Overview¶
Observational datasets are direct measurements of the Earth system from instruments such as weather stations, radiosondes, buoys, radars, and satellites. They provide the closest representation of the actual atmospheric, oceanic, or land state available — but only at the locations and times the instruments operate.
This section demonstrates how to access observational datasets hosted on NCAR’s GDEX and use them in geoscientific workflows.
Characteristics of observational data¶
Highest accuracy at the point of measurement — within instrument error bars
Spatially and temporally incomplete — gaps between stations, satellite swath limits, missing data during instrument outages
Heterogeneous — different instruments have different biases, units, sampling cadence, and quality flags
Limited to what is observed — derived variables (e.g., latent heat flux) are usually unavailable and must come from a model
Datasets in this section¶
HadISST — Hadley Centre Sea Ice and Sea Surface Temperature dataset; monthly 1°×1° SST and sea ice, 1870–present. Used here for El Niño / ENSO analysis.
EOL Radar Precipitation — NCAR Earth Observing Laboratory precipitation radar observations.
Where to start¶
If you are new to working with GDEX observational data, start with the HadISST El Niño notebook, which shows how to load gridded SST observations, define an ENSO index, and visualize ENSO phases.
Related sections¶
See Reanalysis for gridded, gap-filled products that combine observations with model physics.
See Data Fusion for examples that combine observations with reanalysis or simulation output.