Carbon and nutrient stocks of three Fabaceae trees used for forest restoration and subjected to fertilization in Amazonia

Amazonia is crucial to global carbon cycle. Deforestation continues to be one of the main causes of the release of C into the atmosphere, but forest restoration plantations can reverse this scenario. However, there is still diffuse information about the C and nutrient stocks in the vegetation biomass. We investigated the carbon and nutrient stocks of Fabaceae trees (Inga edulis, Schizolobium amazonicum and Dipteryx odorata) subjected to fertilization treatments (T1 – no fertilization; T2 – chemical; T3 – organic; and T4 – organic and chemical fertilization) in a degraded area of the Balbina Hydroelectric Dam, AM Brazil. As an early successional species, I. edulis stocked more C and nutrients than the other two species independent of the fertilization treatment, and S. amazonicum stocked more C than D. odorata under T1 and T4. The mixed species plantation had the potential to stock 4.1 Mg C ha year, while I. edulis alone could stock 9.4 Mg C ha year. Mixing species that rapidly assimilate C and are of significant ecological and commercial value (e.g., Fabaceae trees) represents a good way to restore degraded areas. Our results suggest that the tested species be used for forest restoration in Amazonia.


INTRODUCTION
The global climate is changing; more intense and frequent El Niños and rising temperatures are becoming reality (Malhi et al. 2008, Moss et al. 2010, Davidson et al. 2012, Kintisch 2016).In many ways, deforestation and forest degradation have exacerbated these changes by releasing CO 2 and other greenhouse gases (GHGs) that were previously stored in the forest biomass into the atmosphere (Houghton 2005).Tropical forests have a significant capacity to store C and other nutrients in their biomass, and the C can then be released through burning or decomposition, primarily from conversion into pastures and agricultural fields, which are the primary causes of forest degradation in Amazonia (Houghton et al. 2000, Achard et al. 2002, Karsten et al. 2013).Although some studies have shown that almost 20% of the original Amazonian forest cover has disappeared, forest degradation may be higher depending on the monitoring system (Asner et al. 2005, Davidson et al. 2012).ROBERTO K. JAQUETTI and JOSÉ FRANCISCO C. GONÇALVES Forest restoration plantations have the capacity to recover large amounts of C stocks in a short time while simultaneously producing goods and services (Ruiz-Jaen and Potvin 2011, Houghton 2012, Guariguata and Brancalion 2014), but the choice of species and their silvicultural traits is a fundamental aspect of this practice (Fang et al. 2007).A mixture of different ecological groups (pioneer and late-successional species) is desirable to better use the available resources and therefore enhance C stocks (Rodrigues et al. 2009).
Forest restoration is an effective way to restore biodiversity and C stocks, especially under severe degradation (Soares-Filho et al. 2014), but transcending ecosystem barriers is very difficult as non-adapted species can easily collapse and die (Rowland et al. 2015).The main obstacles to be overcome in the degraded ecosystems of the Amazon are high irradiance and temperatures, periods of drought and, especially, the low natural fertility of the soils that are the major limiting factor for C assimilation and plant growth (Gonçalves and Santos Junior 2005, Santos Junior et al. 2006, Ferreira et al. 2009, Jaquetti et al. 2014, 2016).
Therefore, the study of the use of different fertilization treatments in forest restoration and the responses of native species to these treatments is essential to develop planting designs that are capable of better utilizing available resources to increase stand-level productivity and individual tree growth rates, and the use of Fabaceae trees is important to achieve these objectives in degraded landscapes (Siddique et al. 2008).These species are well adapted to low nutrient availability and arid ecosystems, as they have developed the capacity to fix atmospheric nitrogen (N) through bacterial symbiosis (Macedo et al. 2008, Joslin et al. 2011).Thus, they can rapidly rebuild C and N stocks through rapid biomass production (Tapia-Coral et al. 2005), which in turn is an excellent success indicator of forest restoration projects (Viana et al.

