American marten (Martes americana) are frequently reported to require late-successional, conifer-dominated forests. In the northeastern United States, however, marten use mixed coniferous-deciduous and deciduous-dominated forests, as well as forests defoliated by insects or managed for fiber. Previous research in Maine also has shown that American marten avoid recent and regenerating clearcuts with tree height <6 m when establishing home ranges. However, forest stands with a history of extensive defoliation and tree mortality due to an eastern spruce budworm epidemic are not avoided. Such results at the scale of the forest stand suggest that structural attributes other than stand age and dominant overstory type are important determinants of habitat suitability for marten. Characteristics of forest structure contributed by coarse woody debris, understory vegetation, and overstory vegetation have functional significance to marten, and may be the common currency through which habitat suitability for marten should be assessed.

We compared forest structural attributes between areas receiving different intensities of use by marten and site-level habitat between stands regenerating following budworm defoliation versus stands regenerating following clearcutting. Our objectives were to: (1) identify specific, site-level characteristics associated with patterns of spatial use by marten; (2) identify features of forest structure with functional significance for marten that are lacking in some managed stands; (3) offer recommendations for silvicultural practices that would maximize future habitat quality for marten in managed forests; and (4) propose goals for silvicultural practices that more closely resemble natural disturbance and thus improve habitat quality for marten.

We used radio-location data from resident, adult marten in Maine to identify areas of high useintensity and low use-intensity in an industrial forest managed for pulpwood and in a forest reserve. We also identified unused areas in the industrial forest that had been harvested via clearcutting 10-20 years prior to our study and were adjacent to used areas. Characteristics of coarse woody debris, understory vegetation and overstory vegetation in each area were quantified using standard mensurational techniques. In the industrial forest, overstory features related to stand maturity were most useful for discriminating between areas receiving different intensities of marten use. Used areas had taller trees, higher live-tree basal areas, and greater snag volumes than unused areas. Areas receiving high use were further characterized by relatively high deciduous basal areas relative to areas receiving low use by marten. Ground and understory structure was not limiting in any useintensity category. Few habitat differences were detected between high-use and low-use areas in the forest reserve, suggesting that other factors (e.g., intraspecific competition associated with high marten density) may have been responsible for patterns of marten spatial use there. Based on our results, we suggest stand-level management guidelines to predict when overstory development becomes sufficient for regenerating stands to support marten. These guidelines include live-tree basal area >18 m2/ha (78 ft2/ac), mean tree height >9 m (30 ft) for trees with DBH >7.6 cm (3 in), and snag volume >10 m3/ha (43 ft3/ac). Methods of harvesting that retain some overstory characteristics may be consistent with conservation of marten habitat in managed forests. Our guidelines can be used for developing stand-scale harvesting strategies in forests that support marten; however, companion studies suggest that landscape-scale habitat requirements of marten should be managed for simultaneously.

We used standard mensurational techniques to quantify characteristics of overstory vegetation, understory vegetation, and coarse woody debris in stands clearcut 10-20 years prior to our study that were not used by resident, adult marten. The protocol was repeated in nearby stands that were defoliated by eastern spruce budworms 10-20 years previously and did receive marten use. In contrast to regenerating clearcuts, features associated with stand maturity were retained in budworm-defoliated stands, i.e., defoliated stands had greater volumes of snags, downed logs and root masses, and included taller trees. Although live-tree basal area was similar between stand types, our results suggest that vertical structure provided by large snags can offset limited availability of live trees for marten, particularly where coarse woody debris and understory vegetation is plentiful. Marten habitat quality may be enhanced in stands under even-aged management by retention of cull trees and snags such that basal area of live trees and snags exceeds 18 m2/ha (78 ft2/ac). Unevenaged silvicultural systems, which more closely mimic natural disturbance by defoliating insects, have particular promise as a management tool for maintaining marten habitat. We recommend that future research focus on marten-habitat relationships in stands regenerated by uneven-aged management.