Page History

...

The global attribute "Conventions" must be specified in the netcdf file (e.g. Conventions="CF-1.6"). If this attribute is not present or not valid, then the NetCDF Java library assumes a default convention (different from CF), which can cause unexpected behaviour in the NetCDF import in Delft-FEWS. Also see http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#identification-of-conventions.

...

Gridded data

See for references to the cf-conventions website below.

1. all variables (data and coordinate) must have a unique variable name and the following attributes:
   1. units (required, see units link below)
   2. standard_name (required, see standard_name link below) - should comply with standard names list
   3. long_name (recommended, see long_name links below)
   4. _FillValue and/or missing_value (required if data contains missing values)
   5. scale_factor (optional to compact data)
   6. add_offset (optional to compact data)

see cf-conventions on these:
units                  http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#units, 
standard-name http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#standard-name
http://cfconventions.org/Data/cf-standard-names/26/build/cf-standard-name-table.html
long-name        http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#long-name

1. Data variable for a 2D grid must depend on three dimensions, representing time and two spatial dimensions (in either the order time, y, x or time, x, y).  The names of the dimensions can be anything.

1. 1. regular or rectangular grids - each spatial dimension must have a corresponding 1D coordinate variable with the same name.
      See  http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#idp5553648.
   2. curvilinear grids -  both spatial dimensions are 2D and depend on both spatial dimensions.
      1. Time coordinate variable requires axis "T"
      2. x coordinate variable requires axis "X"
      3. y coordinate variable requires axis "Y"
      4. (optional) analysis_time coordinate, see below
      5. The data variable must have an attribute "coordinates" that refers to the y and x coordinate variables (in that order), e.g. coordinates="y x". See  http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#idp5559280.
      6. Each coordinate variable must have an attribute "axis" that indicates which axis is defined by that coordinate variable.
   3. Time dimension must have a corresponding 1D time coordinate variable with the same name.
   4. Spatial dimensions, for
2. Time coordinate variable must have the attributes
   1. "units" with a valid unit of time - e.g. units="minutes since 1970-01-01 00:00:00.0 +0000"
      See  http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#time-coordinate
   2. "axis" with value "T"
3. Forecast reference time/analysis time/base time (optional), must be stored in a separate coordinate variable that
   1. contains only one value: the forecast reference time, and
   2. standard_name="forecast_reference_time"

The name of this variable must be present in the coordinates attribute of the data variable, e.g. coordinates="analysis_time" or coordinates="y x analysis_time".
If needed, different data variables can point to different forecast reference time variables.
See  http://cfconventions.org/Data/cf-conventions/cf-conventions-1.6/build/cf-conventions.html#scalar-coordinate-variables

1. Time and spatial coordinate variables must have values either in ascending or in descending order.

...

1. Data variable for a 2D grid must depend on three dimensions, representing time and two spatial dimensions (in either the order time, y, x or time, x, y).  The names of the dimensions can be anything.
   1. Time dimension must have a corresponding 1D time coordinate variable with the same name.
   2. Spatial dimensions, for
      1. regular or rectangular grids - each spatial dimension must have a corresponding 1D coordinate variable with the same name (see coordinates link below).
      2. curvilinear grids -  both spatial dimensions are 2D and depend on both spatial dimensions.
         1. The data variable must have an attribute "coordinates" that refers to the y and x coordinate variables (in that order), e.g. coordinates="y x"
         2. Each coordinate variable must have an attribute "axis" that indicates which axis is defined by that coordinate variable.
            1. Time coordinate variable requires axis "T"
            2. x coordinate variable requires axis "X"
            3. y coordinate variable requires axis "Y"
            4. (optional) analysis_time coordinate, see below
2. Time coordinate variable must have the attributes
   1. "units" with a valid unit of time - e.g. units="minutes since 1970-01-01 00:00:00.0 +0000" (time coordinate link below)
   2. "axis" with value "T"
3. Time and spatial coordinate variables must have values either in ascending or in descending order.
4. Forecast reference time/analysis time/base time (optional), must be stored in a separate coordinate variable that
   1. contains only one value: the forecast reference time, and
   2. standard_name="forecast_reference_time"

The name of this variable must be present in the coordinates attribute of the data variable, e.g. coordinates="analysis_time" or coordinates="y x analysis_time".
If needed, different data variables can point to different forecast reference time variables (see scalar coordinate link below).

...

...

...

...

...

...

...

Compacted data

Netcf data variables with attributes scale_factor and add_offset will be treated as compacted data and the values of these attributes will be used to uncompact the values of this variable.

...