2014).
Despite the importance of understanding the responses of native species to different ecological conditions, a significant knowledge gap exists concerning the basic silviculture of tropical species, for both restorative and commercial purposes, in degraded areas of the Amazon (Ferreira et al. 2016).The potential for native tropical trees to assimilate and incorporate C and nutrients into their biomass is equally important, especially in the context climate change, as native tropical trees can mitigate the impacts of deforestation as well as be traded on the global C credit market (Gibbs et al. 2007, Ruiz-Jaen and Potvin 2011, Silva et al. 2011, Visseren-Hamakers et al. 2012).
To increase the understanding of C and nutrient stocks in trees in tropical forest restoration, we investigated the composition of the biomass of three Fabaceae tree species subjected to fertilization treatments (chemical and organic amendments) in a degraded area of central Amazonia.

STUDY SITE, SPECIES DESCRIPTION AND FERTILIZATION TREATMENTS
This study was carried out in a degraded area within the Balbina Hydroelectric Dam Forest Restoration Project (01°55'99"S and 59°24'65"W) in the county of President Figueiredo in Central Amazonia, AM, Brazil.The study area covers a total of 6 hectares and was degraded in 1983 during dam construction, during which the natural vegetation of the area, characterized as "nonflooded dense forest," was completely removed.The blocks of this study were located in gaps between the vegetation, and shaded spots produced by natural islands of regeneration were avoided.The climate of the region is Af (tropical humid), according to the Köppen climate classification system, with 2,280 mm of precipitation annually and a mean annual temperature above 25°C.The soil in the area is classified as yellow oxisol with high acidity and low natural fertility.
The beginning of the experiment was marked by the manual excavation of 30 cm × 40 cm-deep tree-planting holes using a straight spade modified with an iron cable; the distance between the seedlings was 2 m × 2 m.Fertilization treatments were performed fifteen days before planting and involved incorporating 50 g of lime, 150 g of Fosmag ® Bunge chemical fertilizer (São Paulo, SP, Brazil: 4% N, 14% P, 7% K, 11.5% Ca, 2.7% Mg, 10.4% S, 0.07% B, 0.59% Zn and 0.15% Cu) and 15 g of FTE BR-12 ® Nutriplant (São Paulo, SP, Brazil: 7.1% Ca, 5.7% S, 1.8% B, 0.8% Cu, 2.0% Mn, 0.1% Mo and 9.0% Zn) into the soil in all of the fertilization pits except for T1 (control), which was not fertilized throughout the experiment.The fertilization pits were only created during site preparation, so the period following fertilization of the T2, T3 and T4 treatments can be considered the post-planting fertilization period.The species were randomly placed within each block; three individual plants from three different species were fertilized according to one of four fertilization treatments for a total of 12 treatments per block, each of which was 144 m 2 (12 × 12 m) in size.This protocol was replicated three times in each block, resulting in a total of 36 seedlings per block and a total of 216 experimental seedlings.
Individuals of three native Fabaceae species, tonka bean (Dipteryx odorata [(Aubl.)Willd.]),ice-cream bean tree (Inga edulis Mart.) and white faveira (Schizolobium amazonicum Huber ex Ducke), of uniform size (35 cm high for D. odorata, 40 cm for I. edulis and 30 cm for S. amazonicum) and health within the same species were selected, and the following fertilization scheme was used: T1) unfertilized (control), T2) post-planting chemical fertilization, T3) post-planting organic fertilization and T4) post-planting chemical and organic fertilization combined.Two additional applications were performed for the post-planting chemical fertilization treatments (T2 and T4), the first of which was performed four months after planting with the application of 50 g of lime, 150 g of Fosmag ® (the same formulation as above) and 15 g of FTE BR-12 ® Nutriplant (the same formulation as above).The second application was performed 8 months after planting and consisted of 50 g of lime and 150 g of NPK 4-14-8 ® Bunge chemical fertilizer (São Paulo, SP, Brazil: 4% N, 14% P and 8% K).The chemical fertilizers were applied around the seedlings over an area that was 50 cm in diameter.
In treatments T3 and T4, branches and leaves were collected from the lower third of the fallow trees in the area, and the material was then manually chopped with a machete into 10-cm-long pieces and applied to the bare soil without incorporation.The treatment areas were covered with a 25-cmhigh layer of this green material.Two additional applications were also performed in the organic fertilization treatments.The first application was performed 10 days before planting and consisted of 12.4 kg m -2 (124.5 Mg ha -1 ) of fresh mass, and the second application was performed 4 months after planting and consisted of 5.5 kg m -2 (55.2 Mg ha -1 ) of fresh mass.Assuming that each plant covered an area of 4 m 2 , the first application consisted of approximately 13 kg C, 290 g N, 11 g P, 106 g K, 144 g Ca and 41 g Mg per plant, and the second application consisted of approximately 6 kg C, 131 g N, 5 g P, 48 g K, 65 g Ca and 18 g Mg per plant.

