Comparison of silvicultural and natural disturbance effects on terrestrial salamanders in northern hardwood forests

In forested ecosystems timber harvesting has the potential to emulate natural disturbances, thereby maintaining the natural communities adapted to particular disturbances. We compared the effects of even-aged (clearcut and patch cut) and uneven-aged (group cut, single-tree selection) timber management techniques with natural ice-storm damage and unmanipulated reference forest sites on red-backed salamanders. We used cover boards and litter searches to survey for salamanders in northern hardwood forests in New Hampshire, USA. We estimated abundance while accounting for detection probability using the Dail–Madsen open population model. We found significant reduction in salamander abundance in recent group cuts, patch cuts, and clearcuts compared to reference forest sites, and significant but less effect of single-tree selection and ice-storm damage. Our results contribute to the evidence of detrimental effects of even-aged harvests on salamander abundance, but in contrast to most previous research, we also found lower abundance in sites following uneven-aged harvest practices when we accounted for detection probability. To more accurately reflect the total effect of harvests on salamanders, we also employed a parametric, nonlinear hierarchical model to estimate edge effects while accounting for imperfect detection. We found that group cut, patch cut, and clearcut logging reduced salamander abundance 34 m into the surrounding forest. These edge effects can greatly expand the total area affected by logging, especially in the northeastern US where cuts tend to be relatively small. This novel method for estimating edge effects will allow managers to directly calculate the total effects on populations for various size and shape harvesting plans.