Gridded datasets are discrete products created by entities such as NOAA, ECMWF, NASA, Australia Bureau of Meteorology and others. Gridded data products are developed to provide spatially and temporally consistent data coverage and can be used to supplement or replace existing ground-based observations, which can often be sporadic, irregularly spaced or have fragmented historical records.
Gridded datasets are constructed using inputs including ground-based observations, satellite imagery, rainfall radar and oceanographic measurements. These are then used to run physical models of the type used in weather forecasting to create a series of “reanalysis data” interpolated to regular grids. These are available both at the ground surface and sometimes at levels through the atmosphere. Some products provide full global coverage for a range of weather variables, others may only provide regional datasets or single variables though these can be at a higher spatial resolution.