Advancing geostationary infrared sounding, and its application.

1
Advancing geostationary infrared sounding, and
its application.
By 1Gary S. Wade, with input from 1Timothy J.
Schmit, 2James P. Nelson III, 2Jun Li, 2Zhenglong
Li, and 1Robert M. Aune, among others
1NOAA/NESDIS/StAR/CoRP/Advanced Satellite
Products Branch (ASPB) and 2Space Science and
Engineering Center (SSEC), all within the
Cooperative Institute for Meteorological
Satellite Studies (CIMSS) at the University of
Wisconsin-Madison
Presented during visit to the National Weather
Service (NWS) Storm Prediction Center (SPC) in
Norman, OK on 29 May 2009
2
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

3
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

4
All GOES Sounder Spectral Bands
18 UTC on 13 May 2009
5
Infrared electromagnetic spectrum of the
atmosphere
Spectral response functions for the 18 infrared
GOES Sounder bands, separated into long, mid, and
shortwave sections.
Incoming solar radiation peaks in the visible,
near 0.5 um (corresponding to a radiative
temperature of 5780K for the Suns surface)
outgoing terrestrial radiation peaks in the
infrared, near 11 um (corresponding to an average
earth radiative temperature of 255K). from
Weins Law Various gaseous constituents effect
the amount of absorption at the different
wavelengths.
6
All GOES Sounder Spectral Bands
Hourly animation 12UTC on 13 May 2009 12 UTC
on 14 May 2009
7
The Ups and Downs of Progress
8
The Ups and Downs of Satellite Advances
Successful launch
VAS Demo
GUFMEX
EPAC
VAS
NSSFC(KC)
AVE
geo IR sounding
1980 1985
1990
9
Expectations and reality
GOES-R
GOES-R
GOES-R
GOES-R
GOES-R
GOES-8
GOES-4
1980 1994
2015
10
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

11
GOES Sounder Cloud Top Pressure (CTP) DPI
18 UTC on 13 May 2009
12
GOES Sounder Cloud Top Pressure (CTP) DPI
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
13
GOES Sounder Total Column Ozone
Trend of edge patterns of total ozone maxima
(350-400 DU) helps indicate upper level jet
cores (N.B. into S CA and across ME during
overnight hours of 28 May 2008).
14
GOES Sounder Total Column Ozone (TCO) DPI
18 UTC on 13 May 2009
15
GOES Sounder Total Column Ozone (TCO) DPI
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
16
GOES Sounder Total Precipitable Water (TPW) DPI
18 UTC on 13 May 2009
17
GOES Sounder Total Precipitable Water (TPW) DPI
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
18
Statistical assessment of SFOV retrieved TPW
W W W

• For a one year period (01 Sep 2005 31 Aug
  2006), GOES Sounder retrieved Total Precipitable
  Water vapor (TPW) values (in mm), and their first
  guess (GFS) values, were compared with co-located
  (within 11 km) radiosonde observations.

W T T
WTW
Computations generously provided by J. P. Nelson
(CIMSS) and A.S. Allegrino (OPDB).
19
Comparing GOES retrieval trends with other
observations at the ARM-SGP site (1)
Ground-based microwave radiometers at the DOE
Atmospheric Radiation Measurement Southern
Great Plains site provide quality independent
comparison for total column integrated water (aka
TPW).
16 Sep 2006
20
Comparing GOES retrieval trends with other
observations at the ARM-SGP site (2)
14-16 Sep 2006
21
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

22
The retrieval concept (in 25 words or less)

• Perturb the first guess profile (surface reports,
  GFS forecast, and SST),
• and thus, the resulting, calculated radiances
  (via RTE),
• until those calculated radiances match the
  observed radiances
• when satisfied, that perturbed profile is the
  (final) retrieved profile.

