Survival, growth and regeneration of forest species in mining areas in the Eastern Amazonia

Planting forest saplings is the most usual method for restoring areas degraded by mining. Therefore, the objective of this study was to evaluate the survival growth of planted forest species and spontaneous regeneration in post-bauxite mining areas. For this purpose, we sampled plots in recovery areas with ages ranging from five to nine years, in which the establishment, recruitment, mean annual increase in diameter and total height of the species were evaluated. The species were further classified for conservation status and origin. Of the 156 species found, 18 presented less than 25% of survival and were therefore not recommended for use in the areas, while another 22 planted species showed more than 55% survival, indicating that they can be used. Five species were registered with some threat level, another nine species were registered as exotic to Brazil and four to the Amazon. In general, 35 species were classified as suitable for planting, with an acceptable mortality rate and satisfactory growth. Thus, it was possible to select those most suitable for planting the post mining area through an evaluation of the survival rate and growth of the species.


INTRODUCTION
The Brazilian Amazon region is the third largest bauxite ore reserve in the world, with the state of Pará being the main producer of this raw material to obtain aluminium with 90.9% of the national production in 2014 [7]. Mineração Rio do Norte, Norsk Hydro and Alcoa are the main companies that exploit bauxite in Pará, which together contribute an average of 25% of the GDP (Gross Domestic Product) of the municipalities of Oriximiná, Paragominas and Juruti, generating a large contingent of direct and indirect jobs [1].
The bauxite extraction process to obtain aluminium in Brazil comprises the following phases: a) total vegetation suppression (shallow cut); b) topsoil removal; c) soil removal; d) removal of sterile material which does not present any economic use for the company; and e) separation of the bauxite from the tailings [26]. After these activities, topographic terrain reconditioning is conducted with subsequent spreading of the previously removed topsoil throughout the superficial soil layer, and then proceeding to restore the degraded areas according to the recovery plan of degraded areas (PRAD) of the company.
Returning the topsoil after the mining is essential, since it potentiates the natural regeneration due to the great quantity of seeds of forest species [23,27] and improves the physical, chemical and microbiological properties of the soil [10]. The most commonly used method for forest restoration of areas degraded by mining is to plant seedlings of forest species associated with topsoil soil cover [18,27], since these areas present low or no resilience after mineral extraction, especially when the sources of propagules are far from the areas to be restored. However, besides being costly, this method coupled with a high mortality rate of seedlings in the early years in mined areas many times does not guarantee project success [18].
Associated with these problems, the lack of seedlings of different native forest species for planting has become frequent [30]. This is because the seeds of some species present low viability and germination difficulties in nurseries [21], which may be the determining factor for the high mortality rates recorded in the field, especially in the early years where some species present up to one hundred percent mortality [27]. The mortality of seedlings not only increases the restoration cost, but also can expose the soil, which can lead to an erosive process. However, this obstacle can be minimized with the correct selection of species for restoration, considering their adaptability, information on rapid growth [34], production of resources to attract wildlife [4] and high litter production [24,33] for soil cover. In addition, it should also be considered that a total modification of the environment will provide micro sites which are totally different from those prior to suppression, and that these environments may favour a particular group of species [9].
Thus, it is necessary that the selection and planting of species consider information about establishment and growth in sites near the mined areas, increasing the chance of survival of the seedlings after planting, thereby reducing the implementation costs. For this reason, continuous monitoring of the performance of planted species and natural regeneration in mining areas [3] is fundamental. It is possible to diagnose species adaptation, survival and growth through monitoring data, and then use this information to establish groups with similar growth characteristics and suitability for initial implantation [9,27].
In this context, we aimed evaluate the survival, growth and entrance of naturally regenerating forest species after planting in bauxite mining areas in the Eastern Amazon.

