Analytic dataset informing prediction of subterranean cave and mine ambient temperatures

Caves and other subterranean features provide unique environments for many
species. The importance of cave microclimate is particularly relevant at
temperate latitudes where bats make seasonal use of caves for hibernation.
White-nose syndrome (WNS), a fungal disease that has devastated
populations of hibernating bats across eastern and central North America,
has brought renewed interest in bat hibernation and hibernaculum
conditions. A recent review synthesized current understanding of cave
climatology, exploring the qualitative relationship between cave and
surface climate with implications for hibernaculum suitability. However, a
more quantitative understanding of the conditions in which bats hibernate
and how they may promote or mediate WNS impacts is required. We compiled
subterranean temperatures from caves and mines across the western United
States and Canada to: a) quantify the hypothesized relationship between
mean annual surface temperature (MAST) and subterranean temperature and
how it is influenced by measurable site attributes, and b) use readily
available gridded data to predict and continuously map the range of
temperatures that may be available in caves and mines. Our analysis
supports qualitative predictions that subterranean winter temperatures are
correlated with MAST, that temperatures are warmer and less variable
farther from the surface, and that even deep within caves temperatures
tend to be lower than MAST. Effects of other site attributes (e.g.,
topography, vegetation, precipitation) on subterranean temperatures were
not detected. We then assessed the plausibility of model-predicted
temperatures using knowledge of winter bat distributions and preferred
hibernaculum temperatures. Our model unavoidably simplifies complex
subterranean environments, and is not intended to explain all variability
in subterranean temperatures. Rather, our results offer researchers and
managers improved broad-scale estimates of the geographic distribution of
potential hibernaculum conditions compared to reliance on MAST alone. We
expect this information to better support range-scale estimation of winter
bat distributions and projection of likely WNS impacts across the West. We
suggest that our model predictions should serve as hypotheses to be
further tested and refined as additional data become available.