Of course, many other issues (instrument noise,
transmittance function accuracy, number of
vertical levels, surface emissivity calculations,
, and obscuring clouds) come into play with
successfully solving the matrix inversion of the
perturbation form of the radiative transfer
equation, simultaneously for temperature and
moisture.
23
GOES Sounder Retrieval System
24
Current GOES Sounder Operational Uses
25
Information content of atmospheric vertical
profiles
Only a hyperspectral sounder (e.g. an IR
interferometer, such as GIFS, AIRS, or IASI)
appears sufficient to provide vertical
resolutions comparable to those from in-situ
radiosondes. Until a Hyperspectral
Environmental Sounder-type instrument is
restored, geostationary satellite sounding will
remain limited.
26
GOES Sounder Lifted Index (LI) Stability DPI
18 UTC on 13 May 2009
27
GOES Sounder Lifted Index (LI) Stability DPI
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
28
An Improved GOES Sounder Retrieval Algorithm
24 Apr 2007
An improved retrieval algorithm has been
demonstrated in Li, Z., J. Li, W. P. Menzel, T.
J. Schmit, J. P Nelson III, J. Daniels, and S. A.
Ackerman (2008), GOES Sounding improvement and
applications to severe storm nowcasting, Geophys.
Res. Lett., 35, L03806, doi10.1029/2007GL032797.
Better stability evolution

• Improvements were made to the Ma et al.
  (1999)-based version of the GOES physical
  retrieval algorithm by employing
• a regression-retrieved temperature and moisture
  first-guess (versus a forecast),
• a real covariance matrix for the first-guess
  (versus a correlation coefficient matrix),
• a new radiance bias adjustment scheme along
  with inverted cone filtering, and
• the PFAAST transmittance model (for
  calculated radiances).

29
Real time algorithm version comparisons
http//cimss.ssec.wisc.edu/goes/realtime/compare/a
nilatestdpicomn.html
GOES Sounder TPW DPI
http//cimss.ssec.wisc.edu/goes/realtime/compare/a
nilatestdpicomd.html
16 UT on 13 Jan 2009
30
Seeking details in retrieval results from
different GOES Sounder algorithms (1)
TPW DPI derived from Li et al (2008)
Old 3x3 FOV version
TPW DPI derived from Ma et al (1999)
An intermediate SFOV version
From routine real-time web display of DPI
comparisons http//cimss.ssec.wisc.edu/goes/realt
ime/compare/anilatestdpicomd.html , delve
deeper, plotting each retrieval (as well as
radiosondes). Example here is 12 UT on 13 Feb
2009.
31
Seeking details in retrieval results from
different GOES Sounder algorithms (2)
Profile retrieved with Li et al (2008)
vs first-guess
vs raob
Profile retrieved with Ma et al (1999)
Example of TPW here for 12 UT on 13 Feb 2009 at
Cocoa Beach, FL shows both versions moving away
from first guess (13.8mm), towards the radiosonde
value (16.8), but not enough (old Ma at 14.7
new Li at 14.5).
vs first-guess
vs raob
32
An Improved GOES Sounder Retrieval Algorithm (1)
The old (Ma 1999) and new (Li 2008) versions of
the GOES Sounder retrieval (vertical profile)
algorithm are compared with total precipitable
water (TPW) data from 12 UT on 06 May 2009 over
the south central US.
http//cimss.ssec.wisc.edu/goes/realtime/scus/beg
in-scus.html
Profile comparisons (in skewT/logP diagrams) are
shown in following slides for Slidell, LA and
Corpus Christi, TX and support the noted
improvement with the Li 2008 data in
significantly reducing the extreme (erroneous)
TPW values produced by the Ma 1999 version. The
next question, then, is do the Li 2008 retrievals
actually improve upon the first-guess?
(Ma 1999)
(Li 2008)
Derived Product Imagery (DPI) of TPW with no
clouds included.
33
An Improved GOES Sounder Retrieval Algorithm (2)
Low level moistening over rated TPW
Minimal low level change excellent (integrated)
TPW
34
An Improved GOES Sounder Retrieval Algorithm (3)
Extreme moistening unrealistic TPW
Modest compensation slightly moist TPW
35
Comparing old and new GOES Sounder Retrieval
Algorithm DPI over the South Central US (setting)
14 May 2009
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
13 May 2009
36
Comparing old and new GOES Sounder Retrieval
Algorithm DPI over the South Central US (1 -
Total PW)
(Ma 1999)
(Li 2008)
18 UTC on 13 May 2009
37
Comparing old and new GOES Sounder Retrieval
Algorithm DPI over the South Central US (1
Total PW)
(Ma 1999)
(Li 2008)
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
38
Comparing old and new GOES Sounder Retrieval
Algorithm DPI over the South Central US (2
Lifted Index)
(Ma 1999)
(Li 2008)
18 UTC on 13 May 2009
39
Comparing old and new GOES Sounder Retrieval
Algorithm DPI over the South Central US (2
Lifted Index)
(Ma 1999)
(Li 2008)
Hourly animation 12 UTC on 13 May 2009 12 UTC
on 14 May 2009
40
3a. GOES profiles from old vs new algorithm (OUN)
Extreme moistening unrealistic TPW
41
3b. GOES profiles from old vs new algorithm (FWD)
Overall drying unrealistic TPW
42
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

