Facing dire perspectives of biodiversity collapse, loss of ecosystem services, and the threat of climate change, society is urging for sound strategies aiming at the long-term maintenance of the world's ecosystems. Lack of governance on environmental issues, increasing population consumerism, and unfolding impacts on the natural resources are leading to unpredictable volatile effects on food production, water resources, and population health throughout the world (Watson et al., 2000; Stocker et al., 2013; Novais et al., 2016; Ceballos et al., 2017). Led by international agencies (e.g., UN), many countries are now working on global agreements and strategies to mitigate and adapt to the challenges imposed by global change as a response to the aggravation of the environmental crisis (Mace, 2014; Nimmo et al., 2015; Szinwelski et al., 2015).

Commitments within such strategies are the Bonn Challenge (http://www.bonnchallenge.org), whose aims include the restoration of 150 million hectares of the world's deforested and degraded land by 2020, and the New York Declaration on Forests, endorsed at the 2014 Climate Summit, that aims to restore 350 million hectares by 2030. While these initiatives are welcome and very relevant, there has been a hot debate about where and how to run restoration initiatives (see Veldman et al., 2015a,b; Fernandes et al., 2016a). For instance, the recently proposed afforestation of grasslands and savannas can bring more problems than solutions in terms of carbon sequestration (e.g. Berthrong et al., 2012; see also Deng et al., 2017). The introduction of trees where they have not been present historically can have disastrous impacts on local biodiversity and associated ecosystem services (Fernandes, 2016). Thus, these effects must be considered as restoration involves much more than simply planting trees. Ultimately, we must define the identity of the ecosystem in question in order to restore it with native species and bring back its biodiversity and ecosystem services (e.g., Overbeck et al., 2015; Kollmann et al., 2016).

In this paper, we advocate for a well conducted and scientifically oriented restoration of roadsides. We define roadsides as the environment alongside paved roads, which in Brazil generally comprises 15m on each side. These areas have been completely neglected as a target for restoration, despite their high potential. Roadsides occupy a considerable area but are often degraded, suffering from erosion, covered by invasive species, or even left in bare soil. In some cases, roadsides are illegally used for parking, house construction, and increasingly for food production. The latter can have different forms, depending on the socio-economic context of the region: they may be used for small-scale farming, often using slash and burn as the management strategy, or simply be incorporated into adjacent agricultural areas. The construction of paved roads and their management has many impacts on biodiversity and ecosystem services around the world, including wildlife kills, the spread of wildfires and invasive species, chemical and noise pollution, erosion, and watercourses siltation (see Ibisch et al., 2016 for a recent review). Even more importantly, roads are a major facilitator of (often uncontrolled) deforestation, especially in tropical countries (Laurance et al., 2009). In the Brazilian Amazon, for example, 94% of deforestation has occurred within 5.5km of a road (Berber et al., 2014) and represents a great threat to biodiversity (Vale et al., 2008).

In many countries, road construction and maintenance lack the most basic concerns with conservation and impact mitigation. In Brazil, for instance, although there are many state and federal regulations to reduce the environmental impact of roads, there is practically no law enforcement by governmental agencies (Fernandes, 2016). This has led to an unsurprisingly high number of these same agencies allowing the use of noxious invasive species in the revegetation of disturbed roadsides (Hilário et al., 2011) (Fig. 1).

Fig. 1.

Brazilian roadsides along paved highways are the scenario of many environmental problems. (a) Area along MG-10 left alone for many years and later ill-restored with noxious weeds by the Department of Roads of the state of Minas Gerais (DER-MG). (b) Spread of ruderal and invasive species into pristine vegetation from highway MG-10. (c) “Restoration” of highway MG-735 roadside in Minas Gerais (DER-MG) with exotic African grasses and other species. (d) Use of roadside along highway GO-118 (Brasilia-Alto Paraíso) to expand the plantations of soybean and bean. (e) Plantation of corn to the very edge of the RS-472 in Rio Grande do Sul state. (f) Plantation of soybean and plantation/invasion of pine trees along BR 290 in the Pampa, Rio Grande do Sul state.

