Kevin Garrett, Eric Maddy
and Krishna Kumar
Riverside Technology, Inc., JCSDA
Yingtao Ma
AER, Inc, JCSDA
Sid Boukabara
NOAA/NESDIS/STAR, JCSDA
20th International TOVS Study Conference - Lake Geneva, WI - October 30, 2015
Motivation: Increase the number and types of satellite radiometric observations assimilated in NWP
**CrIS**
**IASI**
Megha-Tropiques
SAPHIR
S-NPP ATMS
Inversion
Process
MetOp-A
AMSU/MHS
MetOp-B AMSU/MHS
 Inversion/algorithm
consistent across
Benefits
 Consistent Quality Control
 Characterization of surface state
NOAA-18 AMSU/MHS
 Characterization
of atmospheric
NOAA-19 AMSU/MHS
state
 Linearization of state vector
elements / background adjustment
MIIDAPS
TRMM TMI
GPM GMI
all sensors
 All parameters included in state vector
 Uses CRTM for forward and Jacobian
operators
 Valid
over all surfaces/all-sky
DMSP F16 SSMI/S
conditions
DMSP F17 SSMI/S
 UseDMSP
forecast,
fast regression or
F18 SSMI/S
climatology as first guess/background
GCOM-W1 AMSR2
**MIIDAPS extended to the hyperspectral Infrared for IR only or IR+MW 1DVAR analysis
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
2
Quick overview of MIIDAPS 1DVAR
Development status
Preliminary analysis and forecast impacts
Future work
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
3
Quick overview of MIIDAPS 1DVAR
Development status
Preliminary analysis and forecast impacts
Future work
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
4
Solution
[X]
Reached
Convergence
Observed
TBs
(processed)
Bias Correction
Compare
No
Convergence
Compute
∆X
Simulated
TBs
Covariance
Matrix [B]
Obs Error
[E]
Retrieval mode
Climatology
Assimilation mode
Forecast
Initial State Vector
[X]
K
Update State
Vector
[X]
CRTM
Iterative Processes
T
1
 1

T
J( X) =  (X − X 0 ) × B −1 × (X − X 0 ) +  Y m − Y( X) × E −1 × Y m − Y( X) 
2
 2

ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
(
)
(
)
5
1DVAR Analysis Fields and Derived Parameters
State Vector [X]
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
6
Quick overview of MIIDAPS 1DVAR
Development status
Preliminary analysis and forecast impacts
Future work
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
7
Extended state vector with hyperspectral IR covers trace gases / surface emissivity
1DVAR Analysis Fields and Derived Parameters
CO
CO2
O3
CH4
N2O
Extended to IR
IR expansion
State Vector [X]
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
8
• MIIDAPS atmospheric state vector extended to
trace gas profiles
• MIIDAPS emissivity state vector expanded to IR
channels (AIRS, CrIS, IASI)
• ATMS/CrIS brightness temperatures simulated
using ECMWF analysis (T, Q, CLW), trace gas
climatologies, and various IR/MW emissivity
models
• MIIDAPS applied to simulated data to retrieve
state vector elements for MW only, IR only and
combined IR+MW
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
9
Temp
WV
WV
MW+IR
94% Conv.
MW Only
94.5% Conv.
MW Only
Temp
IR+MW
IR Only
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
ECMWF
10
Use the same MIIDAPS interface
MIIDAPS package is freely available!
MIIDAPS Standalone Mode
Data preprocessing
Optional NWP collocation/guess
CRTM Initialization
Call MIIDAPS
Post processing
DA Interface (GSI)
MIIDAPS Package
MIIDAPS Library
Ingest
readers
FM
FM/
averaging
Guess
Colloc/
regress
CRTM initialization
Collocate guess to obs
MIIDAPS
1DVAR
Call MIIDAPS
Driver
Call CRTM for background calc
Call quality control subroutines
Post
gridding
Bias correction
process
Gross error check
figures
Primary arguments are Tb, guess fields,
Diagnostic file output
geometry, qc/geophysical output
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
Ancillary Data
Covariance
Matrix [B]
Obs Error
[E]
Bias
Coefficients
11
Quick overview of MIIDAPS 1DVAR
Development status
Preliminary analysis and forecast impacts
Future work
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
12
• GSI r46725, T670/254 (GFS/GDAS)
– PRCN: GDAS/GFS Operational configuration
– PR1D: PRCN + MIIDAPS applied to ATMS only
• Summer season
– August 1, 2014-September 10, 2014
• MIIDAPS Applied to ATMS only
– ATMS QC based on MIIDAPS output
– SSMI/S QC still being tuned
– Use of 1DVAR Geophysical outputs still being
explored
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
13
MIIDAPS-based QC Scheme: Sounders
O-B
MIIDAPS-based
(all observations)
GSI QC
forFlags
LEFT:for
SNPP
LEFT:
ATMS
SNPP
Channel
ATMS Channel
5 (52 GHz)
5 (52
and
GHz)RIGHT:
and RIGHT:
Channel
Channel
7 (547GHz)
(54 GHz)
ChiSq > 5
Remove:
Points Passing MIIDAPS QC but Failing Operational QC
All Channels
GWP/RWP > 0.05
Remove :
19-53 Ghz,
89-190 GHz
With
1DVAR GSI
QC
Heritage
ATMS 52.8 GHz
CLW > 0.2
Remove:
19-51 GHz
89-190 GHz
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
With Heritage QC
1DVAR
QC
QC Flags for LEFT: SNPPWith
ATMS
Channel
5 (52 GHz) and
RIGHT: Channel 7 (54 GHz) ATMS 54.5 GHz
With Heritage QC
14
500 hPa Height AC (NH)
200 hPa Wind RMSE (TRO)
Summary of Impacts
• MIIDAPS applied to just ATMS
has neutral impact on
traditionally important metrics
• Exploitation of MIIDAPS on all
MW, or all MW+IR, will show
true utility for QC application
500 hPa Height AC (SH)
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
850 hPa Wind RMSE (TRO)
• Further impact expected from
utilizing MIIDAPS for
hydrometeor and surface
characterization, background
adjustment
15
FCST: 00Z 8/4/2014
1. PR1D 2. PRCN
FCST: 00Z 8/5/2014
1. PR1D
2. PRCN
FCST: 00Z 8/7/2014
1. PR1D 2. PRCN
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
1. PR1D 2. PRCN
FCST: 00Z 8/8/2014
1. PR1D 2. PRCN
FCST: 00Z 8/6/2014
1. PR1D 2. PRCN
FCST: 00Z 8/9/2014
1. PR1D 2. PRCN
16
Quick overview of MIIDAPS 1DVAR
Development status
Preliminary analysis and forecast impacts
Future work
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
17
• Finalize QC implementation for current
sensors and extend to new sensors
a)
b)
c)
d)
– ATMS, SSMI/S, AMSU/MHS, GMI, AMSR2, SAPHIR,
IR sensors
Figure. MIIDAPS retrieved liquid water path and GFS 6hr forecast valid 12Z Jul
• Use of 1DVAR geophysical fields
3, 2014, for Hurricane Arthur event off the U. S. Southeast coast. a)
Displacement of MIIDAPS 1DVAR analysis and GFS forecast; b) 1DVAR liquid
water path; c) GFS 6hr liquid water path; and d) MIIDAPS-GFS liquid water path.
GFS forecast is collocated in space/time to GPM GMI observation points.
– Surface emissivity, hydrometeors
• Explore use of 1DVAR analysis as background
and all-sky radiance assimilation
– Resolve displacement errors
– Linearization
ITSC-20, Lake Geneva, WI, October 28-November 3, 2015
18