Most of the tropical forests of Central Africa are characterised by a remarkable abundance of light-demanding canopy species. A popular hypothesis is that these forests are still recovering from the intense slash-and-burn farming activities that ended abruptly in the 19th century with the arrival of the colonists. Today, it is assumed that the zones occupied by crop fields until the 19th century are covered by forests dominated by light-demanding species. However, this hypothesis of human disturbance has not yet been sufficiently tested using spatial distribution. So, using the 'Kernel Density Estimation' (KDE) tool in the SAGA GIS software, we mapped the density distribution of light-demanding species, subdivided into 3 tree functional types, along transects in the Moni river catchment. We also produced a similar map for a particular shade-tolerant species, 'Gilbertiodendron dewevrei'. The species were then divided into the following groups, known as 'functional tree types': LLP=Long-Lived Pioneer, NPLD=Non-Pioneer Light Demanding, SLP=Short-Lived Pioneer, and STS=Shade-Tolerant Species. At the end of this analysis, a density distribution map of the species of each tree functional type was produced. This map highlights the pockets (zones with a high density relative to the study site average) of tree functional types. For each type of pocket, we selected the pockets with a high density of trees of the group concerned and which were not on the edge between the forest and village crops or fallow land. This is how the location of the permanent plots was determined. Next, we installed a total of ten full forest inventory plots (1 ha each) inside and outside the pockets located by the KDE analysis along the Moni River transects. More specifically, we installed one plot in a pocket of short-lived pioneers (SLP-01), three plots in pockets of long-lived pioneers (LLP-01 to -03), two plots in pockets of NPLD (NPLD-01 and -02), two plots in pockets of the shade-tolerant species Gilbertiodendron dewevrei (GIL-01 and -02) and finally two plots were located in a mixed old-growth forest outside the pockets (MIX-01 and -02). These plots were established (1) for long-term monitoring of biodiversity and forest dynamics; and (2) to see if there is a difference in terms of species composition and abundance of light demanders between the forest inside the pockets and that outside the pockets.

The plots were installed in the Moni catchment along the transects. They are rectangular, 50 m wide (25 m on either side of the transect baseline), and 200 m long, centered on the transect baseline. Each plot was well delimited and divided into 16 square sub-plots (25 m × 25 m). All trees of all species were surveyed using the international standardised inventory protocol for tropical forests (RAINFOR) (Dallmeier 1992; Condit 1998; Phillips et al. 2009; ForestPlots.net et al. 2021). All trees with at least 10 cm of diameter at breast height (DBH) were identified, measured (DBH), and mapped (X and Y coordinates). The points of measurement (POM) of the diameter were marked with paint. For trees with buttresses or any deformations at 1.30 m, the POM was taken about 50 cm above the deformation.