Understanding the dimensions and trajectories of landscape change, the likely impacts of change on systems of conservation interest, and system resilience or adaptive capacity is essential to effective environmental management. For example, ecosystem service provisioning and species habitat qualities are shaped by the distributional dynamics of landscape capacities, uses, and management in the context of stressors such as land use change, climate change, and wildland fire. Remotely sensed data have become sufficiently rich in temporal and spatial dimensions to provide the raw observations for assessing nuanced landscape change including vegetation disturbance and recovery. But challenges have been to develop (1) similarly nuanced metrics of ecological change beyond land use/cover that can be quantified over large areas and (2) robust ecological theory linking observed dynamics to ecosystem management-relevant concepts such as landscape resilience. In response, we developed annual vegetation phenology monitoring data across North America at moderate resolution from MODIS satellite vegetation greenness time-series. We subjected these data to a system dynamics analysis using information theory to characterize landscape complexity and change over multiple phenological dimensions, and landscape behaviors over nearly two decades.
Results/Conclusions
A polar-coordinates method for developing cyclical annual phenology metrics such as key dates (e.g., middle of growing season), seasonality, and productivity reveal fundamental biogeographic gradients as well as land use impacts and transient disturbances. Factor analysis further suggests that these metrics resolve to a few core phenological dimensions. We compile inter-annual changes along these dimensions in transition matrices for multi-pixel landscapes, and use information-theoretic metrics (system entropy, mutual information) and matrix projection to characterize landscape dynamics. These metrics reveal geographic gradients in landscape complexity, dynamism, organization, and predictability resulting from observed regimes of disturbance, recovery, and other changes through time. We explore the utility of this approach for assessing landscape resilience and ecosystem service sustainability, given the dependence of these on maintaining core ecosystem attributes even given ongoing environmental and ecosystem change. We suggest that regional monitoring and assessment for purposes of ecosystem management and conservation planning can benefit from quantitative, ecologically nuanced landscape dynamics assessment grounded in complex systems and resilience theory.