FOREST DYNAMICS IN THE EASTERN AMAZON WITH SPECIAL REFERENCE TO SAPOTACEAE SPECIES

Sapotaceae species are among the most dominant tree species in the Amazon forest and are high valued timbers. This paper aimed to evaluate the effect of Reduced Impact Logging (RIL) over a dense tropical rain forest with special reference to Sapotaceae species, to generate information helping decision making in forest management. In 1997, out of a 1050 ha forest area located in the municipality of Moju, State of Pará, Brazil, 200 ha were selected to be logged applying a logging intensity of 23 m ha and RIL techniques. To monitor forest dynamics, 22 permanent sample plots (11 ha sampling area) were established and all trees with DBH ≥ 10 cm were measured in years 1995, 1998 and 2010. In the whole observation period, mortality and ingrowth of Sapotaceae species were, respectively, 1.5% year and 1.0% year. Trees with full overhead light grew 0.39 cm year-1 and those completely shaded grew only 0.16 cm year. Mortality higher than ingrowth indicates that the population of Sapotaceae species still need a period longer than 13 years to recover equilibrium after RIL, despite showing increased diametric growth rates.


INTRODUCTION
Sapotaceae have great socioeconomic importance in the Brazilian Amazon for including commercial fruit species, like abiu (Pouteria caimito (Ruiz & Pav.) Radlk.and Pouteria macrophylla (Lam.)Eyma), and many timber species like maçaranduba (Manilkara huberi (Ducke) A. Chev.),FLORESTA,Curitiba,PR,v. 45,n. 3,jul (REIS et al., 2013).Survey of commercially exploited forest species, with regard to their recovery, is fundamental to indicate management actions aimed to conservation.In this context, the study of growth, ingrowth and mortality of trees are fundamental instruments to understand the dynamic of exploited species populations.These studies are also base to the development of growth and production models, which make possible the indication of sustainable management methods to produce timber in tropical forests.
Understanding growth of plants in tropical forests helps decision making regarding important silvicultural matters, like optimization of exploitation intensity, silvicultural treatments and periodicity of the cutting cycle, according with the growth rhythm of exploited species (JARDIM et al., 2008;BRAZ et al., 2012).This knowledge is useful as indicator of the forest ecosystem regeneration.In these terms, success of the second cutting is closely connected to restoration of density, which depends on balance between ingrowth and mortality.
Although monitoring of exploited forests in Brazilian Amazon have been running for more than three decades (REIS et al., 2010), a more detailed look over recuperation of managed species is still at an early stage.This work aimed to assess the effect of Reduced Impact Logging (RIL) on a dense tropical rain forest, with emphasis on Sapotaceae species, to generate information possibly assisting decision making on forest management of native species.The hypothesis of this experiment was the following: after logging, Sapotaceae species increase growth and ingrowth, and reduce mortality along time.
Climate of the region is Ami type (hot and humid), according to Köppen classification.Annual rainfall varies from 2.000 to 3.000 mm.Relative humidity is 85%, with average annual temperature of 26 °C.The relief is flat, with formation of dystrophic Yellow Latosol in different textures.Typology of the experimental area is Dense Tropical Rainforest in terra firme (LOPES et al., 2001).
In 1995, before this cutting, 22 permanent stands measuring 50 m x 100 m, were randomly planted and measured, corresponding to 11 ha of sampling area.All trees with Diameter at Breast Height (DBH) ≥ 10 cm were registered and measured.Stands were measured again in 1998 and 2010.Local experts in plant species recognition, known in Brazil as "parabotânicos", determined common names of the plants in loco.Botanic samples collected, were lately identified in the herbarium belonging to Embrapa Amazonia Oriental.Botanical identification reports were filed with the following numbers: 292010, 442010, 522010, 652010, 602011, 762011 and 342012.Mortality rate (% year -1 ) was calculated by the equation proposed by Sheil et al. (1995): where: M: yearly mortality rate; N 0 : number of individuals of the initial population; m: number of dead trees, not considering trees cut in logging; t: time in years.
Casual deaths were considered in the analysis of mortality rate, originated by various factors: pathogens attacks, parasites and herbivores; storms; damage caused by heavy rains or other natural causes (ROSSI et al., 2007a), and anthropic causes too, like damages inflicted during logging and transportation of logs.FLORESTA,Curitiba,PR,v. 45,n. 3,jul Ingrowths were considered as the number of trees that reached or passed the minimum diameter of 10 cm, in each measurement session, starting from the second one (ROSSI et al., 2007b).The ingrowth rate (% year -1 ) was obtained with the equation proposed by Sheil and May (1996): where: I: annual ingrowth; Nt: final number of surviving trees after time t; i: number of ingrowths; t: time in year.
In the analysis of growth, periodic annual increment (PAI) in diameter (cm year -1 ) was used, calculated with the equation: where: IP: difference between final and initial diameters; and t: time in years.
