PARENT SESSION
Contributed Oral Session 6: Biodiversity: Management, Theory, and Techniques
Monday, August 8, 8:00 AM - 11:30 AM, Meeting Room 516 D, Level 5, Palais des congrès de Montréal
Modeling regional plant species richness: Handling habitat age by probabilistic analyses.
Reger, Birgit*,1, Waldhardt, Rainer1, Otte, Annette1, 1 Justus-Liebig-University Giessen, Giessen, Germany
ABSTRACT- We developed a GIS-based spatially explicit model for the estimation of the plant species richness in agricultural mosaic landscapes. The model is based on an area-wide classification of habitat types and on the relative frequencies of the species within these types. The classification of habitat types reflects both the spatial patterns of abiotic site conditions and the recent land use. The model was successfully applied to agricultural used landscape tracts of the Lahn-Dill-Highlands (Germany). As in other marginal European landscapes, in the Lahn-Dill-Highlands the proportion of grasslands has significantly increased within the last six decades. The grassland habitats therefore differ in age. Since in grasslands the relative frequencies of species are affected not only by the recent land use, but also by the habitat age, we decided to additionally consider information on land-cover changes in the model. Spatially explicit data on land-cover changes with respect to several decades may be derived from air photo interpretation. However, area-wide air photo interpretation - the Lahn-Dill-Highlands cover a total area of about 1,200 km2 - is time-consuming and costly. We therefore developed a probabilistic procedure that aims to differentiate three age classes (young: < 20 years, mid-old: 20 - 40 years, old: > 40 years) for each grassland habitat type. We first subdivided the entire landscape into six types of land-cover patterns and dynamics. This classification was carried out by a k-means cluster analysis based on spatially explicit and non-explicit land-cover data that reached back to 1955. Within each of the six types we randomly selected 20 grassland fields of each grassland habitat type. In total, about 1,000 fields were selected. By means of multi-temporal (1945 - 2001) air photo interpretation we classified each of these fields with respect to the above named age classes. Based on this information we estimated the proportions of young, mid-old and old grassland fields for each grassland habitat type. This enabled us to derive probabilities of belonging to the age classes for each grassland field of the entire landscape.
Key words: grassland, habitat type, probability calculation, marginal cultural landscape