ABSTRACT
There is a great need for data harmonization as there is a huge problem of compatibility and comparability between different land-cover (LC) products. Harmonization should be the process whereby differences among existing de‡nitions of land characterization are identi‡ed and clari‡ed and inconsistencies are reduced. However, this is not the reality, since current maps exist mostly as independent and incompatible datasets. This lack of harmonization can be explained by the poor compatibility of LC classi‡cations or legends, which is often an arcane “black box” to anyone outside the immediate group involved in the preparation of legends. By its nature, mapping is a local activity, thus facilitating the tendency to establish unique classi‡cation systems to ‡t local environmental conditions. However, these incompatibilities hamper the aggregation toward broader regional and global datasets. To be able to integrate data from multiple sources, there is a strong need for semantic interoperability.
