Process models in the Earth sciences describe the evolution of geo(bio)physical surface properties in time, independently from remote sensing observations. Examples of such process models on various time scales are, for instance, numerical weather prediction models (NWPs), vegetation growth models, hydrological models, oceanographic models and climate models. Process models in the geosciences usually rely on regular observations at many locations spread over a large area. Traditionally, these observations were mostly made in the field with a variety of instruments. Remote sensing techniques have tremendously increased the capability of spatial sampling and the consistency of the surface parameters measured. RS instruments are mostly sensitive to many physical properties of the surface, some of these may not belong to the set of properties that the user is interested in. Exceptions to this are the mapping of sea-surface temperature, laser altimetry and gravimetry, which are measurements of direct geophysical interest. In the majority of cases, however, there are only indirect relationships between what is observed with the instrument and the physical object properties of interest. In these cases, the use of observation models becomes an attractive option, since these models describe the relationships between all object properties relevant for the observation and the observed remote sensing data.
In progress (GI-N2K)