Brazil: Restoration of a Restinga Sandy Coastal Plain in Rio de Janeiro


Brazil’s population is overwhelmingly centered on the coast, and the resultant anthropogenic pressures have brought severe degradation and widespread habitat loss to coastal ecosystems. Restingas, or sandy coastal plains, are one such ecosystem that has been greatly impacted by encroaching human populations, and this study was undertaken to restore natural vegetation on a remnant restinga in Rio de Janeiro. Although shrubs and trees comprise the majority of a natural restinga plant community, exotic grasses had predominated on the study site and drastically reduced species richness. The project, therefore, began with the removal of grass cover and then proceeded with the introduction of 17 nursery-reared native shrub and tree species that were monitored for growth and survival over a period of 2 years. In spite of the adverse growing conditions typically presented by the nutrient-poor sandy soils characteristic of restingas, 70% of the introduced species showed high rates of survival and considerable growth at the end of the monitoring period. Results of the study suggest that the plantation of nursery-reared native species is a viable option for restoring natural vegetation and ecosystem function to these degraded restingas.

Quick Facts

Project Location:
Av. Nilo Peçanha, 26 - Centro, Rio de Janeiro - RJ, 20020-100, Brazil, -22.90608055634014, -43.17413485634768

Geographic Region:
Latin America

Country or Territory:


Coastal, Dune & Upland

Area being restored:
Not reported

Organization Type:
University / Academic Institution


Project Stage:

Start Date:

End Date:

Primary Causes of Degradation

Deforestation, Invasive Species (native or non-native pests, pathogens or plants), Urbanization, Transportation & Industry

Degradation Description

Restingas of the Brazilian littoral have been affected by human impacts for about 8,000 years (Kneip 1987). Human occupation has recently increased to such an extent that there is a need for conservation of remnant patches and restoration of degraded areas. Restinga vegetation has suffered considerable habitat destruction because most Brazilian big cities are situated on the coast. However, these systems are often not treated as a conservation priority because they have few endemic species (Barbosa et al. 2004).

The city of Rio de Janeiro once had a vegetation complex consisting of Atlantic forest, mangroves, and restingas. The restinga vegetation covered an extensive area on the west side of the city; however, only 0.63% (770.65 ha) of the total area of the municipality is still occupied by this vegetation type, which represents a loss of 30% of the restinga cover in the period from 1984 to 1999 (PCRJ 2000a). Despite this massive reduction in area, the restingas of Rio de Janeiro provide important habitat for endemic species of plants, insects, fish, and lizards, which are threatened with extinction (Vanzolini & Ab’Saber 1968; Araujo & Maciel 1998; PCRJ 2000b).

The restinga targeted for restoration under the project had been subjected to deforestation and fire in the past in order to clear the area for a housing project. Invasive plants occupied the area before the onset of the experiment, such as the exotic grasses Panicum maximum Jacq., Imperata brasiliensis Trin., and Melinis minutiflora P. Beauv, which are locally common throughout urban areas, and the exotic tree Casuarina equisetifolia L. (Casuarinaceae), which is frequently planted in coastal zones in Brazil as a wind shield.

Reference Ecosystem Description

Restinga is both a geomorphological and a botanical term. It applies equally to the sandy plains dating from the Quaternary, mostly from the Holocene, and to the vegetation covering these plains. Restingas differ from dunes in that they are sand marine deposits, whereas dunes are wind deposits. The restinga vegetation is a mosaic of plant communities ranging from creeping types to open scrubs and even forests (Lacerda et al. 1993; Martin et al. 1993).

Most plants and animals inhabiting the restingas of Rio de Janeiro originated in the neighboring Atlantic rainforest and successfully migrated and colonized the geologically younger sandy plains (Rizzini 1979; Araujo 2000). The restinga ecosystem is therefore unique because it comprises a pool of species with high ecological plasticity, since, despite their rainforest origin, they colonized, survive and grow in the dry, resource-poor restingas. This characteristic may be of key relevance in a global change scenario (Scarano 2002).

Restinga vegetation is often species-rich, although plants are subjected to the various constraints imposed by drought, nutrient-poor sandy substrate, wind, salinity, and high soil and air temperatures (Reinert et al. 1997). Paradoxically, it has been shown that few restinga plants are capable of establishing via seeds on bare sand and, therefore, the structure and function of open restinga vegetation relies on a few pioneer nurse-plants that facilitate the entry and establishment of a number of other species (Scarano 2002; Dias et al. 2005).

Remnants of the native vegetation at the project site–such as Cupania emarginata Cambess. (Sapindaceae), Eugenia ovalifolia Cambess. (Myrtaceae), E. rotundifolia Casar. (Myrtaceae), Inga maritima Benth. (Leguminosae – Mimosoideae), Myrciaria floribunda (H. West ex Willd.) Legrand (Myrtaceae), and Ocotea sp. (Lauraceae)–provide a strong indication that the original plant community prior to disturbance was that described as a “˜”˜Myrtaceae thicket” (Lacerda et al. 1993).

