Balsam Fir and American Beech Influence Soil Respiration Rates in Opposite Directions in a Sugar Maple Forest Near Its Northern Range Limit

Abstract

Conifers and deciduous trees greatly differ in regard to their phylogenetics and physiology as well as their influence on soil microclimate and chemical properties. Soil respiration (Rs) in forests can therefore differ depending on tree species composition, and assessments of the variation in Rs in various forest types will lead to a more thorough understanding of the carbon cycle and more robust long-term simulations of soil carbon. We measured Rs in 2019 and 2020 in stands of various species composition in a sugar maple forest near the northern range limit of temperate deciduous forests in Quebec, Canada. Seasonal variations in soil temperature had the largest influence on Rs, but conditions created by the stands also exerted a significant effect. Relative to the typical sugar maple-yellow birch forest (hardwoods), Rs in stands with >20% of basal area from balsam fir (mixedwoods) was increased by 21%. Whilst, when American beech contributed >20% of litterfall mass (hardwood-beech stands), Rs was decreased by 11 and 36% relative to hardwoods and mixedwoods, respectively. As a whole, Rs was significantly higher in mixedwoods than in other forest types, and Rs was significantly higher in hardwoods than in hardwood-beech stands. Sugar maple and American beech at the study site are near their northern range limit, whereas balsam fir is near its southern limit. Rs in mixedwoods was therefore higher than in hardwoods and hardwood-beech stands due to high root activity in the presence of fir, despite colder and drier soils. We estimated that root respiration in mixedwoods was more than threefold that in hardwoods and hardwood-beech stands. The lower Rs in hardwood-beech stands compared to hardwoods points to the lower soil temperature as well as the poor quality of beech litter (low decomposability) as indicated by a generally lower heterotrophic respiration. Other than soil temperature, regression models identified mixedwoods, soil water potential and Mg2+ activity in the soil solution as important predictor variables of Rs with about 90% of its variation explained. Our study shows the benefits of combining forest-specific properties to climatic data for more robust predictions of Rs.