Question: Can a new cost-distance model help us to evaluate the potential for accessibility bias in ecological observations? How much accessibility bias is present in the vegetation monitoring plots accumulated over the last three decades in Great Smoky Mountains National Park?
Location: Great Smoky Mountains National Park, North Carolina and Tennessee, USA.
Methods: Distance, slope, stream crossings, and vegetation density were incorporated into a least-cost model of energetic expenditure for human access to locations.
Results: Estimated round-trip energy costs for the park ranged from 0 to 1.62 × 105 J kg−1. The estimated round-trip energetic expenditure for the surveys ranged from 53 to 1.51 × 105 J kg−1. Their distribution was more accessible than the random expectation. Ten (17%) of the vegetation types in the park are significantly under-sampled relative to their area, and 16 (29%) are over-sampled. Plots in 18 of the 40 vegetation types exhibited a significant positive correlation with accessibility.
Conclusions: The least-cost model is an improvement over previous attempts to quantify accessibility. The bias in plot locations suggests using a least-cost model to test for bias in cases in which human accessibility is confounded with other sources of ecosystem variation.