INFLUENCE OF PIONEER-SPECIES COMBINATIONS ON RESTORATION OF DISTURBED ECOSYSTEMS IN THE ATLANTIC FOREST , RIO DE JANEIRO , BRAZIL

Disturbed ecosystems may have suffered anthropic damage, but are still capable of returning spontaneously to their former conditions (SER, 2002; PINHEIRO, 2004). They can resist spontaneously to outside disturbances KNOWLES, 1999; SER, 2002). However, degraded ecosystems may have their natural mechanisms of homeostatic balance regulation altered, which reduces their capacity to recover (BROWN and LUGO, 1994; NEPST AD et al., 1996; MARTINS, 2001), leading them gradually further from the original point of homeostatic balance, as disturbances occur . INFLUÊNCIA DA COMBINAÇÃO DE ESPÉCIES PIONEIRAS NA RESTAURAÇÃO DE ECOSSISTEMA PERTURBADO DA MATA ATLÂNTICA, RIO DE JANEIRO, BRASIL


INTRODUCTION
Disturbed ecosystems may have suffered anthropic damage, but are still capable of returning spontaneously to their former conditions (SER, 2002;PINHEIRO, 2004).They can resist spontaneously to outside disturbances KNOWLES, 1999;SER, 2002).
However, degraded ecosystems may have their natural mechanisms of homeostatic balance regulation altered, which reduces their capacity to recover (BROWN and LUGO, 1994;NEPSTAD et al., 1996;MARTINS, 2001), leading them gradually further from the original point of homeostatic balance, as disturbances occur.

INFLUÊNCIA DA COMBINAÇÃO DE ESPÉCIES PIONEIRAS NA RESTAURAÇÃO DE ECOSSISTEMA PERTURBADO DA MATA ATLÂNTICA, RIO DE JANEIRO, BRASIL
The restoration activities of degraded/disturbed and degraded ecosystems consist of human intervention to re-establish ecological functions, in order to attain a homeostatic balance similar to the one existing prior to disturbance (MINTER/ IBAMA, 1990;VALCARCEL and SILVA, 2000), and to recover form and function (DIETRICH, 1990;ENGEL and PARROTA, 2003).
The use of reforestation as biological measures for ecosystem restoration is a feasible strategy in tropical environments, because it combines low-cost and high biotic potential (VALCARCEL and SILVA, 2000).The species used in reforestation have an effect on the establishment of new ecological relationships and determinate ecosystem construction, offering means to create different and emergent properties (BROWN and LUGO, 1994;AIDE et al., 1995;CHAPMAN and CHAPMAN, 1996;PARROTA et al., 1997 a, b;PARROTA, 1999;FLORENTINE and WESTBROOKE, 2004).
Practical problems to make functions of reforestation a viable means to reconstruct an ecosystem constructor agents begin with the selection of appropriate species.In location with exiguous levels of environmental attributes, only some species will be capable of surviving and resisting.The use of rustic species initiates the colonization process and development of ecological functions, similar to pioneers (CHAPMAN and CHAPMAN, 1996), generating a hydrological condition (BOSCH and HEWLETT, 1982) capable of supporting more demanding species in the watersheds.
In this study, we evaluated what composition and /proportion of rustic species are capable of generating the best ecological conditions for regeneration at the beginning of the restoration process, on areas with low offer of environmental attributes.

