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

[netCDF #PSZ-497222]: Combination of nested NetCDF files

Hi Brian,

> This is Brian Etherton, out in North Carolina.  Hope all is well with
> you guys out in Colorado, and all in the netcdfgroup.
> I have a quandary, and the information presented to me on the Unidata
> web site is so great as to be overwhelming.  Hence.  My note.
> I have a triple nested WRF run, a 27km outer nest, a 9km middle nest,
> and a 3km inner nest.  The 9km and 3km nests move every time step,
> following a hurricane (in this case, Earl).
> What I want to do is to combine the three NetCDF files valid at a
> moment in time into one file.  In this new file, I want to use the 3km
> resolution data where I have it, then 9km where I have it, and the 27km
> where I have it.
> In essence, I want to remove the 27km data and replace it with the 9km
> data where I can, and remove the 9km data and replace it with the 3km
> data when I can.  I want to maintain the native resolution of my data
> in all instances.  I do not want to degrade everything to the 27km grid.
> My question: what tool would allow me to do this?  I have looked at a few,
> but none has this exact feature.
> Thoughts?

I think you want the capabilities that will be provided by "libcf", 
the implementation of the GRIDSPEC library, which is currently under development
at Unidata and several other institutions.  The "libcf" library provides C and
Fortran (and eventually other language) programming interfaces for support of 
Climate and Forecast (CF) metadata conventions, which include specifications of
coordinate systems and complex mosaic and other advanced grids that software 

Here's the abstract of a GRIDSPEC talk at the upcoming AGU annual meeting in 
December, by 9 authors oincluding Unidata's Ed Hartnett:

  Increasingly earth system models perform computations on grids that
  are not describable as simple, rectangular arrays (e.g. lon by lat),
  instead requiring a mosaic of interacting, logically rectangular
  tiles. Such grids are developed for a variety of reasons that
  include removal of coordinate singularities that may degrade
  numerical reliability in a region of interest (e.g. the north pole
  in an ocean model) and increasing the uniformity of numerical
  precision over the globe. Coupled earth system models, typically
  characterized by independent coordinate reference systems for
  modeling atmosphere, ocean, ice, and terrestrial processes, are
  themselves examples of such mosaic grids. GRIDSPEC is a proposed set
  of conventions to the Climate and Forecast library (LibCF)
  describing data on mosaic grids developed by V. Balaji et
  al. (Geophysical Fluid Dynamics Laboratory). It supports
  unstructured assemblies of structured grids, including the
  cubed-sphere and tripolar meshes. Here we review a GRIDSPEC NetCDF
  format based on host, contact, grid, and data files. We will show
  how mosaic grids can be created from the ground up using a C API and
  the Python Climate Data Anaysis Tools (CDAT) for visualization. As
  an application we use GRIDSPEC to regrid cubed-sphere data onto a
  longitude-latitude grid.

GRIDSPEC includes nested grids, and an early implementation of GRIDSPEC
is currently available in the netCDF software, if configured with
--enable-libcf, but I don't know of any widely used analysis and
visualization applications have been adapted to it yet.  Because it's
still under development, and because GRIDSPEC is only a proposed
extension to the CF standard that hasn't been approved yet, I expect
developers will want to wait until it's nailed down before they adapt
their applications to deal with grids specified by GRIDSPEC.

I'm Cc:ing Ed, in case he has anything to add, but he's out of email contact
until next week ...


Russ Rew                                         UCAR Unidata Program
address@hidden                      http://www.unidata.ucar.edu

Ticket Details
Ticket ID: PSZ-497222
Department: Support netCDF
Priority: Normal
Status: Closed

NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.