1' FY09 GOESR3 Project Review Title Page

1
1. FY09 GOES-R3 Project Review Title Page

• Title ABI Proxy data studies Regional
  Assimilation of SEVIRI Total Column Ozone
• Project Type Product Utilization Proposal
• Status Second Year
• Duration 3 years
• Leads
• R. Bradley Pierce1, NOAA/NESDIS
• Other Participants
• Chris Schmidt1 (CIMSS)
• Todd Schaack (SSEC)

1No Cost
2
2. Project Summary

• This project will use the WRF-CHEM1 regional
  chemical model coupled to the RAQMS global
  chemical analysis2 to evaluate the impacts of ABI
  like Tropospheric Column Ozone (TCO) retrievals
  on Air Quality forecasts. SEVIRI measurements
  will be used as ABI proxy data.
• Results from these studies will provide guidance
  and develop capabilities for operational use of
  GOES-R TCO within the NOAA Operational AQ
  forecasting system (CMAQ/WRF-NAM).
• This study will also demonstrate Committee on
  Earth Observation Satellites (CEOS) atmospheric
  composition constellation concepts contributing
  to the space component of the Global Earth
  Observing System of Systems (GEOSS) Air Quality
  Assessment and Forecasting Near Term Opportunity
  (NTO)

1Grell, G. A., et al., Fully coupled online
chemistry within the WRF model, Atmos. Environ.,
39, 6957-6975, 2005. 2Pierce, R. B., et al.
(2007), Chemical data assimilation estimates of
continental U.S. ozone and nitrogen budgets
during the Intercontinental Chemical Transport
ExperimentNorth America, J. Geophys. Res., 112,
D12S21doi10.1029/2006JD007722.
3
3. Motivation/Justification

• This task supports NOAA Mission Goal(s) to Serve
  Societys Needs for Weather and Water
  Information through development of capabilities
  for assimilating ozone observations from future
  NOAA operational geostationary and polar-orbiting
  research satellites into AQ forecast models and
  evaluating the impacts of these observations on
  air quality prediction capabilities.
• Justification Development and testing of
  methodologies for incorporating satellite based
  ozone observations into National AQ forecasts
  supports the National Weather Services National
  Air Quality Forecast Capability (AQFC) mandated
  by Congress (H.R. Energy Policy Act of 2002
  Senate Amendment).

4
4. Methodology

• Linking global RAQMS ozone analyses1, with ozone
  predictions from the WRF-CHEM regional2 model
  provides a first guess for SEVIRI TCO
  assimilation studies. WRF-CHEM regional AQ
  predictions, initialized with ozone analyses with
  and without SEVIRI TCO, will be used for AQ
  forecast impact studies. (RAQMS/WRF-CHEM Linkage
  completed)
• A baseline analyses with and without SEVIRI TCO
  retrievals will be conducted using existing
  Optimal Interpolation (OI) assimilation
  capabilities implemented into the WRF-CHEM
  regional model. (Baseline WRF-CHEM SEVIRI TCO
  analyses using OI assimilation completed)
• Coupling between RAQMS/WRF-CHEM and the Gridpoint
  Statistical Interpolation (GSI) system will be
  developed. Assimilation of SEVIRI TCO retrievals
  within the regional Gridpoint Statistical
  Interpolation (GSI) system will be tested.
  (RAQMS/GSI coupling completed, testing of
  WRF-CHEM/GSI coupling underway)
• Coupling RAQMS/WRF-CHEM to the Community
  Radiative Transfer Model (CRTM) will be
  developed. SEVIRI TCO assimilation with radiance
  based observation operator will be tested.
  (RAQMS/WRF-CHEM/CRTM coupling completed)
• Surface (EMEP), and ozonesonde (WMO) data
  collected over Europe provide insitu measurements
  for verification.(WMO ozonesonde data acquired)

1August 2006 chemical and aerosol assimilation
data denial experiments have been conducted to
determine the optimal combination of satellite
ozone measurements for RAQMS global ozone
analysis. (FY08 NASA Decadal Survey/GOES-R AWG
activity) 2WRF-CHEM regional model is designed to
be portable to any geographic region.
5
5. Summary of Previous Results

• A 2006 RAQMS global ozone analysis constrained
  with OMI TCO and TES ozone profile retrievals has
  been completed. The GOES statistical TCO
  retrieval has been adapted for SEVIRI TCO
  studies. Initial comparisons are being conducted
  to assess SEVIRI TCO biases and RMS errors
  (Figure 1).

