Joint OSSE

1
Joint OSSE August 28, 2008

• Summary from subgroup meeting
• ADM Aeolus preprocessor
• Meso-Regional OSSEs
• GMAO Observation Simulator
• Evaluation of T511 and T799 NR
• Thomas Jung
• New Projects
• Lidia Cucurull
• Plan for calibration and simulation of
  observation
• Estimated resource requirements
• Future direction of Nature Run

More surface fields in 1x1 for T511 NR?
2
Next meeting 707 available 10/16, 23, 30
11/6,27 12/4,18 We can negotiate Rm 209 for
other date
3
David Tan visited US and demonstrate
installation of L2b processor
Talk_Tan_AeolusDataProcessing_JCSDA_20080710.sml.p
pt Manual is posted JointOSSEs-gtManuals-gtADM/L2BP_
Release1.33_doc/ Start from Sofrware Release Note
AE-RN-ECMWF-L2BP-001_20080229_SRN_Iss1.33.pdf
Relevant material for L2b data
format AE_SAF_KNMI_L2BP_003_Setup_Guide.v1.1.20080
520.pdf pseudoPilot.f At NCEP Greg Krasowski
(JCSDA and work with  Dennis Keiser) will work on
preprocessor.
4
ADM-Aeolus Data Products
Product Contents Processor developerand location Size in MByte/orbit
Level 0 Time ordered source packets with ALADIN measurement housekeeping data MDA (Canada)Tromsø (Norway) 47
Level 1b Geo-located, calibrated observational data preliminary HLOS wind profiles (standard atmosphere used in Rayleigh processing) not suitable for assimilation spectrometer readouts at measurement scale ( 1-5 km ) input for Level 2a/b processing viewing geometry scene geo-location data MDA (Canada) Tromsø (Norway) 10-15 (BUFR) 22 (EE XML Format)
Level 2a Supplementary product Cloud profiles, coverage, cloud top heights Aerosol extinction and backscatter profiles, ground reflectance, optical depth DLR-IMF (Germany) Tromsø (Norway) 12
Level 2b Meteorologically representative HLOS wind observations HLOS wind profiles at observation scale ( 50 km ) suitable for assimilation - temperature T and pressure p (Rayleigh-Brillouin) correction applied with ECMWF (or other) model T and p ECMWF Reading (UK) (and other NWP/research centres) 18
Level 2c Aeolus assisted wind vector product Vertical wind profiles (u and v component) NWP model output after assimilation of Aeolus HLOS wind ECMWF Reading (UK) 22
5
Ongoing ADM-Aeolus Scientific Studies
Title Team
Consolidation of ADM-Aeolus Ground Processing including L2A Products DLR Germany Météo-France, KNMI, IPSL, PSol
Development and Production of Aeolus Wind Data Products ECMWF UK Météo-France, KNMI, IPSL, DLR, DoRIT
ADM-Aeolus Campaigns DLR GermanyMétéo-France, KNMI, IPSL, DWD, MIM
Optimisation of spatial and temporal sampling KNMI Netherlands
Tropical dynamics and equatorial waves MISU Sweden
Rayleigh-Brillouin Scattering Experiment tbd
ESA plans an Announcement of Opportunity AO for
ADM-Aeolus scientific use of data for late 2008
distinct from the AO for Cal/Val
6
KNMI can simulate L1B data which simulate errors
due to the different between NR and DAS model.
However, this seems to be much more complicated
and KNMI and SWA will simulate L2B data. We
still have to discuss about Bufr table for L2B.
relevant materials are AE_SAF_KNMI_L2BP_003_Setup
_Guide.v1.1.20080520.pdf The code to write L2b
data is out_pseudoPilot.f Gert-Jan is planning
to attend LWG meeting in next winter and discuss
with SWA about simulation of DWL. Ad Stoffelen
may be able to visit US in summer 2009. Both SWA
and Arlindo has aerosol simulator for DWL, We
are trying to hold a meeting with SWA and NASA.
7
MESO-Regional scale OSSE

