Review of the NCEP production Suite: Recent Changes and Plans

1
Review of the NCEP production SuiteRecent
Changes and Plans

• Marine Modeling and Analysis Branch
• EMC/NCEP

Hendrik L. Tolman Chief, Marine Modeling and
Analysis Branch NOAA / NWS / NCEP /
EMC Hendrik.Tolman_at_NOAA.gov
WHERE AMERICAS CLIMATE WEATHER AND OCEAN
SERVICES BEGIN
2
Mission Statement

• The Marine Modeling and Analysis Branch is part
  of the Environmental Modeling Center.
• MMAB provides analysis and real-time forecast
  guidance (1 16 days) on marine meteorological,
  oceanographic, and cryospheric parameters over
  the global oceans and coastal areas of the U.S.
• Evaluates quality of retrievals of ocean surface
  data from satellite borne sensors, improves their
  quality as needed, and examines the impacts of
  the data on forecast models.
• Monitors performance of operational guidance
  products.

3
Why MMAB ?

• Traditionally part of the EMC support to the NWS
  (OPC, TPC, Regions, WFOs). Justification
• General support of meteorological mission.
• From weather to Earth System modeling.
• Safety Of Life At Sea (SOLAS) 1974
• NWS responsible for wind waves.
• NOAA SAB on Ocean modeling 2004 and IOOS-MAST
  final report 2008.
• NCEP to be the backbone for ocean modeling within
  NOAA in strong partnership.
• Focus on global scales (NOPP, Navy, .).
• Enable/perform regional/coastal (NOS, IOOS, )
• Enable ecosystem modeling.

4
resources

• MMAB resources and collaborations
• MMAB consists of 91 FTE and 1241 contractors.
• Funding for ocean modeling much improved (HFIP,
  hurricane).
• Collaboration with all groups and branches within
  EMC.
• Global and mesoscale branch
• Reanalysis ice processing.
• SST and ice products.
• CFS ice and surface wind quality assessment.
• Coupling with ocean and waves
• HYCOM in NEMS, .
• WAVEWATCH III in NEMS, .
• Wind speed downscaling over coastlines.

5
resources

• Collaboration conted
• Hurricane group
• Base ocean modeling effort.
• Coupled HWRF (HYCOM, WAVEWATCH III).
• LSM group
• River inflow in oceans (USGS clim, ? NOAH?).
• Surge and inundation modeling.
• Ensembles group
• Waves.
• Coupled modeling (CFS).
• Assimilation / JCSDA
• SST retrievals and products.
• Altimeter SSH retrievals and assimilation.
• Wave and ocean data assimilation (SSH, SST,
  profiles, .)

6
resources

• Many other collaborations within NOAA
• OPC/TPC support of general services, feedback.
• NOS Ocean Modeling backbone, IOOS, storm surge.
• GLERL Great Lakes Modeling (waves, circulation)
• CBOS pilot projects for ecosystems modeling.
• OAR hurricane, ocean, sea spray.
• Collaboration with other government agencies
• US Navy wave and ocean modeling (Backbone).
• USACE wave modeling (possibly surge).
• USGS, . (Ocean Modeling Backbone).
• Other collaborations
• Co-development of WAVEWATCH III SHOM, URI,
  Industry, 7 NOPP teams.
• Coupled modeling URI, .
• Co-development of HYCOM (HYCOM follow on NOPP).

And then some .
7
joint ventures

• Collaborations to consolidate operational models
  on NCO managed compute resources
• NOS ocean circulation, PORTS, tidal and
  inundation models.
• Ecosystems pilot projects.
• NOS/GLERL water quality for Great Lakes.
• Chesapeake Bay pilot projects (sea nettles, HAB).
• Tsunami modeling with PMEL / ATWC and PTWC.
• Enabling organizations outside NCEP to become
  part of NCEP operational environment,

8
Wind wave modeling
9
waves

• Wind wave modeling efforts at NCEP are
  consolidated on the WAVEWATCH IIITM model.
• Developed as NCEP model, but used as a community
  modeling framework.
• Close to open-source approach, but with license
  and trademark.
• Version 3.14 publicly available since July 2009.
• Many external collaborators.
• Federal NRL, FNMOC, USACE, SHOM.
• NOPP Federal 7 teams from academia.
• Non NOPP collaborators (URI, LSU, Alkyon, Delft
  UT, ..)
• Many beta-testers.
• Two wave models based on WAVEWATCH III elements.