(1.02MB).

If scientifically guided and properly restored, these areas could dramatically increase carbon sequestration as well as provide other ecosystem services. These include control and prevention of erosion and landslides, reduction of noise and chemical pollution, containment of invasive species, and conservation and enhancement of native biodiversity – for instance, by providing habitat and by connecting natural ecosystems. To illustrate and provide empirical support to our proposal, we take a hypothetical scenario where all state and federal paved roads in Brazil would have their roadsides properly restored, i.e. according to the native vegetation the roads cross through. The Brazilian case is particularly interesting because the country has set a national commitment to restore 12 million hectares of land by 2030, as part of the Bonn Challenge and the Paris Agreement of the United Nations Framework Convention on Climate Change (Brazil iNDC, 2015). We argue that if roadsides were properly restored instead of being left degraded and unprotected, they could help Brazil meet its carbon sequestration commitment and restoration goals, while additionally contributing to the provisioning of ecosystem services and to the mitigation of a number of negative effects of roads on the environment.

We used the complex and large paved road system of Brazil, which comprises more than 210,000km and spans six biomes. We focused on paved roads because they are known to be more environmentally harmful than unpaved roads (e.g. Nepstad et al., 2001), which generally have low traffic, with a lesser impact on roadsides and no intentional introduction of exotic species (Barbosa et al., 2010). Despite plans for expansion of the road network in Brazil and ongoing road maintenance work, this is currently the best scenario available to perform our calculations. We followed the Brazilian federal legislation on land allotment (Federal Law 6.766/1979), which determines that 15m along each roadside should be set aside. Thus, we calculated the area within the 15m on each side of paved roads from each biome. We excluded urban areas, which we assumed not to be available for restoration, even though very often the environmental damage done during construction can be considered here as well (e.g., Barbosa et al., 2010). Spatial data on paved roads were acquired from the Brazilian National Transport Infrastructure Department (DNIT, http://www.dnit.gov.br), biome data from the Brazilian Ministry of the Environment (MMA, http://www.mma.gov.br) and urban areas from the Brazilian Institute of Geography and Statistics (IBGE, http://www.ibge.gov.br).

We do acknowledge that there are some variables that could be estimated which would make our model more realistic. However, the main aim of this exercise is to provide the first insights into the potential for roadside restoration to enhance environmental quality and carbon sequestration in these areas. We are aware, for instance, that there might be practical or technical reasons why restoring to the immediate edge of the road may not be feasible everywhere. For example, tree roots could damage the pavement, while leaves and debris could prevent roads from efficiently draining during storms, or could fall on vehicles. Also, some locations will not be available for restoration due to intersections or other infrastructures, or because they have been illegally occupied. In addition, traffic signs or road signs erected at the side of roads must be visible to give instructions or provide information to road users. On the other hand, larger roads often contain median strips, usually with herbaceous or shrub vegetation that could also have some potential for restoration actions. In addition, there may be public land adjacent to the roadsides that would be available to be included in the planning. In virtually all road construction in Brazil, vegetation removal along roads is larger than the expected, patios are created for construction equipment, and sand, gravel, or slate are excavated from the ground and used in the road pavement. Here, however, we simplify and assume a scenario where roadsides could be restored throughout their entire width. Given the scale we are working at, it would be difficult to account for potential losses and gains in area, due to specific constraints, simply because data are not available in such a fine resolution for the whole country

