Landscape and site factors interact in the determination of secondary forest richness and diversity in northeastern Costa Rica
In tropical secondary forests, a wide range of variables drive community richness and diversity during succession. The combination of different land uses over time has resulted in secondary forest patches being present in a range of landscape context, each also with its unique environmental conditions. Many studies have emphasized the effects of age and site factors as determinants of the characteristics of secondary vegetation, so in this work we seek to understand the interactions between site conditions and landscape composition and structure. We postulate that this interaction affects succession differently at different recruitment stages (seedlings, saplings and trees). We examined the relative importance of these landscape and environmental variables for determining the diversity and richness of seedlings, saplings and trees of secondary rainforests and predict higher diversity and richness in seedlings. We selected twenty-five 0.25 ha plots, from 26 to 52 years post-abandonment, and measured and identified trees > 5 cm dbh, saplings 1-4.9 cm dbh and seedlings < 1 cm dbh. Soil and climatic variables were characterized, and landscape variables were assessed at different radii around plots for the period of 1986-2011 using GIS.
Saplings had higher species richness (S´) and Shannon diversity per plot, while the seedling layer had the lowest values, contrary to our predictions. Isolation (Euclidean distance) between secondary forest patches declined at 1 and 5 km radii from each plot from 1986 to 2011, while percentage forest cover increased at a 50 m radius around the plots, remained stable at 500 m and increasing at 1 and 5 km radii. Linear regressions showed that seedling S´ was related negatively to isolation at 5 km radius. On the other hand, sapling and tree S´ were related to soil and climatic variables but not to landscape context. No relationship was found between diversity and the predictor variables for any recruitment stage. There are no effects of stand age on any of these response variables. These results indicate that recruitment stages respond differentially to diverse landscape and environmental variables, highlighting the importance of landscape context and spatial scale in the understanding of tropical forest restoration.