Keywords
landscape change, modelling, generic, process-based modelling, species richness
Start Date
1-7-2004 12:00 AM
Abstract
The analysis of landscape change impacts on community composition and dynamics is difficult for species rich plant communities, because of their high complexity. One approach to deal with this challenge are generic process-based models. In these models, the species are described by a common set of parameters and functional responses. Thus, they allow both the integration of knowledge on key processes, and a common description for several ecological patterns. An important aspect of these models are trade-offs in the species’ physiological and life-history traits, which prevent ‘super-species’ that dominate under all environmental conditions. We compare process-based models with two other model types that have been applied to similar ends – statistical habitat models, and phenomenological population models. These process-based models come at the price of an increased number of parameters for an individual species. However, a description of the interactions between species, which has proven difficult to incorporate in statistical habitat models, or requiring excessively many parameters in phenomenological population models, can be included easily. Finally, processed-based models produce a rich set of patterns on several organizational levels that can be compared to empirical observations, and thus be used for model calibration and validation. The approach is illustrated with a case study of Southern African plant communities. The investigated semi-arid landscapes are characterized by high stochastic fluctuations in population sizes. These fluctuations may in the short term mask the effects of environmental or land use change, and models allow to assess likely long-term consequences. Questions pertinent to the management of these landscapes include the effect of grazing on the diversity of the plant communities and the impact of climate change.
Generic process-based plant models for the analysis of landscape change
The analysis of landscape change impacts on community composition and dynamics is difficult for species rich plant communities, because of their high complexity. One approach to deal with this challenge are generic process-based models. In these models, the species are described by a common set of parameters and functional responses. Thus, they allow both the integration of knowledge on key processes, and a common description for several ecological patterns. An important aspect of these models are trade-offs in the species’ physiological and life-history traits, which prevent ‘super-species’ that dominate under all environmental conditions. We compare process-based models with two other model types that have been applied to similar ends – statistical habitat models, and phenomenological population models. These process-based models come at the price of an increased number of parameters for an individual species. However, a description of the interactions between species, which has proven difficult to incorporate in statistical habitat models, or requiring excessively many parameters in phenomenological population models, can be included easily. Finally, processed-based models produce a rich set of patterns on several organizational levels that can be compared to empirical observations, and thus be used for model calibration and validation. The approach is illustrated with a case study of Southern African plant communities. The investigated semi-arid landscapes are characterized by high stochastic fluctuations in population sizes. These fluctuations may in the short term mask the effects of environmental or land use change, and models allow to assess likely long-term consequences. Questions pertinent to the management of these landscapes include the effect of grazing on the diversity of the plant communities and the impact of climate change.
