Documentation regarding free surface pressure initialisation is unclear

[Jon Hill]

As reported on the Fluidity mailing list:

"When reading in a netcdf file for initial condition of free surface,
I found that the values after reading in become one order larger than the original values.

For example,

-2 < Free Surface (in my netcdf file) < 4
in xxx_0_0.vtu file, it is
-20 < FreeSurface < 40

Anyway, I am testing with a manipulation of the netcdf file.
Is it a bug or just my simple mistake somewhere."

The documentation in the schema and the manual needs to tell the user that the initial pressure is formed by:
free surface * gravity magnitude * density
and that they should scale their input appropriately.

[Adam Candy (asc)]

The files relating to this bug are up on scratch at:
/scratch/asc/seto/
Lee recently sent initSWL.grd and seto_0_0.vtu, and there's also the seto.flml from a previous mail. The other files have been generated by me. We don't have everything to run the case - we need the mesh file, so I've asked Lee to send this also. I'll look at the code in more detail to try and identify the issue - this will be easier once we have the mesh information.

[Adam Candy (asc)]

Simon Mouradian pointed out that the source netCDF file may have 'variable attributes' and 'offset' defined, which Fluidity does respect. These do not appear in the headers though (see below), which implies this is not the issue.

> # ncdump -h initSWL.grd
> netcdf initSWL {
> dimensions:
> x = 2161 ;
> y = 1321 ;
> variables:
> double x(x) ;
> x:long_name = "x" ;
> x:actual_range = 126., 144. ;
> double y(y) ;
> y:long_name = "y" ;
> y:actual_range = 28., 39. ;
> float z(y, x) ;
> z:long_name = "z" ;
> z:_FillValue = NaNf ;
> z:actual_range = -0.20319390296936, 0.410419315099716 ;
>
> // global attributes:
> :Conventions = "COARDS/CF-1.0" ;
> :title = "surface" ;
> :history = "grdmath initSWL.grd 0.1 MUL = initSWL.grd" ;
> :GMT_version = "4.5.6" ;
> }
>
> # grep FreeSurface seto_0_0.vtu
> <DataArray type="Float64" Name="FreeSurface" format="binary" RangeMin="-1.5027338993" RangeMax="4.0253471026">
>
> # grep Pressure seto_0_0.vtu
> ...
> <DataArray type="Float64" Name="Pressure" format="binary" RangeMin="-14.741819553" RangeMax="39.488655076">

[Adam Candy (asc)]

Lee has provided the mesh file, so we can now run the simulations he has been finding this issue with.
See /scratch/asc/seto/

[Adam Candy (asc)]

Note the triangle node file needs to be parsed through sed first, like the Med tide example.

[Adam Candy (asc)]

We're also missing the bathymetry data file. I'm looking into whether this can be debugged without this file - we should be able to.

[Adam Candy (asc)]

This issue has been solved. I've found the problem and will commit a fix soon.