The condition of forest ecosystems depends on the temporal and spatial pattern of management interventions and natural disturbances. Remnants of previous conditions persisting after disturbances, or ecosystem legacies, collectively comprise ecosystem memory. Ecosystem memory in turn contributes to resilience and possibilities of ecosystem reorganization following further disturbance. Understanding the role of disturbance and legacies is a prerequisite for maintaining resilience in the face of global change. Several legacy concepts discussed in the peer‐reviewed literature, including disturbance, biological, soil, land‐use, and silvicultural legacies, overlap in complex ways. Here, we review these established legacy concepts and propose that the new terms “material legacy” (individuals or matter, e.g., survivors, coarse woody debris, nutrients left after disturbance) and “information legacy” (adaptations to historical disturbance regimes) cut across these previous concepts and lead to a new classification of legacies. This includes six categories: material legacies with above‐ and belowground, and biotic and abiotic categories, and information legacies with above‐ and belowground categories. These six legacies are influenced by differential patterns of editing and conditioning by “legacy syndromes” that result from natural or human‐manipulated disturbance regimes that can be arranged along a gradient of naturalness. This scheme is applied to a case study of hemiboreal forests in the Baltic States of Estonia, Latvia, and Lithuania, where natural disturbance, traditional clearcut silviculture, and afforestation of abandoned agricultural lands constitute the three main legacy syndromes. These legacy syndromes in turn influence forest response to management actions and constrain resilience, leading to a mosaic of natural, manipulated, and artificial (novel) ecosystems across the landscape, depending on how the legacies in each syndrome affect ecological memory.

In light of the difficulties in stand volume estimation of natural forests, we analyzed height–diameter relationships and derived a set of height estimation equations for volume estimation for naturally developing forest ecosystems, using the Järvselja old-growth and the Laeva commercial forest in Estonia as a case study. This contribution presents an approach to model individual tree height–diameter relationships for Scots pine, common aspen, silver and downy birch, Norway spruce, black alder, gray alder, linden species, European ash, Norway maple, deciduous species and coniferous species in multi-size and mixed-species naturally developing stands in Estonia. Single-tree-level data were collected in 2013. Two methods were used to obtain stand-level data: whole surface inventory and partial surface inventory. To model the height–diameter relationship in naturally developed mixed-species forest stands in order to predict single tree height based on observed diameter at breast height, we applied nonlinear mixed models where we applied the Chapman–Richards and Näslund models as fixed-effects and the influence of the species contribution at the sites as random effects. The fixed-effects followed a set of criteria: (1) height starts at h = 1.3; d = 0; (2) the applied functions are monotonically increasing with a clear inflection point and 3) the fixed-effect model has an asymptotic value) in a naturally developed mixed-species forest.

Changing disturbance regimes and uncertain recovery trajectories under global change have created a need for conceptual models to guide forest management within a sustainability framework. This conceptual article develops on operational framework for ecosystem memory in forest ecosystems. Conventional views of resilience may foster the misleading assumption that ecosystem development follows a fixed trajectory within the boundaries of the natural range of variation. However, climate change may transform resilience mechanisms, causing forest ecosystems to shift beyond their historical range of variation. Under such conditions, the concept of ecosystem memory offers a framework for analyzing and quantifying temporal system properties. The constraints inherited through memory patterns are themselves likely to be modified under changing environmental conditions. A valid interpretation of ecosystem memory requires a clear ontological understanding of memory components as material entities organized under temporal and spatial constraints. The human memory metaphor is an appealing entry point for understanding ecosystem memory but risks erroneously attributing semantic encoding and representational processes to ecological systems, where no such mechanisms exist. Based on a narrative synthesis of 39 publications, supplemented by an updated Web of Science search that identified 396 records, we develop an operational ecosystem-memory framework that distinguishes persistent ecological legacies from functionally active memory by linking ecological imprints, ecosystem engrams, and measurable ecosystem responses. The framework provides a basis for identifying and monitoring measurable legacy variables, including deadwood, retained trees, regeneration, soil properties, and refugial structures, in post-disturbance forest management.