by Stephen Johnson
Contrary to the popular image of deforestation as a clear-cut resulting in the absolute destruction of forests, most deforestation in the tropics takes place piecemeal. As forest is logged or converted to agriculture, patches are often left standing, resulting in a fragmented system of forest patches in a mosaic composed primarily of agriculture. Forests in tropical areas are increasingly highly fragmented, which has significant impacts on biodiversity and environmental conditions within the fragments. However, little is known about the impact of fragmentation on the ability of the forest to store carbon. In an ever-more-fragmented, ever-more-carbon-saturated world, understanding how these remnant forests sequester carbon is critical. Osuri et al. (2014) examined the relationship between rainfall, fragmentation, and carbon storage in fragments and continuous forest in the Western Ghats of southern India. Using linear mixed models and regressions, they found that fragmented forests stored almost 40% less carbon than continuous forests, as a result of trees that were shorter, had less dense wood, and were shorter for a given trunk diameter. Fragmented forests also relied more on large trees to store carbon, while displaying signs of transitioning to a community of less-dense, lower-carbon species.
In order to determine how forest structures differed between fragments and continuous forest, Osuri et al. measured each tree with a diameter greater than 10 centimeters at chest height. They recorded diameter, canopy height, identified the species, and found the density of the wood. From these measurements they calculated the height-to-diameter ratio, the amount of biomass, and thus the amount of carbon present. Using multiple linear regressions, they compared how these factors varied between fragmented and continuous sites, with both fragmentation and mean annual precipitation as possible explanatory factors. They further used linear models to see how the height-to-diameter ratio changed within a species or a community between sites, and how carbon storage was distributed between size classes.
Though the continuous sites received more rainfall than the fragmented sites, precipitation was only primarily responsible for increased tree density in the continuous sites. Other differences were better explained by fragmentation: trees in fragments were 25% shorter, occupied 22% less space per hectare, were 6% less dense, and stored 36% less carbon per hectare. Fragment trees were also shorter for a given diameter, and large trees did the bulk of the carbon storage in the fragments, while in continuous forest storage was more evenly spread across size classes.
Trees in fragments are subject to different pressures than continuous forest trees: fewer trees around means more light, but it also means that high winds aren’t stopped by the combined trees of the forest. In such conditions, growing tall is not only unnecessary, it can be damaging, as it makes the tree vulnerable to being blown-over by wind. Older trees present before fragmentation were still taller, but as the community ages, replacements can be expected to be shorter and smaller. In many fragmented areas, large old trees are lost quickly, brought down by wind and fire coming from open land like pasture adjacent to the fragment. Osuri et al. note that agroforestry systems like shade coffee plantations next to the fragments likely buffer them from these effects, preventing the early loss of old trees. However, while older trees prevent significant carbon losses in the present, as the community transitions to smaller, less dense individuals and species, high carbon losses may be inevitable. Ultimately, in order to mitigate this threat, active efforts will be needed to restore species composition and forest structure, and thus maintain carbon storage capacity.
Osuri, A. M., Kumar, V. S., & Sankaran, M., 2014. Altered Stand Structure and Tree Allometry Reduce Carbon Storage in Evergreen Forest Fragments in India’s Western Ghats. Forest Ecology and Management 329, 375–383. doi:10.1016/j.foreco.2014.01.039.