At the end of the 1980s, the equilibrium view of ecological systems seemed unsatisfactory to a large fraction of ecolo-gists. Most notably, in 1987, D. L. DeAngelis and J. C. Waterhouse developed a general theoretical framework for equilibrium-related concepts that includes both equilibrium and nonequilibrium behavior of ecosystems (see Stability). DeAngelis and Waterhouse point out that the (non)equilibrium behavior of an ecosystem may depend on spatial scale. For example, systems that are unstable at small spatial scales (e.g., paddock scale) may be stable on the landscape scale. This is illustrated above by the example of the grazer-vegetation system in the section titled 'Spatially explicit herbivore dynamics on larger scales'. Metapopulations are another important example for ecological systems that are stochastically dominated at the local patch scale but may show equilibrium at the landscape scale. Finally, hierarchical patch dynamics is a paradigm to describe interacting, equilibrium and none-quilibrium, dynamics at multiple spatial and temporal scales.
The equilibrium versus nonequilibrium debate in rangeland ecology
These advances in general theory were mirrored in rangeland science. One year after the review by DeAngelis and Waterhouse, J. E. Ellis and D. M. Swift published an influential paper on East African pastoral ecosystems which stimulated an ardent debate about equilibrium versus nonequilibrium concepts in arid and semiarid grazing systems. They showed evidence for nonequilibrium dynamics of a pastoralist ecosystem and argued that a fundamental misunderstanding of the ecological dynamics based on equilibrium concepts may lead to inappropriate and failed interventions. The emerging 'new rangeland ecology' posits that traditional, equilibrium-based rangeland models have not taken into account the considerable spatiotemporal heterogeneity and climatic variability of semiarid rangelands, and that mobility, variable stocking rates, and adaptive management are essential for effectively and sustainably utilizing semiarid and arid rangelands.
Central to the debate became the question about the relative importance of biotic and abiotic factors in driving rangeland production. This includes the question if, and how, grazing may affect vegetation dynamics, and if grazing may induce rangeland degradation. This has led to concerns about the ecological consequences of uncritically adopting the nonequilibrium paradigm for management in areas which are not predominantly experiencing nonequilibrium dynamics. The debate also gained heat because of confusion regarding spatial and temporal scales: One problem difficult to overcome is the mismatch between the scales of ecological investigation and those at which ecological processes take place. Since rangeland degradation usually takes place over timescales much greater than those at which management decisions are made, degradation may not be appropriately perceived. An additional problem is that the question of the relative importance of herbivores and stochastic rainfall for vegetation dynamics cannot be answered in general, but only for specific systems based on the biological information on hand and for well-defined spatial and temporal scales. The debate expanded from its initial focus on communal rangelands in sub-Saharan Africa to other continents and tenure systems, such as commercial rangelands in Australia and pastoral regions of Asia.
Specific properties make the dynamics of arid and semiarid communities difficult to analyze, including long lifespans of dominant plants, and episodic and event-driven changes in species composition occurring on long time-scales in response to rare or extreme (rainfall) events. Because of these characteristics, a deeper analysis of these systems requires models which are able to include specific biological information about life-history traits of individual species and to specify the stochastic character of driving events. Theoretical top-down models which are generally not linked to specific spatial and temporal scales are unsuitable for this purpose. Instead, the development of powerful computers facilitated a new approach of bottom-up modeling.
Grid-, rule-, and individual-based simulation models
Although there is little (long-term) field data available on the full dynamics of arid plant communities, attributes of individual plant behavior such as conditions for seed production, recruitment, or mortality are relatively easy to observe at smaller temporal and spatial scales. The basic idea is therefore to incorporate the short-term knowledge in the form of rules or simple mathematical equations into a computer simulation model. In order to investigate community dynamics, these models use external drivers such as rainfall and simulate fate and interactions of individual plants within the community. This way, these models extrapolate from the behavior of individual plants to long-term community dynamics.
Was this article helpful?