CRTF Telecon - 03-25-2015

NOAA Climate Reanalysis Task Force

25 March 2015, 2-3 pm EDT

Recording: https://mapp.adobeconnect.com/p7fo6g5oifb/

Rapporteur: Steve Penny, U. of Maryland

2:00-2:05 Welcome and Introduction, Jim Carton, U. of Maryland

Theme: All Foci

1. Hydrological cycle, salinity, and water vapor

2. Total Heat transport and storage in ocean and atmosphere

3. Polar changes

4. Extreme events in reanalysis

2:05-2:30 Coupled Data Assimilation at NCEP. Suru Saha, NCEP/EMC

Theme: Hydrological cycle and water vapor

2:30-2:55 Improving the representation of aerosol-cloud in NCEP global models. Sarah Lu, SUNY-Albany

2:55-3:00 Task Force business items

                

NOAA Climate Reanalysis Task Force Technical Workshop

4-5 May 2015

NOAA Center for Weather and Climate Prediction, College Park, MD

Draft Prospectus

Notes (in progress...)

Coupled Data Assimilation at NCEP. Suru Saha, NCEP/EMC:

The Climate Forecast System Reanalysis (CFSR) (2010) http://journals.ametsoc.org/doi/abs/10.1175/2010BAMS3001.1

The CFSR includes a 9-hr coupled T574/L64 forecast guess (GFS + MOM4 + Noah)

Atmos Global Data Assimialtion System (GDAS): 6 hour analysis cycle

Global Ocean Data Assimilation System (GODAS): 6 hour analysis cycle

Global Land Data Assimialtion System (GLDAS): once daily at 18Z

The Climate Data Assimilation System version 2 (CDASv2) provided a number of major improvements over the existing NCEP/DOE Reanalysis (R2), including improved resolution (T62 -> T574; 210km -> 27km; 28 -> 64 vertical layers; 7 -> 24 layers above 100 hPa; 6 -> 13 layers below 850 hPa; 40m to 20m lowest layer thickness). The analysis scheme was upgraded from the SSI to the GSI, and assimilation of temperature retrievals was changed to assimlation of radiances, with a large increase in the number of sattelites used.

From the CFS to CFSv2, the ocean model transitioned from GFDL's Modular Ocean Model version 3 (MOM3) to MOM4.0d, going from partially (no representation in the arctic) to fully global. The resolution was increased from 1 degree global to 1/2x1/2 degree (with latitudinal refinement within 10 degrees of the equator to 1/4-degree); 40 Z-levels. (Dave Behringer, EMC)

GLDAS is running Noah LSM under NASA/Land Information System. (Mike Ek, EMC)

The CFSR uses historical concentration of CO2 via the historical TOVS instruments retrofit into the Community Radiative Transfer Model (CRTM).

Linear trends closer are to observations with CFSR vs. the original NCEP/NCAR reanalysis (R1)

The response of precipitation to warming SST is too fast in the NCEP R1 and R2 (due to a lack of SST cooling from atmos/ocean feedback effects). The realistic lag of precipitation was improved with the CFSR, e.g. to be in close agreement with observed lags in the Tropical Western Pacific winter.

Using NOAA Environmental Modeling System (NEMS) and the National Unified Operational Prediction Capability (NUOPC) - the goal is to improve cooperation between NOAA, the Navy and the Air Force. NEMS is a unifying single framework to share comon structures and components to expedite interoperability between NOAA, Navy and AirForce modeling systems. The NUOPC, for example, will help standardize the connections to physics packages to enhance their portability. (Mark Iredell, EMC)

NAM/NMMB and GOCART have been implemented with NEMS into the NCEP operations, with others such as the GFS, etc. to followin the next few years.

A prototype Climate Data Assimilation System (CDASv3) is being constructed:

It will combine: Atmosphere, Ocean, Sea Ice, Wave, Aerosols, Chemistry, and Land

GDAS: The hybrid 3DVar/EnKF with 40-member coupled forecast with Semi-lagrangian dynamics, with 64 levels in the vertical hybrid sigma/pressure coordinates.

T254 EnKF / T574 3DVar

Proposed upgrades to GDAS:

Inline bias correction for correcting for satellite data / radiances. 

Stochastic physics is to replace addictive inflation in the atmospheric data assimilation.