BIOMASS AND CARBON SEQUESTRATION RATE
The biomass in each treatment was determined by randomly selecting one of the three replicates of each treatment in the six blocks (n = 6).The shoot was collected first and set aside, and the roots were subsequently collected by removing the soil within 50-100 cm of the base of each plant until the first roots were found.The surrounding area was then carefully excavated, and the roots were separated from the soil.The component leaves, stems and roots were separated from each plant and taken to the Laboratory of Soils and Plants of the National Institute of Amazonian Research (LSP -INPA); the roots were washed to remove all soil particles.The material was placed in a forced-ventilation oven (65°C for the leaves and fine roots, 105°C for the stems and coarse roots) until the material reached a constant weight, and after drying, the material was weighed using a digital scale (Mettler PM 30-K) with a precision of 0.01 g (Zhao et al. 2001, Schroth et al. 2002).The absolute growth rates in biomass (AGR biomass ) of the leaf, stem and root components were calculated as described by Hunt (1990) (AGR biomass = W2 -W1/t2 -t1), where W1 = initial weight; W2 = final weight; t1 = initial time; and t2 = final time.At harvest, 11 months after planting, I. edulis and S. amazonicum were, on average, 2.5 and 2.2 m in height and 7 and 5 cm in diameter, respectively, both in the T4 treatment, while the D. odorata individuals had mean T2 values of 1.3 m in height and 2 cm in diameter.The C content of each component was obtained by summing the biomass of each component and multiplying the total by a factor of 0.47, which is the current suggestion for tropical angiosperm species by Thomas and Martin (2012).

LEAF NUNTRIENT CONTENT AND ASSIMILATION
To determine the macro-nutrient contents for each treatment, four healthy leaves that appeared fully expanded under full exposure to the sun were pre-selected from the middle third of the plants just before harvesting.The total N content was determined by the Kjeldahl method (Miyazawa et al. 2009), and the phosphorous content was determined by spectrophotometry at 725 nm.The contents of other macro-nutrients (K, Ca and Mg) were determined by atomic absorption spectroscopy (Perkin-Elmer 1100B, Uberlingen, Germany) (Embrapa 2009).

EXPERIMENTAL DESIGN AND STATISTICAL ANALYSIS
The experiment was arranged in randomized blocks and consisted of 12 treatments involving a combination of 2 main factors (species = 3 and fertilization = 4); measurements were obtained from 3 replicates of each treatment.A two-way ANOVA was used to analyze the obtained data, and the means were separated using Duncan's multiple range test at the p < 0.05 level of probability with the statistical program STATISTICA 7.0 (StatSoft Inc., Tulsa, OK, USA).

