A comparison of six models predicting vegetation response to hydrological habitat change.

In The Netherlands several models have been constructed over the last decade facilitating prediction of the occurrence of plant species or vegetation types in relation to hydrological or hydrogeochemical habitat conditions. This paper compares six correlative ecohydrological models: NTM, DEMNAT, MOVE, HYVEG, ICHORS, and ITORS on the basis of the properties, possibilities and constraints of different model concepts. The models were developed for different purposes so that the suitability of the model concepts depends on the areas for which they were constructed and the aim of the user. Major differences between the models are: the expert knowledge and field measurements required, the scale level, ecosystem and habitat conditions for which prediction is made, and the number of input variables. The latter also causes different opportunities of connection with input models. Without comparable validations it was impossible to mark the most appropriate predicting model. The suitability of the model concepts in other areas is illustrated by means of a selection of the most suitable model for the Prairie Pothole area (North America) and for the catchment area of the river Dommel (Belgium/The Netherlands). The development of a set of ecohydrological modules, containing causal relationships combined with expert knowledge and empirical-statistical based black box relationships, between plant species and habitat conditions, is considered a major need in future ecohydrological modelling.