Study area
The study was developed in areas of the Mineração Paragominas S.A. company of the Hydro group, located in the northeast of the state of Pará, in Platô Miltônia 3 (3°15'38"S, 47°43'28"W) at an altitude of 150 m, 70 km from the municipal headquarters of Paragominas-PA.
The predominant vegetation is classified as Submontane Dense Ombrophylous Forest [13] and the main soils are characterized as Yellow Latosols, Yellow Argisols, Plinthsols, Gleysols and Neosols [25]. The climate in the region is "Aw" (tropical humid) according to the Köppen-Geiser classification, and the average temperature is 26.3°C with annual rainfall of approximately 1,800 mm, with the rainy season occurring from January to May with relative air humidity around 81% [2].

Study ecosystem
For selecting the species, we chose five areas under recovery with different years of implantation from 2009 to 2013 were selected (Table 1). A method of planting seedlings of native forest species with an average of 147 species was used in all areas, adopting spacing of 3 × 3 m (1,111 seedlings.ha -1 ). The number of individuals per species was different, as the number of seedlings per species varied according to the practicality of production, availability of seeds and also the absence of operational control at the time of planting. The following procedures were performed in all areas prior to planting seedlings: a) reconditioning the terrain; b) subsoiling with application of natural reactive phosphate (33% total P2O5 and 10% P2O5 soluble in citric acid 2%) at the bottom of the groove; and c) topsoil spreading. Then, 800 kg ha -1 of dolomitic limestone was applied over the total area and 200 g of NPK fertilizer (06 30 06) and micronutrients (0.5% B, 0.5% Cu, 0.5% Zn) were applied per pit. Additionally, 2.5 kg of organic compound were used per pit, consisting of vegetal soil, charred açaí pits, charred rice straw, tanned and ground chicken litter and tanned and ground sheep manure in the proportion of 20% of each component.

Species sampling and classification system
In order to quantify the individuals planted in the five areas, we installed plots of 40 × 50 m (1,000 m²) with subplots of 10 × 10 m (100 m 2 ) to evaluate the natural regeneration of tree and shrub individuals ( Table 1). The number of plots is variable per year of implantation according to the total size of the reclaimed area.

Identification of tree and shrub species
The tree and shrub species were identified at the species level and tagged with aluminium plates. When it was not possible to identify the species in situ, plant material was collected for scientific determination in the Felisberto Camargo herbarium of the Federal Rural University of Amazonia (UFRA). Botanical samples were sought for all registered species to certify the adopted scientific determination. The scientific names were corrected by the portal Reflora -Flora do Brasil 2020 -Algas, Fungos e Plantas (http://floradobrasil.jbrj.gov.br/) [12], adopting the APG IV classification system [5].

Absolute density of planting and natural regeneration
We calculated the absolute density (AD) of the individuals that were planted, and the natural regeneration by the ratio of the number of individuals by the area in which they were sampled [32]

Classification of species according to conservation status and origin
We classified the species according which conservation status [14,20]. Those that are exotic in Brazil and the Amazonia biome were also registered.

Mortality, recruitment and entrance
Mortality of trees and shrubs of a given species was determined by dividing the number of dead individuals by the number of living individuals [32] (Equation 2).
x 100 (2) In which: MRi (%) = mortality rate between two evaluations; ni = number of dead trees and shrubs of the i species between two successive evaluations; Ni = number of live trees and shrubs of the i species in the first evaluation.
In addition, we evaluated the recruitment and entrance of natural regeneration (Equation 3), with all areas being evaluated from May 2013 to November 2017, thus equating to 4.5 years of monitoring. For the recruitment, we considered the new individuals of a species which were already present in the area and the entrance of individuals of new species appearing [15].
In which: RRi (%) = natural regeneration rate of recruitment and entrance; A0 = absolute abundance in the first sampling of the i species; A1 = absolute abundance in the second sampling of the i species.

Increase in diameter and height
The growth data are the results of semi-annual flora monitoring, from which we obtained the DBH values (Diameter at Breast Height, measured at 1.30 m from the soil level for individuals with stem height > 1.30 m), and/or stem diameter at 0.10 m from the soil (for individuals with stem height ≤ 1.30 m) and total plant height (Ht). The DBH was measured with a tape measure and the height using a measuring stick graduated in centimeters of 4 m in size. The last measurement of the growth variables was performed in November of 2017 and the obtained values were then used to calculate the Mean Annual Increase (MAI) in DBH and height (Equation 4 and 5 respectively).

