Automatic interpolation of phenological phases in Germany

The German joint project DeCover 2 is developing a methodological framework to cope with the increasing demand for up-to-date land cover information using remote sensing techniques. New satellite systems like RapidEye provide both data of high geometric resolution and high repetition rates. Because of the Germany-wide diversity of natural conditions, same acquisition dates don't correspond to same phenological phases. Thus, a phenological structuring of the available imagery over the year is needed for the assessment of Rapid-Eye imagery regarding their suitability for the classification and distinction of vegetation classes. On the example of the phenological phase `Yellow Ripeness' of Winter Wheat in 2010, the presented algorithm demonstrates for the total area of Germany how daily phenological phases can be automatically interpolated on demand, in real-time and considering interpolation accuracies. As input, daily provided point data on temperature and phenological phases from the extensive network of the German Weather Service as well as a SRTM digital elevation model are used. The modeling results enable the identification of temporal phenological windows for specific test sites.