The NetCDF data format describes array-oriented scientific data.
NetCDF scientific weather data is stored as gridded datasets. The types of measurement and the dimensions along which they are measured determine the structure of the datasets. The measurements are referred to as variables. Temperature, humidity, and cloud cover are all examples of variables. The dimensions of the measurements typically consist of the location coordinates of the measurement, the time of the measurement, and other parameters, such as the altitude at which the measurement was taken. For instance, atmospheric temperature can be measured at a specific location, but at multiple times and vertical positions. These measurements results in a NetCDF temperature dataset, with measurement grids organized along the spatial, temporal and vertical dimensions.
By default, NetCDF allows multiple sets of variables in a single file, so one file can contain both salinity and temperature measurements for example.
The NetCDF standard itself does not impose any semantic rules or conventions, just like XML files or comma-separated value files.
LuciadLightspeed mainly uses the NetCDF format to decode data that complies with the Climate and Forecast (CF) Metadata Conventions. The CF conventions define metadata that describe what the data in each NetCDF variable represents, and the spatial and temporal properties of this data. The data model is widely used in climate and weather forecast systems and in other geoscience communities, and has been adopted by the Open Geospatial Consortium.