Print PageDespite the importance of tropical forests to the global carbon cycle, ecological controls over landscape-level variation in live aboveground carbon density (ACD) in tropical forests are poorly understood. Here, we conducted a spatially comprehensive analysis of ACD variation for a mainland tropical forest—Barro Colorado Island, Panama (BCI)—and tested site factors that may control such variation. We mapped ACD over 98% of BCI (~ 1,500 ha) using airborne Light Detection and Ranging (LiDAR) and used multiple regression to examine controls over LiDAR-derived ACD, including slope angle, bedrock, soil texture, and forest age.
Results/Conclusions
LiDAR-derived ACD was well-correlated with ground-based measurements of ACD in Panamanian forests of various ages (r2 = 0.77, RMSE = 29 Mg C ha-1, P < 0.0001). Collectively, environmental variables explained 14% of the variation in ACD at 30-m resolution, and explained 33% at 100-m resolution. At all resolutions, slope (linked to underlying bedrock variation) was the strongest driving factor; standing carbon stocks were generally higher on steeper slopes, where erosion rates tend to exceed weathering rates, compared to gentle slopes, where weathering in place produces deep, oxic soils. This result suggests that physiography may be more important in controlling ACD variation in Neotropical forests than currently thought. Although BCI has been largely undisturbed by humans for a century, past land-use over approximately half of the island still influences ACD variation, with younger forests (80 - 130 years old) averaging ~ 15% less carbon storage than old-growth forests (> 400 years old). If other regions of relatively old tropical secondary forests also store less carbon aboveground than primary forests, the effects on the global carbon cycle could be substantial and difficult to detect with satellite monitoring.
