A study reveals that human-driven disturbances in previously undisturbed Amazon rainforest can cause biodiversity losses as severe as those of deforestation. Urgent policy interventions are needed to preserve forest quality. See Letter p.144
As we enter the Anthropocene, a proposed geological epoch shaped by human activity, mankind is driving a global biodiversity extinction crisis1. The conversion of forest to agricultural land is widely considered to be the leading cause of this crisis, especially in the hyperdiverse tropics2, so avoiding deforestation is the predominant strategy for biodiversity conservation3. On page 144, Barlow et al.4 present a landmark field study of Amazonian biodiversity in which they challenge the adequacy of this strategy by demonstrating the striking magnitude of several types of human-associated forest disturbance that are less immediately visible than deforestation.
Many studies have identified the negative effects on biodiversity of individual kinds of disturbance in tropical forests. These include the hunting of large animals5, the selective logging of large, marketable trees6, forest fires7 and the creation of new edges to primary forests (those forests that have never been fully cleared) that, owing to deforestation, are buffeted by the hotter, drier and windier conditions found on adjacent farmland8 (Fig. 1). However, by focusing on only one form of disturbance, such studies may have overlooked much greater conservation losses from the combined effects of forest disturbances.
Barlow and colleagues conducted biodiversity censuses across multiple landscapes and then developed a computational method for evaluating conservation losses (termed the 'conservation value deficit', a numerical value calculated by assessing biodiversity in disturbed primary forests relative to undisturbed ones). This enabled the authors to quantify the direct negative effects of deforestation and those resulting from the plethora of other types of forest disturbance.
The authors assembled an impressive data set collected across a large region of the Brazilian Amazon. They sampled 36 catchments (each 32–61 square kilometres in size) containing small rivers, spanning Belém and Tapajós, two major regions of endemism — areas that contain species that are found nowhere else. Each sample catchment varied in the degree of disturbance: 5 were entirely deforested, whereas the other 31 contained varying amounts of remnant forest, including undisturbed primary forests, and primary forests that had been disturbed by hunting, selective logging or fires, or isolated by surrounding farmland. Sampling the biodiversity across each catchment, Barlow and colleagues encountered a breathtaking total of 1,538 plant species, 460 bird species and 156 dung beetle species.
Their findings make for uncomfortable reading. Even catchments that retained 80% of their forest cover — the maximum that can be required of Amazonian estates under Brazil's Forest Code legislation — lost between 39% and 54% of their conservation value, and about half of this loss is due to disturbance within the remaining forest areas, rather than the losses from conversion to farmland. By extrapolating these disturbance-driven losses across the state of Pará, which represents 25% of the entire Brazilian Amazon, the authors found that conservation losses from disturbance are equivalent to the losses that would result from deforesting 92,000–139,000 km2 of primary forest — an area roughly equivalent to the size of Greece.
They also found that species with higher conservation importance were more negatively affected by forest disturbance. Bird species that were restricted to small regions (with small global range sizes) fared worse than those with larger distributions, suggesting that forest disturbance is homogenizing biodiversity across regions9. Tree species with high wood density declined more than those with softer wood, degrading the ability of disturbed forests to store carbon, thereby driving climate change10.
Barlow and colleagues' study has three limitations worthy of comment, each of which probably means that their assessment of biodiversity loss from disturbance is conservative. First, the extensive human activity in the study regions means that many of their undisturbed forest plots could have suffered from low-level or historical disturbances that were not detected. Although the authors attempted to correct for this potential bias in their analytical approach, if they had sampled truly remote and undisturbed forests instead, densities of the most sensitive species would probably have been higher, and the conservation value deficit would thus have been even more severe in disturbed forests.
Second, the authors' results from two areas with species endemism were extrapolated to a quarter of the Brazilian Amazon, which spans an additional three unstudied areas of endemism. Their analysis therefore fails to fully capture the species differences between regions9. Small-ranged species will probably be more negatively affected than those with larger distributions.
Finally, mammals were not studied, yet they have crucial roles in maintaining healthy ecosystems10. Mammals would be expected to suffer at least as profoundly as the sampled taxa, owing to the severity of hunting in accessible forests located among farmlands or near roads11. More generally, whether findings from the Brazilian Amazon can be extrapolated to other tropical regions is unknown. Barlow et al. have given scientists the impetus and methodological tools to make such assessments, making similar studies a research frontier elsewhere in the tropics.
“The authors’ results underscore the need for a step-change in forest governance.”
This research challenges the governance of tropical forests, and hence the protection of the myriad conservation and ecosystem benefits they provide, which sustain some of the most biodiverse ecosystems on the planet and the livelihoods of millions of people. As the authors acknowledge, avoiding deforestation must remain a key tenet of conservation strategies3. However, their results underscore the need for a step-change in forest governance, with much greater emphasis on the ecological health of retained forest1. In particular, Barlow and colleagues have shown that policies, such as Brazil's Forest Code, that set targets for forest cover without also setting requirements for forest quality are insufficient to prevent substantial conservation losses, and are a slippery slope to greatly impoverished ecosystems10.
To remedy this, agricultural landscapes must be better designed to promote the protection of larger and less-isolated forest blocks. More-stringent regulation and enforcement are needed, both of fire use in agriculture (which frequently spills into forests) and of selective logging — together with clear economic benefits for more sustainably managed agriculture and sustainable logging6. These all require coordination between landowners, policymakers and conservationists across entire landscapes and regions.
In many tropical forest regions, disturbed primary forests that have seen big biodiversity losses are especially valuable for conservation, because large undisturbed forests are rare or completely lacking. Within such regions, these results underscore the necessity of assisting the recovery of disturbed forests or of taking unproductive farmland out of use to restore forest coverage and connectivity. Although the biodiversity extinction crisis could be even worse than currently recognized, by embracing better management strategies, the solutions are still within our grasp.Footnote 1
Lewis, S. L., Edwards, D. P. & Galbraith, D. Science 349, 827–832 (2015).
Newbold, T. et al. Nature 520, 45–50 (2015).
Nepstad, D. et al. Science 326, 1350–1351 (2009).
Barlow, J. et al. Nature 535, 144–147 (2016).
Peres, C. A. Conserv. Biol. 14, 240–253 (2000).
Edwards, D. P., Tobias, J. A., Sheil, D., Meijaard, E. & Laurance, W. F. Trends Ecol. Evol. 29, 511–520 (2014).
Barlow, J. & Peres, C. A. Ecol. Appl. 14, 1358–1373 (2004).
Ferraz, G. et al. Science 315, 238–241 (2007).
Socolar, J. B., Gilroy, J. J., Kunin, W. E. & Edwards, D. P. Trends Ecol. Evol. 31, 67–80 (2016).
Peres, C. A., Emilio, T., Schietti, J., Desmoulière, S. J. M. & Levi, T. Proc. Natl Acad. Sci. USA 113, 892–897 (2016).
Laurance, W. F., Goosem, M. & Laurance, S. G. W. Trends Ecol. Evol. 24, 659–669 (2009).
Related links in Nature Research
About this article
Forest Ecology and Management (2020)
Conservation value of moist evergreen Afromontane forest sites with different management and history in southwestern Ethiopia
Biological Conservation (2019)
Rapid diversity and structure degradation over time through continued coffee cultivation in remnant Ethiopian Afromontane forests
Biological Conservation (2019)
An optimization model for the planning of agroecosystems: Trading off socio-economic feasibility and biodiversity
Ecological Engineering (2018)
Nature Ecology & Evolution (2018)