The following information is valid as of March 1, 2006. Information on the future update to the operational forecast and analysis system will be given whenever it is available.
JMA operationally performs three kinds of objective atmospheric analyses for global, regional and mesoscale forecast models. All of them employ a fourdimensional variational (4DVar) scheme on model coordinates for the analysis of surface pressure, temperature, vector winds and relative humidity. The specifications of the atmospheric analysis schemes are listed in Table 1. Global snow depth analysis is performed every day, and described in Tables 2a.
JMA runs Global Spectral Model (GSM) four times a day (36 hour forecast from 06 UTC and 18 UTC, 90hour forecast from 00 UTC and 216hour forecast from 12 UTC), Regional Spectral Model (RSM) twice a day (51hour forecasts from 00 and 12 UTC) and Mesoscale Model (MSM) eight times a day (15hour forecasts from 00, 03, 06, 09, 12, 15, 18 and 21 UTC). Typhoon Model (TYM) is run four times a day (84hour forecasts from 00, 06, 12 and 18 UTC) whenever a typhoon exists or is expected to be formed in the responsible area of RSMC TokyoTyphoon center. A mediumrange ensemble forecast and a onemonth ensemble forecast models run everyday and once a week, respectively. The basic features of the operational forecast models of JMA are summarized in Tables 3a and b and Tables 4a, b and c, respectively.
Cutoff time  

(global)  2:20 hours for early run analyses at 00, 06, 12 and 18 UTC map time, 11:35 hours for cycle run analyses at 00 and 12 UTC map time, 5:35 hours for cycle run analyses at 06 and 18 UTC map time 

(regional) 
2:45 hours for analyses at 00 and 12 UTC map time, 8:45 hours for analyses at 06 and 18 UTC map time 

(mesoscale)  50 minutes for analyses at 00, 03, 06, 09, 12, 15, 18 and 21 UTC map time  
Initial Guess  
(global)  6hour forecast by GSM  
(regional)  6hour forecast by RSM  
(mesoscale)  3hour forecast by MSM  
Grid form, resolution and number of grids  
(global) 
Gaussian grid, 0.5625 degree, 640 x 320 for outer model Gaussian grid, 1.1250 degrees, 320 x 160 for inner model 

(regional)  Lambert projection, 20 km at 60 N and 30 N, 325 x 257 for outer model grid point (1, 1) at northwest corner and (200, 185) at (140 E, 30 N) Lambert projection, 40 km at 60 N and 30 N, 163 x 129 for inner model 

(mesoscale)  Lambert projection, 10 km at 60 N and 30 N, 361 x 289 for outer model grid point (1, 1) at northwest corner and (245, 205) at (140 E, 30 N) Lambert projection, 20 km at 60 N and 30 N, 181 x 145 for inner model 

Levels  
(global)  40 forecast model levels up to 0.4 hPa + surface  
(regional)  40 forecast model levels up to 10 hPa + surface  
(mesoscale)  40 forecast model levels up to 10 hPa + surface  
Analysis variables  Surface pressure, temperature, wind and specific humidity  
Methodology  Fourdimensional variational (4DVar) scheme on model levels  
Data Used  SYNOP, SHIP, BUOY, TEMP, PILOT, Wind Profiler, AIREP, SATEM, ATOVS, SATOB, Australian PAOB, surface wind data from the scatterometer on the QuikSCAT satellite and MODIS wind data from Terra and Aqua; The RadarAMeDAS (dense network of surface observations including precipitation) composite precipitation data for the regional and mesoscale models; Doppler radar radial wind and precipitable water and rain rate derived from microwave radiometer on SSM/I, TMI and Aqua for the mesoscale model; Typhoon bogussing applied for all analyses 

Initialization  
(global)  Incremental nonlinear normal mode initialization and a vertical mode initialization  
(regional and mesoscale)  Nonlinear normal mode initialization with full physical processes for the first five vertical modes> 
Methodology  Twodimensional Optimal Interpolation scheme 

