We established eight subplots per plot. These subplots consisted of circles with a 20-m radius in densely vegetated habitats (savanna and all forest habitats) and 35.5 m × 35.5 m squares at Helichrysum plots, covering the same sampling area in all habitat types. We quantified vertical vegetation heterogeneity, canopy height and canopy closure on every subplot in the warm dry season (December to March) and in the cold dry season (July to October). To obtain vertical vegetation heterogeneity, we first estimated the vegetation cover in layers at 0, 1, 2, 4, 8, 16, 32 and 64 m above ground and then calculated the Shannon–Wiener diversity index across these eight strata. Canopy height was measured with a laser rangefinder as maximum canopy height above ground. Canopy closure was measured as the mean percentage of closed cells from four spherical canopy densitometer readings taken at the centre of each subplot in the four cardinal directions.

SP6 recorded several variables related to habitat structure on 8 subplots per plot. However, only 4 of these subplots (at two plots only 3) are located within the actual 50mx50m plot, while the remaining 4 subplots (at two plots 5) are located in the same habitat type and as close to the plot as possible. This dataset contains only information derived from subplots which are located within the actual plot, i.e. from 3-4 subplots per plot.Methods on:Closure of canopy area = measured with a canopy densiometer at X subplots with 4 measurements each (north, east, south, west)Vegetation heterogeneity, vertical = vegetation cover is estimated at 0, 1, 2, 4, 8, 16 and 32 m at X subplots per plot within a radius of 20 m; the Shannon-Diversity-Index is calculated over these estimates for each subplot; the minimum Shannon-Diversity-Index of all subplots is givenCanopy height = canopy height is measured with a laser range finder or estimated (if canopy height was smaller than minimum measuring distance for the LRF)