Predicting Landscape Vegetation Dynamics Using State-and-Transition Simulation Models

This paper outlines how state-and-transition simulation models (STSMs) can be used to project changes in vegetation over time across a landscape. STSMs are stochastic, empirical simulation models that use an adapted Markov chain approach to predict how vegetation will transition between states over time, typically in response to interactions between succession, disturbances and management. With STSMs a landscape is divided into a set of simulation cells, each cell is assigned to an initial vegetation state, and the model then predicts how each cell may change from one vegetation state to another over time. Over the years an extensive suite of features have been added to STSMs that allow them to represent a range of dynamics important to landscape modeling, including tracking age-structure, triggering transitions based on past events, setting targets for certain transitions, and varying transition rates over time. STSMs are also now able to represent spatial variability in two different ways: by dividing the landscape into spatial strata, typically defined by one or more important drivers of vegetation change, or alternatively by developing a spatially-explicit STSM, whereby transition events, such as fire or invasion by non-native vegetation, can be simulated to spread across the landscape. Since their introduction in the early 1990s, STSMs have been applied to a wide range of landscapes and management questions, including forests, rangelands, grasslands, wetlands and aquatic communities, over spatial extents ranging from thousands to millions of hectares. Several software tools currently exist to support the development of STSMs; the most recent of these products, called the Path Landscape Model, is the latest in a lineage of STSM development tools that includes both the Vegetation Dynamics Development Tool (VDDT) and the Tool for Exploratory Landscape Analysis (TELSA).