Coupled human-environment systems (HESs), such as human-impacted landscapes, have complex dynamics that must be understood to facilitate effective conservation and land-use management. Progress has been made in modelling these systems, with previous work showing that adding spatial structure or coupling a human system to the environment system can qualitatively alter model behaviours. However, there has not been a systematic analysis of how these features interact to impact model dynamics. This is of particular interest when modelling mosaic ecosystems, where multiple land-states coexist in close spatial proximity. This leads us to our motivating question: how are mosaic ecosystems impacted by interactions between human and spatial dynamics?
We compare a non-spatial differential equation (DE) model of a human-impacted mosaic ecosystem, to an analogous spatially-explicit agent-based model (ABM). These models couple forest-grassland transitions to the opinion dynamics of a human population where individuals have a preferred land-state. However, in the ABM, land-state transitions depend on local cover levels in a spatially-structured landscape, unlike the homogeneous landscape of the DE model. We analyse how spatial structure and human influence impact system dynamics, focusing on how these features interact with environmental and anthropogenic conditions to generate mosaic landscapes.
Results/Conclusions
In the absence of human influence, the addition of spatial structure reduces the region of parameter space in which we observe forest-grassland mosaics. As well, spatial structure facilitates the persistence and spread of forest, even when it is initially the minority land-state. Incorporating human influence has disparate impacts on dynamics for spatial and non-spatial models. For both models, we observe that sufficiently strong human influence may preclude the occurrence of mosaics, instead resulting in landscapes dominated by a single land-state. However, in the absence of spatial structure, human influence can also destabilize mosaics through persistent large amplitude oscillations in forest cover. These initial findings suggest there may be significant variation in the behaviour of models for mosaic ecosystems, depending on structural choices.
Our work contributes to a more complete understanding of how the interaction of spatial structure and human influence impact the dynamics of mosaic ecosystem models, and can be used to inform future HES model construction and analysis. Beyond the field of mathematical ecology, an improved understanding of system dynamics will play a part in increasing confidence in HES model predictions and facilitate their use as a tool for policy-makers.