1. Boundary condition for regional analysis
2. Regional NR add further uncertainty . Regional
   NR could divert from global NR
3. Regional OSSE for the best use of limited
   resources
4. Choice between WRF var and GSI
5. Recycling regional forecast or not

8

• GSI support Plan at DTC.
• Ming Hu (NOAA/ESRL)
• "DTC does have plan to release and support the
  use of GSI for communityusers. Here is a little
  details on the progress of the plan
• GSI has been installed on NCAR IBM and GSD linux
  cluster.
• GSI is running well GSD Linux cluster
• . GSI users guide is ready for NCEP and GSD to
  review.
• DTC is planning to do GSI tutorial on January
  2009.

9
Namx Low resolution regional run template by
Yochiaki Sato Vapor/jcsda/save/wx23sy/Namx/Scrip
ts.Base (cycle) Vapor/jcsda/save/wx23sy/Namx/Scri
pts.GlbGes(GSI only) Vapor/jcsda/save/wx23sy/N
amx/llprep.sh Vapor/jcsda/save/wx23sy/Namx/prep
_exp.sh gt copy tools gt gt 1. mkdir
GPFS_VOLUME/save/USER/Namx gt 2. cd
GPFS_VOLUME/save/USER/Namx gt 3. copy the
attached files into GPFS_VOLUME/save/USER/Na
mx gt Original scripts are in HAZE/jcsda/save/wx
23sy/Namx gt 4. gunzip -c Scripts.Base0829.tar.gz
tar xvf - gt gt modify the baseline script
gt gt 7. edit llprep.sh gt please change lord
leveler parameters for your account gt
account_no, class, group, and gpfs volume name gt
8. run the script (sh llprep.sh) gt gt set
experiment environment gt gt 9. edit
prep_exp.sh gt please configure your
experiment gt EXP_NAME, VOL_NAME, S_DATE,
E_DATE gt 0. run the script (sh prep_exp.sh) gt
10
Squall line / cold front case event from T799 NR
Chris Hill Mississippi State University
T799 NR case event May 3 06 UTC MSU
Out of season storm over Mississippi. Provide
good sample for MSU.
11
MSU OSSE update

• Investigating GSI and WRF-VAR for assimilation of
  synthetic observations
• ATMS (and later CrIS) radiance simulation by
  NESDIS
• Conducting baseline MM5 simulation
• T799 data serve as initial conditions and
  boundary conditions
• 9-km and 3-km domains
• will attempt to extract synthetic observations
  (i.e. RAOB, SFC)

12
Evaluation of Tropics in T511 Oreste Reale
Reale_Evaluation_ECMWF_T511NR.ppt

• The NR shows a stronger than climatology TEJ, but
  a very credible and realistic decrease of
  easterly speed with the progress of the monsoon
  season. As a consequence, the environment becomes
  conducive to more development
• First Nature Run to simulate one entire season
• Twelve tropical cyclones develop
• Realistic variability of tracks
• Most intense reaches 957 hPa
• Binary vortices, looping and singularities are
  observed (good from OSSE perspective)

13
Concluding remarks on the African Monson region
and tropical Atlantic (T511NR)

• A synoptic assessment of the NR over the AM
  region and the tropical Atlantic shows an overall
  very realistic African Monsoon, AEJ and wave
  activity
• Several weak tropical and sub-tropical systems
  are present, together with major tropical
  cyclones
• In spite of a tendency of creating several early
  recurvers, it can be stated that the NR, given
  the resolution limitation, does have a very good
  representativeness of tropical cyclone track
  variability in the Atlantic, as it would occur in
  an active season
• This Nature Run represents a very promising tool
  to perform OSSEs over the tropical Atlantic

