|Medium-term dynamics of tree species composition in response to silvicultural intervention intensities in a tropical rain forest
AL de Avila, AR Ruschel, JOP de Carvalho, L Mazzei, JNM Silva, JC Lopes, MM Araujo, CF Dormann, J Bauhus
Appendices. Supplementary Results and Figures
Figure A1. Sample based rarefaction and extrapolation to 80 sampling units in the initial census (1981 for the logged plots and 1983 for the control area) and considering all individuals with DBH ≥ 5 cm.
Figure A2. Time series responses of poles (a) (5 cm ≤ DBH < 10 cm) and trees (b) (DBH ≥ 10 cm) for each treatment (C: Control, L: Logging only, LLTI: Logging and light thinning intensity, LMTI: Logging and medium thinning intensity, LHTI: Logging and high thinning intensity). The time-constrained clustering was applied by using multivariate regression trees (MRT). Clusters that were identified in the analysis are indicated by i, ii and iii. Interventions of logging (1982) and thinning (1993-1994) are shown by downward arrows above the time axis. Details of the analyses are given in Tables 2 and 3.
Figure A3. Non-metric multidimensional scaling diagram of 41 plots grouped by treatments for the first (1981) and the last census (2012) for poles (a and b) and trees (c and d). The analysis of variance (function adonis from vegan in R) using distance matrices at the plot level indicated that treatments spatially differed in species composition in all measurements, except for trees in 1981 (c). The control treatment was not included in 1981 (a and c) because the first assessment of the control plots occurred in 1983. Ordination stress and P‑values from the analysis of variance are presented in each graph.