An integration framework for linking avifauna niche and forest landscape models.

Paul Pickell,Tara Hudiburg,Eric S. Walsh

doi:10.1371/journal.pone.0217299

Abstract

Avian cavity nesters (ACN) are viable indicators of forest structure, composition, and diversity. Utilizing these species responses in multi-disciplinary climate-avian-forest modeling can improve climate adaptive management. We propose a framework for integrating and evaluating climate-avian-forest models by linking two ACN niche models with a forest landscape model (FLM), LANDIS-II. The framework facilitates the selection of available ACN models for integration, evaluation of model transferability, and evaluation of successful integration of ACN models with a FLM. We found selecting a model for integration depended on its transferability to the study area (Northern Rockies Ecoregion of Idaho in the United States), which limited the species and model types available for transfer. However, transfer evaluation of the tested ACN models indicated a good fit for the study area. Several niche model variables (canopy cover, snag density, and forest cover type) were not directly informed by the LANDIS-II model, which required secondary modeling (Random Forest) to derive values from the FLM outputs. In instances where the Random Forest models performed with a moderate classification accuracy, the overall effect on niche predictions was negligible. Predictions based on LANDIS-II simulations performed similarly to predictions based on the niche model’s original training input types. This supported the conclusion that the proposed framework is viable for informing avian niche models with FLM simulations. Even models that poorly approximate habitat suitability, due to the inherent constraints of predicting spatial niche use of irruptive species produced informative results by identifying areas of management focus. This is primarily because LANDIS-II estimates spatially explicit variables that were unavailable over large spatial extents from alternative datasets. Thus, without integration, one of the ACN niche models was not applicable to the study area. The framework will be useful for integrating avifauna niche and forest ecosystem models, which can inform management of contemporary and future landscapes under differing management and climate scenarios.

Highlights

The structure and composition of forest ecosystems are expected to shift with climate-induced changes in precipitation, temperature [1], fire [2], carbon mitigation strategies [3,4], and biological disturbances [5]
The Flammulated Owl realized niche model was implemented because the transferability assessment suggested it was an acceptable fit for the Northern Rockies Ecoregion of Idaho
The overall performance of the RF did not appear to affect the predicted Flammulated Owl distribution; the two distributions were 99.6% correlated

Summary

Introduction

The structure and composition of forest ecosystems are expected to shift with climate-induced changes in precipitation, temperature [1], fire [2], carbon mitigation strategies [3,4], and biological disturbances [5]. Climate change induced declines in tree species occurrence [6], shifts in forest carbon stocks [1], increases in forest mortality events [7], and increases in forest burn area [8] have been predicted. Forest composition and structure are integral to biodiversity [9], and the climate induced changes are likely to have wildlife biodiversity implications [10] especially for avifauna [11]. Though climate models predict increases in area burned or fire intensity, which may increase habitat suitability for woodpeckers [13], tree species composition shifts via climate change may pose adaptation constraints on them [14]. Integrating the feedbacks between wildlife, habitat, and forest processes in a modeling framework could improve our understanding of climate induced wildlife habitat changes and subsequent climate-forest adaptive management

Methods

Results

Discussion

Conclusion