Long-term vegetation monitoring for different habitats in floodplains

A floodplain-restoration project along the Danube between Neuburg and Ingolstadt (Germany) aims to bring back water and sediment dynamic to the floodplain. The accompanied long-term monitoring has to document the changes in biodiversity related to this new dynamics. Considerations on and results of the vegetation monitoring concept are documented in this paper. In a habitat rich ecosystem like a floodplain different habitats (alluvial forest, semi-aquatic/aquatic sites) have different demands on the sampling methods. Therefore, different monitoring designs (preferential, random, systematic, stratified random and transect sampling) are discussed and tested for their use in different habitat types of the floodplain. A stratified random sampling is chosen for the alluvial forest stands, as it guarantees an equal distribution of the monitoring plots along the main driving factors, i.e. influence of water. The parameters distance to barrage, ecological flooding, height above thalweg and distance to the new floodplain river are used for stratifying and the plots are placed randomly into these strata, resulting in 117 permanent plots. Due to small changes at the semi-aquatic/aquatic sites a transect sampling was chosen. Further, a rough stratification (channel bed, river bank adjacent floodplain) was implemented, which was only possible after the start of the restoration project. To capture the small-scale changes due to the restoration measures on the vegetation, 99 additional plots completed the transect sampling. We conclude that hetereogenous study areas need different monitoring approaches, but, later on, a joint analysis must be possible