To calculate the area of roadsides that can be restored inside each biome we used the package raster v2.5-8 (Hijmans and van Etten, 2012) implemented in R (R Development Core Team, 2010). We calculated the amount of carbon that can be sequestered by properly restored roadsides based on previous estimates made for the native vegetation of each of the six Brazilian biomes (see Milton et al., 2015). We used the estimated values of 150tCha−1 for forest environments for the Amazon and Atlantic Rain Forest biomes (Malhi and Grace, 2000; Amundson, 2001; Baker et al., 2004; Chave et al., 2008; Lewis et al., 2009), 50tCha−1 for savanna/grassland ecosystems as the Cerrado, Pampa, and Pantanal biomes (Watson et al., 2000; Grace et al., 2006; Ward et al., 2014), and 15tCha−1 for the xeric Caatinga biome (Amundson, 2001; Grace, 2004). In addition, although Brazil has not yet regulated its carbon market, we estimated the monetary value of carbon that could be traded, considering the unit values adopted by the main climate finance initiative in Brazil (the Amazon Fund), and the minimum and maximum carbon prices of existing initiatives in the world: US$1/tCO2e and US$131/tCO2e, respectively (World Bank et al., 2016). In this calculation, the C was converted to CO2, since the biophysical value sequestered is in the first unit and the market value in the second unit. This transformation was performed using the ratio of molecular weights (44/12). We also calculated, based on estimates by Benini and Adeodato (2017), the restoration costs for each biome, using a currency exchange of 1 Brazilian Real to 0.30 US Dollars. As restoration costs vary depending on site conditions and restoration technique used, we used the minimum and maximum cost associated with natural regeneration under good environmental conditions and active restoration under poor environmental conditions, respectively: US$54.00 to US$5247.60ha−1 for the Amazon, US$55.50 to US$6381.30ha−1 for the Atlantic Forest, US$54.00 to US$6683.70ha−1 for Cerrado, US$54.30 to US$5984.40ha−1 for the Caatinga, US$94.50 to US$8547.60ha−1 for Pantanal and US$54.30 to US$7755.00ha−1 for the Pampa. In theory, we would have to subtract current costs of maintenance and management of roadside area from this value, but these data are not available. We are aware that our calculations are based on relatively coarse estimates regarding: (1) total area to be restored, (2) carbon sequestration potential, (3) carbon prices, and (4) restoration costs. On the other hand, the carbon estimates we use are most likely quite conservative. For example, for soil stocks, estimates exist for some regions that exceed the values we used (e.g. Paiva and Faria, 2007 for Cerrado; Tornquist et al., 2009 for Pampa). This should offset possible uncertainties in the other parameters used in the calculations, or limitations for restoration due to technical issues.

The 30m (15m on each side) alongside federal and state paved highways in Brazil represent an area of ∼585,600ha. Of this, approximately 3.3% (19,500ha) are within urban areas. Excluding these urban areas, a total of 566,100ha would potentially be available for restoration. This corresponds to only ∼4.7% of Brazil's commitment under the Paris Agreement but could sequester 55.3 million tons of carbon if properly restored (Table 1, UNFCCC, 2015). The greatest potential is in the Atlantic Forest biome, which combines the greatest road density (40% of the total area to be restored) with the highest carbon stock potential (Fig. 2), followed by the Cerrado and the Caatinga (∼30% and ∼17%, respectively). Our crude estimate is that the negotiated amount of carbon could range between US$ 202 million and US$ 26.5 billion. The estimated cost of restoration ranged from US$ 32 million to US$ 3.7 billion (Table 1) and, therefore, the choice of restoration strategy can be crucial for cost-effectiveness, especially under low carbon market values. Considering the values adopted by the Amazon Fund, the negotiated value could reach up to US$ 1 billion.

Fig. 2.

Distribution of paved state and federal roads within Brazilian biomes: Amazonia forest, Caatinga xeric shrubland, Cerrado savanna, Pantanal wetland, Atlantic Forest, and Pampa grassland.

(0.63MB).

In Brazil, there is the opportunity to, with a single measure – roadside restoration – restore 566,100ha of currently underused land, increasing carbon sequestration in a magnitude that would be enough to offset one-fourth of the country's annual emissions from vehicles (MMA, 2014), contributing to carbon stock increases at a regional scale. The restoration costs, despite significantly overlapping with potential negotiated carbon price, can be at least partly offset if the carbon stocks were to be negotiated.

The authors declare no conflicts of interest.