Submitted to: Journal of Applied Ecology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 1/26/2011
Publication Date: 3/15/2011
Publication URL: http://handle.nal.usda.gov/10113/58040
Citation: Bestelmeyer, B.T., Goolsby, D., Archer, S.R. 2011. Spatial perspectives in state-and-transition models: A missing link to land management? Journal of Applied Ecology. 48:746-757. Interpretive Summary: State-and-transition models (STMs) are usually based on data gathered at specific points in space. Here we show how mappable data linked to point data may be used to better understand transitions in ecological systems. We focus on three distinct processes that can be mapped: the spatial pattern of a driver (such as grazing management), the spatial dependence in the effects of that driver (such as when grazing management has differing outcomes on different soils or in different climate zones), and the spatial elements of feedbacks (such as when a local transition has of-site effects, such as via altered runoff or erosion that affects a downslope site). We offer a framework for ecologists and land managers to understand the combined effects of these three processes and their interactions.
Technical Abstract: Conceptual models of alternative states and thresholds are based largely on observations of ecosystem processes at a few points in space. Because the distribution of alternative states in spatially-structured ecosystems is the result of variations in pattern-process interactions at different scales, we argue that data and models need to explicitly consider ecosystem processes in the context of the spatial pattern of drivers, spatial dependence in responses to drivers, and the spatial structure of feedbacks. Here, we review the literature on spatial patterns and processes in the context of state transitions (or regime shifts) and draw upon examples from semiarid ecosystems to illustrate these linkages. We suggest three spatial perspectives be used to expand conceptual state-and-transition (S&T) models to landscape systems. First, S&T models should represent how spatial patterns of natural and anthropogenic drivers interact to initiate transitions. This would account for often observed situations wherein contrasting states occur along gradients of disturbance intensity or in different management units. Second, spatial dependence in response to drivers and triggers should be represented. This would help account for observed differences in ecological resilience tied to inherent spatial variation in environmental variables (e.g., soil and landform attributes) that mediate driver effects and lead to variation in the likelihood of state transitions. Third, the nature of feedbacks, how they propagate over time and space, and how they are conditioned by patterns of spatial dependence and drivers should be represented. At fine scales, changes in the nature and intensity of feedbacks may be associated with predictable changes in patch structure (e.g., fragmentation or coalescence processes foreshadowing thresholds). At coarser scales, feedbacks may be associated with the spatial extent and arrangement of land cover states, broad-scale resource redistribution (e.g., wind/water erosion/deposition) or land surface/climate interactions. S&T conceptual models with a spatial component would make use of readily-obtainable patch-scale data to predict how the likelihood of transitions might vary across landscapes and to forecast changes leading to degradation or recovery. Such predictions would be valuable in the design of monitoring schemes aimed at adjusting management to avert crossing degradation thresholds and to take advantage of environmental conditions required to restore desired states.