A "Layered" dataset is a dataset on a grid that is cartesian in the horizontal (X-Y) axes, but has an arbitrary vertical displacement in the vertical (Z) direction. The WRF-ARW grid and the ROMS grid are examples of such a grid; WRF and ROMS datasets use a terrain-following grid rather than a Cartesian grid. VAPOR support for Layered data can be used to visualize any dataset that is defined on a similar grid; i.e. a grid that is Cartesian in the X and Y dimensions, but for which each node can be displaced in the Z direction.
Support for converting WRF-ARW data is described in Populating a VDC.
The following are required to visualize layered grids in VAPOR, when the data is not from WRF-ARW:
- To specify the metadata, use vdfcreate instead of using wrfvdfcreate. (See the vdfcreate man pages.) The gridtype must be set to “layered” in the vdfcreate command arguments.
- The .vdf file must contain a variable named "ELEVATION". This variable must be then be added to the VDC, using raw2vdf for example, to specify the vertical (Z) displacement of each node in the grid. For each (X,Y) coordinate, the variable ELEVATION must be nondecreasing in Z.
- If you would like to be able to show the terrain elevation in your dataset, so that you can visualize images, 2D data, or Barbs mapped to the terrain, then you will need to provide a 2D (horizontal) variable, that indicates the height above sea level at each point of the terrain surface. (Such a variable is named "HGT" in a WRF VDC.) In MOM, ROMS, and POP datasets, there is ordinarily a variable indicating ocean depth, however this variable must be negated in order to be used as height.
- The vertical (Z) data extents in the metadata must indicate the range of values of ELEVATION that will be visualized. With WRF data this range usually goes from the top of the bottom layer of the data to the bottom of the top layer of the data. Data extents are specified as arguments to the vdfcreate command.
- All of the variables in the data need to be on the same grid. If, for example, your data includes variables on a staggered grid, these variables must be interpolated or truncated to the same grid as is used for the ELEVATION variable. The ncdf2vdf data conversion routine supports interpolating variables from staggered grids.
