OOS 3-6 - The zoogeochemistry of soil trampling by large mammalian herbivores: Does moose marching matter?

Long-term herbivore-exclusion studies have demonstrated strong net effects of large mammalian herbivores (LMH) on ecosystem carbon and nutrient cycling, but the direction and magnitude of effects are highly context dependent. Improved mechanistic understanding is needed to predict LMH effects in diverse environmental contexts, and to predict effects over ecosystem and Earth-system spatial extents. Mathematical models have played an important role in clarifying potential mechanisms of LMH zoogeochemistry, based mostly on their role as primary consumers and recyclers of organic and inorganic matter via defecation and urination. However, theoretical development of non-consumptive effects of LMH, such as physical trampling of vegetation and soil, has been slow. Trampling is nearly ubiquitous among terrestrial mammals, but the relative importance and potential mechanisms of trampling in diverse environments remain poorly understood, despite frequent calls for greater attention by theorists and empiricists alike.
In this talk, we present insights from a mathematical model of herbivore trampling effects on ecosystem nitrogen cycling, and an associated empirical study of moose trampling in a Newfoundland boreal ecosystem. The model tracks nitrogen through a simplified ecosystem with both biotic (plants, detritivores, and dead organic matter) and abiotic (soil inorganic nitrogen) nitrogen pools. With a combination of analytical and graphical analyses, we explore context-dependent effects of herbivore trampling on soil functioning. We find trampling can result in net increases, decreases or no change in N-mineralization rate, driven primarily by the sign and relative magnitude of trampling direct effects on intrinsic N-mineralization efficiency of detritivores, and on detritivore mortality. In contrast to some consumptive mechanisms, our results also suggest the pace of soil N-cycling prior to trampling has little influence over the direction of the trampling effect on N-mineralization, though the magnitude of simulated trampling net effects were greater in slow-cycling systems (e.g. boreal forests) than in fast-cycling systems (e.g. grasslands).
To empirically explore some of these findings, we take an observational approach using natural moose trails to isolate trampling effects in two distinct boreal communities (forest and heath). Using path analysis, we evaluate trampling direct effects on vegetation (composition, diversity, and height) and soil properties (bulk density, temperature, moisture, pH, SOM, %C, %N) and resulting indirect effects on microbial abundance and composition (via PLFA analysis) and N-mineralization rates. We conclude with a roadmap to further incorporating parallel consumptive and non-consumptive effects into future empirical studies and theoretical development of LMH zoogeochemistry in space and time.