An Overview of the North American Regional Reanalysis

1
An Overview of the North American Regional
Reanalysis

• W. Ebisuzaki, F. Mesinger, G. DiMego, E. Kalnay,
  K. Mitchell, M. Ek, R. Grumbine, D. Jovic, P.
  Shafran, and J. Woolen
• http//wwwt.emc.ncep.noaa.gov/mmb/rreanl/inde
  x.html

2
Motivation for North American Regional
Reanalysis (NARR)

• Create hi-resolution reanalysis for the North
  American domain
• Address some of the weakness of the N/N
  Reanalysis, i.e., near surface temperature,
  winds
• Hydrology in the North American domain
• Make a more useful reanalysis (for people who
  want point values and live on the surface
• 
  Expectations
• more applied meteorological applications
• more uncertain about how non-meteorologists will
  use the data

3
High Lights of NARR

• ETA model and assimilation system (EDAS)
• 32 km resolution over the North and Central
  America
• R-2 horizontal boundary conditions
• new land surface model (NOAH)
• precipitation assimilation
• 8x daily analyses from October 1978 onwards
• 5 TB for basic set of analyses and fluxes
• Oct 1978-Nov 2002 done

4
NARR domain

5
180 km ( CDAS) vs 32 km (NARR) resolution
topography

6
ETA / NOAH LAND-SURFACE MODEL UPGRADES 24 Jul 01
- assimilation of hourly precipitation --
hourly 4-km radar/gage analysis (Stage IV) - cold
season processes(Koren et al 1999) -- patchy
snow cover -- frozen soil (new state variable)
-- snow density (new state variable) - bare
soil evaporation refinements -- parameterize
upper sfc crust cap on evap - soil heat flux --
new soil thermal conductivity (Peters-Lidard
et al 1998) -- under snowpack (Lunardini,
1981) -- vegetation reduction of thermal cond.
(Peters-Lidard et al 1997) - surface
characterization -- maximum snow albedo
database (Robinson Kukla 1985) -- dynamic
thermal roughness length refinements -
vegetation -- deeper rooting depth in forests
-- canopy resistance refinements
NOAH LSM tested in various land-model
intercomparison projects, e.g., GSWP, PILPS 2a,
2c, 2d, 2e, Rhone, and (near-future) DMIP.
7
January 1997 Precipitation Results
8
Precipitation

• Several sources of precipitation
• CONUS data with PRISM (Mountain Mapper) to
  improve orographic effects
• Canada
• Mexico
• CMAP (combination of satellite and gauge data)
  over oceans CMAP is blocked
• Near central areas of hurricanes (7.5 by 7.5 deg)
• Observed precipitation gt 100 mm/day
• A 15-degree 'blending belt' between 27.5 and 42.5
  N, with no CMAP north of 42.5 N

9
Comparison of N/N Reanalysis and NARRmodel

• Global
• Spectral
• 250 km resolution
• 28 sigma levels
• Land sfc (Pan and Marht, 1987)
• One soil type (sandy loam)
• Two soil levels
• Simplistic snow depth, fn(T)
• 6 hour analyses time steps

• N. Am. Sector
• Grid point
• 32 km resolution
• 45 pressure levels
• NOAH model (2002)
• Realistic soil types
• 4 soil layers
• Snow depth, modeled/observed
• 3 hour analyses time steps

10
Comparison of N/N Reanalysis and NARRinput
observations

• Sondes, aircraft
• TOVS temperature retrievals
• Satellite winds
• Land sfc prs
• Ocean sfc prs, q, t, winds
• Snow cover old is low res
• Reynold's SST
• Sea ice from R. Grumbine

• Sondes, aircraft, profilers
• Direct assimilation of radiances
• Satellite winds
• Land sfc prs, q, winds
• Ocean sfc prs, q, t, winds
• Precipitation
• USAF snow depth (hi res)
• Reynold's SST lakes
• Sea ice from R. Grumbine lakes