Figure 1 August 1st 2006 OMI V8 TCO retrieval
(left), RAQMS 18Z TCO Analysis (middle), and
SEVIRI 18Z TCO retrieval (right). RAQMS ozone
assimilation includes OMI TCO and TES ozone
profiles from NASA Aura Satellite. TCO
enhancement over central Africa is associated
with tropospheric ozone column enhancements due
to biomass burning.
6
6. Expected Outcomes

• Guidance for incorporation of GEOS-R ABI data
  into National AQ prediction.
• Testing and Evaluation of GSI regional ozone
  assimilation capabilities.
• Testing and Evaluation of CRTM SEVIRI/ABI forward
  modeling capabilities.
• Development of observation operators for ABI
  statistical TCO retrieval.
• Demonstration of CEOS Atmospheric Composition
  Constellation concepts.
• Demonstration of Integrated Observed-Modeled Air
  Quality Fields for GEOSS Air Quality Assessment
  and Forecasting Near Term Opportunity

7
7. Major Milestones

• FY08
• Conduct August-September 2006 WRF-CHEM regional
  baseline (no SEVIRI TCO) simulation using RAQMS
  global chemical analysis for lateral boundary
  conditions.
• WRF-CHEM model ported to CIMSS
• Initial testing over CONUS domain has been
  completed
• WRF-CHEM anthropogenic emissions inventory for
  SEVIRI domain (Africa/EU) has been developed
  (Need to build daily biomass burning emission
  inventory for SEVIRI domain)
• RAQMS/WRF-CHEM SEVIRI simulation initiated
• Conduct August-September 2006 RAQMS regional SDF
  SEVIRI TCO assimilation.
• Upon completion of baseline (no SEVIRI TCO)
  simulation
• Evaluate impact of assimilation of SEVIRI TCO
  retrieval through comparison with WMO ozonesonde
  and EMEP surface network.
• WMO ozonesonde data acquired
• RAQMS August 2006 global chemical analysis has
  been evaluated over Africa
• Develop coupling between the WRF-CHEM regional
  model and the regional GSI system.
• GSI has been ported to CIMSS
• WRF-ARW/Regional GSI has been tested (without
  ozone assimilation)
• Global GSI SBUV ozone assimilation has been
  tested with RAQMS

8
7. Major Milestones First Year Results

• A 2006 RAQMS global ozone analysis data denial
  study has been completed. Comparisons with August
  2006 WMO sondes show excellent (lt10 bias)
  agreement in stratosphere and reasonable (lt20
  low bias) agreement in the troposphere at
  Ascension Island (Figure 2).

RAQMS vs WMO sondes August 2006 (5 sondes)
OMI TCO August 2006
RAQMS TCO August 2006
Ascension Island
WRF-CHEM 36km SEVIRI domain
Figure 2 August 2006 OMI V8 TCO retrieval
(left), RAQMS TCO Analysis (right), and RAQMS/WMO
ozonesonde comparison (middle). RAQMS ozone
assimilation includes MLS ozone profiles from
NASA Aura Satellite. TCO enhancement over
central Africa is associated with tropospheric
ozone column enhancements due to biomass burning.
9
7. Major Milestones (cont)

• FY09
• Conduct August-September 2006 RAQMS regional GSI
  SEVIRI TCO retrieval assimilation
• Evaluate impact of GSI SEVIRI TCO assimilation
  through comparison with baseline, SDF, WMO
  ozonesonde, and EMEP surface network.
• Develop RAQMS/CRTM linkage for forward modeling
  of SEVIRI radiances.
• CRTM ported to CIMSS
• CRTM/RAQMS linkage has been tested with
  AIRS/SEVIRI radiance comparisons
• Develop radiance based observation operator for
  SEVIRI statistical TCO retrieval

10
7. Major Milestones First Year Results (cont)

• Comparisons between clear sky synthetic (RAQMS)
  and observed (AIRS, SEVIRI) brightness
  temperatures on August 16th, 2006 show a 2K warm
  bias at 1050cm-1, consistent with the
  underestimate in analyzed UT O3 mixing ratios
  relative to WMO ozonesondes (Figure 3).

