In addition to measurements we also use observations to describe the weather. Observations can be:
- Private sensors
As measurements cover only a very small part of the world, any other way to obtain data and weather information are interesting and useful to validate and improve our meteoblue results.
Observations can be used in the validation process of the model, and for the post-processing, to verify if an event actually happened or not, and to compare some data at a very local scale or for very localised phenomena (such as wind gusts, turbulences, thunderstorms, hail, tornados). This allows to improve the description of reality and to validate the models, should systematic differences occur. Contrary to the measurements, observations cannot be used as initial variables for modelling, as they are often not systematic, nor accurate enough and are only reported some time after the event.