********************************* RECENT MESSAGES SENT TO THE POM USERS GROUP (Note: only selective messages of broader interest are listed) -------------------------- Date: 05/18/2006 From: ezer -------------------------- Although this has been discussed before, following on questions on Pressure Gradient Errors and Hydrostatic Consistency, here are some clarifications (see Mellor et al., JTECH, 1994, Mellor et al., JTECH, 1998, also review of many PG related papers is given in Ezer et al., OM, 2002). "Hydrostatic Consistency" parameter usually is maximum (and most restrictive) near the bottom where abs(zz) is max and dz is small, say, HC=abs(zz(kb-1)*(h(i+1,j)-h(i,j))/(dz(kb-1)*0.5*(h(I,j)+h(I+1,j)))) However, according to Mellor et al., (1994), just requiring HC<1 can be a misleading concept since PG errors decrease with decreasing dx (thus dh), but also with decreasing dz, in fact, PGerr=0 for HC=0 AND HC=1 (when the two terms of the PG cancel each other). Models usually are not unstable even if HC is somewhat larger than 1 (say < 10 or so). Smoothing topography so dh/h < 0.4 and using rmean may be sufficient for most cases (Mellor et al., 1998), but users should check each application or try higher order schemes (by Chu, Shchepetkin and others) if desired. -------------------------- Date: 05/17/2006 From: ocerug@atlantis.org.br -------------------------- Mellor (1993) in \"The pressure gradient conundrum of sigma Coordinate Ocean Models\" presents the equation (1). How can i calculate each term in this equation? Could someone give me some numerical example? When i have a linear slope this index (equation 1) decreases toward the bottom using sigma=1.In this case the term dxH is constant for the entire slope. Is the term dsigma equal to dz(kbm1})for the bottom, isn´t it? Is this index more restritive for the bottom? Is this correct? -------------------------- Date: 05/10/2006 From: wwwwjjjj@mail.ustc.edu.cn -------------------------- > I can\'t find my output file! > In the code POM2k.f,I find in line 827 and line 828,there was a statment > as follow: > c call write_netcdf(netcdf_file,1) ! > *netCDF* > c call write_netcdf(netcdf_file,2) ! *netCDF* > But I can\'t find the subroutine \"write_netcdf\"! > > Answer (TE): Subroutine write_netcdf is in pom2k.n. In the version now online the CDF output is not used, you will need to uncomment "include pom2k.n" to reactivate it. -------------------------- Date: 05/01/2006 From: lyo@princeton.edu -------------------------- > A number of people have asked about WAD in POM. I have posted 2 > publications [oey*wad*pdf] that you can get from ftp://scylla.princeton.edu/pub/lyo/pom/ > These will also be posted on the POM\'s Pub. List. > Leo, I put your papers on the pom web online-publication, which can be found from main pom page or: http://www.aos.princeton.edu/WWWPUBLIC/htdocs.pom/PubOnLine/POL.html Tal. -------------------------- Date: 03/31/2006 From: glm@splash.princeton.edu -------------------------- Dear POM Colleagues, On the POM web site you will find a new manuscript entitled: A surface wave model for coupling with numerical ocean circulation models by G. L. Mellor and M. A. Donelan This describes a new surface wave model that can be coupled to pom2k.f. The model itself is in the contrib_code section in a folder called WAVES. For those interested, you will need to download the 3 files in the folder. I suggest you run the wave model alone before coupling to pom2k.f. It is set up for an idealized elliptical basin for which calculations are shown in the manuscript along with other cases. In the not too distant future, we will have a coupled pom2k.f (what to call it?), but those who have a model setup for a particular domain and application may wish to couple themselves. George Mellor P.S. In the manuscript, on p.14, change the long expression For H_M = . . . to the more simple H_M = (greek) gama D. On p. 18 change gama from 0.6 to 0.7. -------------------------- Date: 03/30/2006 From: tomasz.dabrowski@nuiqalway.ie -------------------------- I have a question related to ECOMSED. I am experiencing difficulties with \'hot starting\' ECOMSED. The model becomes numerically unstable and crashes during a second hot start run when I split the run into 6 hrs intervals (after approximately 17 hrs since the beginning of the cold start run). It is a tidal+surge model of the N-E Atlantic. It crashes after approximately the same time regardless of the date when simulation starts (regardless of the current pressure and wind conditions). Initially, I thought it a was machine-related problem and had something to do with precision when writing and reading the startup/restart file. However, to my surprise, when exactly the same model run is split into 12 hrs intervals or the model is run continuously, all calculations go fine for any time period. The time step I use is 300s, and 6*3600 = 21600s is divisible by 300s. The model is intended to be used as the operational model and 6 hrs runs have to be performed. I have carried out extensive tests to ensure that the problem here is purely due to the time interval between subsequent hot starts. It is something I cannot understand and I will greatly appreciate any help in this regard. Below are the specs of the platform I use and compiler flags: 32 bit Linux IBM XSERIES_255 server Intel(R) Xeon(TM) MP CPU 3.00 GHz FC = pgf77 FFLAGS = -Msave -byteswapio -fast LFLAGS = -Ktrap=fp -byteswapio -fast -o Kind Regards, Tomasz Dabrowski -------------------------- Date: 03/16/2006 From: whuang@eng.fsu.edu -------------------------- A Graduate Assistantship is available to support a Ph.D. student in the Civil and Environmental Engineering Department of Florida State University in USA. Major research task is to integrate a hydrological (rainfall runoff) model and an estuarine hydrodynamic model to study the impact of storm runoff and freshwater inputs on water quality in rivers and estuaries. The graduate assistantship is expected to start in the fall semester 2006. Admission Requirements: TOEFL >550, GRE(verbal +quantative)> 1100, GPA >3.0 (on 0-4 scale) MS degree Background desired: hydrodynamics,hydrology, Fortran and computer programming, hydrodynamic and/or hydrological modeling. Contact: Dr. Huang: whuang@eng.fsu.edu -------------------------- Date: 02/22/2006 From: typoon_zj@163.com -------------------------- I run POM in LINUX system.I have downloaded the latest version of pom2k from the web, and want to generate curvilinear grid in the directory GRID-DATA, in which the files are as follows: GRID.f, gridcom, gridborder,params, rungrid, TS.dat, WIND.dat, TOPO.dat. Will you please tell me how can I modify the subroutines in GRID.f to generate curvilinear grid? Thank you very much. Best wishes. sincerely from jing Answer: You only need to change gridborder for your particular region, grid.f is already a curvilinear generation program. Tal.