Both canopy and understory traits act as response–effect traits in fire‐managed forests

Abstract

AbstractCommunity‐level shifts in the distributions of plant functional traits associated with environmental change are expected to influence ecosystem functioning. However, few studies have identified traits that both respond to environmental change and affect ecosystem properties, thus limiting potential to scale the effects of environmental change through the community level. We measured canopy and understory plant functional traits, characterizing the most abundant functional trait value (community‐weighted mean; CWM) and the functional diversity (FD), across a soil resource gradient in fire‐managed mixed‐deciduous forests to determine how traits both respond to a disturbance‐resource gradient and affect stocks of active and stable soil organic carbon (SOC) fractions. We expected that understory traits would respond mainly to fire and canopy traits would respond mainly to soil resources. We further hypothesized that fire and resource conditions affect SOC stocks (1) through mass‐ratio, by influencing trait abundance; (2) through non‐additive effects, by influencing the FD of plant communities; or (3) directly, through either combustion or environmental controls on SOC stocks. Understory traits responded to soil resource conditions and fire, whereas only canopy CWM leaf dry matter content (LDMC) varied with resource conditions; no canopy traits varied with fire. Among the response traits, canopy CWM LDMC and diversity in the maximum height of the understory were related to SOC stocks, suggesting they play dual roles as response and effect traits. SOC stocks were primarily associated with mass‐ratio effects from canopy leaf traits and secondarily with non‐additive effects from the canopy and understory. There were also strong, fraction‐dependent patterns in SOC stocks with fire disturbance. Repeatedly burned forests characterized by resource conservative traits (i.e., high canopy CWM LDMC) had a higher relative proportion of active SOC, whereas unburned forests characterized by resource acquisitive traits (i.e., high canopy CWM leaf nitrogen content) had a higher relative proportion of stable SOC. Our results suggest that canopy community‐aggregated leaf traits and diversity in understory size traits can act as both response and effect traits in disturbed forests. Predicting forest SOC stocks using a response–effect trait framework will thus require knowledge of both canopy and understory trait distributions, as well as disturbance history.