A key goal of ecology is improving our understanding of how environmental conditions shape species’ spatial distributions. From a management perspective, the ability to predict current and future distributions of important subsistence resources and potentially harmful invasive species will improve monitoring and conservation efforts. However, attempts to model species’ distributions are often hampered by the availability of spatially unbiased species records across the landscape. We will present two studies that bridge this gap through the use of the USDA Forest Inventory and Analysis (FIA) vegetation dataset.
In the first study we utilized FIA records collected throughout the Tanana River Basin of interior Alaska to predict shifts in the distribution and abundance of the subsistence species bog blueberry (Vaccinium uliginosum) under future climate scenarios. Second, we assessed the common practice of building regional-scale species distribution models using solely climatic predictors through an examination of the relative contributions of abiotic and biotic predictors to models for the distributions of six common invasive plants in Western U.S. forests across five spatial resolutions. Existing tests of this practice for predicting invasive species’ distributions are rare and present contrasting conclusions but are of substantial importance for informing monitoring and control efforts on managed lands.
Results/Conclusions
The first study revealed that presence of V. uliginosum is anticipated to decrease dramatically under future climatic conditions, particularly under high emissions scenarios. On average, presence within 25 km of interior Alaskan communities is predicted to decline by 38% and cover by 6% as V. uliginosum shifts towards cooler high-elevation areas. Our results suggest that this subsistence resource may become less abundant in interior Alaska under climate change and emphasize geographic regions that may warrant intensified monitoring.
In the second study, the mean deviance explained by models constructed using solely abiotic predictors (D2 = 0.26 ± 0.03) did not differ significantly from those containing abiotic and biotic predictors (D2 = 0.29 ± 0.03) at all spatial resolutions, supporting the practice of constructing regional-scale SDMs using abiotic predictors alone. However, the mean deviance explained by biotic predictors (D2 = 0.14 ± 0.02) was unaffected by decreasing resolution, suggesting that biotic conditions may be correlated with invasive species distributions even at resolutions typical of SDMs.
These studies highlight the potential of the FIA vegetation dataset to support examinations of theoretical and applied questions in macroecology and land management by providing unbiased, spatially comprehensive records of species presence at the landscape scale.