Changes between Version 30 and Version 31 of ADAGUC


Ignore:
Timestamp:
Jul 26, 2013 1:31:54 AM (5 years ago)
Author:
maartenplieger
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • ADAGUC

    v30 v31  
    33The ADAGUC netCDF driver reads and writes files which comply with the ADAGUC Data Products Standard [http://adaguc.knmi.nl ADAGUC website] (Atmospheric Data Access for the Geospatial User Community). ADAGUC files are created by using the NetCDF4 API. The NetCDF4 API has the ability to store files in the HDF5 file format while remaining backwards compatible with previous versions of netCDF. The ADAGUC driver supports the creation of NetCDF3 and NetCDF4 files. In NetCDF3 mode the driver writes NetCDF files, in NetCDF4 mode the driver writes HDF5 files. ADAGUC netCDF files follow the Climate and Forecast (CF) 1.2 and ISO:19115 metadata conventions. Besides these conventions the format provides space for specific product metadata.
    44== The file structure ==
    5 The ADAGUC internal file structure contains variables which can be subdivided in three types: variables to store the metadata, variables to store the dimension scales and variables to store the data. A schematic overview of the ADAGUC file structure for the different types is given in the file structure figure (below on the left). The three types of variables (metadata, dimension scales and data) are indicated with A, B and C respectively.
     5The ADAGUC internal file structure contains variables which can be subdivided in three types: variables to store the metadata, variables to store the dimensions and variables to store the data. A schematic overview of the ADAGUC file structure for the different types is given in the file structure figure (below on the left). The three types of variables (metadata, dimension and data) are indicated with A, B and C respectively.
    66
    77
    88''' (A) Metadata ''' - Metadata is stored in attributes which are assigned to a variable. The variables ‘product’, ‘iso_dataset’ and ‘projection’ are used to store the metadata.
    99
    10 ''' (B) Dimension scales ''' - Dimension scales are used as coordinate variables and time intervals. They provide information about geolocation and time to the data variables. The dimension scales follow the  [http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.2/ NetCDF Climate and Forecast (CF) Metadata Conventions]. Dimension scales are lists containing coordinates or time intervals. The dimension scale figure shows geographic raster data with the dimensions time, lat and lon. Its size is, in this example, 4x6x12. This means that there are 4 indexes for time, 6 for latitude and 12 for longitude. According to the CF conventions the dimensions should have the order time, lat, lon.
    11 In case the raster data uses a geographic coordinate system, the required dimensions are time, lat and lon which represent the center of the pixel. When the raster data is projected the dimension scales x and y are included which represent the projected coordinates. The dimension scales lat and lon become a function of y and x (see the right part of the file structure), providing the latitude and longitude at the location of y and x, respectively. In this case y and x contain the projected coordinates, while lat and lon represent the geographic latitudes and longitudes in degrees at location y, x. For detailed information see the [http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.2/ NetCDF Climate and Forecast (CF) Metadata Conventions].
    12 
    13 ''' (C) Data ''' - This part contains the data in the file structure. The data is stored in variables which are functions of the dimension scales. For geographic raster data the variable is a function of time, lat and lon. For projected raster data the variable is a function of time, y and x, as can be seen in the illustration of the file structure.
     10''' (B) Dimensions ''' - Dimensions are used as coordinate variables and time intervals. They provide information about geolocation and time to the data variables. The dimensions follow the  [http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.2/ NetCDF Climate and Forecast (CF) Metadata Conventions]. Dimensions are lists containing coordinates or time intervals. The dimension scale figure shows geographic raster data with the dimensions time, lat and lon. Its size is, in this example, 4x6x12. This means that there are 4 indexes for time, 6 for latitude and 12 for longitude. According to the CF conventions the dimensions should have the order time, lat, lon.
     11In case the raster data uses a geographic coordinate system, the required dimensions are time, lat and lon which represent the center of the pixel. When the raster data is projected the dimensions x and y are included which represent the projected coordinates. The dimensions lat and lon become a function of y and x (see the right part of the file structure), providing the latitude and longitude at the location of y and x, respectively. In this case y and x contain the projected coordinates, while lat and lon represent the geographic latitudes and longitudes in degrees at location y, x. For detailed information see the [http://cf-pcmdi.llnl.gov/documents/cf-conventions/1.2/ NetCDF Climate and Forecast (CF) Metadata Conventions].
     12
     13''' (C) Data ''' - This part contains the data in the file structure. The data is stored in variables which are functions of the dimensions. For geographic raster data the variable is a function of time, lat and lon. For projected raster data the variable is a function of time, y and x, as can be seen in the illustration of the file structure.
    1414 
    15 The ADAGUC format currently supports X, Y and time dimensions, stored as dimension scales in the file according the CF Conventions. The driver supports two ways of time sub-setting. The first is by making selections in time by choosing band numbers, the second is by providing the time in the file name. By using gdalinfo on the ADAGUC file the available band numbers and dimensions are listed. With this information a subset can be read with for example:
     15The ADAGUC format currently supports X, Y and time dimensions, stored as dimensions in the file according the CF Conventions. The driver supports two ways of time sub-setting. The first is by making selections in time by choosing band numbers, the second is by providing the time in the file name. By using gdalinfo on the ADAGUC file the available band numbers and dimensions are listed. With this information a subset can be read with for example:
    1616{{{
    1717gdalinfo ADAGUC:thefile.nc:variable:time=20
    1818}}}
    1919 [[Image(ADAGUC_raster_file_structure_small.jpg, 250)]] [[Image(ADAGUC_NetCDF_dimension_scales.jpg, 310)]]
    20       Left: filestructure; Right: dimension scales
     20      Left: filestructure; Right: netcdf dimensions
    2121----
    2222== GDAL Creation options ==