43
Hyper-spectral sounding today (from a polar
orbiter) - AIRS
The Atmospheric InfraRed Sounder (AIRS) flies
onboard the NASA Aqua satellite. AIRS is a
grating spectrometer, sensing nearly 2400
channels in the IR (within 3.7 to 15 um). An
accompanying instrument, the Advanced Microwave
Sounding Unit (AMSU), measures earth-atmosphere
radiation at microwave regions (at markedly lower
spatial resolution).
Aqua
13.5 km
Satellite remote sensing for regional/mesoscale
4-D sounding of the atmosphere
geostationary advantage high temporal and good
horizontal resolution polar
orbiting advantage better vertical resolution
and cloud penetration
See http//airs.jpl.nasa.gov/technology/how_AI
RS_works/how_AIRS_works_detail/
44
AIRS granules (ascending only) for one day
for 10 Jun 2008
AIRS Granule
AIRS Start Time
45
Vertical moisture resolution with hyper-spectral
AIRS
46
Retrieved temperature at 700 hPA at 20 UT on 10
Jun 2008
(K)
(K)
GFS forecast (first guess for GOES and AIRS)
GOES Sounder GFS
AIRS GFS
AIRS single field-of-view retrievals in both
clear and cloudy skies
(from Li, Weisz, et al 2009)
47
Retrieved humidity at 700 hPA at 20 UT on 10 Jun
2008
g/kg
g/kg
GFS forecast (first guess for GOES and AIRS)
GOES Sounder GFS
AIRS GFS
AIRS single field-of-view retrievals in both
clear and cloudy skies
(from Li, Weisz, et al 2009)
48
Comparing TPW from AIRS and the GOES Sounder
AIRS (Aqua)
20 UT - 09 Apr 2009
Ma(1999)
Li(2008)
GOES-12
49
Comparing Lifted Index from AIRS and the GOES
Sounder
AIRS (Aqua)
20 UT - 09 Apr 2009
Ma(1999)
Li(2008)
GOES-12
50
Assimilating GOES Sounder products into a
numerical model and generating forecast imagery
(1)
The CRAS (CIMSS Regional Assimilation System)
model starts initializing 12 hours earlier, with
3-hourly GOES insertions, within an NCEP GFS
background for producing the initial 00-hour
analyses.
GOES Sounder SFOV DPI
03 UT 05 Oct 2006
Cloud top pressure (CTP)
51
Assimilating GOES Sounder products into a
numerical model and generating forecast imagery
(2)
Forward radiative transfer equations (RTE) are
used to compute the expected image radiative
temperatures, at a given wavelength, from the
forecast temperature and moisture profiles
predicted by the CRAS.
McIDAS display on web
18 UT 05 Oct 2006
OR
On AWIPS, comparing GOES observed imagery with
CRAS forecast imagery, both valid at the same
time.
18 UT 06 Oct 2006
52
CIMSS Regional Assimilation System (CRAS)
North America CRAS providing Forecast imagery for
NWS AWIPS
Resolution 45 km Sigma levels 38 Grid
Size 280x180 Time-step 200 seconds Forecast
length 84 hours Initialization 12-hr spin-up,
44 scan inserts Output 3 hourly Initial
times 00/12 UTC Start times 0115/1315
UTC Initial conditions 1/2 degree GFS, 6 hr
Forecast Bndry conditions 6 hourly, one degree
GFS Forecasts Input obs Surface, RAOBs GOES
sounder 3-layer PW, cloud-top pressure Verificatio
n RAOBs, Surface, GOES Note Multiple
high-resolution nests can be placed anywhere in
the 45 km grid.
CRAS TEMPESTUS HODIE Tomorrows Weather Today
53
Small scale view of GOES Sounder SFOV LI DPI