Classification of species in terms of planting suitability
We categorized planted species and natural regeneration into four mortality classes [27], in which the mortality of a species could: a) very low -when the mortality rate is lower than the relative value of the mortality mean of all species sampled subtracted from the standard deviation; b) low -relative mortality below or equal to the mean and equal to or above the upper limit of the very low class; c) high -relative mortality above the mean and equal to or below the lower limit of the very high class; and d) very high -relative mortality above the sum of the mean plus standard deviation. We excluded exotic species from Brazil from this classification, as well as those which obtained a total of less than five individuals, providing a higher reliability of the evaluations. We also categorized the MAIDBH and MAIH of each species into four classes: a) very low; b) low; c) high; and d) very high, following the same criteria used in the mortality description. The survival rate is considered the variable of greater interest in relation to the MAI, since some species present slow growth in relation to others. For this reason we attributed weights to the variables, with mortality receiving a weight of 2, and the MAIDBH and MAIH a weight of 1. Thus, we classified the MAI as: "very low = 1", "low = 2", "high = 3", "very high = 4". The values with a weight of 2 for the mortality rate: "very high = 2", "high = 4", "low = 6" and "very low = 8". We calculated the weighted value by the sum of the values assigned to each variable, and the higher the value, the better the performance of the species and, therefore the more suitable it is for the initial planting.
Therefore, we classified the species considered suitable for planting, meaning with satisfactory performance, those with "very low" to "low" mortality rate and "very high", "high" and "low" MAI in diameter and height. We also considered the low category for MAI in diameter and height in virtue of two species to considerably raise the average value of these variables. In addition, some species show limited growth ecologically.

Planting density and natural regeneration
In total, we found 125 species at the planting and 40 the natural regeneration, with 40 common species between the planting and the regeneration, therefore resulting in 156 species in general. Of these, the largest number of species was recorded in the area implanted in 2009, while the lowest in the area implanted in 2013 (Table 2). The absolute density of planted species was higher in the area with the shortest restoration time; in contrast, the number of regenerating species was small when compared to the other areas ( Table  2). We also observed that there was no increase in the number of species in the natural regeneration with the advance of the restoration, which can be explained as a function of the topsoil storage time, which was possibly different between the areas. The topsoil storage time presents an inverse relationship with germination, because the seeds end up germinating early, being able to also die buried due to an absence of oxygen and still be preyed upon by microorganisms [11].
Of the five areas, those of seven and eight years presented the lowest densities of planted species. However, this low density was supplied by natural regeneration, especially in the area of seven years. This can be considered an indicator of the local vegetation reestablishing and consequently of the successional processes, beginning the stratification of a future forest [16].

Conservation status of species
Of the 156 species found in the five areas, six (3.85%) presented some conservation status [14,20] (Table 3). These species are considered important for conservation of the flora of the Amazonia biome and for this reason we suggest that they be planted initially or in the vegetation enrichment phase, except for Amburana cearensis (Allemão) A.C. Sm. which does not occur naturally in the north of Brazil, and Handroanthus impetiginosus (Mart. ex DC.) Mattos, which presented an elevated mortality percentage, and should therefore be planted as a vegetation enrichment strategy. Plantations which aim at conserving endangered species should be carried out near areas of natural occurrence, seeking to recompose natural populations and the germplasm bank [28].
Exotic species are not recommended for restoring forest of degraded areas, as some of them may inhibit the growth and natural regeneration of native species or even occupy extensive areas, transforming the structure and consequently ecosystem functions. Among the exotic species in Brazil, Leucaena leucocephala is considered aggressive, inhibiting forest succession [29]. In addition to being exotic and invasive, Mangifera indica as well as L. leucocephala should be eradicated and are not recommended for implantation in other areas.
Regarding the exotic species of the Amazonian Biome, Libidibia ferrea appears in four of the five studied areas, with 95% survival (38 individuals), therefore constituting one of the species with the highest survival percentage, being adapted to the bio-edaphic climatic conditions of the region. This species is considered to be multipurpose and naturally occurs from Piauí to Rio de Janeiro [17], and is also recommended for restoring degraded areas in the state of Rio de Janeiro [22].

