Estimating occurrences of tree-related microhabitats of retained trees in harvested conifer-planted forests considering imperfect detection

Tree-related microhabitats (TreMs) indicate potential habitats for multiple taxa associated with trees and are considered viable options for evaluating the ecological importance of trees. There are limited studies on TreM occurrence in Asia and the role of imperfect detection of TreMs in occurrence estimates, which pose challenges for forest practitioners surveying TreMs. In this study, we first surveyed TreMs in 13 species of broadleaved trees retained within harvested conifer-planted forests in northern Japan to examine the TreM occurrence and the factors that influence it. We then focused on seven species of broadleaved trees and conducted a multiple-visit survey to examine the detection probability of each TreM. We found that “bark loss” was the most frequent TreM in trees retained in harvested areas, followed by “insect galleries and bore holes”, “dead branch”, and “ivy and lianas”. TreM richness increased with tree size and varied among tree species, with higher richness observed in Juglans mandshurica, Phellodendron amurense, Ulmus davidiana, and Ulmus laciniata. The estimated detection probability ranged from approximately 0.4 to 0.7, with higher values in bark loss and in ivy and lianas, and lower values in “insect galleries and bore holes”, “trunk-rot hole”, and “dead top”. The detection probability likely depends on the characteristics of the TreMs themselves, such as survey ease, rather than on the surveyed tree characteristics (tree size and shape). Overall, our results demonstrate that TreM occurrence and richness are strongly influenced by tree size and species, while detection probability varies substantially among TreM types.