Network architecture of plant-fungal root endosphere communities across a productivity and growing season length gradient in a sparsely vegetated alpine ecosystem

Monica Brady and Emily Farrer, Ecology & Evolutionary Biology, Tulane University, New Orleans, LA, Clifton P. Bueno de Mesquita, DOE Joint Genome Institute, Berkeley National Laboratory, Berkeley, CA, Dorota L. Porazinska, University of Florida, Gainsville, FL, Marko Spasojevic, Department of Evolution, Ecology, and Organismal Biology, University of California Riverside, Riverside, CA, Jane G. Smith and Katharine N. Suding, Institute of Arctic and Alpine Research, University of Colorado, Boulder, CO, Steven K. Schmidt, Ecology and Evolutionary Biology, University of Colorado, Boulder, CO

Background/Question/Methods
A fundamental question in community ecology is how communities change across environmental gradients. However, most work on this topic has focused on single trophic levels. The use of network analyses to study multiple levels of ecological communities is a powerful tool for understanding the structure and dynamics of communities. Ecological networks are not often used to analyze communities over gradients but rather look at communities at a single site. Here, we test how two environmental gradients, productivity and growing season length, measured by soil moisture and snow depth respectively, influence the network structure of plant-fungal root (ITS) communities. Our study system is a high-elevation alpine habitat that exhibits variability in productivity and growing season length over landscape. To test for differences in network structure, we constructed bipartite networks and evaluated network architecture as measured by specialization, modularity and nestedness. We investigate the mechanisms behind any shifts in network structure across the gradients by assessing whether compositional change, species richness, or abiotic conditions per se may be driving the shifts. We hypothesize that network complexity will increase with longer growing seasons and higher productivity and that changes in community composition will best explain changes in network structure across both gradients.
Results/Conclusions
We find that there are visual differences in network structure between low, medium, and high productivity and growing season length. Network complexity decreases at shorter growing season length and at low productivity. Across the two gradients, the networks exhibited high modularity, indicating that the communities are highly organized into clusters with little interaction between the clusters. In addition, all networks exhibited relatively low nestedness, lower than would be expected by chance, meaning that taxa with few connections did not tend to have a subset of the connections that taxa with more connections had. Further, all networks exhibited moderate interaction specialization. For fungal composition, productivity explains 2.4% of the variation and growing season length 1.1% of the variation. For plant composition, productivity explains 10% of the variation and growing season length 4% of the variation. In addition, for plants there is a strong linear relationship between species richness and both productivity and growing season length. However, for fungi there was no relationship between richness and either of the two gradients. Overall, results indicate that network complexity changes with growing season length and productivity, but that community composition or species richness may not best explain network structure changes over the gradients.