14
Summary of Tropics of T799 NROreste Reale
Reale_T799_October_Nature_Run.ppt

• AEJ is 40 weaker than climatology
• Atlantic TC activity contains some highly
  suspicious tracks
• Eastern Pacific seems to present excessive
  proliferation of weak TCs
• The intensity of the strongest ATL systems is not
  superior to T511
• Different behavior in different basins
• Structure of some intense system not very
  satisfactory in terms of scale and size of
  eye-like feature

15
Preliminary conclusions

• The increased resolution does not necessarily
  provide stronger confidence in a much better
  Nature Run for the tropics
• Representation of TC activity does not
  immediately appear superior to the T511
• Caution should perhaps be used in adopting this
  NR for applications centered on future
  instruments targeting hurricanes
• Further investigation is needed

16
Summary Extratropical Cyclones 31R1 _at_ T511 (DJFM)
Thomas Jung Jung_ECMWF_Nature_Run_TJ.ppt

• Synoptic activity (extratropics)
• Generally well represented
• Distribution of storm frequency quite realistic
• Problems in the Gulf Stream/Kurushio area?
• Too many systems in the eastern Atlantic
  (blocking)

17
Summary Tropical Cyclones 31R1 _at_ T511 (DJFM)

• T511 model produces quite realistic frequency of
  occurrence of TS/Hurricanes in the Atlantic.
• North African Monsoon too strong northward
  shifted.
• AEJ is too weak (problem at higher resolution).
• Perhaps slightly too much AEW activity.
• Strong problems in the western tropical Pacific
  (too little atmospheric upwelling). Problem
  larger at higher resolution.

18
Some Other Issues Model Climate 31R1 _at_ T511

• Amplitude of the MJO is underestimated (better in
  recent cycles but still an issue).
• Convectively coupled Kelvin waves in the tropics
  are underestimated (better from 32R3 onwards).
• Too little (much) precipitation over the tropical
  continents (oceans) (improved in recent cycles).
• No QBO
• Euro-Atlantic blocking significantly
  underestimated (large improvements in 33R1).
• Anti-cyclonic circulation bias in the North
  Pacific (fixed in 32R3)

19
Tropical Cyclones in Recent Model Cycles

• Generally there is an increase in the
  number/intensity of tropical cyclones/hurricanes.
  
• Particularly lower resolution versions perform
  better than older cycles (e.g., US landfall).
• Large increases in the western tropical Pacific
  (partly due changes in the large-scale
  atmospheric circulation). Too large?

20
Sensitivity to Horizontal ResolutionThomas
JungJung_ECMWF_Nature_Run_TJ.ppt

• Short-range and medium-range forecasts suggest
  that T799, if anything, produces stronger
  hurricanes than T511.
• Resolution studies, however, suggests, that some
  aspects of the tropical climate (i.e., beyond the
  medium-range) of 31R1 deteriorate when increasing
  horizontal resolution (T159-gtT511). So it may be
  possible that T799 performs worse than T511.
• In the extratropics the largest changes occur
  when going from T95 to T159. Rather little
  changes occur beyond T159 (T159-gtT511). Hence, it
  seems reasonable to assume that T511 and T799
  perform similar.

21
Colder surface in Both poles (T799NR) Michiko
Masutani
Masutani_T799_T511.080410.ppt
Transient eddy kinetic energy become activs in SH
toward summer Both Poles are colder in T799 NR
10 day average sfc skin temperature
22
Arctic in T511 NRNikki Prive
Prive_arctic_ice_080221.ppt

• Stratospheric cold pole - without input from
  obs, model develops abnormally cold polar region
  in winter with strong polar vortex. Common
  problem with global models.
• Surface temperatures are strangely high over the
  arctic
• Disconnect with temperatures above surface layer?
  
