Level and pattern of overstory retention shape the abundance and long-term dynamics of natural and created snags

• Published in: Forest ecology and management Vol. 526; p. 120575
• Main Authors: Halpern, Charles B., Rossman, Allison K., Hagar, Joan C.
• Format: Journal Article
• Language: English
• Published: Elsevier B.V 15-12-2022
• Subjects: Aggregated retention; Dispersed retention; Regeneration harvest; Snag decay; Snag longevity; Snag recruitment; Aggregated retention; Dispersed retention; Snag decay; Snag longevity; Regeneration harvest; Snag recruitment
• ISSN: 0378-1127; 1872-7042
• DOI: 10.1016/j.foreco.2022.120575

•Pre-harvest variation in snag density persisted for two decades after harvest.•Level and pattern of retention had strong effects on snag carryover through harvest.•Level and pattern had weak effects on long-term survival, decay, and recruitment.•Aggregated retention sustained greater density and diversity of snag decay states.•Aggregates mitigate loss of large-diameter snags that are critical to biodiversity.
Standing dead trees, or snags, serve myriad functions in natural forests, but are often scarce in forests managed for timber production. Variable retention (VR), the retention of live and dead trees through harvest, has been adopted globally as a less intensive form of regeneration harvest. In this study, we explore how two key elements of VR systems — level (amount) and spatial pattern of live-tree retention — affect the carryover and post-harvest dynamics of natural and artificially created snags. We present nearly two decades of data from the DEMO Study, a regional-scale experiment in VR harvests of Douglas-fir-dominated forests in the Pacific Northwest. Snag losses to harvest were greater at 15 than at 40% retention (67 vs. 47% declines in density) and greater in dispersed than in aggregated treatments (64 vs. 50% declines). Densities of hard and tall (≥5 m) snags were particularly sensitive to low-level dispersed retention, declining by 76 and 81%, respectively. Despite these losses, post-harvest densities correlated with pre-harvest densities for most snag size and decay classes. In contrast to initial harvest effects, snag densities changed minimally over the post-harvest period (years 1 to 18 or 19), with low rates of recruitment offsetting low rates of loss. Post-harvest survival of snags was greater at 15 than at 40% retention (79 vs. 69%), as were rates of decay (68 vs. 52% of hard snags transitioned to soft). However, pattern had no effect on either process. Snag recruitment did not vary with retention level or pattern at the scale of the 13-ha harvest unit, but was several-fold greater in the 1-ha aggregates (14.3–27.8 snags ha−1) than in the corresponding dispersed treatments (4.2–5.3 snags ha−1). Snag size (diameter) distributions showed greater change in dispersed than in aggregated treatments, reflecting greater loss of smaller snags and recruitment biased toward larger snags. Created snags showed uniformly high survival (97%), irrespective of treatment, but rates of decay were greater at lower retention. If a goal of VR is to sustain snag abundance and diversity through harvest, emphasis should be placed on minimizing initial losses, either by reducing the intensity of felling in areas of dispersed retention or locating forest aggregates in areas of greater initial snag density, diversity, or incipient decay.