• The SFOV resolution is evident (the nominal
  sub-satellite 10 km FOV being 11-16 km over the
  Midwest). However, practical confidence remains
  stronger with the larger pattern and the temporal
  trend.

21 UT 13 Apr 2006 03 UT 14 Apr 2006
54
Specific objective for Dynamic
NearcastingExpand the benefits of valuable
moisture Information contained in GOES Sounder
Derived Product Images (DPI)

• GOES Sounder products images already are
  available to forecasters. Products currently
  available include
• - Total column Precipitable Water (TPW)
• - Stability Indices (LI, CAPE)
• - 3-layers Precipitable Water (PW) . . .
• Build upon GOES Strengths
• Derived Product Images (DPI) of soundings
• speeds comprehension of information
• Data improves upon model first guess
• HOWEVER, some operational realities exist
• - DPIs used primarily as observations
• - No predictive component
• - 3-layer DPIs not used in current NWP models
• - Products often obscured when needed most
• - Cloud development/expansion obscures IR
  observations

Improvement of GOES DPI over NWP guess
Need to add a DPI predictive capability to close
these information gaps
13 April 2006 2100 UTC 900-700 hPa GOES PW 0-Hr
nearcast
55
Moving GOES data from Observations to Forecasts
NearCasts 0 Hour Projection for 2100UTC From
13 April 2006 2100UTC GOES Derived Product
Images 900-700 hPa PW? 700-300 PW?

Are these Moisture Changes Real ?
Validation of rapid moisture change in NE Iowa
using GPS TPW data
?
0 hr Lagrangian NearCasts of GOES
DPIs Valid 2100 UTC
Hail-Storm Path ?
VALIDATION of TPW
2115 UTC Sat Imagery
GOES Observed 900-300 PW 186 24 GPS TPW 25
Mid-Level Drying
Mid-Level Drying
Validation of Local NE Iowa moisture
maximum with GPS TPW data
13 April 2006 2100 UTC 700-300 hPa GOES PW 0
Hour NearCast
13 April 2006 2100 UTC 900-700 hPa GOES PW 0
Hour NearCast
56
Moving GOES data from Observations to Forecasts
NearCasts 2 Hour Projection for 2300UTC From
13 April 2006 2100UTC GOES Derived Product
Images 900-700 hPa PW? 700-300 PW?

Are these Moisture Changes Real ?
Validation of rapid moisture change in NE Iowa
using GPS TPW data
?
?
2 hr Lagrangian NearCasts of GOES
DPIs Valid 2300 UTC
Hail-Storm Path ?
VALIDATION of TPW
2315 UTC Sat Imagery
GOES Observed 900-300 PW 205 25 GPS TPW 28
Mid-Level Drying
Validation of Local NE Iowa moisture
maximum with GPS TPW data
13 April 2006 2100 UTC 700-300 hPa GOES PW 2
Hour NearCast
13 April 2006 2100 UTC 900-700 hPa GOES PW 2
Hour NearCast
57
Moving GOES data from Observations to Forecasts
NearCasts 4 Hour Projection for 0100UTC From
13 April 2006 2100UTC GOES Derived Product
Images 900-700 hPa PW? 700-300 PW?