The GDAS system will be upgraded to a Hybrid 4D-EnVAR, (ie. no tangent linear or adjoint model is used). Can still perform an outer loop, as in the traditional 4DVAR. Potentially improved initialization of the forecast, due to 4D IAU (incremental analysis update) , which allows for the prescription of an incremental 4D trajectory, instead of a single time-level of analysis increment. (Daryl Kleist, UMD/EMC)

GODAS: Ocean will be GFDL MOM 5.1 and SIS for ocean and sea ice, respectively, using the current CFSv2 coupler. Depending on NEMS compatibility, the system may transition to the GFDL MOM 6 / SIS2 system. Resolution will be upgraded to 1/4x1/4 degree global (1440x1020) with a possible increase from 50 to 75 vertical levels, using 2m resolution near the surface. Near surface innovations of SST and surface winds will be assimilated into the ocean GCM.

For the Ocean DA, the Hybrid-Gain 3DVar/LETKF will be used http://journals.ametsoc.org/doi/abs/10.1175/MWR-D-13-00131.1

The Near Surface Sea Temperature (NSST) model (Xu Li & John Derber) will do assimilation of direct radiance measurements. The NSST will receive bulk surface temperature estimates from the MOM OGCM.

http://www.cpc.ncep.noaa.gov/products/ctb/meetings/2012/CFSv2/Li_X_Incl…

Aerosols: Inline GOCART for aerosol coupling (Sarah Lu)

Waves: Inline WaveWatchIII for wave coupling (Arun Chawla, Henrique Alves, Hendrik Tolman; EMC)

Wave coupling: GFS/Ocean/WaveWatchIII

Coupling of waves back to the ocean is being worked on by GFDL and will possibly be available in MOM6 (including Langmuir mixing and stokes drift) (note: a prototype already exists with MOM5, e.g. see http://journals.ametsoc.org/doi/abs/10.1175/JCLI-D-13-00583.1)

Wave DA: GFS and LETKF (Arun Chwala & Henrique Alves)

GLDAS: (Land) Inline Noah Land Model for land coupling

Upgrades to the Global Land Data Assimilation System (GLDAS) in CDASv3

New datasets: near real-time land data use.

Sea-ice observations: remote/in situ, real-time or climatology, fraction and thickness. (Xingren Wu; EMC)

Expected upgrades to GDAS December 2015: 

Hybrid 4D-EnVar, with hourly time bins (7 time levels per analysis)

80-member ensemble

T1534/L64 deterministic run with T574/L64 ensemble

Temporally adaptive thinning of data

Cloudy microwave radiances

Upgrade to CRTM 2.2

Additional aircraft and AMV data

Bias correction of aircraft temperature observations

NSST incorporated into forecast and analysis, using radiances, to improve diurnal cycle and surface forcing

Plans for CFSv3 include increased coupling, increased resolution and increased computational options, better description of extremes (hurricanes and storm surges), producing a better daily cycle is necessary for 2m Temperature, interface, etc.

Toughest challenges: Analysis of sea-ice, improvement in modeling, coupling of sea-ice to freshwater and impact on the AMOC

Improving the representation of aerosol-cloud in NCEP global models. Sarah Lu, SUNY-Albany:

include aerosol effects on radiation and clouds. Help improve assimilation of satellite observations.

Larger cloud droplets => less reflective

Smaller cloud droplets => more reflective

Already in place for NRL, ECMWF, NCEP, UKMetOffice

International cooperative aerosol prediction.

NEMS GFS Aerosol component (NGAC)

NCEP’s global in-line aerosol forecast system.

Prognostic aerosol capability

many proposed upgrades to the physics suites

New microphysics in GEOS-5

Transitioning from one to two-moment scheme

Adopting GEOS-5 aerosol cloud package into the CFS (GFS)

GOCART/MAM7

Explore resources vs. complexity

Need ohs-based diagnostic package.

Developing aerosol DA capability in the hybrid EnKF-GSI is underway

Downstream: providing 3D-aerosol fields for SST retrievals

Suru:

now: prototype CDASv3

later: decide ocean/sea-ice model

technical development will continue for about a year, then the testing starts to get ENSO right, etc.

Workshop: May 4-5, at NCWCP, College Park, MD

Workshop prospectus gives what could be covered. There is an umbrella focus on reanalysis (technical focus). Success equals the EMC informed of the taskforce activities. A written document: NOAA OAR technical report, will be created based on the workshop and discussions. We intend to get community feedback on the fully coupled system.