1. Plant-fungal associations strongly influence forest carbon and nitrogen cycling. The prevailing framework for understanding these relationships is through the relative abundance of arbuscular (AM) versus ectomycorrhizal (EcM) trees. Ericoid mycorrhizal (ErM) shrubs are also common in forests and interactions between co-occurring ErM shrubs and AM and EcM trees could shift soil biogeochemical responses. Here we test hypotheses that the effects of ErM shrubs on soil carbon and nitrogen either extend or are redundant with those of EcM trees.

2. Using regional vegetation inventory data (>3,500 plot observations) we evaluated the frequency, richness, and relative abundance of ErM plants in temperate forests in the eastern United States and examined their relationship with EcM plant cover. We then used surface soil (7 cm) data from 414 plots within a single forest to analyze relationships between ErM plant cover, relative EcM tree basal area, and soil carbon and nitrogen concentrations while accounting for other biogeochemical controls, such as soil moisture.

3. At both scales, we found a positive relationship between ErM and EcM plants, and the majority of ErM plants were in the shrub layer. Within the forest site, ErM plants strongly modulated tree mycorrhizal dominance effects. We found negative relationships between EcM relative basal area and soil carbon and nitrogen concentrations, but these relationships were weak to negligible in the absence of ErM plants. Both EcM relative basal area and ErM plant cover were positively associated with the soil carbon-to-nitrogen ratio. However, this relationship was driven by relatively lower nitrogen for EcM trees and higher carbon for ErM plants. As such, the functional effects of ErM plants on soil biogeochemistry neither extended nor were redundant with those of EcM trees.

4. Synthesis. We found that ErM shrubs strongly influenced the relationship between tree mycorrhizal associations and soil biogeochemistry, and the effects of ErM shrubs and EcM trees on carbon and nitrogen were functionally distinct. Our findings suggest that ErM shrubs could confound interpretation of AM versus EcM tree effects in ecosystems where they co-occur but also bolster growing calls to consider mycorrhizal functional types as variables that strongly influence forest biogeochemistry.

We collected soil and vegetation data from 420 plots arrayed across a 3,213-ha second-growth, mixed-hardwood forest in Connecticut, USA (41°57’ N, 72°07’ W) to test the effects of ericoid mycorrhizal (ErM) plant cover on surface carbon and nitrogen concentrations across a tree mycorrhizal dominance gradient. The plot arrangement consisted of 42 transects established in parallel pairs located 50 m apart. Each transect included ten 4-m radius plots located every 20 m. In May­­–July 2013, we estimated percent cover of all understorey ErM plant species within each 4-m radius plot and estimated total basal area of trees ≥1.37 m in height from the plot center using variable radius plot sampling with a 2.3 m² ha^-1 basal area factor angle gauge. For each tree tallied during VRP sampling, we also identified the species and measured DBH at 1.3 m. In 2014, we collected and pooled 20 2-cm diameter, 7-cm deep soil cores within each plot after removing the litter layer. We also measured soil temperature and moisture (volumetric water content) in the field on three occasions during the 2014 growing season (May, July, and September) and recorded five measurements per plot at each sampling occasion. Soils were processed at Yale University and analyzed at the Yale Analytical and Stable Isotope Center. 

The U.S. National Park Service (USNPS) data used in this manuscript were collated by Robert Warren II and are publicly available from the USNPS Vegetation Inventory and Monitoring Program website: https://www.nps.gov/im/vmi-products.htm. Plots collated from this larger dataset met the following criteria: (1) occurred in the eastern U.S., (2) were deciduous forest types, and (3) had a plot size of 20 x 20m (400 m²). Although this dataset is publicly available, and our co-author group did not collect it, we included the subset of USNPS plots used here given these selection criteria.

This dataset is comprised of three tabs in a single excel file. See the "metadata" tab for information pertaining to the variables measured and analyzed. The "CT_data" tab includes values for all the soil variables analyzed and reported on from the Connecticut site. The "USNPS_data" tab includes the percent vegetation cover for each plant taxon by stratum/vegetation layer for each of the U.S. National Park Service plots analyzed and reported on in the manuscript.