Survival of Saplings in Recovery of Riparian Vegetation of Pandeiros River (MG)

This study monitored the survival of saplings planted according to different recovery models in a riparian forest of the Pandeiros river (Januária, MG). The models consisted of planting the saplings in lines of 2 or 4 m with presence (T2S and T4S, respectively) or absence of direct seeding (T2 and T4, respectively). We planted 16,259 saplings of 17 botanical families, 32 genera and 33 species. The saplings, in general, presented a survival rate after one year of 34.4% (±1.8). The species with highest survival rates were Jacaranda brasiliana, with 85.0% (±13.5) of survival, Anadenanthera colubrina, with 70.1% (±7.0), and Triplaris gardneriana, with 69.3% (±9.1). Survival did not vary between the models tested, probably due to the short evaluation period (12 months).


INTRODUCTION
The Cerrado, a region of considerable biological diversity, is identified as one of the richest and most threatened ecosystems in the world, with a large number of restricted and specialized endemic species (about 44% of the flora), which are therefore susceptible to extinction (Scariot et al., 2005).This biome is considered a conservation priority for global biodiversity, since 55% of its original area was deforested or transformed by human action (Klink & Machado, 2005).The relatively flat surface and the presence of fertile soils promotes the highest and fastest rates of agricultural growth in Brazil, being considered the last agricultural frontier of the country, attracting a significant part of the national agribusiness (Klink & Machado, 2005).
In this context, the riverine formations (ciliar, riparian or gallery forests) that occur in the Cerrado show the effects of this agricultural expansion and represent environments significantly disturbed and degraded by anthropic action.Riparian forests protect and maintain water resources, since they are formations associated with watercourses, standing out for their richness and genetic diversity and acting as a physical barrier between terrestrial and aquatic systems (Hughes et al., 2008;Roni & Beechie, 2013).These forests act to filter water flowing into the watercourses along which they are found, developing conditions conducive to infiltration and significantly reducing the possibility of water contamination by sediment, fertilizer residues and agricultural pesticides (Ribeiro & Schiavini, 1998;Roni & Beechie, 2013).Additionally, they are important ecological corridors for the movement of fauna and gene flows (Martins, 2007).Thus, given these particular characteristics that highlight their importance, forests bordering streams and springs, are protected by federal legislation and are considered Permanent Preservation Areas (PPAs), according to the Forest Code, Law (number 12.651 of May 25 th 2012), that is, an ecological reserve that cannot undergo alteration, and whose vegetation must remain intact in its original state (Brasil, 2012).
Even when protected by law, PPAs are not exempt from the impacts of human activity, both in the surroundings and in loco (Rezende, 2004).In this sense, the recovery of degraded areas has become a priority to combat the environmental degradation processes to which natural areas are exposed (Martins, 2007).However, to recover degraded areas, it is necessary to establish a model developed from studies of phenology, details regarding the biology and germination of species, and growth analysis, amongst others, that are fundamental for vegetation rehabilitation (Hughes et al., 2008;Andel et al., 2012).Thus, techniques in combination such as sapling planting and direct seeding, which aim to guard against soil erosion and assist local revegetation, can offer good outcomes, which help to ensure the successful recovery of degraded areas (Martins, 2007;Roni & Beechie, 2013).
In order to restore vegetation and ecological processes, several models for the recovery of degraded areas have been tested (Andel et al., 2012;Pereira & Rodrigues, 2012).Most studies are based on successional models (McClain et al., 2011) with species of different ecological groups (Gonçalves et al., 2005;Pereira & Rodrigues, 2012) or by facilitating species (Beltrame & Rodrigues, 2008).However, according to Tilman (1988), in nutrient-deficient environments (such as disturbed and degraded environments), competition between species may be a determining factor for their establishment.Despite the importance of competition for seedling/sapling survival and growth, studies with recovery models of degraded areas based on distances between saplings to test the effect of competition, are scarce.
The Environmental Protection Area (EPA) of Pandeiros River was established by Law 11,901 of September 1 st 1995, with the objective of conserving the water resources and significant biological diversity present at the site (Nunes et al., 2009).The Pandeiros River EPA is located in northern Minas Gerais, an ecotonal area between the Cerrado and Caatinga domains, forms a junction between riparian forest, dry forest, cerrado (savanna) and veredas (palm swamps), occurring in alternating areas that may show abrupt vegetation changes within relatively small areas (Nunes et al., 2009).Among these physiognomies, the riparian forests, characterized by accompanying springs and water courses (Andel et al., 2012;Roni & Beechie, 2013) form ecological corridors that highlight the botanical diversity present, as well as protecting and sheltering terrestrial and aquatic fauna (Roni & Beechie, 2013).
Despite the protection afforded by Law (both PPAs and EPAs), there are several factors, mainly livestock and fires that impact the riparian forest of Pandeiros river compromising local biodiversity conservation and making the forest sparse and even wholly absent (Nunes et al., 2009;Rodrigues et al., 2009).In this sense, this study aimed to monitor the survival of saplings of native species planted according to different recovery models for the riparian vegetation of the Pandeiros River, to evaluate the most effective rehabilitation model for riparian forests in the region.