11
Comparison of N/N Reanalysis and NARRdates

• Software 1994
• Analyses 1948-present

• Software 2002
• Analyses Oct 1978-Nov 2002
• Nov 2002 in development

12
Operational ETA vs NARR

• 32 km
• 45 levels
• Frozen physics
• Better precip obs
• Better lake obs

• 12 km
• 60 levels
• Evolving physics
• Precip assim for CONSUS
• Lakes SST, ice-real time

13
Operational ETA vs NARRUser Output RR is more
friendly

• RR 1 grib table vs 3 grib tables for OPN
• (some software cannot handle multiple
  tables)
• RR Winds are N/S vs grid-relative winds in OPN
• (some software cannot handle grid-relative
  winds)
• RR new variables moisture fluxes, etc
• RR redundant variables removed ex dew pt, rel
  q, specific q
• RR added new BL levels, removed others
• RR BUFR files have QC and increments
• RR many monitoring plots

14
Why NARR should have better surface values

• 1) higher horizontal orography 32 km vs 250 km
• need to resolve orography
• 2) better land surface model
• to get 2m temp and q .. need better sfc
  fluxes
• 3) assimilated precipitation
• better model precipitation .. clouds,
  radiative flux
• better soil moisture .. latent and sensible
  heat flux
• 4) gridded snow depth is similar to NARR
  resolution
• a problem with NCEP's global reanalyses
• 5) assimilate more near surface data (winds, q)
• 6) 3 hour assim cycle, better resolve diurnal
  cycle

15
(No Transcript)
16
(No Transcript)
17
(No Transcript)
18
(No Transcript)
19
(No Transcript)
20
(No Transcript)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
(No Transcript)
25
(No Transcript)
26
(No Transcript)
27
(No Transcript)
28
(No Transcript)
29
(No Transcript)
30
Principal Output Datasets

• Observations with QC marks, increments (BUFR)
• plots obs locations, fit to sondes, fit to sfc
  obs,
• fit to global reanalysis, various NARR fields
• binary restart files (to rerun model, high
  resolution), very big, small usage
• GRIB analyses and first guess on model grid big,
  no interpolation errors
• GRIB analyses and first guess on Lambert
  conformal 32-km grid, big, easier to use,
  interpolation error
• Merged GRIB analyses and averaged/accumulated
  flux/precip and some 1st guess fields for water
  budgets, Lambert Conformal grid (AWIP 221), 5 TB
  for 8x daily for entire period. This will be
  work-horse data set.

31
Data Distribution

• NCAR
• SDSC (San Diego Super Computing Center, UCSD)
• U. of Maryland
• NCDC-NOMADS (NOAA Operational Model Archive
  Distribution System)

32
NOMADS

• Partnership of many organizations including core
  collaborators CDC, COLA, FSL, GFDL, LLNL, NCAR,
  NCDC, NCEP, PMEL, UNIDATA
• On-line access to the merged RR data set
  (analyses fluxessome first guess)
• NCDC 7 TB on-line storage, web servers, x-TB
  tape storage
• NCEP ¾ TB on-line, web servers, software
  upgrades
• Software
• on-line plotting package
• on-line slice-and-dice of GRIB files
• GrADS-DODS
• Status
• plotting package will be enhanced by NCEP
• downloading of GRIB files will be enhanced by
  NCEP
• GrADS-DODS needs enhancements
• NOMADS-NCDC has experience running software

33
I can't download the 5 TB merged data set!

• Download what you want
• Select
• 1) variables
• 2) levels
• 3) times
• 4) domain
• 5) resolution 32 km, 64 km, 96 km
• Downloading can be scripted

34
Summary

• Oct 4, 2003 Finished Oct 1978 to mid-Nov 2002
• no more inland lake data
• no 2003 Canadian precip
• currently planning the real-time NARR
• will use different datasets as some
• precip/lake data are not available in
  real time
• currently planning subsets and archive
• Dec 2003 internet-2 will be available at NCDC and
• distribution to archive sites can start
• Until NARR appears at archive sites, you can look
  at
• 1 year of merged data at NCEP.
• http//wwwt.emc.ncep.noaa.gov/mmb/rreanl/inde
  x.html