MATERIALS AND METHODS
The study area (22º45'50"-22 o 51' S and 43 o 26' -43 o 36'10" W) is located on the north slope of the Madureira-Mendanha range, Nova Iguaçu county, metropolitan region of Rio de Janeiro, in southeastern of Brazil.The vegetation is a semi-deciduous Atlantic Forest, where different disturbances occurred from the beginning of the nineteenth century to the middle of the twentieth century, during the coffee, sugar-cane, and orange growing cycles.The crops were succeeded by grasslands, managed with fire, which resulted in clear degradation tendencies.
Annual precipitation (1.212 mm) is intense and concentrated in summer (BARBIERE and KRONEMBERGER, 1994;MATTOS et al., 1998).This heavy rainfall prevented the acquisition of environmental attributes for the ecosystem recovery.The climate is classified as Tropical from altitudes (Cwa) occurring at altitudes higher than 300 m, with temperatures between 17 and 22º C; and as Tropical (Aw) occurring at altitudes between 30 and 300 m, with temperatures between 22 and 24º C. The Litholic Neosoil (PALMIERI, 1980), includes environments with shallow soils and uniform aspect of terrain, with similar hydrological characteristics along the series.These environments that match these effects present edaphic and climatic limitations with minimization of pedogenetic soil development, water retention, and establishment of forest species.
The plot located in the least-disturbed ecosystem, Forest, was the most developed in terms of homeostatic balance and served as a parameter for restoration actions.The most disturbed site, Grassland, has poor environmental attributes and lacks any human conversational activity.At this site, ephemeral but intense cycles of growth and latency of the grass Panicum maximum Jacq.Are observed, conferring high flammability on the ecosystems, as well as erosion process and impoverished biodiversity, all characteristics that lead to inertial degradation tendencies.Three 12year-old reforestation projects were used as Biological Measures (BM's), of restoration and were evaluated as strategies to reverse degradation tendencies.They where planted in 1994, and used different densities of pioneer species: BM1 used 82% of anthropic pioneer species, where 73% were Mimosa caesalpiniifolia Benth, and 9% were Eucalyptus citriodora Hook.and 18% were native species.BM2 used 91% of anthropic pioneer species, where the proportions were: 9%; 82% and; 9% respectively.BM3 used 25%, in a proportion of: 15% and 10% respectively, and 75% were distributed among 7 native species (Table 1).
Grassland represents areas where biotic components of ecosystem are less resilient and present inertial degradation tendencies.Forest represents a resilient environment with a natural tendency toward restoration.The BM's represent treatments where the use of planted species can influence biotic components, modifying the levels of homeostatic balance, and defining trends toward degradation and/or restoration for 12 years (Figure 1).The inventory of spontaneous regeneration in the BM's was carried out in February 2005, in the rainy season, utilizing the point method (MANTOVANI and MARTINS, 1991).The points where located so as to achieve representativeness of the plant community, which was determined by the collection curve adjusted by the best adjustment to the linear or logarithmic regression.The parameters used to characterize regeneration were : a) Relative Density (RD), b) Relative Vigor (RV), c) Relative Frequency (RF), Importance Value (IV), d) Jaccard similarity index (C), and e) Shannon's diversity index (H') (MAGURRAN, 1988;RICKLEFS, 2003).The plants surveyed where identified by consulting the herbarium (RBR) of the Universidade Federal Rural do Rio de Janeiro and specialized bibliography.
Characteristics of the Atlantic Forest such as high levels of biodiversity and endemism (MYERS et al., 2000) can contribute to the restoration of disturbed lands, because they improve functional complementarities of species optimizing results from restoration strategies (VALCARCEL and SILVA, 2000).In latitudes with low biodiversity offer, physical measures (constructions) and physical-biological measures (constructions with organic materials) of recuperation are the most important means of restoration and /rehabilitation (TRAGSA, 1994).In tropical ecosystems, the use of native species that facilitate the return of biodiversity is recommended.The choice of the species at the beginning of a restoration project is important, as observed in studies of regeneration in the Amazonian forest (PARROTA and KNOWLES, 1999), where reforestation with native species allowed the admission of 141 species, and with commercial species (40) and with primary forest (157), and the diversity of planted species facilitated the appearance of many spontaneous species (PARROTA, 1995;PARROTA et al., 1997a).
On abandoned pasture lands, manipulation of structural complexity during the restoration process facilitates the arrival of seed-dispersing organisms that use different vegetation gradients as perches and shelter, thus intensifying the plant-disperser interaction and therefore the quality of restoration (MIRITI, 1998).Managing this ecosystem complexity so as to improve biodiversity in restoration projects involves the use of zoocoric species, because a large part of dispersion is affected by animals, which participate in about 50-90% of arboreal forest species dispersion (SILVA, 2003).This process increases genetic flows between populations, enhancing long-term sustainability during the restoration activities (BLEHER and BÖHNING-GAESE, 2001;SILVA, 2003;DARIO, 2004).
BM3 was also the treatment with the largest amount of regeneration from zoocoric arboreal species such as Filicium decipiens Thw, Inga fagifolia Willd., Eugenia sp., and, Sorocea sp.In environments under the phase of construction, the habit and function of each species contribute to the restoration sustainability.The composition and interaction between species develop distinct environmental attributes, modifying the species richness and ecosystem complexity.
A major difference between the BM's was due to the canopy-cover effect, which vary with composition and richness of the planted species.At BM3, the great richness of planted species may have affected the regeneration results, because it facilitates the entrance of light, filtrates radiation by alternation of their deciduous caducifoly period, modifying the soil physical-chemical conditions, improving the regeneration process.Studies with planted species at the Cerrado site indicated different soil properties according to the litter composition (GARAY et al., 2004).
The Forest ecosystem contained 12 arboreal species: represented by Tabebuia crysotricha (Mart.)Standl., Astronium graveolens Jacq., Erythroxylum pulchrum A. St. Hil., Croton urucurana Baill., Eugenia cuprea (O.Berg.)Willd., Eugenia fumicifolia Dc., Casearia sylvestris SW., Brosimum guianensis (Aubl.)Huber, and others; where most, are secondary species.At treatment BM1, BM2, and BM3, most of the arboreal individuals from spontaneous colonization were autochthonous species come from the recruitment of planted species.However, six allochthonous arboreal species that had not come from planted species were found: Inga fagifolia Willd., Filicium decipiens Thw , Myrocarpus fastigiatus Allemao, Sorocea sp., Eugenia sp., and Sp27, which were not identified because they were seedlings on the initial development phase.Studies comparing seed bank in a forest, agriculture and grassland sites identified a higher number of viable arboreal seeds on the forest stand, predominance of invasive species on agriculture site and grasses on the pasture (GASPARINO et al., 2006).
The diversity of arboreal spontaneous regeneration among the five treatment areas reflects the strong influence of the species used in the reforestation in 1994 (Table 2).At BM1, even the low richness of the planted species (4) allowed 75% of arboreal regeneration from the planted species and 25% from allochthonous species, and it was the area that herbaceous species and grasses were best suppressed.BM1 showed 4% of similarity with Forest (Table 3), indicating that it had begun to reverse tendencies, constructing a sustainable ecosystem.BM2, with richness of 3 planted species, showed 85% of Eucalyptus citriodora regeneration, sharing no species with Forest.BM2 was the one with the most similar species to Grassland (31%) in regeneration , representing the least sustainable treatment over a medium term.BM3 had the highest richness of planted species (9), and reached 50% of regeneration from allochthonous species and a 5% species similarity with Forest.
