Hello all,
We are in the process of switching standard model output at CMOP over to
CF compliant netcdf. As we do so, we are looking at whether we can
retain our data in projected coordinates or whether we need to switch to
using global lat-lon coordinates. We have run into an ambiguity in the
handling of directional data in a projected coordinate system in CF.
If I have model output velocities generated in a projected coordinate
system, should velocity data that is relative to the projected
coordinate system?
Take the canonical lambert conformal conic example (example 5.7 at
http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.0/ch05s06.html). If
I add a variable:
float VelN(time, y, x);
VelN:standard_name ="northward_sea_water_velocity"
VelN:coordinates="lat lon";
VelN:grid_mapping = "Lambert_Conformal";
would it then be understood (by CF compliant software) that this is the
y-component of velocity within the context of the projection, which will
differ from the true Northward component of velocity at most points
within the projection? Or does using the name "northward_X" mean that
the velocity should be the true Northward component of velocity?
thanks,
Charles Seaton
Research Associate
Center for Coastal Margin Observation and Prediction (CMOP)
Oregon Health and Science University
cseaton at stccmop.org
Received on Wed Feb 06 2013 - 19:55:12 GMT