RAQMS AIRS SEVIRI
Figure 3 August 16th 2006 SEVIRI 9.6um BT (left)
and RAQMS/AIRS/SEVIRI radiance comparison
(right).
11
7. Major Milestones Second Year Plans (cont)

• Radiance based observation operator for SEVIRI
  statistical TCO retrieval (TOZ)1

• Where
• A, B, and C are the regression coefficients
• Tb is IR BTs (band j1,n)
• Ta is temperature (layer l1,101)
• ? is the local zenith angle
• M is the month
• LAT is the latitude of the satellite pixel
• Approach
• Use CRTM to calculate SEVIRI Tbj from
  RAQMS/WRF-CHEM first guess
• Generate first guess TCO retrieval (TOZfg) from
  synthetic radiances and model temperature
  profiles
• Analysis increment is TOZ-TOZfg
• 1Xin Jin et al., Retrieval of Total Column Ozone
  From Imagers Onboard Geostationary Satellites,
• IEEE TRANSACTIONS ON GEOSCIENCE AND REMOTE
  SENSING, VOL. 46, NO. 2, FEBRUARY 2008

12
7. Major Milestones (Second Year)

• FY08 (outstanding due to delay in acquiring
  SEVIRI TCO retrievals and biomass burning
  emissions)
• Conduct August-September 2006 WRF-CHEM regional
  baseline (no SEVIRI TCO) simulation using RAQMS
  global chemical analysis for lateral boundary
  conditions.
• Baseline simulation completed
• Identification of significant low bias in
  WRF-CHEM upper tropospheric ozone over Africa
• Attributed to lack of lightning NOx source in
  WRF-CHEM
• Conduct August-September 2006 RAQMS regional SDF
  SEVIRI TCO assimilation.
• Cloud clearing algorithm developed
• Diurnal bias correction algorithm developed
• SDF SEVIRI TCO assimilation studies completed
• FY09
• Conduct August-September 2006 RAQMS regional GSI
  SEVIRI TCO retrieval assimilation
• GSI SEVIRI TCO assimilation testing initiated
• Need to incorporate lightning NOx emissions into
  WRF-CHEM before further assimilation studies are
  conducted
• Evaluate impact of GSI SEVIRI TCO assimilation
  through comparison with baseline, SDF, WMO
  ozonesonde, and EMEP surface network.
• Incorporation of lightning NOx emissions will
  require rerunning baseline and SDF assimilation
  studies

13
7. Major Milestones Second Year Results

• NESDIS SEVIRI fire emissions have been acquired
  and incorported into WRF-CHEM emissions
  preprocessor (Figure 4).

NESDIS African BB emissions are consistently
larger than WRF BB emissions during the time
period considered. This leads to increased
tropospheric ozone production.
Figure 4 August 11th 2006 WRF Biomass Burning
(BB) Organic Carbon (OC) emissions vs NESDIS BB
OC emissions
14
7. Major Milestones Second Year Results

• SEVIRI Total Column Ozone (TCO) has been acquired
  and algorithms for bias correction and cloud
  clearing have been developed.

• Enhancements in SEVIRI TCO near cloud edges
  require additional cloud clearing. Hourly
  averaging of 15 minute clear SEVIRI TCO
  retrievals reduces artifacts associated with
  cloud edge effects.
• Diurnal and spatial bias correction (based on
  differences between RAQMS and cloud filtered
  SEVIRI data) was implemented to remove scan angle
  and zenith angle dependencies in SEVIRI TCO.

Figure 5 August 11-23 2006 18Z SEVIRI (left) and
Bias Corrected SEVIRI (right) TCO
15
7. Major Milestones Second Year Results

• Statistical Digital Filter (SDF) assimilation
  studies have been completed and results have been
  compared to SHADOWS ozonesonde data to determine
  optimal approach for assimilation of SEVIRI TCO.

Description
Experiment
nosdf - no assimilation
runsdfratiof - assimilation with
original SEVIRI data               
(Increment WRF and RAQMS domain)sdfbias
- assimilation with bias corrected SEVIRI
data              (Increment WRF
domain) sdfbiasratiof - assimilation with
bias corrected SEVIRI data              
(Increment WRF and RAQMS domain)dates run
from August 11 to 23 th  2006
16
7. Major Milestones Second Year Results

• Statistical Digital Filter (SDF) assimilation
  studies have been conducted and results have been
  compared to SHADOWS ozonesonde data to determine
  optimal approach for assimilation of SEVIRI TCO.

No Assimilation
RAQMS MLS Analysis
17
7. Major Milestones Second Year Results

• Statistical Digital Filter (SDF) assimilation
  studies have been conducted and results have been
  compared to SHADOWS ozonesonde data to determine
  optimal approach for assimilation of SEVIRI TCO.

Bias Corrected Assimilation
RAQMS MLS Analysis
18
7. Major Milestones Second Year Results

• Summary of SDF SEVIRI TCO Assimilation
  Experiments vs SHADOWS ozonesonde measurements.

