COS 82-3 - Long-term studies of a subtropical grassland: Grasslands change slowly

Background/Question/Methods

We studied patterns of beta diversity in a subtropical grassland in South Africa over 60 years. Burning and mowing treatments were applied in a factorial design to 132 plots (each 15 x 20 yards) that were established in 1950. Plots were burned annually, biennially, triennially, or left unburned (control). The same plots were mowed in early summer, in late summer, early plus late summer, or left unmowed (control). We examined changes in beta diversity over time, with a focus on replacement and richness differences. Four indices of beta diversity were calculated for comparison: Jaccard and Sørensen (qualitative), Ružička and Bray-Curtis (quantitative). We used simplex (triangle or ternary) plots to examine the relative importance of β diversity, replacement, and richness differences. We also performed classic beta regressions with a mean model with a logit link and multivariate distance-based redundancy analyses (dbRDA) to examine the relative effects of mowing and burning on replacement and richness differences.

Results/Conclusions

There was no significant effect of richness differences in the four indices of beta diversity, but a strong effect of replacement. Grass beta diversity did not show significant changes among burning and mowing treatments until 1988, 38 years after the start of the experiment. Of the 17 species that showed important changes post-1988, eight species changed in all censuses. Thus, changes in grassland composition took a long time to manifest themselves, but there were consistent changes once they occurred. Using a classic approach, beta regression showed that there was a significant effect of mowing but no significant effect of fire return intervals or time of burning. In a multivariate approach, using dbRDA, the beta regression results were affirmed: there were changes in beta diversity that demonstrated the primary effects of summer mowing frequency caused by a decline in species richness with two mows per annum, and secondarily of burn frequency. This is the first study to examine beta diversity in a subtropical grassland and we found that mowing was the primary driver of plant community change.