Forest showed a high level of homeostatic balance, without the presence of grasses or colonization by planted species, unequivocal evidence of selfdetermination capacity.The absence of grasses at this site shows the maturity of this ecosystem, where the herbaceous species observed were: Crypthantus sp., Bromelia antiacantha Bertol., Oceocladis maculata (Lindl) Lindl., and Philodendron sp.The herbaceous species found in the BM's and Grassland were ruderal species, which is a strong competitive distinctiveness in relation to other species, with intense cycles of plant growth and senescence, producing a dry biomass that is highly flammable in hot and dry periods, inhibiting the competition by wood species.
The BM1 presented the highest species diversity among the reforestation areas, with a diversity index with values close to the native diversity of Forest.Although similar in species richness, these sites have different taxonomy and ecological functions in the succession process and site occupation.The BM3 favored ingress and colonization of allochthonous arboreal species from other seed sources, showing that, even though it is located at the same distance from seed sources as the other treatments, it offered the best conditions to facilitate functional connectivity and genetic flow, even at 634 m distance from Forest and isolation by Pasturelands.This information emphasizes the importance of using different species to guarantee the restoration sustainability in disturbed ecosystems.At BM1, the regeneration of Mimosa caesalpiniifolia Benth, Eucalyptus citriodora Hook and, Leucaena leucocephala (Lam.) de Wit were predominant, indicating a smaller increment of emergent properties, and that the sustainability of this restoration measure may be jeopardized in the medium term.BM2 showed a diversity index lower than Grassland, making it evident that E. citriodora contributed little to the generation of emergent properties and establishment of spontaneous colonization in the first 12 years, slowly acting to reverse degradation tendencies.These results differ from those found for a seasonal forest in Minas Gerais (SOUZA et al., 2007), which supported the use of Eucalyiptus grandis as an anthropic pioneer in a Tabela 2 -Índice de diversidade e riqueza da regeneração espontânea nas parcelas amostradas e porcentagem de regeneração de espécies arbóreas plantadas (autóctones) e não plantadas (alóctones).Table 2 -Spontaneous regeneration diversity index and richness in the sampling plots evaluated and the percentage of arboreal regeneration for planted (autochthonous) and non-planted (allochthonous) species.
R. Árvore, Viçosa-MG, v.33, n.5, p.927-936, 2009 degraded ecosystem rehabilitation.Grassland had two arboreal species regenerating; however, Eugenia jambolana Lam. was a sprout from a burned tree and Pisidium guajava L. regenerated from a cut tree, evidencing a continuous human disturbance, accelerating the trend toward inertial degradation.These environments have an imminent risk of fire, where, the low offer of environmental conditions together with fire management as pastures renovation, halt the successional mechanisms and the self-sustainability process in ecological restoration.
The Jaccard similarity index, applied among species from all treatments can indicate tendencies toward spontaneous regeneration (MAGURRAN, 1988) and demonstrates the sustainability of restoration treatments.The areas under reforestation were somewhat similar for species composition ranging from 35 to 50% (Table 3); they were closer to the Grassland, with species similarity ranging between 21 and 31%.The BM3 (5%) was the area most similar to the Forest, followed by BM1 (4%).The BM2, with predominance of 82% of E. citriodora Hook did not show similarity of species with Forest, indicating little emergent property acquisition that could aid the entrance and establishment of propagules over 12-year periods, even with the artificial control of fire.
The low similarity between species from the Forest, located 640 m far from the reforestation areas may indicate: a) a lack of ecological attributes that permit colonization and establishment of the most demanding stages of succession; b) Internal forces from ecosystems have stronger degradation tendencies than the sum of efforts from vegetation to construct ecosystems, indicating that using only biological measures does not guarantee the sustainability of restoration over a medium to long-term; d) there are different levels of homeostatic balance, and the ones created by the effects of reforestation do not move on a clear way from inertial degradation tendencies in ecosystems with poor environmental condition attributes.The ecosystems may have changed so far from their original level of homeostatic balance that even after 12 years of restoration measures, it was observed the admission of only 5 % of native plant species.
The BM1 and BM3 had one arboreal species in common with Forest, showing a functional connectivity with the seed source.Forest did not present similarities with Grassland, demonstrating that these environments have opposite evolutionary tendencies even though they are only 655 m apart from each other.The variation of similarity in the BM's, from 35 to 50%, shows the difference between treatments, where the choice of species affected restoration quality.
Amongst the similar species, the grass Panicum maximum Jacq.was notable for appearing in four of the five treatments as the species with the highest Value of Importance (Fig. 2).The BM1 was the measure that best suppressed the grass P. maximum, which might be related to the high litter cover in winter, which reduces the chances of small seeds to reach the soil, or because in summer, when grasses increase their biomass and grow vigorously, M. caesalpiniifolia provides offers great amount of shadow, reducing the density of P. maximum.At BM2; the E. citriodora did not play the same role and the grass P. maximum had an exceptionally high IV, influencing directly the regeneration process and species interactions in this plot.Vines were recorded in all the treatment areas; and in the BM's, they were represented by three species (Vigna vexillata (L.) A. Rich., Clitoria macrophylla Wall., and Arrabidea sp.).Although vine plants increase organic matter and the structural complexity of the understory, they can strangulate some arboreal individual trees and even slow the overall development of trees.This effect was observed at BM3 with Senna sp., which suffered 100% of mortality.
The sustainability achieved may be jeopardized by other factors such as competition, seed and seedling predation, distance from seed source, and exotic species invasion (NEPSTAD et al., 1991;AIDE and CAVALIER, 1994;GUARIGUATA et al., 1995;NEPSTAD et al., 1996).Most ruderal species have physiological mechanisms that optimize energy utilization and resistance to climatic stress and poor environmental condition of ecosystem attributes (LARCHER, 1977).These attributes confer an advantage on their establishment which can retard restoration activities.Grasses, for example, constitute one of the major problems for spontaneous regeneration, because it is an aggressive colonizer that competes for nutrients and water, impeding the establishment of native species (NEPSTAD et al., 1991;VIEIRA et al., 1994;NEPSTAD et al., 1996;FLORENTINE and WESTBROOKE, 2004;SOUZA and BATISTA, 2004).Grasses form a vegetal covering mantel that prevents seeds from reaching the soil (WHELAN et al., 1991;AIDE et al., 1995), and even if the seeds reach the soil, their viability is impaired by high  levels of predation (NEPSTAD et al., 1991;WHELAN et al., 1991).Associated with these problems, there is a risk of biomass combustion, because grasses grow and dry rapidly, providing good conditions for fires, incinerating species that spontaneously colonize the area, volatilizing nutrients and humic acids, and altering the physical-chemical properties of the soil.All these difficulties favor grasses, so they become abundant in disturbed ecosystems.