10
waves
Tolman et al, Dec. 8. 2009
NCEP production suite review, 10/50
11
waves

• WAVEWATCH III conted
• Present capability
• Global multi-grid or mosaic model.
• Hurricane wave models (to be transitioned).
• Great Lakes wave models (NAM or NDFD winds).
• Assimilation of altimeter and buoy data (to be
  transitioned).
• Global wave ensemble
• Joint effort with FNMOC, NAEFS/NUOPC.
• 2009 implementations limited by moratorium
• Great Lakes NDFD driven model.
• Following slides on ensembles and on
  implementations in the pipeline for FY2010.

12
wave ensembles
old
cycle
2008/03/28 t06z 48h forecast Old (2007) system
severely underestimates spread, new system much
more realistic, with 48-72h of memory of
perturbations.
cycle, BC
13
wave ensembles

• Wave model with cycling initial conditions and
  consistency in wind perturbations between cycles
  is first in the world to show realistic
  spread-error relationship.
• No separate wave perturbations needed?
• To be combined with FNMOC.
• We will have to address compatibility of system.

Wave height
Deter. Model error
Old ens. spread
New ens. spread
14
waves

• Global wave model upgrade (Q2-Q3)
• Generation of GRIB2 data on-the-fly as they
  become available, instead of at the end of the
  model run.
• TPC requirement.
• Removes legacy mode of operations.
• Extending computations from 77.5N to 83N.
• AK regional requirement.
• Code unification to model version 3.14.
• Bug fix for boundary data for West Coast Coastal
  grid.
• Increase of internal spectral resolution
  (tentative).
• Needed for future physics upgrades, but minor
  impact on present model.
• Spectral resolution of point output data will not
  change, until both global and hurricane model can
  be updated in lockstep.

15
waves (global)

• Extended model computations for Arctic Ocean.

16
waves (global)

• West coast bug fix

17
waves (global)

• Internal spectral resolution increase.

18
waves (global)

• Internal spectral resolution increase.

19
waves

• Hurricane wave model upgrade (Q3)
• Develop mosaic grids for hurricane wave models.
• Include high-res grids and shallow physics from
  global model.
• Retire old NAH and NPH hurricane wave models.
• Unification of wave model version 3.14 (see
  previous).
• Add output fields consistent with present global
  model
• Mainly wind-sea and swell separation.
• Increase of internal spectral resolution
  (tentative, see previous).

20
waves (hurricane)

• New hurricane wave model grid mosaic

21
waves (hurricane)

• Increased coastal hurricane wave model resolution

NAH
new
22
waves

• Additional plans FY2010
• Great Lakes NDFD driven model fixes in processing
  of NDFD winds (Q3).
• Deal with inconsistency between consecutive grid
  masks in NDFD winds.
• Flagging of missing data in SMS.
• Increase of spectral output resolutions of salt
  water models (Q4).
• Transitioning of data assimilation to multi-grid
  model.
• Limited resources, no hard time line yet.
• Adding FNMOC data to ensemble
• No time line yet, waiting for operational data
  feed from FNMOC.

23
waves

• Coastal wave modeling requirements have been
  formulated within the NWS (OSIP 06-093).
• Based on pilot projects spearheaded by WFO
  Eureka.
• Responsibilities to transition to NCEP using
  localized WAVEWATCH III
• Spiral development with local-central model.
• Local off-cycle capability.
• AWIPS-II or local workstation.
• Now funded by NOS through HFIP.
• Toward coupled wave-surge capability in
  collaboration with NOS.
• WAVEWATCH III ADCIRC capability.
• NOS ADCIRC model for East Coast for boundary
  data.

24
waves

• Miscellaneous issues that we are working on
• Coupling efforts have started.
• HWRF HYCOM - WAVEWATCH
• ESMF (NEMS) version being developed (NRL).
• NOPP project (NOAA-USACE-ONR, 2010-2014) to
  address deep water and shelf physics (2M/year).
• Approx. 2012 model physics upgrade.
• Approx. 2014 model physics upgrade.
• Shallow water (mosaic) applications require
  alternative grid and solution approaches (sub-km
  scale resolution)
• Generalized curvilinear grids (NRL).
• Unstructured grids for coastal areas, and
  coupling to surge (SHOM).

