Driving factors of forest growth: a reply to Ferry <i>et al.</i> (2012)

Abstract

Summary 1. In a recent paper, we analysed the effects of climate, soil and logging disturbance on tree and forest growth (Toledo et al. 2011a). We took advantage of one of the largest data sets in the Neotropics, consisting of 165 1‐ha plots and over 62 000 trees distributed over an area of c. 160 000 km 2 , across large environmental gradients in lowland Bolivia. The main findings were that climate was the strongest driver of spatial variation in tree growth, whereas soils had only a modest effect on growth and that the effect of logging disappeared after a few years. 2. Ferry et al. (2012) suggest that we underestimated the disturbance effects on growth because of a supposedly wrong coding of Time After Logging (TAL) for unlogged plots. Although we have good biological reasons why we coded TAL like we did, we checked Ferry et al. ’s suggestions for recoding and found no differences in variables that significantly explained tree and forest growth. We agree, however, that for future research, it is important to go beyond simple descriptors such as time after logging and basal area logged, to better describe the variation in logging impact found in areas under forest management. 3. Ferry et al. claim that we did not define basal area growth properly. We believe this is a semantic issue, as we clearly defined basal area growth as the net change in basal area. This net basal area change in Bolivian forests is indeed relatively high compared to other studies, which may be attributed to the higher soil fertility and biogeographic differences in species composition and their traits. 4. Synthesis . Many apparent discrepancies in the ecological literature arise because tropical forest ecologists tend to see the world from the perspective of their ‘own’ forest (despite clear biogeographic differences) and try to capture the same ecological processes using different variables and measurement protocols. To advance our understanding and go beyond single‐case studies, we need to assemble large databases, quantify forest dynamics and disturbances in similar ways, be aware of differences among forests and analyse environmental dose–response curves.