Are these Moisture Changes Real ?
Validation of rapid moisture change in NE Iowa
using GPS TPW data
?
?
?
4 hr Lagrangian NearCasts of GOES
DPIs Valid 0100 UTC
Hail-Storm Path ?
VALIDATION of TPW
0115 UTC Sat Imagery
GOES Observed 900-300 PW 224 26 GPS TPW 30
Mid-Level Drying
Validation of Local NE Iowa moisture
maximum with GPS TPW data
13 April 2006 2100 UTC 700-300 hPa GOES PW 4
Hour NearCast
13 April 2006 2100 UTC 900-700 hPa GOES PW 4
Hour NearCast
58
Moving GOES data from Observations to Forecasts
NearCasts 6Hour Projection for 0300UTC From
13 April 2006 2100UTC GOES Derived Product
Images 900-700 hPa PW? 700-300 PW?

Are these Moisture Changes Real ?
Validation of rapid moisture change in NE Iowa
using GPS TPW data
?
?
?
6 hr Lagrangian NearCasts of GOES
DPIs Valid 0300 UTC
Hail-Storm Path ?
?
VALIDATION of TPW
0315 UTC Sat Imagery
GOES NearCast 900-300 PW 10 3 13 GPS TPW
16 ?13
Mid-Level Drying
Validation of Local NE Iowa moisture
maximum with GPS TPW data
13 April 2006 2100 UTC 700-300 hPa GOES PW 6
Hour NearCast
13 April 2006 2100 UTC 900-700 hPa GOES PW 6
Hour NearCast
59
Moving GOES data from Observations to Forecasts
Vertical Moisture Gradient (900-700 hPa GOES PW -
700-500 hPa GOES PW) 6 Hour NearCast for
0300UTC From 13 April 2006 2100UTC
Formation of Convective Instability
6 hr Lagrangian NearCasts of GOES
DPIs Valid 2100 UTC
Vertical Moisture Difference .
0315 UTC Sat Imagery
0315 UTC Sat Imagery
Differential Moisture Transport Creates
Vertical Moisture Gradients
13 April 2006 2100 UTC 700-300 hPa GOES PW 6
Hour NearCast
13 April 2006 2100 UTC 900-700 hPa GOES PW 6
Hour NearCast
60
Temporal and spatial evolution of GOES Sounder LI
DPI
Sequence of GOES Sounder Derived Product Imagery
(DPI) of Lifted Index (LI) stability, following
clearing of frontal cloud cover, supports
convective development across Oklahoma and Texas
on 10 Feb 2009. Does the instability evolution
provide any new, value-added information for a
forecaster?
61
Moving GOES data from Observations to Forecasts
Vertical Moisture Gradient (900-700 hPa GOES PW -
700-500 hPa GOES PW) 7 Analyses Plus 6-Hour
Nearcast from 1100UTC 10 February 2009
62
Rationale to approach the GOES Sounder

• Start from good base (sounder(s) operating
  nominally.
• Follow the strengths temporal monitoring
  moisture.
• Deal with the weaknesses vertical resolution
  clouds.
• Build confidence synergize with other
  (satellite) data.
• Share your critiques what is helpful / what is
  not.

