Lianas reduce tree growth, reproduction, and survival in tropical forests. Liana competition can be particularly intense in isolated forest fragments, where liana densities are high, and thus host tree infestation is common. Furthermore, lianas appear to grow particularly well during seasonal drought, when they may compete particularly intensely with trees. Few studies, however, have experimentally quantified the seasonal effects of liana competition on multiple tree species in tropical forests. We used a liana-removal experiment in a forest fragment in southeastern Brazil to test whether the effects of lianas on tree growth varies with season and tree species identity. We conducted monthly diameter measurements using dendrometer bands on 88 individuals of five tree species for 24 months. We found that lianas had a stronger negative effect on some tree species during the wet season compared to the dry season. Furthermore, lianas significantly reduced the diameter growth of two tree species but had no effect on the other three tree species. The strong negative effect of lianas on some trees, particularly during the wet season, indicates that the effect of lianas on trees varies both seasonally and with tree species identity.

Sheet1: Annualized monthly mean relative tree growth, obtained by multiplying the 3-month season data by 4, and the 6-month season data by 2.

Sheet2: Tree accumulated relative growth per year

Data was obtained using dendrometer bands installed at the breast height on 15 to 20 individuals of five tree species. The dendrometer bands were made of stainless steel ribbons (12.7 x 0.15 mm, width and thickness, respectively) with 0.2 mm precision. For half of the individuals of each tree species, we cut all lianas within a 10 m radius. We measured tree growth monthly for 24 months, starting in September 2012. Initially, we installed 10 dendrometer bands for each species, 5 for trees in the liana removal treatment (LR) and 5 for controls (NLR). In September 2013, we added dendrometer bands to 40 additional individuals of the five-target species. We quantified the change in stem diameter based on the initial diameter of each tree individual. The total mean relative growth was calculated as a percentage of the initial size for tree species with liana removals (LR) and controls (NLR).