by Stephen Johnson
Habitat loss is the primary threat to the survival of most tropical biodiversity. Typically, this habitat loss is driven by deforestation for agricultural use. However, deforested landscapes are rarely homogenous fields with low diversity; most often, forest fragments are left embedded in a matrix of varying types of agriculture, from open field monocultures, to pastures and forest-mimicking shaded plantations. The process of fragmentation has a significant negative effect on the biodiversity present in the area; however, fragments are often able to support a variety of species, as are some types of agriculture, such as agroforestry. Less is known about the capacity of such landscapes to sequester and store carbon. What little has been done has focused on carbon in agroforestry systems, with promising, though mixed, results. These conflicting results have led to a debate among conservationists about the best way to protect natural ecosystems. Some argue that agriculture should be intensified on smaller amounts of land, allowing large, continuous blocks of land to be left untouched (land-sparing). Others advocate incorporating natural elements (trees, shrubs, etc.) into agriculture, to make it more hospitable to biodiversity (land-sharing). However, most arguments one way or another have been theoretical, with little supporting empirical evidence. What evidence does exist is typically spatially limited, with few landscape-level studies being performed. To help resolve this debate, Gilroy et al. (2014) present the first integrated study of both multi-taxa biodiversity and ecosystem carbon storage capacity. They quantified carbon contained in forest fragments compared to the same size plots in continuous forest, and surveyed birds and dung beetles in each habitat type. Based on the data collected, they used land-allocation models to predict how biodiversity and carbon would respond under various forest cover scenarios. They found that the greatest amount of biodiversity and carbon could be protected in landscapes in which agriculture was intensified on a smaller amount of land, with large continuous forest blocks left unharmed.
To determine the amount of carbon present in the landscape, Gilroy et al. established plots in forest fragments and advanced secondary forest, as well as control plots in primary forest. They measured the diameter and density of the trees in each plot, as well as the mass of leaf litter. They also estimated root biomass through root-to-shoot ratio data previously established in the literature. Through allometric measurements in the literature, they used the density and diameter to determine the amount of carbon in each type of ecosystem. Gilroy et al. also established larger plots to sample the biodiversity. They performed point counts at three points in each plot to visually assess the bird species present. They also placed baited pitfall traps to capture dung beetles, which are considered an indicator taxon for other biological groups. The variance in carbon stocks was assessed using analysis of covariance tests, and biodiversity was examined using multi-dimensional scaling algorithms and logit-link functions, which assess the community structure and probability of occurrence at each site. The researchers then used these data to construct a set of landscape models. Holding production area constant, they examined how various landscape configurations would affect carbon and biodiversity present. The land-sparing scenarios involved continuous blocks of primary forest or blocks of secondary forest. The land-sharing scenarios they used involved productive areas sprinkled with forest fragments. After running the models many times, they determined which configurations resulted in the highest values. They found that carbon storage was significantly lower in forest fragments compared to primary forest. Advanced secondary forest had lower carbon stocks than the primary forest, but contained more than forest fragments. Biodiversity occurrence followed the same pattern as carbon storage. The models found that land-sparing scenarios resulted in the highest carbon storage and biodiversity protection, with little difference between primary forest and secondary forest blocks.
Land-sparing versus land-sharing has been a hotly debated topic in the conservationist community. However, little empirical evidence has been offered to support one side or the other. Gilroy et al. produced one of the first landscape-level models calibrated with field data to examine broad patterns in biodiversity and ecosystem services. The results indicate that forest fragments are of relatively little value in productive systems. Rather than attempt to maintain fragments, conservationists should look to improve future outcomes by promoting increased yields in intensified plots, while leaving large continuous blocks of forest untouched. Such systems can be successful both in previously undisturbed systems in which new agriculture is forming, and in old agricultural landscapes where forest blocks will be composed of secondary regrowth.
Gilroy, J.J., Woodcock, P., Edwards, F.A., Wheeler, C., Medina Uribe, C.A., Haugaasen, T., Edwards, D.P., 2014. Optimizing carbon storage and biodiversity protection in tropical agricultural landscapes. Glob. Change Biol. 20, 2162–2172. doi:10.1111/gcb.12482