Calibration of Ellenbergs indication values on measured soil characteristics
Kalibratie van Ellenbergs milieu-indicatiegetallen aan werkelijk gemeten bodemfactoren
26 May 2012, PDF |
48 pages |
Alkemade JRM , Wiertz J , Latour JB
RIVM Report 711901016
A multi-stress model has been developed in the Netherlands on a national scale to forecast changes in plant species composition due to acidification, eutrophication and dessication. This model, called SMART-MOVE, consists of: a soil module (SMART2) used for calculating changes in groundwater level, pH and nutrient availability, and a vegetation module, consisting of regression equations that describe the relationships between the probability of occurrence and environmental factors. These environmental factors represent average Ellenberg indication values for pH, availability of water and nutrient availability. Salinity was also included since in large parts of the Netherlands salinity is of major importance for species composition. Essential input data for these modules, such as groundwater level and seepage flux, can be calculated with the National Groundwater Model, for example. In this study relationships were calculated between the Ellenberg indication values and the relevant soil factors: pH, average spring groundwater table, biomass production, nitrogen production, concentrations of several nutrients and chloride concentration for both terrestrial and aquatic systems. Where it was possible to use a sigmoid model, to cope with the original ordinal and limited scale of indication values, the explained variance increased by 5-10%. In the regression analyses, almost 7000 vegetation releves from a variety of ecosystems could be used, satisfying relationships with Ellenberg values were found for pH, average spring groundwater table and biomass and nitrogen production. These relationships are used to connect the soil module with the vegetation module and quantify the confidence of the model outcomes. Average Ellenberg indication values can be concluded to be succesful as estimates for the abiotic conditions in models like SMART-MOVE.