Project Goals

To provide information on the best choices of trees and shrubs for restoring this and other restinga areas in the region.


The project does not have a monitoring plan.

Description of Project Activities:
At the outset of the project, individuals of C. equisetifolia were removed with a chainsaw, and invasive grasses were mechanically removed and buried circa 20 cm deep in the soil. The planting holes had dimensions of 30 x 30 x 30 cm and were separated from each other by 2.5 m along a given line; and lines were separated from each other also by 2.5 m. Between lines, holes were dug at a distance of 5 m from each other and at a distance of 1.25 m from the lines. Although native vegetation in such sites is a closed thicket, a uniform, spaced plant distribution pattern was chosen in an attempt to reduce the effect of competition between neighbors. The planting holes each received 20 L of an organic compound produced from urban garbage. Species were chosen based on two criteria: first, the occurrence of the species in the Myrtaceae thicket of the restingas on the west side of Rio de Janeiro (Araujo & Henriques 1984) and second, the seedling availability. Seventeen species of trees and shrubs were chosen belonging to 10 botanical families"”Anacardiaceae: Tapirira guianensis Aubl.; Bignoniaceae: Tabebuia chrysotricha (Mart. ex DC.) Standl.; Bombacaceae: Pseudobombax grandiflorum (Cav.) A. Robyns; Erythroxylaceae: Erythroxylum ovalifolium Peyr.; Euphorbiaceae: Pera glabrata Baill.; Leguminosae - Caesalpinioideae: Chamaecrista ensiformis (Vell.) H.S. Irwin & Barneby, Senna australis (Vell.) H.S. Irwin & Barneby, and S. pendula (Humb. & Bonpl. ex Willd.) H.S. Irwin & Barneby; Myrtaceae: Eugenia ovalifolia Cambess., E. rotundifolia Casar., E. sulcata Spring. ex Mart., E. uniflora L., and Myrcia cf. multiflora (Lam.) DC.; Ochnaceae: Ouratea cuspidata (A. St.-Hil.) Engl.; Rubiaceae: Tocoyena bullata Mart.; and Sapindaceae: Allophylus puberulus Radlk. and Cupania emarginata Cambess. During the rainy season between 23 and 30 March 1998, 4,700 seedlings belonging to the 17 chosen species were planted on the site. Once planted, seedlings were individually tied to bamboo sticks to protect them from wind damage and facilitate visibility, so as to avoid damage during maintenance activities. During the first 30 days after planting, seedlings were irrigated to field capacity on all days when no rain occurred. Invasive grasses were removed with a backpack brushcutter six times from March 1998 to June 1999. Logistics did not allow further control of invasive grasses. Surviving plants were counted 130 days after planting. Survival counts were undertaken at this stage to assess if the replacement of seedlings was necessary. To allow acclimation, monitoring was started 90 days after planting. A total of 337 plants were monitored: 20 individuals of each species, which were selected randomly and labeled with numbered aluminum tags, except for P. glabrata, which had only 17 surviving individuals.

Ecological Outcomes Achieved

Eliminate existing threats to the ecosystem:
Low mortality was observed after planting, confirming initial expectations. Only 5.2% of the 4,700 plants introduced had died 130 days after planting. Thus, no replacement of monitored plants was necessary. After 2 years, at the final monitoring, a survival rate of 81.9% was recorded for the 337 plants monitored, and 12 of the 17 species studied had survival rates higher than 80%. Almost 50% of the species introduced had 100% survival at the end of the experiment. Nine out of 17 species had some mortality during the 2 years of monitoring. Only one individual of each of Chamaecrista ensiformis, Tabebuia chrysotricha, and Tapirira guianensis died, and in all cases, death occurred in the summer, between October 1998 and February 1999. Erythroxylum ovalifolium's mortality was concentrated between June 1998 and February 1999. Mortality also occurred under the high summer temperatures for Eugenia ovalifolia and E. rotundifolia, always between October 1998 and April 1999, except for one E. ovalifolia plant that died during the following summer. Senna pendula showed a distinct pattern: mortality was zero until April 1999, but from then on, 40% of the plants died, and the remaining plants showed signs of senescence at the end of the experiment. High mortality was also found for E. sulcata, which showed a steady increase in the number of dead individuals from June 1998 to October 1998 (50%); and by the end of the experiment in March 2000, this species showed 65% mortality. Similarly, Myrcia cf. multiflora had 65% mortality by March 2000, with mortality always being higher during the summer periods. Nine of the 17 species studied (Allophylus puberulus, C. ensiformis, E. ovalifolium, E. ovalifolia, E. sulcata, E. uniflora, M. cf. multiflora, S. pendula, and T. chrysotricha) showed no significant increase in height. Individual negative height increments were common. No data were collected on wind, but wind is believed to be partly responsible for this pattern because shoots of many seedlings were broken, even though plants were tied to bamboo sticks. Height increase was largest for Pseudobombax grandiflorum, Cupania emarginata, S. australis, and Tocoyena bullata. Averaging the values of increment in height of the eight species that showed statistically significant positive values (C. emarginata, E. rotundifolia, Ouratea cuspidata, Pera glabrata, P. grandiflorum, S. australis, T. guianensis, and T. bullata) resulted in a mean annual relative growth rate (RGR) for height of 11.32 cm/yr. All species monitored showed a significant increase in basal diameter, except for M. cf. multiflora. Again, P. grandiflorum showed the highest increment in basal diameter, along with T. guianensis. Canopy area growth was highest for S. australis and lowest for C. emarginata. The latter, along with M. cf. multiflora, E. sulcata, and S. pendula, showed no significant increase in canopy area. Decreases in canopy area were usually due to temporary deciduousness or breaking of branches. Senna australis showed the best growth performance overall. The species canopy area growth was the highest, increasing on average 1.6 times per month. Pseudobombax grandiflorum had the second best performance due to high height and basal diameter increment, whereas M. cf. multiflora and E. sulcata had the worst performance. They had no significant growth in any of the parameters analyzed over the 2 years of the experiment.