25
Ocean Modeling
26
ocean

• Present ocean modeling capability at NCEP
• 1/12º resolution Real Time Ocean Forecast System
  for North Atlantic Ocean. (RTOFS-Atlantic, HYCOM
  based).
• Deterministic 5-6 day forecast.
• Assimilation of most available data with in-house
  assimilation schemes.
• Run once per day.
• Emerging global HYCOM capability.
• MOM3-4 (GFDL) ocean model as part of the coupled
  Climate Forecast System (CFS).
• Supporting MOM4 based Global Ocean Data
  Assimilation System (GODAS).
• Test version of HyCOM-HWRF coupled hurricane
  model.
• Emerging NOS capabilities at CCS.

27
RTOFS-Atlantic upgrade

• Due to moratorium and lack of tools to rapidly
  update initial conditions of RTOFS,
  RTOFS-Atlantic has been severely degraded for
  much of FY2009.
• Nov. 2009 upgrade has alleviated most problems
  (see following slides.
• Systematic changes in operations needed and
  identified.
• Agreement with NCO to allow for re-initialization
  of RTOFS-Atlantic by EMC on regular (quarterly)
  basis.
• Re-engineering of assimilation mechanics
• 1-2 week assimilation cycle.
• Improved and interactive QC of data.
• Synthetic data?

Short term
Long term
28
RTOFS-Atlantic upgrade

• 2009 Upgrade details
• QC criteria for SSH
• Add check to climatology to suppress model drift.
• Modified SSH mask for the Middle Atlantic Bight
  (MAB)
• Removing systematically questionable SSH data
  (see following slides).
• Add pseudo-observations in the MAB beneath the
  SSH mask to preserve the mean Gulf Stream
  position (see following slides).
• Updated tidal elevations and transport boundary
  conditions from TPX07.
• The mean SSH is shifted by a value of -0.25m
  for the model-west (South Atlantic) boundary.

29
RTOFS-Atlantic upgrade

• Mask modifications

• Pseudo data

30
RTOFS-Atlantic upgrade
Impact of the modified SSH quality control
algorithm on the entire basin in upgraded model
(left panel) compared to SSH in old model (right
panel).
31
RTOFS-Atlantic upgrade
Idem, surface currents.
32
RTOFS-Atlantic upgrade
Transports for the old model (right panel) run
were double of those measured by AOMLs cable
from Florida to Bahamas. New model over-predicts
modestly In contrast Navys results tend to
under-predict.
33
RTOFS-Atlantic upgrade
Potential Temperatures from model (left panel)
and observations (right panel) at WOCE section A03
34
RTOFS-Atlantic upgrade
Salinities from model (left panel) and
observations (right panel) at WOCE section A03
35
RTOFS-Atlantic

• RTOFS-Atlantic plans.
• Planned FY2010 Q3 model upgrade
• Adding Jason-2 and minor tweaks to assimilation.
• Robustness of system
• Need at least 3 altimeters to accurately
  constrain model.
• Model less likely to drift, likely to drift
  slower.
• Jason-2 already used in EMC re-initialization,
  hence small impact on results expected.
• Longer term plans
• Work with Navy and Academia on more unified
  assimilation system.
• Re-engineering from slide 24.
• Add data (e.g., synthetic profiles).

36
Coupled HWRF- HYCOM system
RTOFS-Atlantic (HYCOMBasin) O(417 km)
IC
BC
GFS O(25 km)
Data Assimilation (SST, SSH, TS) using 2D/3DVAR
meld
Feature Model Wind
SST
ATMOS. HWRF O(9 27 km)
OCEAN HYCOMRegional O(814 km)
Wind-stress Heat Fluxes Precipitation
Atm. Pressure
37
(No Transcript)
38
ocean
Real time testing for 2008 hurricane season
(Ike). Realistic Oceanic Simulation and Response
to a Storm!
39
HWRF-HYCOM

• In parallel real-time testing for 2009 hurricane
  season.
• Models Coupled HWRF-HYCOM (HyHWRF), where HWRF
  is the 2009 version (will do with 2008 ops
  version).
• Issues with 2008 HWRF version that greatly
  complicate coupling.
• Approach Real-time in parallel to Operational
  HWRF
• Cases
• 1. Hurricane Bill (Category 4)
• 2. Tropical Storm Danny
• 3. Tropical Storm Erika
• 4. Hurricane Fred (Category 3)
• 5. Tropical Storm Grace
• 6. Tropical Storm Henri

