Spatio-temporal structural equation modeling in a hierarchical Bayesian framework. II. What controls acid grassland vegetation?

Spatial and temporal pin-point plant cover monitoring data are fitted in a structural equation model in order to understand and quantify the effect of selected environmental and land-use drivers on the observed variation and changes in the vegetation of acid grasslands. The important sources of measurement- and sampling uncertainties have been included using a hierarchical model structure. Furthermore, the measurement- and sampling uncertainties are separated from the process uncertainty, which is important when generating ecological predictions. Generally, increasing atmospheric nitrogen deposition leads to more grass-dominated acid grassland habitats at the expense of the cover of forbs. Sandy soils are relatively more acidic, and the effect of soil type on the vegetation includes both direct effects of soil type and indirect effects mediated by the effect of soil type on soil pH. Both soil type and soil pH affected the vegetation of acid grasslands. Even though only a small part of the temporal variation in cover was explained by the model, it will still be useful to quantify the uncertainties when using the model for generating local ecological predictions and adaptive management plans.