This data was used to investigate how the age and size of beaver disturbances act as predictors for primary wood-boring beetle abundance and species richness around beaver-altered habitat patches. To do so, we sampled beetles around 16 beaver-disturbed and unaltered watercourses within the Kouchibouguac National Park (Canada) and modeled beetle demographical responses to site conditions and their physical characteristics, distance from the watercourse, deadwood biomass, and the geographical location of the sites.

The study was conducted in Kouchibouguac National Park (lat. 46° 48’ N, long. 64° 54’ W), a 238-km² federal park located in eastern New Brunswick, Canada. We used 16 sites previously identified as beaver-engineered (BEA), where watercourses encompassed beaver activity material (i.e., beaver dam, pond and/or meadow). We determined the age (in years; 12 to 44 years old) and the perimeter (in m; 135 to 889 m) of the disturbed area through aerial photography. We paired BEA sites with control sites (CON), which each included a watercourse without any sign of beaver activity within 100-m radius from the CON periphery. Hereafter, we will use the term “site condition” as a general concept encompassing both BEA and CON. At each site, we ran a 50-m transect oriented perpendicularly to the watercourse. For BEA, each transect began at the limit of the flooded area. The transects for meadows began at the end of habitat patches linked to ancient flooding. Transects in BEA and CON intersected the same type of habitats since we used beaver ponds with a sharp perturbation edge, preventing any bias of wetlands (not directly linked to beaver activity) buffering the beaver-modified habitat.

Flight-intercept traps (FIT) were installed at 5, 20, and 50 m along the transects. Each trap was composed of two perpendicularly inserted 30.5 x 61 cm acrylic sheets under which a styrene funnel was attached. A plastic cup was placed at the bottom of every funnel and contained a mixture of 70% ethanol and a drop of dishwashing soap. Two clothespins were pinned on the base of an acrylic sheet to stabilize the FIT and a round styrene cover was placed over the FIT to prevent debris and rain from accumulating in the collecting cup. Each trap was suspended at 1.6 m above ground, from the center of the acrylic sheets, on a rope attached to two trees that were less than 4 m apart. Placing the traps between the trees rather than against the trunk of a given tree species allowed for the sampling of all saproxylic beetles and not only that of the beetles attracted to a given host species.

From the beginning of June to the end of August in 2015 and 2016, traps were emptied bimonthly to recover wood-boring beetles. We only selected beetle taxa known to actively bore into fresh dead wood material (i.e., Ptinidae, Lymexylidae, Cerambycidae and Buprestidae families, and Scolytinae, Conoderinae and Cossoninae subfamilies). Beetles were classified in two trophic categories, being coniferous borers or deciduous borers.

The dataset includes the pooled volume in our sampling plots of all woody debris (snags and logs) of the same essence (coniferous or deciduous) for fresh debris (class 1–2) and old debris (class 3-4-5), our ordinal numeric variables (i.e., distance from the watercourse, perimeter of the disturbance, disturbance age, site coordinates), and five dependent variables (i.e., Anisandrus sayi abundance, coniferous-borer abundance/richness, deciduous-borer abundance/richness). Anisandrus sayi abundance was analyzed separately from other deciduous borers because of its sheer abundance compared to other species in this study and other work carried out in New Brunswick, Canada.