Characterization of study area
This work was developed at the Pandeiros River EPA, Januária, northern Minas Gerais, Brazil.The region presents several phytophysiognomies, resulting from the transition between the Cerrado and Caatinga biomes, such as the Deciduous Seasonal Forest (dry forest), cerrado sensu stricto, riparian forests, floodplain areas and veredas (Nunes et al., 2009).Additionally, the riparian vegetation of the Pandeiros River presents a junction of riparian forest, dry forest, savanna, and palm swamps, exhibiting high tree diversity (Rodrigues et al., 2009;Menino et al., 2012;Veloso et al., 2014).
The climate of the region, according to Köppen, is Aw, with average annual temperature of 21 to 24 °C and rainfall of 900 to 1,200 mm, and rainfall concentrated in the months of November to January (Azevedo et al., 2014).During the study period, the average temperature was 24.3 °C and 23.6 °C, the maximum average temperature was 31.9 °C and 31.3 °C, minimum average temperature was 18.4 °C and 17.5 °C and total precipitation was 1,302.6 mm and 1,877.4mm, throughout the period from December 2009 to December 2010 and from January 2011 to December 2011, respectively.The rains were concentrated during the summer, with peaks in December/2009, March and December/2010 and March and December/2011 (Figure 1).The highest temperatures were observed in January and October/2010 and February and September/2011.The climatic data were obtained from the Meteorological Station of Januária, of the National Meteorological Institute (INMET 2016).
Geomorphologically, the EPA of Pandeiros River is located in the São Fransiscana Depression and São Francisco Planalto, with a geological process of the Urucuia and Santa Helena Formations (sedimentary material) and limestone layers of the Bambuí Series (Jacomine, 1979).The predominant soil is the typical Dystrophic Red-Yellow Latosol, with a moderate A horizon and clayey texture, cerrado phase, with smooth to wavy relief, where there are spots of typical Eutrophic Haplic Cambisol with moderate A horizon and clayey texture added to the Argisol Red-Typical eutrophic yellow, with moderate A and clayey texture, deciduous forest phase, with flat to soft undulating relief (UFV, 2010a, b).For the experiment implementation, the areas destined for the revegetation were the PPAs, demarcated by the Brazilian Forest Code (Brasil, 1965), in the year of implantation of the experiment (2009).The experiment was conducted on three farms: Agropecuária Ouro Preto -AGROPOP (15° 36 "S and 44° 42" W), Traçadal (15° 32 "S 44º 43" W) and Pandeiros (15° 39 "S 44° 39" W).All farms present extensive cattle breeding as an economic activity, where the areas used were deforested for cultivation of Brachiaria sp.Experimental plantations were carried out in December 2009, at AGROPOP, January 2010, at Traçadal, and December 2010 at Pandeiros farms.

Species selection, sapling production and recovery models
The native species used for recovery planting were species that occur in the region and were selected from floristic studies conducted in the riparian vegetation of the Pandeiros river, near the study areas (Rodrigues et al., 2009;Veloso et al., 2014).About 10 matrices for each species were marked for seed collection, close to the study areas, but from different forest remnants, to guarantee genetic diversity (Martins, 2007).Seed collection and sapling production were undertaken in partnership with the Instituto Estadual de Florestas (IEF), using the IEF forest nursery in Januária, as well as the Plant Ecology Laboratory of the Universidade Estadual de Montes Claros (UNIMONTES).
Four different recovery models were implemented, with four replications each, consisting of 1.0 ha plots (50 m wide perpendicular to the river × 200 m long parallel to the river).Each plot was demarcated using treated eucalyptus fence and barbed wire.Treatment of the area was also undertaken, consisting of harrowing, with a tractor, and applying organic matter (chicken manure or mamona cake, used 0.5 t per area) manually.Harrowing was realized to decrease Brachiaria sp.grasses from the planting site, to incorporate the organic matter and for soil turning.
To plant the saplings, a fixed spacing of 2 m between the planting lines was applied and saplings were planted in 20 cm deep pits.The treatments (T), or implanted recovery models, consisted of planting saplings at two distances of either 2 or 4 m, with direct sowing -S (T2S and T4S, respectively) or no direct sowing (T2 and T4, respectively).The direct sowing in the soil was done between the saplings, with three seeds per hole, using holes of approximately 5 cm made with a hoe.The species used in direct seeding and the amount of seeds used (by weight) are described in Table 1.The experiment was carried out during the rainy season, aiming to reduce transplanting stress, since during this period the water supply would be sufficient for the establishment and growth of saplings and seed germination.
A total of 16,259 saplings were planted, distributed in 33 species, 32 genera and 16 families (Table 1), ranging from 20 cm to 1 m in height, all of them over four months old.After planting, the saplings were numbered with aluminum plates and tied by a nylon thread, and the survival of these saplings was monitored every four months for a year.

