Decision support for evaluating landscape departure and prioritizing forest management activities in a changing environment

We evaluated changes (hereafter, departures) in spatial patterns of various patch types of forested landscapes in two subwatersheds ("east" and "west") in eastern Washington, USA, from the patterns of two sets of reference conditions; one representing the broad variability of pre-management era (similar to 1900) conditions, and another representing the broad variability associated with one possible warming and drying climate-change scenario. We used a diagnostic set of class and landscape spatial pattern metrics to compare current spatial patterns of test subwatersheds against the two sets of reference conditions. in a companion decision support model built with the EMDS modeling system, we considered the degree of departure in the subwatersheds, relative to the two sets of reference conditions along with two additional criteria (vulnerability to severe wildfire and timber harvest opportunity), to determine the relative priority of landscape restoration treatments, and the potential for timber harvest to underwrite the treatments. In the decision support model, the current spatial pattern conditions of physiognomic types, cover types, forest structural classes, and those of late-successional and old forest patches of the two subwatersheds were compared against the two sets of reference conditions. The degree of departure in spatial patterns of physiognomic conditions was moderate in both subwatersheds in the pre-management era and climate-change comparisons. The situation was similar for the cover-type departure analysis, but spatial patterns of cover types increased in similarity to the reference conditions in the western subwatershed under the climate-change scenario. Spatial patterns of structural conditions showed a high degree of departure in both subwatersheds when compared to either set of reference conditions, but similarity improved in the eastern subwatershed under the climate-change scenario. Spatial patterns of late-successional + old forest structure were strongly similar to the broad envelope of conditions represented by the pre-management era reference in the western and moderately similar in the eastern subwatershed, but declined in both subwatersheds when compared with the climate-change reference conditions. When the degree of departure in spatial patterns of all patch types was considered along with vulnerability to severe wildfire and timber harvest opportunity, the eastern subwatershed rated higher priority for landscape improvement using either set of reference conditions. We conclude by considering uncertainties inherent in the analysis approach, types of sensitivity analysis needed to investigate model performance, and broad implications for forest managers. (C) 2008 Published by Elsevier B.V.