Mortality, recruitment and entrance
Species represented by more than four planted individuals presented a high mortality rate, reaching 100%, such as Annona mucosa, Euterpe oleracea, Parkia gigantocarpa and Khaya ivorensis ( Table 4). The initial planting of these species should be avoided, since the environmental conditions are not appropriate to their ecological characteristics. The high mortality rate exposes the soil to erosive processes, consequently promoting compaction and scarcity of initial organic matter [26]. In addition, it is important to consider intraspecific competition for resources, and those species that obtain better interaction with the environment tend to survive.
Schizolobium parahyba var. amazonicum and Bixa orellana were planted in the five evaluated areas, with 2,752 and 1,121 individuals respectively. These species have frequently been used because of the ease to obtain seeds and fast growth; however, the individuals of these species recorded an expressive mortality in this work rate (> 94%) ( Table 4), indicating that they should not be planted in similar features. Although S. parahyba var. amazonicum is a legume tree with an indication of microbiological association that promotes soil improvement, the obtained data indicate that its planting is not recommended in severely degraded areas, since the species has a tall and sparse crown [31], thus sparingly covering the soil and leaving it susceptible to breaking its stem in open areas. In addition, their seeds and fruits are not considered attractive to the wild fauna, which is an important characteristic that must be considered with the criterion for choosing the species.
Of the species we found at the regenerating, Trema micranta obtained a mortality rate of approximately 100%, and Solanum species resulted in more than 80% (Figure 1). These species have substantial initial contribution to restoring environments through litter deposition which increases the amount of organic material and improves the soil fertility and then by dying, thereby fulfilling its cycle in the successional process. For these reasons, we suggest that topsoil be managed in such a way as to increase the natural regeneration of these species.
Some regenerating species presented a number of individuals higher than ten and occurred in three or more areas, including Croton matourensis with 228 live individuals (0.53 ind ha -1 ) and a survival rate of 68% after 4.5 years of monitoring. It is a species that produces litter in abundance, aiding in water retention capacity [19], soil protection and promotes soil fertility by the availability of nutrients.
Our results showed that, only three species from planting were established in the natural regeneration, namely: Handroanthus ochraceus (Cham.) Mattos, Handroanthus sp. and Byrsonima crassifolia (L.) Kunth. were found in all the evaluated areas and is popularly known as murucizeiro. The fruits of this species serve as a food source for birds and large mammals which probably disperse the seeds in the most distant areas [8].

Increase in diameter and height
The average MAIDBH values varied from 1.20 ± 0.99 to 0.84 ± 0.61 cm yr -1 for the areas of nine and seven years of age, respectively ( Figure 2). When analysed only the planted individuals, the average ranges from 0.94 ± 0.46 and 0.69 ± 0.39 cm yr -1 . Even though it is smaller than in general, these values of the planted species are superior to the 0.48 ± 0.41 cm yr -1 found at Mineração Rio do Norte -MRN, in Oriximiná after 13 years [27]. We found that natural regeneration species present higher average values than planted species because these plants are already adapted in these areas with investment of their resources initially for height growth and later in diameter [6]. We also infer that lateral growth is usually accompanied by crown area growth, thereby covering the soil, and consequently protecting it against the elements.
Of the species with the highest survival rates, the MAIDBH and MAIH values were 0.83 cm yr -1 and 0.49 m yr -1 , respectively, for 22 species (Table 5). Of these, Senegalia polyphylla was one of the species planted in the five areas, with a total of 155 individuals (Table 5). Despite this, we observed that there has not been natural regeneration under the canopy of this species, with the subforest practically without vegetation, thus proving to probably be allelopathic, and with detailed studies on this therefore being recommended. There was more than 95% survival registered for Libidibia ferrea, Inga laurina, Astronium lecointei, Clitoria arborea, Guazuma ulmifolia, Astronium graveolens, Andira surinamensis, Hymenaea courbaril, and Chrysophyllum sparsiflorum (Table 5). However, except for L. ferrea and C. arborea, the absolute number of individuals for the other species was less than ten, with no guarantee that they will actually achieve satisfactory growth and adaptation performance.

