Abstract
Tropical deforestation is responsible for around one tenth of total anthropogenic carbon emissions, and tropical protected areas (PAs) that reduce deforestation can therefore play an important role in mitigating climate change and protecting biodiversity and ecosystem services. While the effectiveness of PAs in reducing deforestation has been estimated, the impact on global carbon emissions remains unquantified. Here we show that tropical PAs overall reduced deforestation carbon emissions by 4.88 Pg, or around 29%, between 2000 and 2012, when compared to expected rates of deforestation controlling for spatial variation in deforestation pressure. The largest contribution was from the tropical Americas (368.8 TgC y−1), followed by Asia (25.0 TgC y−1) and Africa (12.7 TgC y−1). Variation in PA effectiveness is largely driven by local factors affecting individual PAs, rather than designations assigned by governments.
Highlights
Tropical deforestation is responsible for around one tenth of total anthropogenic carbon emissions, and tropical protected areas (PAs) that reduce deforestation can play an important role in mitigating climate change and protecting biodiversity and ecosystem services
Using empirical relationships between canopy cover and above ground biomass (AGB), we estimated the influence of PAs on tropical forest carbon emissions resulting from reductions in deforestation rates
The fraction of variance in remaining canopy cover in non-PA areas explained by the Generalized Additive Models (GAMs) for the Americas was 95.2%, for Africa 99.0%, and for Asia 92.5%
Summary
Designations and controlling for spatial variation in deforestation pressure. Using empirical relationships between canopy cover and AGB, we estimated the influence of PAs on tropical forest carbon emissions resulting from reductions in deforestation rates. Deforestation rates within PA borders are significantly lower than outside[25,26], and they are especially important for forest conservation in developing countries[27]. Because of non-random PA locations, and because deforestation rate varies among countries due to political and socioeconomic factors[9], PA effectiveness may be overestimated[28], and so statistical models of forest loss in unprotected (non-PA) areas are commonly used to control for this bias[18,20,29]. We analysed the difference (rd) between observed and expected remaining forest cover in PAs to control for any biases in PA location. The expected remaining cover was determined from Generalized Additive Models (GAMs) of forest loss in non-PA regions. Positive rd indicates that remaining forest cover in PAs is greater than expected for a particular location, and negative rd indicates that remaining forest cover is less than expected
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
