Secondary forests comprise an increasing area of the tropics and play an important role in global carbon cycling. We compare above-ground biomass accumulation of both planted and naturally regenerating trees, as well as C in the top soil layer, in three restoration treatments replicated at 14, six to eight year old restoration sites in southern Costa Rica. Restoration strategies include: control (no planting), planting tree islands, and conventional, mixed-species tree plantations. We evaluate the importance of past land-use, soil nutrients, understory cover, and surrounding forest cover in explaining variation in above-ground biomass accumulation (ABA) rate across sites. Total ABA and planted tree ABA rate were highest in plantations, intermediate in islands, and lowest in control treatments, whereas ABA rate of naturally regenerating trees did not differ across treatments. Most ABA in plantations (89%) and islands (70%) was due to growth of planted trees. Soil carbon did not change significantly over the time period of the study in any treatment. The majority of across-site variation in both total and planted tree ABA rate was explained by duration of prior pasture use. Tree growth in the first two years after planting explained approximately two-thirds of the variation in ABA rate after 6–8 years. Soil nutrient concentrations explained relatively little of the variation in planted or naturally recruiting ABA rate. Our results show that planting trees substantially increases biomass accumulation during the first several years of forest recovery in former agricultural lands and that past-land use has a strong effect on the rate of biomass accumulation. Planting tree islands is a cost-effective strategy for increasing ABA and creating more heterogeneous habitat conditions than tree plantations. We recommend small scale planting trials to quickly assess potential biomass accumulation and prioritize sites for ecosystem service payments for carbon sequestration.

Data were collected in May-July of each year.  Trees were recorded as dead (0), alive (1), or cut (c); cut seedlings were accidentally cut during clearing of ruderal vegetation or trails.  Tree height was measured to the nearest centimeter using a measuring tape or measuring poles (2004-2007) and a laser range finder (2008-2012).  DBH was measured to the nearest mm using a DBH tape.  Tree DBH and cover were measured on all living trees in island treatments and on randomly selected trees in the plantation treatment. Canopy cover was measured by measuring the distance of two perpendicular lines to the nearest 5 cm and using the formula for an ellipse (see Holl et al. 2011).