Long‐term subtropical grassland plots take a long time to change: Replacement is more important than richness differences for beta diversity

We studied β diversity of grasses in a subtropical grassland over 60 years in South Africa. We examined the effects of burning and mowing on 132 large plots. We sought to determine the effects of burning and mowing, and mowing frequency, on the replacement of species and the species richness. We conducted the study at Ukulinga, research farm of the University of KwaZulu‐Natal, Pietermaritzburg, South Africa (29°24′E, 30°24′S) from 1950–2010. Plots were burned annually, biennially, triennially, and a control (unburned). Plots were mowed in spring, late summer, spring plus late summer, and a control (unmowed). We calculated β diversity, with a focus on replacement and richness differences. We also used distance‐based redundancy analyses to examine the relative effects of replacement and richness differences on mowing and burning. We used beta regressions to test for the effect of soil depth and its interactions with mowing and burning. There was no significant change in grass beta diversity until 1995. Thereafter, there were changes in β diversity that demonstrated the primary effects of summer mowing frequency. There was no significant effect of richness differences but a strong effect of replacement post‐1995. There was a significant interaction between mowing frequency and soil depth in one of the analyses. Changes in grassland composition took a long time to manifest themselves and were unapparent prior to 1988. However, there was a change in sampling strategy prior to 1988, from point hits to nearest plants, that may also have influenced the rates of changes in replacement and richness differences. Using β‐diversity indices, we found that mowing was more important than burning that burning frequency was unimportant, and there was a significant interaction effect between mowing and soil depth in one of the analyses.