Domain and Grids  Global, 1 x 1 degree regular latlon grids 
First guess  Combination of USAF/ETAC Global Snow Depth monthly climatology (Foster and Davy, 1988) with previous analyzed snow depth anomaly 
Data used  SYNOP and METER snow depth data observed in past one day 
Frequency  Daily 
Basic equation  Primitive equations 

Independent variables  Latitude, longitude and sigmapressure hybrid coordinates and time 
Dependent variables  Winds (zonal, meridional), temperature, specific humidity, surface pressure 
Numerical technique  Spectral (spherical harmonics basis functions) in horizontal, finite differences in vertical Leapfrog, semiLagrangian, semiimplicit time integration scheme. Hydrostatic approximation 
Integration domain  Global in horizontal, surface to 0.4 hPa in vertical 
Horizontal resolution  Spectral triangular 319 (TL319) roughly equivalent to 0.5625 x 0.5625 degrees latlon Gaussian grid, 640 x 320 
Vertical resolution  40 unevenly spaced hybrid levels 
Time step  15 minutes 
Orography  GTOPO30 data set spectrally truncated and smoothed 
Gravity wave drag  Longwave scheme (wavelengths > 100 km) mainly for stratosphere Shortwave scheme (wavelengths approximately 10 km) only for troposphere 
Horizontal diffusion  Linear, fourthorder 
Vertical diffusion  Stability (Richardson number) dependent, local formulation 
Planetary boundary layer  Mellor and Yamada level2 turbulence closure scheme Similarity theory in bulk formulae for surface layer 
Treatment of sea surface  Climatological sea surface temperature with daily analyzed anomaly Climatological sea ice concentration 
Land surface and soil  Simple Biosphere (SiB) model 
Radiation  Twostream with deltaEddington approximation for shortwave (hourly) Table lookup and kdistribution methods for longwave (every three hours) 
Convection  Prognostic ArakawaSchubert cumulus parameterization 
Cloud  Prognostic cloud water, cloud cover diagnosed from moisture and cloud water 
Forecast time  90 hours from 00 UTC, 216 hours from 12 UTC and 36 hours from 06 and 18 UTC. 
The model specifications are the same as those of the high resolution GSM (Table. 3a) except for the following ones.
Horizontal resolution  Spectral triangular 159 (TL159) roughly equivalent to 1.125 x 1.125 degrees latlon Gaussian grid, 320 x 160 

Time step  20 minutes 
Number of members  51 members 
Initial state perturbation  Breeding of Growing Modes (BGM) method (25 independent breeding cycles in 12hour periods) 
Perturbed area  Northern hemisphere and tropics (20S90N) 
Initialization  Nonlinear normal mode initialization with full physical processes for the first five vertical modes 
Forecast time  216 hours from 12 UTC 
Basic equation  Primitive equations 

Independent variables  xy coordinate on a Lambert projection plane and sigmap hybrid coordinate 
Dependent variables  Wind components of xy direction, virtual temperature, natural log of surface pressure and specific humidity 
Numerical technique  Euler semiimplicit time integration, double Fourier for horizontal representation and finite difference in the vertical representation 
Projection and grid size  Lambert projection, 20km at 60N and 30N 
Integration domain  East Asia centering on Japan, 325 x 257 transform grid points 
Vertical levels  40 (surface to 10hPa) 
Lateral boundary  051 hours forecast by GSM runs 
Orography  Envelope orography, GTOPO30 data set is smoothed and spectrally truncated 
Horizontal diffusion  Linear, secondorder Laplacian with targeted moisture diffusion 
Moist processes  Large scale condensation + Prognostic ArakawaSchubert convection scheme + middle level convection + shallow convection 
Radiation (shortwave) (longwave)  Every hour Every hour 
Cloudiness  Diagnosed from relative humidity, maximum overlap 
Gravity wave drag  Shortwave scheme for lower troposphere 
PBL  MellorYamada level2 closure scheme for stable PBL, nonlocal scheme for unstable PBL, and similarity theory for surface boundary layer 
Land surface  Ground temperature predicted with the use of four levels in the ground Evaporability dependent on location and season 
Surface state  Observed SST (fixed during time integration) and sea ice distribution Climatological values for evaporability, roughness length and albedo Snow cover over Japan analyzed every initial time 
Forecast time  51 hours from 00 and 12UTC 
Basic equations  Fully compressible nonhydrostatic equations 

