilation of AIRS profiles yields
  differences of 1 2oC from background
• A version of WRF including AIRS
• position intensity of low in case study
  improved with use of AIRS profiles
• 6h QPF improved at all precipitation thresholds
  with use of AIRS profiles
• Future plans involve running analysis/forecast
  with AIRS radiances using GSI/WRF configuration
  (W. McCarty)

11
The end result of this work is to provide
important information to aid forecasters.
Assistance from local offices in assessing the
usefulness of the AIRS profiles and model runs
with AIRS data would be invaluable in helping us
understand the impact of AIRS data. QUESTIONS/DIS
CUSSION