Establishing protected areas and maintaining biodiversity in managed forests are the main methods to conserve forest habitats and their biodiversity. The habitat characteristics that affect forest biodiversity in both protected and managed forests occur on different spatial scales, with the smallest scale consisting of structures at the level of single trees and smaller. Under the influence of dynamic processes (ecological succession, natural and anthropogenic disturbances) small-scale structures may be strongly affected, altering their contribution to local biodiversity. In this thesis, I studied two small-scale structures in northern boreal forests: deciduous trees (mainly European aspen (Populus tremula)) in conifer-dominated forests and woodpecker-made breeding cavities. These structures provide resources for multiple species. However, their abundance could be affected by dynamic processes, also in protected areas. I examined the temporal dynamics of these structures through long-term (16-30 years) monitoring. Specifically, I studied the recruitment of deciduous trees in managed forests, aspen demography in old-growth forests, the effect of tree-and stand-level factors on cavity persistence, and the potential to detect aspen for conservation purposes via remote sensing. My main findings were: 1) Recruitment of deciduous trees is enhanced by prescribed burning. Mammalian herbivory impacts recruitment significantly, although its effect depends on forest age. 2) Living aspens declined by 37% in protected old-growth forests over an 18-year period and recruitment rate was low despite an abundance of saplings. Recruitment primarily occurred on forest edges. 3) Tree species, size, and condition affect the persistence of woodpecker-made cavities, with cavities lasting longer in coniferous, healthy, and larger trees. For aspen, cavities in small trees also persisted a long time. 4) Multispectral drone images can be used to detect scattered mature aspen trees in oldgrowth forests, most ideally in late spring. My results emphasize that dynamic processes in forests can alter forest characteristics that are important for biodiversity quickly. Small-scale structures, that are essential to maintain local biodiversity, underwent strong changes in just a few decades. My thesis highlights the importance of continuous monitoring of biodiversity-rich habitat structures to uncover major changes in the ability of protected areas to sustain biodiversity. Remote sensing is potentially a valuable monitoring tool, also for tree-level small-scale forest structures.