Dispersal and nutrient limitations of decomposition above the forest floor: Evidence from experimental manipulations of epiphytes and macronutrients

Decomposition is a major component of global carbon cycling. However, approximately 50% of wood necromass and a small proportion of leaf litter do not contact the forest floor, and the factors that regulate the decomposition above the forest floor are largely untested. We hypothesized that separation from soil resources causes slower decomposition rates above the forest floor. Specifically, we tested whether slower decomposition results from decreased nutrient availability (the nutrient limitation hypothesis) and/or microbial dispersal limitation (the dispersal limitation hypothesis) in the absence of soil resources. We tested these hypotheses by combining experimental manipulations of epiphytes and macronutrient fertilization with elemental analyses and community metabarcoding (fungi and prokaryotes). Specifically, we compared wood stick and cellulose decomposition among three treatments: an unaltered trunk section, an epiphyte mat, and a 'removal treatment' where an epiphyte mat was removed to test the effect of soil resources. We also performed a factorial fertilization experiment to test the effects of nitrogen (N) and phosphorus (P) on the decomposition of suspended cellulose. Decomposition rates were fastest on the epiphyte mats, intermediate in the removal treatment and slowest in the controls. Phosphorus addition increased decomposition rates in the fertilization experiment, and greater P concentrations, along with some micronutrients, were associated with increased rates of decomposition on the epiphyte mats and in the removal treatments. Locally dispersed fungi dominated the wood stick communities, indicating that fungal dispersal is limited in the canopy, and fungal saprotrophs were associated with increased rates of decomposition on the epiphytes. These experiments show that slowed decomposition above the forest floor is caused, in part, by separation from soil resources. Moreover, our findings provide support for both the nutrient limitation and dispersal limitation hypotheses and indicate that mechanisms regulating canopy-level decomposition differ from those documented on the forest floor. This demonstrates the need for a holistic approach to decomposition that considers the vertical position of necromass as it decomposes. Further experimentation is necessary to quantify interactions between community assembly processes, nutrient availability, substrate traits, and microclimate. A free Plain Language Summary can be found within the Supporting Information of this article.