[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: VariableStandarized

Unidata Support wrote:


Hi John-

I got the following from Rich Signell


To: address@hidden
From: Rich Signell <address@hidden>
Subject: IDV s-coordinate problem when elevation is packed



Here's an teeny bug:
I discovered that IDV 1.1 does not compute the
vertical coordinate correctly when the sea surface height
variable (used in the formula terms) is stored as a packed
variable (nc_short w/ scale_factor & add_offset) instead of
the usual floating point.   It could be this holds for the
bottom depth as well -- I didn't check.


I'm guessing that the solution is to change:

       etaVar         = ds.findVariable(etaName);
       sVar           = ds.findVariable(sName);
       depthVar       = ds.findVariable(depthName);

to:

       etaVar         = new VariableStandardized(ds.findVariable(etaName));
       sVar           = new VariableStandardized(ds.findVariable(sName));
       depthVar       = new VariableStandardized(ds.findVariable(depthName));

or should it be:

       etaVar         = new VariableStandardized(ds.findVariable(etaName), true
);
       sVar           = new VariableStandardized(ds.findVariable(sName), true);
       depthVar       = new VariableStandardized(ds.findVariable(depthName), tr
ue);

with the true being used to set missing values to NaN?

yes, you're right, thanks for the report.

both forms are equivilent, since useNaNs = tru is default.
this is fixed systematically in netcdf-java version 2.2, as all variables in a NetcdfDataset automatically handle the scale/offset. 
I will fix in the version 2.1.