Looking at the differences between treatments on raw data, I found that seismic lines in cut-blocks have a substantial higher stocking level than lines in uncut forest for height category 2, but smaller differences for height category 1 (figure 12). More precisely, when forest harvesting occurs together with seismic lines, the mean stocking level ~ 135% higher than lines in mature forest for height category 2, and only ~ 50% higher for height category 1. Moreover, I found that differences in stoking levels between seismic lines and adjacent forest in cut-blocks is almost null for height category 1 and ~ 30% for height category 2 (higher in adjacent forest). Since I want to compare the total trees densities in height category 2 with the Alberta's reforestation standard, I present my results in figure 12 as number of stems per hectare. Comparing stoking levels with the Alberta regeneration standard (5 000 n/ha) I found that 70% of seismic lines in cut-blocks reached the benchmark level (on average ~ 80% higher), while only 30% of lines in the uncut mature forest this trees

density.

For question (i) the model showed a positive effect of harvesting and a negative effect of seismic lines on total stems density for both height categories. For height category 1 I didn't find any significant effects of the predictor factorial variables HARVESTING and LINES. Conversely, for height category 2 HARVESTING showed a significant positive effect on total trees regeneration while LINE present a negative non-significant effect (figure 13).

Analyzing the total stems densities (trees and shrubs) against the presence/absence of forest harvesting and seismic lines for different species group, I found substantial differences between treatments only for deciduous species (figure 14). Specifically, the stems regeneration inside the line in the cut-blocks plots was ~ 130% higher than seismic lines in uncut forest, but ~ 47% lower than the adjacent forest in cut-blocks.

From the model fitting I found that HARVESTING and LINE don't show a significant effect for both Coniferous and Shrubs species, while for deciduous species, both predictor variables present an almost significant effect on stems regeneration (figure 15). Specifically, harvesting showed a positive marginal effect on response variable, while the presence of seismic lines a negative effect.

In Alberta's boreal forest seismic lines represent one of the largest contributions to landscape fragmentation, with negative impacts on wildlife populations (Schneider et al. 2010, Dabros et al., 2018). Identifying favorable conditions for a natural reforestation of the line is now a priority for a cost-effective implementation of conservation resources. A key question that is part of this issue is whether forest harvesting ‘erases’ seismic lines by promoting natural tree regeneration through the re-establishment of early seral conditions typical of post-harvest cut-blocks or whether residuals effects from the line persist. Here, I found that 70% of seismic lines within cut-blocks reached the density target from The Alberta Reforestation Standard, while 30% of seismic lines in the mature uncut forest reached the same benchmark standards. Moreover, stocking level inside the line within harvests was on average 80% higher than in the uncut forest. I hypothesized that when a clear cut occur after the line has been cut, regenerating factors provided by the harvesting (e.g., higher sunlight and reduction of competition; Picchio et al., 2020) might drive the reforestation inside the line. My results suggest that harvesting does provide the conditions for the structural regeneration of seismic lines (see question (i) and figure 12), even when natural regeneration does not occur in adjacent unharvested sites. This suggests that harvesting practices are facilitative and thus effectively ‘erase’ seismic lines, even though factors, such as soil compaction, seem to be limiting. Indeed, when different species groups are considered separately, seismic lines show a significative and substantial negative effect on stems regeneration (see question (ii) and figures 14) with deciduous species that represent the most abundant species in cut-blocks but with lower density inside the line. This might suggest that, when the regeneration of the cut-block "erase" seismic lines, the higher soil compaction inside the line delay stems establishment, resulting in lower density of high stems after 10 years. Looking at stem density of other groups (coniferous and shrubs) I found small stems densities in every scenarios with no significant differences between treatments. This suggest that, despite shrubs and coniferous are generally scarce in the studied areas compared to deciduous species, the presence of cut-blocks doesn’t seem to negatively affect their regeneration. Finally, I found that the species composition in cut-blocks tend to be highly dominated by aspen that represent the most abundant species in both seismic lines and adjacent forest. The high dominance of aspen is likely explainable by the reproductive strategy of this species that allow this species to successfully establish in the post-harvest area. Aspen is clonal tree species that regenerates via root suckering after disturbances (Frey et al., 2003). Due to this reproductive strategy, the higher sunlight in the post-harvest areas (and the consequent warming of the soil), represents a significant positive factor for the emergence of aspen suckers, and it likely explain the high densities found in my cut-block sites. Conversely, an interesting result is the higher density of paper birch inside the seismic lines. The regenerations density of this species was significantly higher in both lines from mature forest and cut-blocks, suggesting the presence of favorable conditions for this species inside the corridor. Nevertheless, I believe that the reasons at the base of this pattern cannot be explained by my data, and further investigations are required.

Overall, I found that forest harvesting, and seismic lines present a negative interactive effect on stands regeneration when species are considered separately, suggesting that the hypothesis 2 (see introduction) is more supported by my data. Nevertheless, seismic lines seem to sufficiently regenerate when a clearcut occurs, with the connectivity of the line that is successfully interrupted without the need of an active restoration treatment. Although forest harvesting is not an immediate solution for sensitive species like caribou (early seral habitat negatively impacts caribou), my results suggest that areas that have forest harvesting will contribute in the long-term towards landscape-level structural restoration of linear features without more directed active restoration need.