SEAM - Spectral Element Atmosphere Model

I   Conversion from Fortran 77 to Fortran 90

(a) Comparison of Fortran 77 and Fortran 90 code on Eagle (e) and Seaborg (s) for Held-Suarez forcing expeiments using sigma (s) or hybridsigma (hs) coordinates. The time averages are done for days 200-1200 from every-5-day , 'instantaneous' model output interpolated to the latitude-longitude grids. 

Eagle (1000-day average) Seaborg (1000-day average) Difference (between Eagle & Seaborg)
s77e   [.pdf] [.eps]
s90e   [.pdf] [.eps]
hs77e [.pdf] [.eps]
hs90e [.pdf] [.eps]
s77s   [.pdf] [.eps]
s90s   [.pdf] [.eps]
hs77s [.pdf] [.eps]
hs90s [.pdf] [.eps]
s77e-s77s    [.pdf] [.eps]
s90e-s90s    [.pdf] [.eps]
hs77e-hs77s [.pdf] [.eps]
hs90e-hs90s [.pdf] [.eps]
Difference: (a) F77 & F90:
s77e-s90e    [.pdf] [.eps]
hs77e-hs90e [.pdf] [.eps]
(b) Sigma & HybridSigma:
s77e-hs77e   [.pdf] [.eps]
s90e-hs90e   [.pdf] [.eps]
Difference: (a) F77 & F90:
s77s-s90s    [.pdf] [.eps]
hs77s-hs90s [.pdf] [.eps]
(b) Sigma & HybridSigma:
s77s-hs77s   [.pdf] [.eps]
s90s-hs90s   [.pdf] [.eps]
Legend:

s (as in s77/s90): Sigma
hs: HybridSigma
e: Eagle (16 node x 4 procs/node)
s (as in 77s/90s): Seaborg (4 node x 16 procs/node)

(b) Comparison of Fortran 77 and Fortran 90 code on Eagle (e) for Held-Suarez forcing experiments using hybrid (h) coordinates. The time averages are done for days 200-1200 from every-5-day, 'instantaneous' model output interpolated to the latitude-longitude grids.
 
Fortran 77 (1000-day average) Fortran 90 (1000-day average) Difference
h77e   [.pdf] [.eps] h90e   [.pdf] [.eps] h77e-h90e    [.pdf] [.eps]

(c) Comparison of Fortran 90 code and Fortran 90 code with *.h files (common blocks) replaced by modules. The two simulations on Eagle (e) are identical for Held-Suarez forcing experiments using sigma (s) coordinates. The time averages are done for days 200-1200 from every-5-day, 'instantaneous' model output interpolated to the latitude-longitude grids.
 
Fortran 90 (1000-day average) Fortran 90 with modules (1000-day average) Difference
s90e   [.pdf] [.eps] sm90e   [.pdf] [.eps] s90e-sm90e    [.pdf] [.eps]

(d) The Concurrent Version System (CVS) is used to maintain the source code. A brief description of CVS can be found in webpage http://www.mmm.ucar.edu/mm5old/misc/cvs.html.


II  Expeiments with topography

Held-Suarez forcing experiments with topography and different model resolutions using sigma coordinates. The every-5-day, 'instantaneous' model output (interpolated to the latitude-longitude grids) is first interpolated to the pressure levels before statistical analyses. The time averages are done for the 3rd year or days 840-1200 of integration.

Coarse mesh
Medium mesh with topography
Fine mesh with topography
with topography: sm90s-cm-topo [.pdf] [.eps]
sm90c-mm-topo      [.pdf] [.eps]
sm90s-fm-topo      [.pdf] [.eps]
without topography: sm90s-cm     [.pdf] [.eps]
F77: s77c-mm-topo [.pdf] [.eps]
F77: s77s-fm-topo [.pdf] [.eps]


III  Expeiments with local mesh refinement and topography

Held-Suarez forcing experiments with local mesh refinement (LMR) over the continental US and with topography. Mesh 1 [.pdf] [.eps] and 2 [.pdf] [.eps] are two examples of LMR over the continental US [25N-50N, 125W-75W]. Experiments are done using sigma coordinates and Fortran 90 code with modules. The every-5-day, 'instantaneous' model output (interpolated to the latitude-longitude grids) is first interpolated to the pressure levels before statistical analyses. The time averages are done for the 3rd year or days 650-1000 of integration.

LMR without topography LMR with topography
sm90s-lmr [.pdf] [.eps]
sm90s-lmr-topo [.pdf] [.eps]


IV Experiments with 'real' data and coupling with NCAR's CAM physics

(a) Held-Suarez forcing experiments with 'real' data: 'Instantaneous' sea level pressure (PSL) at days 0, 5, and 10.


Held-Suarez forcing only
Held-Suarez forcing and topography
CAM_SEM
Psl [.pdf] [.eps]
Psl [.pdf] [.eps]
CAM_EUL
Psl [.pdf] [.eps]
Psl [.pdf] [.eps]

(b) Coupling SEAM with NCAR's CAM physics - 'Aqua-planet' Experiments: 'Instantaneous' sea level pressure (PSL) at days 0, 5, and 10.


Aqua_planet experiments
CAM_SEM
Psl [.pdf] [.eps]
CAM_EUL
Psl [.pdf] [.eps]

 

V. Presentations

(a) A poster at the 8th Annual CCSM Workshop, Breckenridge, Colorado, 24-26 June 2003.

(b) A presentation at the CCSM AMWG meeting, Boulder, Colorado, 9-10 March 2004.

(c) A poster at the 2004 SciDAC PI meeting, Charleston, SC, 22-24 March 2004.

 

Updated 18 Mar 2004. Comments or questions: Houjun Wang.