Data analysis
The survival rate of the individuals was determined by the ratio between the number of surviving individuals and the total number of individuals planted for each species and treatment throughout the four evaluations.To detect variations in survival of individuals between recovery models and between planted species, Variance Analysis (ANOVA) was used in GLM (Generalized Linear Models) procedure, followed by Tukey's post-test in the R (R Development Core Team, 2013) statistical package.

RESULTS
The total survival percentage showed significant variation between the evaluation months (df = 3, F = 75.98,p < 0.001).The highest survival rate of the planted saplings was 70.3% (±1.8), after three months of evaluation, with values of 64.0% (± 1.8) after six months, and 51.0% (±1.9) nine months after planting.After 12 months of evaluation, survival dropped to 34.4% (± 1.8), showing that survival decreased significantly during the first year of experimental planting (Figure 2).However, there was no significant variation (df = 3, F = 2.58, p > 0.05) in survival rates between the models tested.The mean survival percentage was numerically higher in the model with sapling and direct sowing at 2 m distance between saplings -ST2 (57.3 ± 1.9%) and lower in the model with saplings at 4 m distance and absence of sowing -T4 (50.2 ± 2.0%).On the other hand, there was a decrease in survival in all treatments tested, over time, as observed between areas and between months of treatment.
The survival rate also showed significant variation between the study areas (df = 2, F = 109.83,p < 0.001).The area that presented the highest survival rate was AGROPOP (76.6 ± 1.7%), followed by Traçadal (47.2 ± 2.1%) and Pandeiros (46.2 ± 1.3%), which showed the lowest survival rates (Figure 3).Survival varied between the study months in the areas studied, but decreased over time in the three areas.

