Plans and Progress

1
Plans and Progress Data Preparation CFS
Reanalysis
2
CISL Research Data ArchiveData for Atmospheric
and Geosciences Research
                              
CISL Research Data ArchiveData for Atmospheric
and Geosciences Research
                              
CISL Research Data ArchiveData for Atmospheric
and Geosciences Research
                              
CISL Research Data ArchiveData for Atmospheric
and Geosciences Research
                              
NCAR Data Support Section (and Roy Jenne)
supplied high quality observations for R1, R2,
ERA40, JRA25, NARR, MERRA, CFSRR
Services
Services
Services
Services
ftpService
ftpService
ftpService
ftpService
Datasets
Datasets
Datasets
Datasets
OtherResources
SpecialProjects
OtherResources
SpecialProjects
OtherResources
SpecialProjects
OtherResources
SpecialProjects
Staff
Documentation
Staff
Documentation
Staff
Documentation
Staff
Documentation
Software
Software
Software
Software
Left to Right Chi-Fan Shih, Will Spangler, Joey
Comeaux, Bob Dattore, Steven Worley, Dave
Stepaniak, Hua Ji, Doug Schuster, Gregg Walters,
Cecilia Banner
3
CFSRR Input Datasets Available from NCDC Archives
TOVS, SSM/I GOES-IR FROM CLASS
COADS
A number of NCDC datasets provided via NCAR
4
The History of Conventional Data Preparation for
Reanalyses at NCEP, 1994-2007

• TEMP PILOT AIREP PIREP ACARS SATOB SYNOP
  MARINE BOGUS
• R1/R2 merged data sources from NCAR, NCDC, NCEP,
  JMA, and ECMWF
• R1/R2 merged qcd datasets were packaged and
  delivered to the JMA-25 project
• R1/R2 datasets were reprocessed and enhanced by
  NCEP/NCAR for ERA-40
• NCEP obtained the ECMWF MARS archive of
  conventional data, 1979-1994
• NCAR improved versions of key datasets, ie,
  COADS, ON29/RAOBS, ON124
• R1/R2 and new datasets were merged for the North
  American Regional Reanalysis
• The latest versions of all source datasets were
  merged for the MERRA project
• Much effort was made by MERRA people to find data
  gaps and other problems

5
(No Transcript)
6
List of Radiosonde/Pibal Datasets Collected at
NCEP for Reanalysis
1920 1930 1940
1950 1960 1970 1980 1990
2000 ---------------------------------------
---------------------------------------------
------------ basua-----1948-2005------------
--------------
cards-----1946-1966------
-----------------------
---------------------- china-----1954-1962
---------------------------------
---------------------------- countries-194
6-1993------------------------
----- france-
---1948-1979--------------------------
----------------
gate------1974-1974------------------------
---------------------------------------
---- jma-------1978-1994------------------
--------------------------------
---- lie-------1967-1967------------
---------------------------------------
---------------- mars------1979-1994-------
----------------------------------------
-------- mit-------1958-1963-
-----------------------------------
--------------------------- navyship--1985
-1996------------------------------------
----------------------- navyua--
--1966-1973------------------------------
------------------------------- o
n20------1962-1972------------------------
-------------------------------
---- on29------1973-2000-------------------
---------------------------
permships-1973-1980-------------
----------------------------------
--------------- ptarmigan-1950-1961--------
-------------------------------
-------------------- ruships---1947-1978--
-----------------------
----------------- rusice----1954-
1978-------------------------------
----------------- rusnp----
-1950-1991----------------------------
------- ru
sua-----1960-1990-------------------------
----------------
--- safrica---1961-1967--------------------
----------------------------------
-------- scherhag--1954-1962--------------
---------------------------------
-------------- td52------1922-1971-----
---
-------------------- td53------1919-1969---