See - http//cimss.ssec.wisc.edu/goes/rt/
63
Example of real-time GOES Sounder Derived Product
Imagery Focused on Wisconsin
GOES total precipitable water (TPW) DPI at 1700
UTC on 25 Aug 2003
GOES (lt40mm) evidently drier than first-guess
GOES (40-50mm) markedly moister than first-guess
(Shouldnt such differences impact forecasts ?)
64
Fundamental CIMSS research striving to make
quality real-time GOES Sounder radiance
observations into practical useful information
for weather forecasting
Atmospheric continuity and evolution are clearly
evident in multi-spectral animation.
Where will clouds be? Comparison between
observed imagery (bottom) and forecast imagery
(top) builds confidence in how well the CRAS
model is assimilating retrieved GOES Sounder
cloud and moisture information.
Where will forecast (GFS) moisture need to be
modified, monitoring trends, to provide a better
forecast for convection (as across Texas)?
Differences between retrieved GOES Sounder TPW
and the GFS forecast values are plotted over the
GOES TPW Derived Product Imagery (DPI).

1800 UT 2 Apr 2004
65
Acknowledgements and Disclaimer

• - Credit for the GOES SFOV retrieval
  implementation (ongoing) includes
• at SSEC J. Nelson, J. Li, Z. Li, A. Schreiner,
  C. Schmidt, P. Menzel
• at NESDS/StAR ASPB T. Schmit
• at NESDIS/STAR OPDB J. Daniels, A. Allegrino,
  G. Gray
• at NESDIS/OSDPD PIB C. Holland
• - Credit for the AWIPS capability at
  CIMSS includes
• at SSEC J. Gerth, R. Dengel, J. Robaidek, K.
  Strabala, S. Lindstrom, S. Wanzong
• at NWS MKX - K. Rizzo, J. Craven, K. Licitar
  ARX - D. Baumgardt GRB G. Brusky
• - Credit for CRAS and nowcasting
  applications includes
• at CIMSS R. Aune (ASPB), R. Petersen
• - Credit for AIRS processing includes
• at SSEC A. Huang, E. Weisz, J. Li, K.
  Strabala, K. Bah

   The views, opinions, and findings contained in
this message    are those of the author and
should not be construed as an    official
National Oceanic and Atmospheric Administration
   or U.S. Government position, policy, or
decision.
Any errors remain my responsibility while any
progress noted, is only possible by the
cooperation and effort of my colleagues.
66
Launch Schedule
67
The GOES-R Advanced Baseline Imager
ABI Current Spectral
Coverage 16 bands 5 bands Spatial
resolution 0.64 mm Visible 0.5 km
Approx. 1 km Other Visible/near-IR 1.0
km n/a Bands (gt2 mm) 2 km Approx. 4
km Spatial coverage Full disk 4 per
hour Scheduled (3 hrly) CONUS 12 per
hour 4 per hour Mesoscale Every 30
sec n/a Visible (reflective bands) On-orbit
calibration Yes No
68
The additional bands on the Advanced Baseline
Imager (ABI) allow new or improved products
Cirrus Clouds
Clouds, etc
Aerosols
Vegetation
0.64 ?m
0.86 ?m
1.38 ?m
0.47 ?m
Snow, Cloud phase
Fog, Fires, Clouds, etc
Water Vapor, Precip
Particle size
1.61 ?m
2.26 ?m
3.9 ?m
6.19 ?m
WV, Upper-level SO2
Volcanic ash, Cloud phase
Total Ozone
Water Vapor
6.95 ?m
7.34 ?m
8.5 ?m
9.61 ?m
Low-level moisture
Surface features, Clouds
Clouds, Precip, SST
Cloud heights
10.35 ?m
11.2 ?m
12.3 ?m
13.3 ?m
69
ABI bands via NWP simulation (CIMSS AWG Proxy
Team)
70
GOES-R Product List (Total 68) Product Set
Number 1-4
ABI Advanced Baseline Imager
Continuity of GOES Legacy Sounder Products from
ABI
SEISS Space Env. In-Situ Suite
EXIS EUV and X-Ray Irradiance Sensors
GLM Geostationary Lightning Mapper
Magnetometer
SUVI Solar extreme UltraViolet Imager