CARBON SEQUESTRATION RATE
Both factors, species and fertilization, were found to be statistically significant for all components.Under the T4 treatment, I. edulis (3768.7 g C plant -1 year -1 ) accumulated almost 4 times more C in the entire plant than S. amazonicum (962.8 g C plant -1 year -1 ) and more than 16 times that in D. odorata (226.5 g C plant -1 year -1 ).Under the T3 treatment, the value was 23 times higher for I. edulis (2215.3g C plant -1 year -1 ) compared to D. odorata (95.3 g C plant -1 year -1 ) and almost 7 times higher relative to S. amazonicum (328.3 g C plant -1 year -1 ) (Table I).The greatest difference between S. amazonicum and D. odorata was exhibited under the T4 treatment, where the former stocked greater than 4 times more C in the entire plant.Differences between the two species under the T2 treatment were not significant.
In terms of fertilization, both I. edulis and S. amazonicum responded better to the simultaneous application of both chemical and organic fertilizers (T4), while D. odorata assimilated more C under the T2 treatment.Under T4, I. edulis assimilated almost 47 times more C in its biomass (3768.7 g C plant -1 year -1 ) compared to the unfertilized T1 treatment (80.7 g C plant -1 year -1 ), and S. amazonicum assimilated 13 times more C (962.8 g C plant -1 year -1 ) compared to the T1 (73.0 g C plant -1 year -1 ) treatment (Table I).

POTENTIAL FOR CARBON SEQUESTRATION WITH FOREST RESTORATION
We present four possible restoration scenarios based on our data: each species planted by itself and all three species planted together.Individually, I. edulis could potentially stock more C than the other two species, and when the T4 treatment data were extrapolated over hectares, I. edulis (9.4 Mg C ha -1 year -1 ) could assimilate 16 times more C than D. odorata (0.6 Mg C ha -1 year -1 ) and almost 4 times more C than S. amazonicum (2.4 Mg C ha -1 year -1 ) (Figure 1).Under the T2 treatment (chemical fertilization), which is the most common fertilization scheme in restoration forest plantations, I. edulis (5.0 Mg C ha -1 year -1 ) could assimilate more than 8 times more C compared with the other two species (Figure 1).
If we consider the three species in combination, the T2 and the T4 treatments were able to assimilate 14 times (2.1 Mg C ha -1 year -1 ) and 29 times (4.1 Mg C ha -1 year -1 ) more C, respectively, than the T1 treatment (0.14 Mg C ha -1 year -1 ) (Figure 1).Means of six plants (± SD); mean values followed by the same letters did not differ significantly between treatments at p ≤ 0.05 based on Duncan's test.Within columns, uppercase letters compare the effects of the different fertilization treatments on each species, and lowercase letters compare the three species within the same fertilization treatment.ROBERTO K. JAQUETTI and JOSÉ FRANCISCO C. GONÇALVES

LEAF NUTRIENT CONTENTS AND ASSIMILATION
Significant differences in leaf nutrient contents were found among both species and fertilization treatments.I. edulis had higher leaf N values than the other two tested species under all fertilization treatments, and for D. odorata and S. amazonicum, the highest values were observed under the T2 treatment (24.2 mg N g -1 and 21.5 mg N g -1 , respectively) (Table II).I. edulis had the highest leaf P values, and D. odorata had the lowest values.
The highest K values in the leaves were found for D. odorata under the T2 treatment.I. edulis had a higher Ca content than the other two species, and D. odorata had lower values than the other two species under all fertilization treatments (Table II).Additionally, I. edulis had higher leaf Mg values, but diff erences between I. edulis and S. amazonicum were only found under the T3 treatment.
As a N-fi xing species, I. edulis stocked a greater proportion of its N in the leaves, so the higher biomass production combined with the higher N content of I. edulis under the T4 treatment (78.6 g N plant -1 year -1 ) resulted in the assimilation of almost 7 times more N than S. amazonicum (11.7 g N plant -1 year -1 ) and 20 times more than D. odorata (3.9 g N plant -1 year -1 ) (Table III).If we consider an I. edulis monoculture in the available space, this species alone could assimilate 196.5 kg N ha -1 year - 1 in the leaf tissues, while S. amazonicum could assimilate 29.2 kg N ha -1 year -1 and D. odorata 9.7 kg N ha -1 year -1 .
I. edulis also stored more P than the other two species.The highest proportional P values for the species were found under the T4 treatment (3.5 g P plant -1 year -1 ), i.e., 8.7 kg P ha -1 year -1 in leaf biomass.I. edulis also had higher K values under T4, reaching 20.3 g K plant -1 year -1 (Table III), which amounted to an additional 50.7 kg K ha -1 year 1 .
Finally, I. edulis also stocked more leaf Ca and Mg per plant than the other two species; its highest Ca (26.5 g Ca plant -1 year -1 ) and Mg (4.6 g Mg plant -1 year -1 ) values were found under the T4 treatment (Table III).S. amazonicum had higher rates of Ca (6.5 g Ca plant -1 year -1 ) and Mg (1.1 g Mg plant -1 year -1 ) assimilation under the T4 treatment.