------------------------ td54------1943-1
974----------------------
-------------------- td90------
1943-1962----------------------
------------------------------ twe
rle----1975-1976--------------------------
---------------------------------------
-- usafua----1973-1980--------------------
----------------------------------
-------- usctrl----1946-1969---------------
--------------------
------------- ------------------------------
---------------------------------------------
---------------------
1920 1930 1940 1950 1960
1970 1980 1990 2000
7
A quantitative estimate of duplication in source
archives
Conventional observation counts. R3 is current
collection of source archives
8
(No Transcript)
9
Except for TOVS, these datasets form the basis of
the CFSRR conventional obs 1979-1997 After 1997
the NCEP CDAS and OPS archives are used as
conventional data sources
VTPR
TOVS
TEMP/PILOT
SYNOP
AIRCFT
COADS
SATOB
PAOB
10
A BUFR data structure was designed for R1 to
record feedback information in the observations
which is still used in operations and the CFSRR
as well
analysis
1st guess
ob error
event stack
The diagram illustrates how processing
information such as QC results, and observation
modifications are recorded in the BUFR files
during the execution of the reanalysis. One
temperature level is shown as an example.
11
TOVS, ATOVS, AQUA Archives for CFSRR TOVS
HIRS/2 and MSU 1-B orbital files were obtained
from NESDIS/SAA spanning all satellites for the
years 1978-1997 for the NARR project. Synoptic
1-C BUFR brightness temperature files were
created from the 1-B archive. The NESDIS/NCDC
CLASS server has since become operational and is
a great resource for satellite data. Available
data is constantly being expanded. Some data not
available from CLASS may be obtained directly
from NCDC. For CFSRR, the NARR HIRS/2 and MSU
1-B archives are enhanced, and new holdings
extended through to the present. The complete set
of SSU orbit files were added to the archives, as
well as AMSU-A, AMSU-B, HIRS/3, HIRS/4, MHS,
AIRS, and AMSR-E data. Data sets were obtained
through CLASS and NCEP archives. New codes to
convert raw orbit files to calibrated synoptic
BUFR BT files are created for MSU, and SSU.
Improvements include new calibrations for MSU
NOAA-10-14 (Cheng-Zhi Zou), and more use of the
imbedded QC data found in the orbital files.
12
CFSRR TOVS-ATOVS-AQUA Satellite Data Timelines
NCEP
CLASS
AIRS AMSR-E AMSU-A AMSU-B HIRS/2 HIRS/3 HIRS
/4 MHS MSU SSU
1980 1990
2000
2010
13
Defense Met Sat Program Special Sensor Microwave
Imager SSM/I Sensor Data Records
SDR orbits
ssmit synoptic BUFR files
spssmi ocean surface wind speed BUFR files
14
Availability of GOES data in CLASS is shown in
the following table. 
High density SNDR and IMGR data is available in
CLASS starting with GOES-8 in 1994. With some
help from NESDIS/NCDC with cloud processing, we
could hope to go at least that far back. There
was an IR sounder (VAS) on GOES-4-7 satellites,
but whether anything can be done with them for
reanalysis is currently undetermined.
15
NESDIS V8 SBUV TOTAL AND PROFILE
OZONE RETRIEVALS, 1978-2007
V8 is the most recent version of the
backscattered ultraviolet (BUV) algorithms that
have undergone three decades of progressive
refinement. Its predecessors, V7 for total ozone,
developed about in 1998, and V6 for profile
ozone, developed in about 1990, have been used to
produce the acclaimed TOMS total ozone and
SBUV(/2) ozone profile time series. V8 will
correct several small errors in V7 and V6 that
were discovered by extensive error studies using
radiative transfer models and by comparison with
ground-based instruments. The V8T uses only two
wavelengths (318 nm and 331 nm) to derive total
ozone. Other wavelengths are used for diagnostics
and error correction. Experience with TOMS and
SBUV/2 suggests that the algorithm is capable of
producing total ozone with rms error of about 2,
though these errors are not necessarily randomly
distributed over the globe. The errors typically
increase with solar zenith angle and in presence
of heavy aerosol loading.
16
Subset of the SBUV ozone retrievals selected for
CFSRR
a ascending node d
descending node
17
This shows when data types enter the CFSRR stream
TEMP/PILOT SYNOP AIRCRAFT MARINE SATOB BOGUS
VADWND PROFILR METAR ERSWND
MSONET
SSMIWND
V8 OZONE
QKSWND
ACARS
1980 1990
2000
2010
AIRS
GOESND 5X5
AMSRE HIRS4 MHS
HIRS MSU SSU
GPSIPW
HIRS3 AMSUA AMSUB
GEOIMR
GOESND 1X1
18
Tropical Cyclone Synthetic Wind Data 1978-2002
CSFRR will use Mike Fiorinos TC Bogus wind
dataset as an aid to locating storm vortices in
the GSI analysis. The synthetic winds consist of
a central location surrounded by four wind
profiles.
19
SEA ICE for CFSRR NASA GSFC recomputed (for
instrument drifts) from Oct 1978 to approx 1996,
when it will turn over to the NCEP ice. The
NCEP ice will be recomputed with present analysis
methods, land masks, algorithms, etc. from
1995/10/01 to 2000/02/25 from prior grids of
brightness temperatures.  From 2/25/2000 to
present using archived SSM/I passes. The GLERL
Great Lakes analysis will be used for the
historic coverage of those lakes.  NIC Great
Lakes will continue this towards the present.
GSFC analysis does not cover all ice covered
areas of our interest. Those (great lakes,
peripheral (to the arctic) lakes and seas) will
be filled by similar method to NARR --
identifying proxy points and filling from there. 
Proxies will be found from the 2/25/2000 to
present period, where there is full data
available and the procedure can be continued into
the future. Half degree ice will be produced
consistently throughout the period (vs.
Reanalysis 1 switching from Nomura 1 degree to
NCEP 1 degree to NCEP 0.5 degree).  Starting with
2/25/2000, a 5' lat-long grid will also be
available. Where there are temporal gaps,
they'll be filled by linear interpolation in
time.  This is particularly a problem for the
Great Lakes.  It is not currently a concern, with
3 SSM/I instruments flying, and AMSR-E also
providing comparable data for any area which can
be analyzed by automated means.
20
Ingest of Snow Analyses for CFSRR