CONCLUSION
The role of species as ecosystem constructors at the five treatment areas were distinct and reflected evolutionary adjustments of these ecosystems, where the approximately 100-year-old Forest achieved a rearrangement of species that propitiates effective entrance and retention of rainwater from precipitation, guaranteeing the maintenance of the functional structure.The BM's showed different degrees in the reconstruction process and distinct levels of sustainability.Grassland presents a tendency toward inertial degradation, making it very difficult the spontaneous reversion of this trend.
The quality of disturbed ecosystems restoration depends on the composition of rustic species at the beginning of the reforestation process implantation.

Figure 1 -
Figure 1 -Ecosystem evolutionary tendencies, and the role of reforestation projects as strategies to reverse degradation.The dark arrows represent possible paths of evolution ways of each Biological Measure over time, according to the biotic interactions among the planted species.Figura 1 -Tendências evolutivas do ecossistema e o papel dos reflorestamentos na reversão da degradação.As setas escuras representam os possíveis caminhos evolutivos de cada medida biológica ao longo do tempo, de acordo com as interações bióticas entre as espécies plantadas.

Table 1 -
Species utilized in different proportions in the biological measures implanted on an environment with low offer of attributes at Nova Iguaçu, RJ.

Table 3 -
Similarity of spontaneous regeneration species in treatments surveyed in ecosystems with poor environmental conditions, Nova Iguaçu, state of Rio de Janeiro.