Pulling on multiple threads of evidence to reveal tree species coexistence

Understanding drivers of species coexistence is a central challenge in ecology (1, 2). Coexistence cannot be observed directly, and while species co-occurrence in time and space is necessary for coexistence, it is not sufficient to prove coexistence (3, 4). Species exclusion from a region is potentially observable, but can take decades to occur, and still might just occur stochastically (5, 6). Thus, ecologists generally use theory to identify indirect observations that are indicative of mechanisms driving coexistence or exclusion (7-9). Various methods have been developed to indirectly infer coexistence, each of which requires different data, none of which are usually conclusive on their own, and are not typically combined (e.g. (9-12)). Here, we combine three different approaches which are normally used in isolation to build a case about a lack of coexistence of multiple hardwood species. Importantly, though each method has its flaws, we find that all three point to the same conclusion which strengthens the case for a lack of coexistence. First, in an experimental planting of three mature tree species we found no relationship between productivity and species diversity, which has been hypothesized to occur due to a lack of niche differences among species. Second, we used modern coexistence theory to calculate niche and fitness differences for each pair of species, which confirmed the high niche overlap among species and competitive dominance of one species. Third, we used the United States Department of Agriculture Forest Inventory and Analysis data to examine co-occurrence patterns of our species across thousands of natural forest stands and found that, these three species were distributed randomly throughout the USA, and do not significantly co-occur as would be expected if coexistence occurred naturally. Given that these independent methods agree; we take this as strong evidence about a lack of coexistence.