Increase in diameter and height of regenerating species
Of the regenerating species in the planting areas, five had more than ten individuals and a mortality rate of less than 50% and MAIDBH that ranged from 0.70 to 2.08 cm yr -1 and MAIH from 0.67 to 1.41 m yr -1 ( Table 6). Three of these species occur in all evaluated areas, and Byrsonima crassifolia was recorded in planting and natural regeneration. The highest MAI values for Vismia guianensis and Croton matourensis were found in DBH and height, and the MAI value for C. matourenses is more than twice the mean value recorded for planted areas (Figure 2), and well higher than the mean values of the regenerating species, in which the MAI in diameter and height were respectively 1.37 cm yr -1 and 0.98 m yr -1 .
In a fragment of secondary forest near our study area found V. guianensis and C. matourensis were as being among the ten most representative species, proving to be abundant in the seed bank and adapted to the local conditions [18].

Species indicated for planting
Of the total species we found with more than four individuals, 35 (22.4%) species were considered suitable for planting in environments degraded by bauxite mining in sites with the same characteristics of the studied areas. These species were selected according to the highest weighted values (WV), calculated by the sum of the weights of mortality and MAI in diameter and total height (Table 7).
Very low (VL) and low (LO) mortality species recorded percentage values lower than 13.38% and between 13.38 and 51.57%, respectively. Of the species we are indicating seven were classified in the class of "very low mortality", indicating high adaptability, and therefore they should be prioritized in future plantations with characteristics like the studied areas.
Values we considered high in relation to MAIDBH (greater than or equal to 1.04 and less than or equal to 1.85 cm yr -1 ) were recorded for seven species, while only C. matourensis was classified as having a very high value (higher than 1.85 cm yr -1 ), also presenting the largest MAIH with 1.41 m yr -1 . Increasing values higher than the general average with high values of standard deviation led to few species being classified as having high to very high increases.
We note that species with unsatisfactory performance should not be totally excluded, especially those of the secondary or shade-tolerant ecological groups, but that their introduction in the areas is in a stepwise manner, meaning for enrichment activity after the initial planting.
Factors not directly associated with the species but which may directly affect their establishment in the areas should be jointly considered with the soil preparation and the appropriate period for planting, thereby guaranteeing survival rate and growth which ensure recovery of the areas.

CONCLUSIONS
The mortality rate was low to very low for 35 of the 156 evaluated species, and the growth in diameter and height was low to very high, and thus are indicated as priorities for the initial planting in new areas in which the goal is to reestablish the native vegetation. In addition, the cost of new projects will decrease with the production of seedlings and replanting, further increasing the efficiency of the restoration, making it successful.
We noted that high mortality rates (> 75%) for 21 species, forming a group of species which are not recommended for planting, at least in the implantation phase of the areas, while the survival rate for 32 species was higher than 60%.
We recorded thirty-six (36) species in the spontaneous natural regeneration, three of which were recruited and the others entered the sample during the monitoring, namely: Solanum crinitum, Solanum sp., Trema micrantha, Vismia guianensis, and Croton matourensis were recorded in all the studied areas.
Six species are on official endangered species lists, with growth rates and mortality according to the criteria adopted in the study, and thus their maintenance and inclusion in new areas should be fostered as a strategy for enriching and conserving recovery areas. Another nine species are on exotic species lists in Brazil and four in the Amazon Biome, and their planting is not recommended for ecological restoration to avoid biological invasions in the future which might compromise the restoration objective.

ACKNOWLEDGMENTS
The Universidade Federal Rural da Amazônia (UFRA) and its field technicians during the activities. The Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq) for granting a scholarship granted to the first author (141945/2019-2) and the Hydro through the Biodiversity Research Consortium Brazil-Norway (BRC). This article is number BRC0013 in the publication series of the Biodiversity Research Brazil-Norway.