Independent variables  xy coordinate on a Lambert projection plane and terrainfollowing height coordinates and time 
Dependent variables  Momentum components in three dimensions, potential temperature, pressure, mixing ratios of water vapor, cloud water, cloud ice, rain, snow and graupel 
Numerical technique  Finite discretization onto the ArakawaC type staggered coordinates, horizontally explicit and vertically implicit time integration scheme, fourth order horizontal finite differencing in flux form with modified advection treatment for monotonicity 
Projection and grid size  Lambert projection, 5km at 60N and 30N 
Integration domain  Japan, 721 x 577 grid points 
Vertical levels  50 (surface to 21800m) 
Time step  24 seconds 
Initial fields  4DVAR analysis with mixing ratios of cloud water, cloud ice, rain, snow, and graupel derived from preceding forecasts considering consistency with the analysis field of relative humidity 
Lateral boundary  0318 hours forecast by RSM runs for MSM runs from 03 and 15UTC 0621 hours forecast by RSM runs for MSM runs from 06 and 18UTC 0924 hours forecast by RSM runs for MSM runs from 09 and 21UTC 1227 hours forecast by RSM runs for MSM runs from 12 and 00UTC 
Orography  Mean orography smoothed to eliminate the shortestwave components 
Horizontal diffusion  Linear, fourth order Laplacian + nonlinear damper Targeted moisture diffusion applied to the grid points where excessive updrafts appear 
Moist processes  Threeice bulk cloud microphysics + KainFritsch convection scheme Lagrangean treatment for the fall of rain and graupel 
Radiation (shortwave)  Every 15 minutes 
Radiation (longwave)  Every 15 minutes 
Cloudiness  Diagnosed from relative humidity with maximum overlap assumed 
Gravity wave drag  No parameterization scheme included 
PBL  Diffusion processes based on diagnosed turbulent kinetic energy, considering nonlocal effect by adjusting mixing length Similarity theory adopted for the surface boundary layer 
Land surface  Ground temperature predicted using a fourlayer ground model Evaporability prediction scheme included 
Surface state  Observed SST (fixed during time integration) and sea ice distribution Climatological values of evaporability, roughness length and albedo Snow cover over Japan analyzed every 6 hours 
Forecast time  15hours from 00, 03, 06, 09, 12, 15, 18 and 21 UTC 
The model specifications are the same as those of the RSM (Table. 4a) except for the followings:
Independent variables  xy coordinate on a Lambert (Mercator) projection plane for target typhoon north (south) of 20 N and sigmap hybrid coordinate 

Integration domain  Center of domain is set at the median of expected typhoon track in the Western North Pacific, 271 x 271 grids, surface to 17.5 hPa 
Projection and grid size  Lambert (Mercator) projection, 24 km at the typhoon center when the center of target typhoon is north (south) of 20 N 
Vertical levels  25 
Forecast phenomena  Typhoon track and intensity 
Lateral boundary  084 hour forecast by early GSM runs (00 and 12 UTC) 690 hour forecast by early GSM runs (06 and 18 UTC) 
Orography  Mean orography, GTOPO30 data set (30" x 30") spectrally truncated and smoothed 
Planetary boundary layer  MellorYamada level2 closure scheme and similarity theory for surface boundary layer 
Typhoon bogussing  Symmetric vortex generated using manually analyzed central pressure and radius of 30 kt winds with gradientwind balance assumed in the free atmosphere, Ekmanfrictional inflow and compensating outflow added in PBL and in upper levels, respectively (The vortex is blended with the global analysis in combination with asymmetric components taken from previous TYM forecasts, when available) 
Forecast time  84 hours from 00, 06, 12 and 18 UTC, maximum two runs 