DISCUSSION
The potential C stocks of tropical trees have been receiving the attention of the scientifi c community due to the role that tropical forests play in the global C cycle (Gibbs et al. 2007).Because of this function, degraded areas in Amazonian ecosystems can be used to establish restoration forest plantations that can provide both ecosystem services (C stocks) and goods (stem production) (Lamb et al. 2005, Guariguata andBrancalion 2014).Furthermore, combining species that exhibit rapid growth and biomass production with later-successional species that are capable of providing good stem quality can benefi t local communities (Rodrigues et al. 2009).
The early successional species I. edulis can produce large quantities of biomass that, combined with the ability of the species to fi x atmospheric

Species
Means of six plants (± SD); mean values followed by the same letters did not differ significantly between treatments at p ≤ 0.05 based on Duncan's test.Within the columns, uppercase letters compare the effects of the different fertilization treatments on each species, and lowercase letters compare the three species within the same fertilization treatment.

TABLE III
Macronutrient assimilation rates in the leaf tissues of the three tested species in a forest restoration plantation in an area degraded by the Balbina Hydroelectric Dam, AM -Brazil.T1 -no fertilization, T2 -chemical fertilization, T3 -organic fertilization, and T4 -organic and chemical fertilization.It seems that the ecological group of a species determines its biomass production and C storage; early successional species such as I. edulis have higher photosynthetic rates and C assimilation and can thus generate more biomass, especially in the early stages of development (Jaquetti et al. 2014).As the plantation matures, late-successional species can incorporate more C into their biomass, especially the stem component.

