Ensemble modeling to predict habitat suitability for a large‐scale disturbance specialist

Abstract

To conserve habitat for disturbance specialist species, ecologists must identify where individuals will likely settle in newly disturbed areas. Habitat suitability models can predict which sites at new disturbances will most likely attract specialists. Without validation data from newly disturbed areas, however, the best approach for maximizing predictive accuracy can be unclear (Northwestern U.S.A.). We predicted habitat suitability for nesting Black-backed Woodpeckers (Picoides arcticus; a burned-forest specialist) at 20 recently (≤6 years postwildfire) burned locations in Montana using models calibrated with data from three locations in Washington, Oregon, and Idaho. We developed 8 models using three techniques (weighted logistic regression, Maxent, and Mahalanobis D2 models) and various combinations of four environmental variables describing burn severity, the north–south orientation of topographic slope, and prefire canopy cover. After translating model predictions into binary classifications (0 = low suitability to unsuitable, 1 = high to moderate suitability), we compiled “ensemble predictions,” consisting of the number of models (0–8) predicting any given site as highly suitable. The suitability status for 40% of the area burned by eastside Montana wildfires was consistent across models and therefore robust to uncertainty in the relative accuracy of particular models and in alternative ecological hypotheses they described. Ensemble predictions exhibited two desirable properties: (1) a positive relationship with apparent rates of nest occurrence at calibration locations and (2) declining model agreement outside surveyed environments consistent with our reduced confidence in novel (i.e., “no-analogue”) environments. Areas of disagreement among models suggested where future surveys could help validate and refine models for an improved understanding of Black-backed Woodpecker nesting habitat relationships. Ensemble predictions presented here can help guide managers attempting to balance salvage logging with habitat conservation in burned-forest landscapes where black-backed woodpecker nest location data are not immediately available. Ensemble modeling represents a promising tool for guiding conservation of large-scale disturbance specialists.

Patterns of red tree vole distribution and habitat suitability: implications for surveys and conservation planning

“…Predictions from our ensemble model are more likely to result in reliable predictions when the model is applied to novel areas than any of the singular models, consistent with the findings from Latif et al. (). The ensemble model could only be applied to the portion of the study area (~63%) that had consistent predictions of tree vole presence and absence among the three models.…”

“…The ensemble model could only be applied to the portion of the study area (~63%) that had consistent predictions of tree vole presence and absence among the three models. Predictions from the ensemble model may be sufficiently accurate to guide management, and areas where predictions from the three models were not consistent may warrant further evaluation or data collection (Latif et al 2013). In these cases, it may be best to use one of the singular models, such as BEST9, or to rely on data on the primary drivers of all models and field studies: older forest, large trees, and geographic location.…”

Drifting baited stereo‐videography: a novel sampling tool for surveying pelagic wildlife in offshore marine reserves

“…Model projections outlined the tip of the canyon head and its eastern and northern slopes as regions of elevated suitability for a diversity of pelagic organisms, a finding consistent with both v www.esajournals.org expectations from simulations of flow dynamics and existing knowledge of topography-controlled primary productivity patterns in the area (Rennie et al 2009b), and which also mirrors the documented behavior of other predator species such as pygmy blue whales (McCauley et al 2004, Rennie et al 2009a). Because various competing hypotheses regarding the underlying ecology of a focal species may be equally suitable in explaining the patterns observed in the data, building ensemble predictions is a useful endeavor that can help overcome the uncertainties of model selection (Latif et al 2013). The congruence between our two approaches evinced a seemingly trivial effect of visual biases on predicted distributions under the conditions of the present pilot study (as was the case for Gormley et al 2011).…”

Interacting and non‐linear avian responses to mixed‐severity wildfire and time since fire

“…, Latif et al. ). Following fire, vegetation composition and structure change rapidly (Moreira et al.…”

“…Bird responses to fire severity and time since fire may be complex, and the inclusion of nonlinear and interacting terms may elucidate these more nuanced responses. Additional information on the responses of target species can help inform evidence-based land management decisions (Guisan et al 2013, Latif et al 2013). Following fire, vegetation composition and structure change rapidly (Moreira et al 2003), and different bird species may perceive high-quality habitat at different points across this range of vegetation conditions (Rose andSimons 2016, Tingley et al 2016).…”