Understory vegetation plays an important role in maintaining biodiversity, soil nutrient cycling and carbon stocks, yet the ability of understory functional traits to predict forest productivity after harvesting disturbance is unknown. Our objective was to investigate the utility of individual traits (the community-weighted trait mean) and combined information from multiple traits (functional diversity indices) of understory plants in predicting tree diameter growth of black spruce (Picea mariana (Mill.) B.S.P.) after two harvesting methods (clearcutting and careful logging). We used vegetation and tree growth data from a network of field plots (171 plots on 43 sites) established in black spruce dominated forests across Ontario and Quebec, in northeastern Canada. Functional traits (11 traits) of dominant understory species that reflect plant morphology, regeneration strategy, and resource utilization were collected. We then used generalized linear mixed models (GLMMs) to model the response of tree diameter increment to functional diversity indices, community-weighted trait means and their interactions with harvesting method. Our results showed that: 1) at the functional diversity level, functional richness had a positive relationship with tree growth in clear-cut forests, while functional evenness showed a negative and functional divergence a positive relationship with tree growth in careful logging forests; 2) at the individual trait level, understory species with relatively large morphological trait values often promoted tree growth in both forest types, with stronger relationships in careful logging than in clear-cut forests. For regeneration traits, heavier seed weight or permanent seed bank persistence of understory plants had negative relationships with tree growth, and those relationships were either only found in careful logging forests or did not depend on harvesting type. A positive relationship was found between vegetative propagation species and tree growth, which was stronger in careful logging than in clear-cut forests. Species preferring either humid or xeric habitat had positive relationships with tree growth; both shade-tolerant and shade-intolerant species also positively correlated to tree growth. In conclusion, our study shows the important roles of understory traits in predicting tree growth based on both niche complementarity and mass-ratio hypotheses. Moreover, disturbance severity influenced the balance or degree of the positive and negative interactions that occurred between understory plants and black spruce tree growth. We suggest that identifying the threshold of harvesting disturbance severity that can activate facilitation interactions between understory plants and crop trees would be an important topic for future study.

Changes in the light availability in forests generated by diversified retention patterns (e.g., clear cut, partial harvest) have been shown to strongly filter the plant species present. Modified soil microsite conditions due to post-harvest site preparation (e.g., mechanical site preparation, prescribed fire) might also be an important determinant of plant diversity. The objective here was to detect how retention pattern and post-harvest site preparation act as filters that explain the understory functional diversity in boreal forests. We also assessed whether these effects were dependent on forest attributes (stand type, time since fire, and time since harvest). We retrieved data from seven different studies within 101 sites in boreal forests in Eastern Canada. Our data included forests harvested with two retention patterns: careful logging and clear cut, plus unharvested control forests. Three post-harvest site preparation techniques were applied: plow or disk trenching after careful logging, and prescribed fire after clear cut. We collected trait data (10 traits) representing plant morphology, regeneration strategy, or resource utilization for common species. Our results demonstrated significant variation in functional diversity after harvest. The combined effect of retention pattern and site preparation was the most important factor explaining understory diversity compared to retention pattern only and forest attributes. According to RLQ analysis, harvested forests with site preparation favored traits reflecting resistance or resilience ability after disturbance (clonal guerilla species, geophytes, and species with higher seed weight). Yet harvested forests without site preparation mainly affected understory plant species via their light requirements. Forest attributes did not play significant roles in affecting the relationship between site preparation and functional diversity or traits. Our results indicated the importance of the compounding effects of light variation and soil disturbance in filtering understory diversity and composition in boreal forests. Whether these results are also valid for other ecosystems still needs to be demonstrated.