Factors limiting recovery of the ecosystem:
Overall growth rates at the site were low, as indicated by the values obtained for RGR for height (11.32 cm/yr) of the eight species that showed significant positive growth values. Slow growth rates are most likely related to nutrient-poor soils, high light intensities, and the wind action common to all restingas (Araujo & Oliveira 1988; Scarano et al. 2001). It is likely that introduction of larger seedlings, the use of more fertilizers, and more frequent removal of invasive grasses would improve overall survival and growth, but under the conditions of the present experiment, the performances of the Myrtaceae Eugenia rotundifolia, E. ovalifolia, E. sulcata, and particularly of Myrcia cf. multiflora would not recommend their use for restoration projects. This was unexpected given that the natural vegetation that used to occupy the study site prior to damage is named Myrtaceae thicket. This low performance deserves further investigation, but it could be related to age and size of seedlings and initial planting season.

Socio-Economic & Community Outcomes Achieved

Key Lessons Learned

Senna australis, P. grandiflorum, C. ensiformis, and T. guianensis showed the best results, and their use is recommended in open restinga areas subjected to wind and high temperatures. Eugenia uniflora and T. bullata, C. emarginata, O. cuspidata, and P. glabrata showed excellent survival and an intermediate reasonable growth, as did E. ovalifolium, which had a 15% mortality rate, however. Allophylus puberulus and T. chrysotricha had reduced growth but high survival, and could also be used with relative success in restoration initiatives in similar restingas. The Myrtaceae, E. rotundifolia, E. ovalifolia, and E. sulcata, and particularly M. cf. multiflora, had a very poor overall performance.

Senna australis showed the highest UVI of the species studied. The rapid soil cover by the canopy of this species indicates that it might be useful to avoid or reduce grass invasion because grass invaders in the restingas, such as the ones removed prior to the start of the experiment, are typical sun plants. Moreover, it is likely that this species also plays an important role in nutrient cycling, due to an apparently high leaf turnover that results in a thick layer of dry leaves in its understory. However, the use of S. australis in restoration programs requires caution because this plant shows a profuse branch formation close to the soil level, which may mechanically hinder seedling establishment under and around its canopy. Pseudobombax grandiflorum had the second best performance, despite its small canopy cover and deciduousness, which are the disadvantages to take into consideration in areas where invasive grasses are present. In such cases, it would appear that the best use of this species would be in association with a species with higher potential for canopy cover.

Although only a small sample of the broad range of restinga species were assessed in this study, the results obtained provide an optimistic perspective for the possibility of successful restinga restoration, particularly from a biodiversity viewpoint. This example of the successful introduction of native woody plants in Brazilian restingas suggests that this type of approach may restore both biodiversity and ecosystem processes in tropical coastal habitats of Brazil and elsewhere. This may, in turn, prove useful for carbon sequestration as well.

Studies on plant colonization and succession in natural restinga areas have often suggested that restingas are a fragile habitat, where plant germination and establishment depend on specific nurse plants, such as bromeliad and Clusia sp. (see Scarano 2002, for review). This project shows that restoration practitioners can play ex situ the role played in situ by the nurse-plants–that is, grow restinga plants in nurseries until they can be safely transferred to the field (Eliason & Allen 1997). Future experiments using introduction of nurse-plants should cast further light on the adequacy of restoration strategies for the restingas.

Sources and Amounts of Funding

Funding and logistical support was provided by the Fundaúão Parques e Jardins, Secretaria Municipal de Meio Ambiente, Prefeitura da Cidade do Rio de Janeiro.

Other Resources

Zamith, Luiz R. and Fábio R. Scarano, 2006. Restoration of a restinga sandy coastal plain in Brazil: survival and growth of planted woody species. Restoration Ecology 14(1): 87-94.

Primary Contact

Organizational Contact