Get documents about "GMAO Satellite Data Assimilation
Document Sample
GMAO Satellite Data Assimilation
Michele Rienecker
Max Suarez, Ron Gelaro, Ricardo Todling, Emily Liu
Yanqiu Zhu, Ivanka Stajner, Meta Sienkiewicz, Rolf Reichle
Christian Keppenne, Robin Kovach
Global Modeling and Assimilation Office (GMAO)
NASA/Goddard Space Flight Center
JCSDA SSC Meeting
May 30-31, 2007
Global Modeling & Assimilation Office
http://gmao.gsfc.nasa.gov
• Atmospheric Assimilation:
• Atmospheric Assimilation:
• NCEP’s GSI
• NCEP’s GSI
• AIRS
• AIRS
• Data impacts - Adjoint tools
• Data impacts - Adjoint tools
• MLS Ozone
• MLS Ozone
• Land Surface: EnKF
• Land Surface: EnKF
• Ocean: EnKF
• Ocean: EnKF
• Ocean Color: SEIK
• Ocean Color: SEIK
2
GEOS-5 Atmospheric Data Assimilation System
Ricardo Todling, Max Suarez, Larry Takacs, Emily Liu
AGCM Analysis
Finite-volume dynamic core Grid Point Statistical Interpolation (GSI)
Bacmeister moist physics Direct assimilation of satellite radiance data
Physics integrated under the JCSDA Community Radiative Transfer Model
Earth System Modeling
Framework (ESMF) (CRTM) for most current instruments in space
Catchment land surface model GLATOVS for TOVS (HIRS2, MSU, SSU) on
Prescribed aerosols board of TIROS-N, NOAA-06,…, NOAA-12
Interactive ozone Variational bias correction for radiances
∂q n
Assimilation
∂t total
= dynamics (adiabatic ) + physics (diabatic ) + ∆q
Apply Incremental Analysis
Total “observed change” Model predicted change Correction from DAS
Increments (IAU) to reduce
00z 03z 06z 09z 12z 15z 18z 21z 00z 03z 06z
shock of data insertion Analysis
IAU gradually forces the Raw analysis (from GSI)
Background (model forecast)
model integration throughout Assimilated analysis
IAU (Application of IAU)
the 6 hour period
3
The next System - 4D-VAR
Background state
Cost Function
Analyzed state
4
Progress in 4D-VAR Development (Tremolet & Todling)
1. Trajectory Model: GEOS-5 with full physics
2. Model Adjoint: FV core with simple physics
3. Extension of GSI components for 4D-VAR
• Observation windowing flexibility
• Observation handling (higher temporal-resolution bins)
• Computation of time-dependent departures (OmF’s)
• Preliminary version of model-analysis interface
• Options for minimization algorithm
4. Fine ⇔ Coarse mappings: ESMF
5
MERRA
http://gmao.gsfc.nasa.gov/merra/
MERRA System EMPHASIS ON WATER CYCLE
Global Precipitation,
Evaporation, Land Hydrology,
1/2° × 2/3° × 72L to .01 mb
Cloud parameters and TPW
1979-present
GSI Analysis with IAU GLOBAL HEAT AND WATER BUDGETS
Parallel AMIP run FOR ALL PROCESSES
DIURNAL CYCLE FROM HOURLY 2-D
FIELDS
MERRA 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07
Stream 1
Stream 2
Stream 3
ROSB
G5-AMIP
Spinup years
Reduced Observing System Baseline (ROSB)
6
Date & Instrument
In-house Radiance Data Bufr Library
Level-1B
Processing Orbit Binary Files
Bufr Table
Calibration
Emily Liu
In house data processing to support HIRS2 MSU SSU
Modern Era Retrospective-analysis for
Research and Applications (MERRA)
HIRS3 HIRS4 AMSUA AMSUB MHS
Level-1b TOVS/ATOVS radiance data
were converted to calibrated radiance in
BUFR format with appropriate quality
controls Quality
Data available from 1979 to present Control
Data blacklists from ECMWF ERA40, Collect Data
JMA25 reanalysis, and GMAO GEOS-4 Orbit BUFR Files 6 hour window
reanalysis (CERES) for further data
screening
Synoptic BUFR Files
Can reprocess the radiance data if Four output files per day for Combine
calibration coefficients can be estimated each instrument type BUFR Files
from a better technique such as SNO
(simultaneous nadir overpass)
Thinned Warmest
Receiving full spatial resolution AIRS
and AMSU-A data from NESDIS
Processing full resolution data set into
thinned and warmest data sets in
BUFR format
Full Resolution 7
The GEOS-5 ADAS Validation: Precipitation
GEOS-5
8
Adjoint tools for Observation Impact Studies
Ron Gelaro
Efficient estimation of sources of forecast error and observation
sensitivity (observation impact)
• determined with respect to observational data, background fields
or assimilation parameters, all computed simultaneously
• useful for designing intelligent data selection strategies and
guiding future observing system design
Forecast
eb
observations
t
assimilated as
ec ea
for Forecast
u nd
Error gro
xb ba
ck ast Error
DAS rec
s fo
l ysi
xa ana xv
Sensitivity to