DISCUSSION
The first year of planting is considered a period of sapling adaptation to the adverse conditions present at the site being revegetated, and is therefore when seedlings present higher mortality (Mcdonald et al., 2003;Raman et al., 2009).The decrease in survival rates over the first year of evaluation may also be related to the climatic conditions of the region, since the rains are concentrated in the months of November to January (Azevedo et al., 2014).The saplings were introduced during the rainy season (December and January), after which they went through a long period of drought during autumn and winter (April to October), until the arrival of the next rainy season in the summer.Therefore, water deficits may have led to the death of the plants, since low soil moisture is a limiting factor for the development of saplings (Paiva & Poggiani, 2000).McDonald et al. (2003), in a study with native species implanted in a revegetation area in Jamaica, Although fenced, the three study areas were affected by the presence of cattle within the plots, a factor that may have contributed to the decreased survival rates of the saplings, especially in the Pandeiros and Traçadal areas, where this influence was more intense.This fact was due to the possibility of animals entering via the river, during periods of drought, and when the fence was damaged.In addition, several studies (Figueiras, 1990;Martins et al., 2004;Rossi et al., 2010) cite the aggressiveness and high competitiveness of African grasses introduced into Brazil, such as Brachiaria species with native plants.In a study by Figueiras (1990), which classified African grasses according to aggressiveness in terms of competition with native flora, eight species of the Brachiaria genus were classified as "very aggressive" and "moderately aggressive", seriously affecting natural regeneration.Thus, exotic grasses used in pasture formation may be an aggravating factor in the mortality of seedlings planted in recovery projects in areas impacted by cattle breeding.
In addition to competition and climatic conditions, survival may be related to factors such as soil type and drainage (McDonald et al., 2003).According to Braga (2011), in an experiment to evaluate the performance of saplings and the development of natural regeneration in different recovery models in the AGROPOP area, half of the plots of this area are located in places originally influenced by dry forest formation with very nutrient rich soils.The soil analysis for this area, according to Braga (2011), found high levels of phosphorus and a higher proportion of clay in these plots.Several authors discuss the positive influence of soil on plant establishment (Alpert et al., 1999;Gurevitch et al., 2009).These same authors cite the importance of phosphorus as a limiting nutrient for the development of plants, as a constituent in several molecules and in cell membranes, acting directly on plant metabolism, and essential for the initial development of seedlings.Moreover, clay is a component that ensures greater soil moisture, providing water to the plants for longer periods during the dry season (Gurevitch et al., 2009) thus ensuring their survival.
The species with the highest survival rates (J.brasiliana, A. colubrina, T. gardneriana, S. spectabilis and C. ferrea), are characterized as pioneer species, heliophytes, fast growing, and are used in heterogeneous reforestation and recovery of degraded areas (Lorenzi, 1992(Lorenzi, , 1998;;Carvalho, 2003Carvalho, , 2010)).Lorenzi (2002) points out that because of the rusticity and adaptation to dry lands, the various angico species are recommended for rehabilitation, as they grow well in poor and degraded soils, as well as having very rapid growth, being ready for field planting in less than four months.Barbosa (1980) pointed out that saplings of the genus Anadenanthera supported relative humidity ranging from 1% to approximately zero, losing 85 to 90% of the initial water content, and recovering turgidity when rehydrated, showing the adaptation of species to the irregular precipitation regime.Among the 10 species with the best development (high survival and rapid growth) found in the recovery of gallery forests on rural properties in the Distrito Federal, Rezende (2004) indicates J. brasiliana and A. colubrina (38.5% and 21.9%, respectively).These species maintained a high adaptive plasticity when confronted by the adverse conditions found in the field (Rezende, 2004), which can also be observed in this study.It is of paramount importance to consider the adaptability of the species to the place where they will be introduced, since, in degraded areas, adverse conditions may decrease species survival (Rezende, 2004;Martins, 2007).
Although Inga vera is a heliophyte, pioneer or initial succession species (Carvalho, 2008), in the present study, the saplings of this species that were planted showed 100% mortality after one year.Probably, the mortality of the species is related to water scarcity during the dry period in the study area, since according to Carvalho (2008), Inga vera occurs naturally in moist, frequently humid soils and even in swampy areas.Neri et al. (2011) emphasized that the identification of native species, with good development in degraded areas, is an important step in the implementation of recovery projects.These authors also argued that the correct choice of species for revegetation in Cerrado areas should take into account their need to adapt to low soil fertility, possible water deficits or flooding and an ability to compete with invasive weeds; factors that may determine the survival and growth of saplings.
Studies involving edaphic and environmental variables are necessary for a better understanding of 9/11 Survival of Saplings in Recovery... Floresta e Ambiente 2018; 25: e20150212 the survival of the seedlings within the tested recovery models.Such studies are scarce, since there are few studies investigating the recovery of degraded areas that use native plants and that propose recovery models to be tested, which makes it difficult to compare the methods analyzed here.

CONCLUSION
The average survival of the saplings planted was low, probably due to the adverse conditions found in the degraded sites, such as long periods of water stress characteristic of the region and possible external interference, such as competition from Brachiaria sp. or trampling and grazing by domesticated animals.Some species are more resistant, such as Jacaranda brasiliana, Anadenanthera colubrina, Triplaris gardneriana, Senna spectabilis and Caesalpinia ferrea, and may be suitable for revegetation in degraded riparian forests in the Pandeiros river region.Additionally, the models tested should be reevaluated to give sufficient time for planted individuals to compete with their peers.

Figure 1 .
Figure 1.Mean monthly of minimum, average and maximum temperature and total monthly precipitation from December 2009 to January 2011, obtained at the Meteorological Station of Januária, Minas Gerais (Source: INMET, 2016).

Figure 2 .
Figure 2. Average survival percentage of saplings implanted in the restoration of riparian vegetation of Pandeiros river (Januária, MG), for restoration model and evaluation time.Restoration models: sapling planting with distance between rows of 2 m without (T2) and with (T2S) direct seeding and in lines of 4 m without (T2) and with (T2S) direct seeding.

Figure 3 .
Figure 3. Average survival percentage of saplings planted for the rehabilitation of riparian vegetation of Pandeiros river (Januária, MG), for restoration areas and evaluation time.

Table 1 .
Native species used in direct seeding (weight in grams) and sapling planting (number of individuals) in the recovery of riparian vegetation of Pandeiros river (Januária, MG).Fagundes NCA, Braga LL, Silva WA, Coutinho CA, Neves WV, Souza RA et al.

Table 2 .
Average survival percentage in time and standard deviation (in parentheses) of the species implanted in the recovery of riparian vegetation of Pandeiros river (Januária, MG).Popular names, species' authors name and their respective botanical families are presented in Table1.Fagundes NCA, Braga LL, Silva WA, Coutinho CA, Neves WV, Souza RA et al.