README

• Configuration and Dependencies

  1. compiler
     • make sure there is ifort install
     • enable compiler option in run.sh
  2. netcdf library
     • make sure there is netcdf library installed in the system
     • modify the include path in pom08_iosub.f for netcdf include accordingly. usually in /usr/include/netcdf.inc
       + netcdf library is best to be compiled in the fortran variant used

• Run the program with run.sh

  1. run the model with run script(execute run.sh)
  2. after successful execution, 3 files will be generated. pom08.x(the executable), out/pom08.nc(netcdf output), and out/pom08.log(model output messages)
  3. results will be in the out folder pom08.nc and pom08.log

• Dependencies

  • netcdf library version at least 4.1.2

• cite this model with Mellor, G.L., M.A. Donelan, and L. Oey, 2008: A Surface Wave Model for Coupling with Numerical Ocean Circulation Models. J. Atmos. Oceanic Technol., 25, 1785-1807, https://doi.org/10.1175/2008JTECHO573.1

• hosting source code for pom08 per request from Leo Oey

• see the pom08.f source for more comments

History

• the followings are concatenated from pomchanges and moreREADME from http://shoni2.princeton.edu/ftp/glm/

As of Feb. 28, 2008, I have made changes based on symetry analysis. Thus, for a straight channel with jm=11, properties should be symmetrical about j=6. They are now to 3 or 4 significant places if cor(i,j)=0.

For mode=3, vertical u(i,j,k) profiles are noisy, but not so for cor(i,j)=1.e-4 or when aam(i,j,k) (a la Smagorinsky) is replaced by aam(i,j,k)=aam_init.

In deep regions where q2's are very small, length scales are meaningless.

The directory, PROBLEMS, contain various problem definitions so that if one wishes to purge pom08.f of definitons in favor of one's own problem, they can be stored in the directory.

Herewith is a list of changes to pom08.f or pom08.c

2/4/08 In the initialization of s.r. wave, change tht_wnd=0 to tht_wnd(i,j)=0.

Change data statement cdmax/.0025 to cdmax/.005 The former is too restrctive; e.g., for laboratory scale waves.

In s.r. specaves add zeta,zetf to save list; e.g., save init,kpDd,F1d,F2d,F3d,F4d,zeta,zetf 2/13/08 In the do 8000 loop, -tpx0(i,j) and -tpy0(i,j) has been included in with (wusurf(i,j)-wubot(i,j)) and (wvsurf(i,j)-wubot(i,j)) respectively. 2/18/08 In s.r. advave, + Sxx, +Sxy, +Syy should be removed. 2/28/08 Changes in s.r.advct, s.r.advave, profu, profv.profq and s.r. wave and its supporting s.r.'s. Replace these s.r.'s OR the whole pom08.f 3/7/08 In s.r. seamount, there is an obvious syntax error in defining ele(j) and elw(j). 4/28/08 Some changes in s.r. profq. Best to swap the s.r. 1/23/10 Some changes in s.r. cur2wave and s.r. advct. Best to swap the s.r.'s 1/1/11 In s.r. wave, change save init to save dtw2,dtinv in s.r. advct, call wave2cur is redundant and has been commented out. 2/25/12 Stokes drift has been removed from velocities in profu, profv and profq. Stokes drift is governed by wave properties and should not be mixed in profu, profv or participate in TKE production. 3/8/12 A number of changes in profq particularly in calculating prod. Best to swap old profq for the new. 8/16/12 specavs has been chnged so that that which was a surface delta function is now distributed. S.R. wave2cur has been changed slightly; there are 2 sign changes and the additional delta function has been removed. It's distributed version is now incorporated in Sxx and Syy.

Feb. 12, 2014, changes were made in recognition that the term that was a delta function in the stress radiation is now distributed across the water column. See papers in the previous directory. Thus the declaration of Srad0 in S.R. cur2wave should be deleted and the minus sign, just above, after Srad(i,j)= should also be deleted. Srad0 can be deleted (or not) everywhere else in the code.