t − 6h 00Z t +24h
initial state
Sensitivity to t=24h
observations
GEOS-5 Model Adjoint
GSI Analysis Adjoint
9
GEOS-5 used to Evaluate Impact of AIRS in NWP
Emily Liu, Ron Gelaro, Yanqiu Zhu
Forecast Skill vs. Time
NH Data from most AIRS
channels improve
Channel Index
numerical weather
Control forecasts
Control + AIRS
Some AIRS
SH channels
degrade the
forecast
Forecast Error Reduction (J/kg)
NH
Control AIRS brings slightly positive impact on forecast skill in
Control + AIRS without Northern Hemisphere; clear positive impact in Southern
moisture channels Hemisphere. But forecast skills are increased when
moisture channels from AIRS are not included
10
Diagnosing impact of hyper-spectral observing systems
GEOS-5 July 2005 00z Totals
AMSU-A (15 ch) AIRS (153 ch) Negative
Impact
H2O Channels
Channel
Channel
-7.0 δe (J/Kg) 0 -0.6 Forecast Error Reduction (J/Kg)0
δ e (J/Kg)
…several AIRS water vapor channels currently degrade the 24h forecast
in GEOS-5…
11
Comparison with OSEs
GEOS-5 Observation Impact: Comparison with OSEs Ron Gelaro and Yanqiu Zhu
16
15
multiple OSEs
14
24h Forecast
Error Energy 13 control
no airs
(J/Kg) 12
no raobs
11 no amsua16
10
9
8
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
July 2005 00z
25
20
control observation impact
15
Observation
10
Count (millions)
5
0
ft
bs
nd
s
a
ce
mi
a
s
b
s
es
-5
u
ind
su
su
air
cra
hir
su
ms
δe
rfa
rao
go
ss
sw
am
am
am
tw
-10
air
sp
su
qk
sa
s_
(J/Kg) -15
eo
-20
-25
-30
12
Assimilating AURA/MLS ozone
Meta Sienkiewicz and Ivanka Stajner
Zonal mean ozone 9/30/2004 00UTC
SBUV daytime only – no data near South SBUV only
Pole due to high solar zenith angle
MLS orbital limit ±82º
MLS only
NOAA 16 SBUV
MLS Ozone hole develops
in MLS assimilation Ozone partial pressure (mPa)
13
Global assimilation of AMSR-E soil moisture retrievals
Rolf Reichle
Assimilate retrievals of
surface soil moisture
from AMSR-E (2002-06)
into NASA Catchment
Validate with USDA SCAN stations
model (GEOS-5)
(only 23 of 103 suitable for validation)
Soil moisture [m3/m3]
Anomaly time series correlation Confidence levels:
Reichle et al. coeff. with in situ data [-] Improvement of
JGR, 2007 (with 95% confidence interval) assimilation over
N Satellite Model Assim. Satellite Model
Surface soil moisture 23 .38±.02 .43±.02 .50±.02 >99.99% >99.99%
Root zone soil 22 n/a .40±.02 .46±.02 n/a >99.99%
moisture
Assimilation product agrees better with ground data than satellite or model alone.
14
Modest increase may be close to maximum possible with imperfect in situ data.
Volumetric soil moisture (m3m-3)
Kumar, Reichle, et al. (2007), Adv. Water
Resources, submitted. 15
Forecast skill (ACC) from CGCMv1
Forecast skill (ACC) from CGCMv1
Heat content anomaly in upper 300m
Heat content anomaly in upper 300m
1993-2006
1993-2006
EnKF OI-TS
1-month lead
3-month lead
6-month lead
16
The impact of Argo - preparing for Aquarius
Christian Keppenne and Robin Kovach
17
GMAO’s Collaborations with JCSDA Partners
Atmosphere:-
Atmosphere:-
•• GSI - NCEP
GSI - NCEP
•• Adjoint tools - NRL
Adjoint tools - NRL
•• AIRS
AIRS
•• Ozone
Ozone
•• Aerosols
Aerosols
•• OSSEs (emerging) - NCEP, NESDIS, et al
OSSEs (emerging) - NCEP, NESDIS, et al
Land Surface:-
Land Surface:-
• EnKF development
• EnKF development
• LIS implementation for Catchment and Noah LSMs
• LIS implementation for Catchment and Noah LSMs
Ocean:-
Ocean:-
• EnKF and MvOI development for MOM4 - NCEP
• EnKF and MvOI development for MOM4 - NCEP
• Altimetry with online-bias-estimation
• Altimetry with online-bias-estimation
• Ocean color
• Ocean color
18
GMAO - Near-term Plans
Atmosphere:-
Atmosphere:-
•• Development of 4Dvar
Development of 4Dvar
•• Contribute to OSSE capability
Contribute to OSSE capability
•• AIRS (QC) - IASI - CrIS
AIRS (QC) - IASI - CrIS
•• Ozone - GOME-2 - OMPS
Ozone - GOME-2 - OMPS
•• Real-time MLS
Real-time MLS
•• MODIS Winds - VIIRS
MODIS Winds - VIIRS
•• CO, CO2 (OCO)
CO, CO2 (OCO)
Land Surface:-
Land Surface:-
• EnKF: Surface Temperature and Snow
• EnKF: Surface Temperature and Snow
• LIS implementation for Catchment and Noah LSMs
• LIS implementation for Catchment and Noah LSMs
Ocean:-
Ocean:-
• MOM4: retrospective analysis for seasonal forecast
• MOM4: retrospective analysis for seasonal forecast
• Surface Salinity
• Surface Salinity
• Ocean color: removing instrument biases
• Ocean color: removing instrument biases
19
Related docs