Species
In a study of tropical species at La Selva Biological Station in Costa Rica, Redondo-Brenes and Montagnini (2006) found similar C sequestration rates in a mixed, eleven-year-old plantation with mean values of 5.1 Mg C ha -1 year -1 , and other values similar to those of our T4 treatment (4.1 Mg C ha -1 year -1 ) were also found by Arias et al. (2011) in Costa Rica with mean values of 5.4 Mg C ha -1 year -1 for the six study species at the same spacing.Higher C sequestration was found in a Populus spp.plantation in China after 10 years of establishment; the greatest C sequestration occurred at a spacing of 3 x 3 m (1.111 individuals ha -1 ), at which the species exhibited mean values of 7.2 Mg C ha -1 year -1 (Fang et al. 2007).
According to Saatchi et al. (2007), C stocks in natural Amazonian forest ecosystems range from 50 to > 200 Mg C ha -1 ; in our experimental region, values range from 125 to 150 Mg C ha -1 .In the coastal Atlantic Forest of Brazil, C stocks in the aboveground live biomass can range from 94 to 127 Mg C ha -1 depending on the altitude (Villela et al. 2012). Feldpausch et al. (2004) found high C sequestration rates in regenerating secondary forest in the Central Amazon with mean values of 11 Mg ha -1 year -1 , which is similar to that of I. edulis in T4 (9.4 Mg ha -1 year -1 ).This is expected because secondary forests are primarily composed of early successional species.With the sequestration rates found in our study, it would take approximately 30 years to rebuild the natural C stocks of the degraded area, but new studies indicate that highly productive species, such as Clitoria fairchildiana, Inga edulis and Acacia mangium, have the potential to reduce the time needed to reach natural C stocks from 30 to 7 years or even less, depending on the composition of the species.
In the Amazon, where low soil fertility plays a crucial role in the development and biomass production of tree species, it is essential to choose the appropriate fertilization treatment by considering the costs and benefits of the various methods (Quesada et al. 2012).Although the organic fertilization treatments in our study provided better results for at least two of the three species (probably due to changes in soil humidity and temperature combined with chemical improvements), this method still has limitations (Ferreira et al. 2009).A large amount of biomass must be present in the surrounding area for incorporation, and the human labor requirements are also a limiting factor that makes organic fertilization more expensive than chemical fertilization (Denich et al. 2005, Joslin et al. 2011).An alternative approach to overcome these limitations is to intercrop N-fixing Fabaceae trees that will act as a nursery for the development of non-N-fixing trees (Nichols et al. 2001, Nichols andCarpenter 2006).
In highly degraded areas, choosing N-fixing species such as I. edulis is a great way to restore the soil N and C contents, which are directly related to the success of forest restoration (Hobbs andHarris 2001, Viana et al. 2014).The rapid recovery of other nutrients, especially K, P, Ca and Mg, is fundamental to creating the conditions necessary for the reestablishment of natural regeneration in degraded areas (Feldpausch et al. 2004).It is also important to note that larger amounts of biomass and C were found to be stored in leaves compared to stems and roots in this study, which can be explained by the duration of the experiment since later plantings will allocate more C to the stem than to leaves.Thus, this pattern is expected to change as a plantation matures.
The tropical climate conditions in Amazonia, which include significant amounts of precipitation, favor plant development, but there is a dearth of basic data about the development of tropical species in degraded areas (Santos Junior et al. 2006).Our study shows that tropical species exhibit significant C storage and stem production potential, but additional studies are necessary to determine the best silviculture practices, such as the combination and arrangement of species, the fertilization treatment (different combinations), spacing and other aspects, that will facilitate better management and make Amazonia forest plantations more competitive with others worldwide.

CONCLUSIONS
Higher sequestration of C and other macronutrients by early successional species is important for successful forest restoration, but mixing other, later-successional species can provide long-term benefits to fulfill restoration goals.As an early successional, N-fixing species, I. edulis stocked more C and other nutrients than the other two tested species.However, S. amazonicum and D. odorata are also of great importance in forest restoration, as they can provide good stem quality and promote the cycling of other nutrients such as K and Mg.Overall, the three study species showed significant C and nutrient sequestration in a degraded area and are recommended for use in forest restoration plantations in the Amazonian region.

Figure 1 -
Figure1-C sequestration potential of the three tested species and a mixed plantation under the diff erent fertilization treatments based on the total biomass (aboveground and root biomass).T1 -no fertilization, T2 -chemical fertilization, T3 -organic fertilization, and T4 -organic and chemical fertilization.
(Joslin et al. 2011, Lojka et al. 2012) followed by the same letters did not differ significantly between the treatments at p ≤ 0.05 based on Duncan's test.Within the columns, uppercase letters compare the effects of the different fertilization treatments on each species, and lowercase letters compare the three species within the same fertilization treatment.ROBERTO K. JAQUETTI and JOSÉ FRANCISCO C. GONÇALVES N, makes it a promising candidate for forest restoration plantations(Joslin et al. 2011, Lojka et al. 2012).S. amazonicum and D. odorata, which are later-successional species, are also of great ecological importance as they provide long-term biomass production and C storage in plantations.