Ida still to be included
40
HWRF-HYCOM

• Typical example Hurricane Bill
• Small but systematic improvements for HyHWRF

Green NCO operational HWRF (coupled to
POM) Black parallel HWRF (version 2009) coupled
to POM Red parallel HWRF (version 2009) coupled
to HYCOM
41
HWRF-HYCOM

• Atypical example Hurricane Fred
• H209 poor, HyHWRF (with H209!) good.

Green NCO operational HWRF (coupled to
POM) Black parallel HWRF (version 2009) coupled
to POM Red parallel HWRF (version 2009) coupled
to HYCOM
42
HWRF-HYCOM

• Overall 2009 parallel results
• Two populations due to several short weak systems

Green NCO operational HWRF (coupled to
POM) Black parallel HWRF (version 2009) coupled
to POM Red parallel HWRF (version 2009) coupled
to HYCOM
43
HWRF-HYCOM

• HWRF-HYCOM plans.
• Continue testing in to determine possible FY2010
  implementation.
• Complicated by stress issues in HWRF model.
• Underlying HYCOM reached reasonable maturity.
• HYCOM system much more complex than POM.
• HR issues being addressed.
• Continue tuning of two-way coupling.
• Adding data assimilation to coupled model
  (presently in RTOFS-Atlantic only).
• Combine three sub-domains in single domain.
• Economy versus complexity.
• Working with URI on coupled HWRF-HYCOM-WAVEWATCH
  III model.

44
Ocean

• RTOFS-Global
• Adopting 1/12 global HYCOM from navy (NAVO/NRL).
• We have been able to run code here.
• We have established data transfer of initial
  conditions from NAVO (Not yet hardened).
• Operationalization at NCEP delayed 4-6 months due
  to lack of disk space in initial CCS transition.
• Building low-resolution RTOFS-Global for
  application in NEMS. FY2010 pre-ops deliverables
• NEMS coupler for HYCOM with full regridding.
• Initialization from upscaling of RTOFS-Global-HR.
• Initialization by using GODAS.
• Initial assessment of resolution (hor. - vert. -
  ML).

45
Sea Ice and SST
46
sea ice
Northern and Southern Hemisphere daily ice
concentration products.
47
SST
Used as a separate product, and for validating
ocean modeling.
48
Ice and SST

• MMAB (Bob Grumbine) now becoming the focal point
  for all EMC ice products.
• Previously some lack of communication between
  MMAB, CFS group and JCSDA contributions.
• Recognized need for different products for
  real-time (OPC, EMC modeling) or CFS products
  (climatological consistency).
• More consistent approach between products, and
  between SST products and ocean modeling.

49
new instruments

• Ice
• AMSR-E Implemented May, 2009
• Higher resolution instrument (12.7 km, vs. 25.4
  for SSMI)
• Improved Team2 algorithm.
• Replaced corrupted F-15 SSMI.
• Now only instrument (F-13 SSMI removed from
  operations December, 2009).
• SST
• METOP-A replaced NOAA-17 June, 2009.
• NOAA-19 will replace NOAA-18.
• AMSR-E will be added to SST analysis.
• See through clouds ? improve currency of analysis
• But lower (50 km footprint) resolution than
  AVHRR.
• Keeping our eyes on GOES.

50
new products

• SST grids include
• Weekly 1 degree OI
• Daily 1 degree OI
• Daily 0.5 degree RTG
• Daily 5' RTG (unsmoothed)
• NEW 15' Reynolds OIv2

51
New Processing

• New and improved land mask for Ice and SST
• Especially better representation of lakes
• 5' RTG will be unsmoothed
• Physical retrievals blended with Navy retrievals
  near coast
• Land issues in physical retrievals
• Better lake temperatures
• Diurnal layer used in GSI assimilation
• And to be added to GFS, CFS/GODAS
• Q Interest in diurnal layer temperatures
  themselves?
• Request
• Feedback on realism of SST products in lakes and
  near shore http//polar.ncep.noaa.gov/develop/sst
  /inland

52
Questions ?