Twenty-five years of community phylogenetic change in tallgrass-prairie remnants

Background/Question/Methods

There have been severe losses of tallgrass prairie habitat in the Midwestern US since European settlement. But even where remnants are protected, fragmentation, fire suppression, invasive species, and other drivers have caused ongoing losses of species diversity and undesirable changes to community composition. We asked whether this degradation was reflected in changes to community-phylogenetic measures of diversity and composition—metrics that enrich understanding of community change by factoring in evolutionary history. We used a vegetation dataset comprising 41 remnant tallgrass prairies that were sampled in 1976 and again in 2001; detailed fire records for this 25-year interval were available for 34 of the sites. A supermatrix phylogeny of the 463 taxa found in these sites was constructed using rbcL, ITS1-5.8S-ITS2, and trnL–trnF sequence data. Community phylogenetic analyses were performed using the R package picante and other tools. Our goals were to investigate: (1) the community phylogenetic structure and diversity of sites over time and at different spatial scales (site vs. plot level), (2) whether species that were clear ‘winners’ or ‘losers’ over time (significantly increased or decreased in prevalence and abundance, respectively) showed a non-random phylogenetic signal, and (3) the effects of fire management on patterns of change.

Results/Conclusions

Overall, there was non-random phylogenetic structure at the site scale, with strong evidence of ‘clustering’ (species being more closely related than expected by chance) in both 1976 and 2001 (p < 0.001). There was greater variability in structure at finer scales, with species’ co-occurrence in plots phylogenetically random in most sites (80% of sites in 1976, 63% in 2001) but clustered (7%, 20%) or overdispersed (less closely related than expected by chance; 12%, 17%) in others. The measure of phylogenetic diversity we used (phylogenetic species variability, PSV) varied across sites over a relative range of 140% (PSV = 0.28−0.66), but these differences were not explained by sampling year or sites’ moisture availability or successional stage (p = 0.13-0.75). There were clear winner and loser species over 25 years of change, but both groups were randomly distributed across the phylogeny. Fire management appeared to stabilize phylogenetic diversity, with PSV varying widely in infrequently burned sites but varying less and converging on intermediate levels as fire frequency increased. Our results indicate that fire management plays an important role in regulating not only taxonomic and functional composition of prairies, but also their phylogenetic diversity.