• Low level inversion too weak - impacted by many
  processes
• What is energy balance over ice?
  SHF/LHF/longwave/cloud impacts?
• Sea ice and SST are applied from 2005/6 fields,
  not interactive with model atmosphere

23
Plot by Yoshiaki Sato Quick satellite view
datTB(ix,iy)datTB(ix,iy)(1.0-datC(ix,iy,iz))dat
T(ix,iy,iz)datC(ix,iy,iz) datC CC, datT T
IR
T799 NR 2006041106
24
New interests and projects
25
OSSE capability for GNSS Radio-Occultation (RO)
observations

• Lidia Cucurull
• JCSDA

Cucurull_OSSE_080828.ppt
26
Summary of the activities at the University of
Utah Zhaoxia Pu, University of Utah (
Zhaoxia.Pu_at_utah.edu)

• Evaluation of the ECMWF natural runs
• 1. Winter storms in T511
• 13 major storms are identified over
  the Northern American Area
• further Evaluation is in progress
• 2. Hurricanes in both T511 and T799
• Regional OSSEs (Obs DWL)
• A few issues 1) model and data
  assimilation systems
• 2) boundary
  condition issues
• 3) hurricane
  intensity (resolution and validation issues)

27
Yves J. Rochon Atmospheric Science and
Technology Directorate Environment Canada
My interest is essentially in conducting
sensitivity analyses in data assimilation
(currently global as oppose to regional). This
work has only just begun. Am currently conducting
an OSSE experiment consisting of (1) observation
simulation using a nature run from our global
model (set here for 120x240 horiz. resolution)
and (2) conducting a few assimilations of this
simulated dataset with and without one of the
observation sources. The simulated obs consist of
obs as used in weather forecasting (e.g.
radiosondes, AMSU-A among others) plus a couple
of additional stratospheric obs sources. For this
first trial, the nature run is simply a model run
covering 2.5 months and the assimilation uses
this same model (realizing that this is far from
the best choice but it's a first try) starting
from different initial conditions than the model
run. Will see what that gives. At some point in
the future, I would like to take advantage of the
nature runs to be made accessible from your joint
OSSE program.
28
OSSE at Northrop Grumman
Project Description The primary objective of the
effort is to provide NASA with the capability to
quantitatively assess the benefits of
observational data  produced by proposed future
observing systems. The research effort  planned
would look into extending the Sensor Web
Simulation that NGIT is  working on with NASA to
incorporate a rigorous end-to-end simulation of
 future sensors and observing platforms, thereby
allowing an accurate  assessment of the impact of
the operational data they produce on global
 weather forecasts. This supports the OSSE that
NASA is pursing by allowing quantitatively-based
decisions on the design and implementation  of
future observing systems. What NGST brings to
this collaborative  effort is EVEREST, the
Environmental Product Verification and
Remote Sensing Testbed. EVEREST is a
"world-class" end-to-end modeling and  simulation
testbed for assessing the performance of remote
sensing  systems. Initially developed to
establish the pre-launch performance of the
NPOESS weather satellite system, EVEREST consists
of comprehensive  environmental scenes, databases
and validated state-of-the-science  models for
physics-based phenomenology, radiative transfer,
sensors,  spacecraft, and data product retrieval.
On the NPOESS program, EVEREST  has been a
critical element in allowing NGST and the
government to  predict the quality of the data
products produced by the NPOESS
 sensor/spacecraft system, including the impact
of changing sensor,  spacecraft, algorithm and
other system design parameters. It was also used
during the PDRR phase of the GOES-R program to
help establish  system requirements for the
program. It is currently being used to ensure
that the as-built NPP/NPOESS system will perform
as specified.
29
Resource requirement T382 one node GSI 4
hour/day (4 cycle) one 5 day forecast 2.5
hour ? 4 cycle per day, 5 day fcst twice a day
with one nodes 9 hr/day T126 2 node GSI
40min/day (4 cycle) one 5day fcst 10min 4 cycle
per day, 5 day fcst twice a day with 2node
1hr/day GOES-R 20km 10 times thining -gt 60 km
resolution 15MB/hr 360MB/day150GB 20km
resolution 3.6GB/day1.5 TB/year 4lm
resolution 3.6x2590GB/day33TB/year