• Current Ops T382 GDAS ingests 2 daily snow
  analyses
• global Air Force analysis (Air Force non-grib
  format, one file per hemisphere)
• N. Hemisphere NESDIS 4-km analysis (known as IMS)
  in grib format
• ops job-step (known as SNOW2MDL) by George Gayno
  of EMC merges the two snow analyses above and
  maps merged anal toT382 GFS grid (or any GFS
  grid)
• CFS RR will assimilate both Air Force and IMS (as
  in Ops) for those years that both are available
  (1997-present) only Air Force analysis before
  1997
• Using SNOW2MDL job step, modified to account for
  data availability below
• 23-km resolution until about 2005 (date TBD),
  then 4-km resolution thereafter
• IMS 23 km analysis available from Feb 1997
  through the present
• IMS 4 km analysis available from Feb 2004
  through the present
• George Gayno has modified his job SNOW2MDL to be
  able to use only Air Force analysis prior to the
  1997 start of NESDIS IMS analysis
• Air Force analysis available for all years needed
• 1979 to 1998 in Grib format from COLA
• M. Fennessy of COLA obtained Air Force analyses
  for above period from NCDC/ETAC and then gribbed
  it with help from EMC
• 2002 to present in Air Force non-grib format used
  in present ops
• (via EMC archive of Jesse Meng
• 1999-2001 (to be extracted from NCO run history
  files)
• Jan-Oct 1999 period may need special attention
  due to CRAY fire impact on archives
• Option 1 Use archived GDAS snogrb file

21
Some SST Options for CFSRR

• OI 1/4 Degree Daily Analysis from NCDC
• AVHRR only satellite source thru May 2002
• Option of including AMSR June 2002 onward,
  greatly improving satellite coverage in cloudy
  regions (especially important for daily analysis)
• May not be available in time
• Currently 1985-present, undergoing revision
  (perhaps available by October 2007, per Dick
  Reynolds)
• Nov 1981-Dec 1984, would need to first compute
  daily AVHRR SST retrievals. (Pathfinder Project
  if funded, or Reynolds himself)
• OIv2 1-degree Weekly Analysis
• Lower resolution and no AMSR satellite data
• More consistent
• Available Nov 1981 present (AVHRR period only)
• If necessary, consider spinning backwards to
  required start date, using satellite quality
  field as initial first guess to retain some
  gradients

22
Some data sets we still hope to get for CFSRR
Reprocessed GMS winds used in JRA-25 (NCAR
DSS) ERS-1 and ERS-2 winds prior to mid-1997
(NCAR DSS) Reprocessed GPS Occultation data from
COSMIC (NCAR) GOES SOUNDER 1x1 radiances (CLASS)