Snag dynamics in northern hardwood forests under different management scenarios

• We compare snag dynamics in Control, Harvest Only, and Harvest Plus Snags scenarios.
• Tree dbh, species, and stand management status were related to snag retention.
• Girdling was generally effective for snag creation in northern hardwoods.
• Managed stands lost snags on average, significant for all sizes but not large snags.
• Plus Snags mitigated some snag losses, largely due to more natural snag recruitment.

Snag retention is increasingly being incorporated into forest management guidelines. Questions remain, however, in northern hardwood systems regarding factors affecting retention in actively managed stands, the effectiveness of snag creation, and the net effects of snag creation and timber harvesting on snag numbers and sizes. To address some of these questions, we examined the dynamics of natural and created snags within mature, even-aged northern hardwood forests under seven different management scenarios: three Harvest Only treatments, three Harvest Plus (created) Snags treatments, and one untreated Control. We found tree diameter, tree species, and stand treatment status (i.e., managed or Control) to be related to the retention of natural snags, created snags, or both. Snags were less likely to remain standing if they had smaller diameters, were species with relatively rapid decay rates, or were found in stands that had been logged. We found girdling trees to be an effective method of dead wood creation, although trees took longer to die than we expected. At least 84% of girdled trees had died in most stands within 4.5 years of girdling, and 30–77% of girdled trees were still standing 5.5 years after treatment. Comparison of net effects of snag creation and timber harvesting among treatments showed that managed stands, on average, experienced net snag losses compared to untreated Controls. These losses were statistically significant for all snag sizes, but not for large snags alone (i.e., dbh ⩾25.4 cm). Active management prescriptions that included snag creation demonstrated the potential to mitigate snag losses, with the extent of mitigation varying with the type of management. Surprisingly, mitigation was primarily driven by significantly greater natural snag recruitment in Plus Snags treatments, potentially due to competition from girdled trees that had not yet died. Our results may help inform the development of snag management guidelines in even-aged, second-growth northern hardwood systems for forest managers who are interested in enhancing the structural complexity of these forests.