Quantitative methods to help build state and transition models
Formulation of a state and transition model
Bioregional planning, assessment, and monitoring of natural resources are improved with a prediction of vegetation spatial pattern at the landscape scale. Understanding vegetation factors and processes is a necessary prerequisite to predict future patterns of vegetation in landscapes. Toward that end, there is renewed interest in implementing models of vegetation dynamics to assess the effect of human activities on ecosystems and help manage landscapes. In range science, traditional approaches have proven inadequate for certain types of rangeland. Traditional range management was based on the hypothesis that the replacement of one type of plant cover by another was the most rational and reliable way to detect overgrazing and grazing value of a range site was determined by the stage of succession represented.
A large body of empirical evidence has accumulated in recent years documenting cases where the assumptions of the traditional range succession model do not hold. In turn, this evidence has stimulated the development of theoretical explanations of when and why the assumptions are invalid. In particular, vegetation changes in response to grazing are often inconsistent or irreversible. Studies increasingly document conditions where vegetation did not change markedly after removal of livestock, where it has not moved toward the presumed climax stage of succession, or where biomass increased without changes in species composition. Vegetation dynamics observed in California's Mediterranean grasslands are not well described by linear succession models and range condition changes do not parallel successional changes. For example, grasslands in Mediterranean climate areas, which have been converted from perennial to annual systems under grazing, have not reverted to perennial grass dominance with the exclusion of grazers.
In range ecosystems, particularly in semi-arid environments, the abiotic environment often dominates and masks interactions and effects of the biotic elements. Most of Californias rangelands exhibit strongly seasonal and irregular rainfall patterns that drive germination, floristic composition and forage biomass accumulation. This dominance of the biotic by the abiotic reduces opportunities for application of the traditional succession model to predict range dynamics. Single disturbance events due to weather, fire, grazing, or management, or combinations of such events can change rangelands in ways that are not consistent with the traditional range succession model. Model failure hinders not only the research and predictive capability of range science but also progress toward sustainable management. When substantial reductions in stocking rates fail to produce the results announced by the theory, the most common management response is to reject any further scientific advice. Similarly, when ranges are rated in poor condition because they are dominated by exotic species, the general public wrongly assumes that they are continuing to deteriorate and that further reduction or removal of livestock will improve the situation.
In 1989, Westoby, Walker and Noy-Meir proposed state and transition models as an alternative to the traditional range management model. These models provide a framework to abstract and summarize knowledge about range dynamics and are a promising way to synthesize our understanding of Californias hardwood rangelands. However, a functional approach is needed for models to be of practical use. Because they are more realistic and better able to describe ecosystem dynamics and management interactions, state and transition models have the potential to improve communication between scientists, planners, land managers and the general public. Unlike the range succession model, state and transition models do not require specific assumptions except for the idea that ecosystems can have multiple stable states. For example, a given rangeland could be described with a greater or a lesser number of states and transitions, depending on the type and goals of management and on the amount of existing knowledge. The transitions between states may be caused by natural disturbances (e.g., weather, fire, herbivory) or by management actions (e.g., grazing, burning, wood harvest, elimination or introduction of plant species, fertilization). Very often a particular combination of both types of causes is needed to trigger a transition. Transitions may occur rapidly (fire) or over a period of many years (woody plant recruitment). However, in either case, the system has crossed a threshold and cannot persist halfway through a transition.
State-and-transition model for Blue oak (Q. douglasii)
dominated vegetation situated in the part of the Sierra Nevada
foothills and adjacent valley floor receiving less than 600 mm of yearly precipitation on average (abiotic domain A).
Typically, state and transition models have been implemented through simple printed flowcharts complemented by catalogs of states and transitions. However, state and transition models can also be implemented on a computer using expert system methodologies. Computerized implementation and linkage to a geographic information system allows testing of the model through spatially explicit simulations (Plant et al). Spatially explicit models based on state and transition models could be very useful to rank alternartives at the scale of a landscape.
Back to Projects