Result
Experiment
nosdf - Large negative biases at
200mb sdfratiof - Majority of
adjustment occurs in stratosphere sdfbias
- WRF column adjustment improves 200mb bias
but introduces lower
stratospheric bias sdfbiasratiof - Majority
of adjustment occurs in stratosphere

• Conclusions
• Bias corrected adjustment of WRF O3 has largest
  impact.
• WRF-CHEM 200mb biases degrade assimilation study.
  

19
7. Major Milestones Second Year Results

• Lack of WRF-CHEM Lightning NOx source is most
  likely cause of WRF-CHEM underestimate of upper
  tropospheric (200mb) ozone over Africa.

RAQMS 200mb Lightning NOx Source (2006-08-23, 18Z)
WRF-CHEM OLR (2006-08-23, 18Z)
NO formation from lightning discharges within
convective systems is a dominate source of NOx
(NONO2) in the tropical free troposphere (Levy
et al. J. Geophys. Res., 104, 26279-26306, 1999)
20
7. Major Milestones Second Year Results

• Lack of WRF-CHEM Lightning NOx source is most
  likely cause of WRF-CHEM underestimate of upper
  tropospheric (200mb) ozone over Africa.

RAQMS 200mb NO2 (2006-08-23, 18Z)
WRF-CHEM 200mb NO2 (2006-08-23, 18Z)
Down wind transport of Lightning induced NO2
significantly impacts South African upper
troposphere (including Niarobi and Reunion
ozonesondes)
21
7. Major Milestones Second Year Results

• Lack of WRF-CHEM Lightning NOx source is most
  likely cause of WRF-CHEM underestimate of upper
  tropospheric (200mb) ozone over Africa.

RAQMS 200mb O3 (2006-08-23, 18Z)
WRF-CHEM 200mb O3 (2006-08-23, 18Z)
Lack of upper tropospheric ozone production down
wind from Lightning NOx emissions leads to
underestimates in WRF-CHEM upper tropospheric
ozone over Equatorial and Southern Africa.
22
7. Major Milestones Second Year Results

• Lack of WRF-CHEM Lightning NOx source is most
  likely cause of WRF-CHEM underestimate of upper
  tropospheric (200mb) ozone over Africa.

Need to incorporate Lightning NOx source
into WRF-CHEM before further assimilation studies
are conducted. Will use Price and Rind (Mon.
Weather Rev., 122, 1930-1939, 1994) scheme
currently implemented into RAQMS.
23
7. Major Milestones Second Year Results

• Testing of WRF-CHEM SEVIRI ozone assimilation
  using the National Center for Environmental
  Prediction (NCEP) Gridpoint Statistical
  Interpolation (GSI) assimilation system has been
  initiated.

Total Column
WRF Column
TCO analyses are similar between the GSI and SDF.
However, GSI shows significantly larger ozone
column within the WRF domain (below 50mb)
GSI
SDF
24
7. Major Milestones (cont)

• FY10
• Conduct August-September 2006 RAQMS regional GSI
  SEVIRI TCO assimilation with radiance based
  observation operator
• Conduct August-September 2006 data denial (with
  and without SEVIRI TCO) AQ forecast studies.
• Evaluate impact of GSI SEVIRI TCO retrieval
  assimilation on European AQ forecasts through
  comparison with WMO ozonesonde and EMEP surface
  network observations.
• Interface with GOES-R Aerosol and Air Quality
  AWG.

25
8. Funding Profile (K)

• Summary of leveraged funding
• GOES-R AWG Task Use RAQMS global/WRF-CHEM
  regional air quality modeling systems to develop
  proxy data sets for GOES-R ABI aerosol
  retrievals.
• Leverage development of RAQMS/WRF-CHEM/CRTM
  linkages under GOES-R AWG task.
• FY09 GOES-R AWG high resolution (4km) proxy data
  set will use GOES-R3 36km SEVIRI ozone/aerosol
  assimilation for initial conditions.

26
9. Expected Purchase Items

• FY08
• (48K) STAR CIMSS Grant for 1 person at 25 time
  from OCT 07 to SEP 08
• Personnel support (including benefits, IT
  charges, overhead, etc) 38K
• Linux workstation, PGI compiler, IDL license
  (10K)
• FY09
• (50K) STAR CIMSS Grant for 1 person at 25 time
  from OCT 08 to SEP 09
• Personnel support (including benefits, IT
  charges, overhead, etc) 40K
• Software charges 2K
• Equipment (RAID disks) 8K
• FY10
• (52K) STAR CIMSS Grant for 1 person at 25 time
  from OCT 09 to SEP 10
• Personnel support (including benefits, IT
  charges, overhead, etc) 42K
• Software charges 2K
• Equipment (RAID disks) 8K

27
(No Transcript)