In the analysis of growth, crown illumination was divided in three categories: emerging Crown or Completely Exposed to light (CE), Partially Illuminated Crown (CPI) and Completely Shadowed Crown (CCS).Furthermore, for this analysis, the presence of Climbers (With Climbers -CC) or absence of Climbers (Without Cimbers, SC), were considered, because these variables have influence on growth of trees.In the evaluation of these variables, only alive trees, standing and with complete crowns were considered.
Diametric growth rate, ingrowth and mortality analyses were conducted considering the entire community, in other words all the species found in the sampling plot, while each Sapotaceae species was considered separately.
Analysis of variance and test-t were performed for independent samples, at 5% of probability for diameter PAI, with relation to illumination of the crown and infestation of climbers, and Tuckey test was run at 5% of probability to compare means.To applicate test-t and Anova, Kolmogorov-Smirnov normality test at 5% of probability was applied.When normality was not verified, the not-parametric Wilcoxon test (independent) at 5% of probability was applied.
Significance of statistics for ingrowth and mortality between the different periods (1995-1998 and 1998-2010) were verified by the chi-square test (χ 2 ) at 5% of probability.
Data were processed with the Tropical Forests Monitoring Software (MFT) developed by Embrapa Amazônia Oriental.The software Bioestat 5.3 processed statistical analyses.
Sapotaceae presented higher mortality than the ingrowth rates (negative balance) in the three considered periods (Table 1).From 1995 to 1998 it was 2.8% year -1 (M) and 2.1% year -1 (I); and from 1998 to 2010 there was a severe decline, 1.2% year -1 (M) and 0.8% year -1 (I).In the entire period (1998 to 2010), rates were 1.5% year -1 (M) and 1.0% year -1 (I).Greatest mortality than ingrowth indicates that observation time is still short to detect signs of population restoration with the same rate of the cuttings performed, this means that population of the family is not yet stable.
As it was expected, mortality and ingrowth rates were high soon after logging and became smaller as time went by, with ingrowth rate overcoming mortality rate in the second period of observation.This is because these pioneer species, with short lifetime, die when the canopy starts to close.Several authors (OLIVEIRA et al., 2005;AZEVEDO et al., 2008) already observed this behavior.Mortality rate was greater in diameter classes 15, 35, 85 and 105 cm, in the period from 1995 to 1998 (Figure 1), considering the whole community.Mortality was caused mainly by damage from logging, which represented alone 58.7% of total mortality.Comparing the first period to the second (1998 to 2010), mortality rate decreased, demonstrating decline of the dynamics, probably caused by the canopy's closure, despite high rates in 75 and 115 cm classes, due to natural causes.Diameter classes of Sapoteceae trees that presented greatest mortality rates in the first period (1995 to 1998) were 15, 35 and 65 cm (Figure 1).This rate diminished in the second period, as happened to the whole community, where classes with highest mortality were 15 and 35 cm.The main mortality cause in the first period was logging, that directly damaged trees.In classes 15, 35 and 65 cm, logging damages were responsible respectively for 64.1%, 50% and 54.5% of the total losses of Sapotaceae trees.
Considering all the Sapotaceae species, in the whole period, 17 presented positive balance between ingrowth and mortality, five negative and 16 had ingrowth equal to mortality, being static for 14 species and dynamic for two.Groups of species formed by Pouteria spp and Botanically Undetermined species, presented greater mortality due to inclusion of trees with uncertain identification in the two groups.
Sapotaceae species with negative balance between ingrowth and mortality (1995 to 2010) were P. laurifolia, M. paraensis, P. macrocarpa, M. acutangula and M. huberi.Among the cited species, only M. acutangula from 1995 to 1998, M. paraensis and P. macrocarpa from 1998 to 2010 (Table 1) presented positive balance or at least ingrowth equal to mortality (dynamic equilibrium).
Mortality of P. laurifolia in the period from 1995 to 1998 was mainly caused by damage from logging operations (64.3% of mortality).Mortality of M. paraensis and M. acutangula was 100% caused by logging.M. huberi presented 42.9% of mortality caused by logging consequences.P. macrocarpa had 100% of mortality due to natural causes.Out of the five species with highest density in 2010 (Reis et al., 2013), only M. huberi showed negative balance.P. virescens had negative balance from 1998 to 2010.M. guyanensis had 4.5% of ingrowth and showed no mortality from 1995 to 1998.
RIL caused different reactions concerning mortality and ingrowth, with positive balance considering the whole community, and negative when considering only Sapotaceae species (Table 1).Despite only five species presented negative balance, these are the dominant species of the family in the area, and were damaged by logging, besides the low ingrowth in species with positive balance and in species that continued with the same stock.Thus, RIL must consider recuperation (silivicultural treatments) in the cutting cycle, for species that continue with negative balance between ingrowth and mortality, to equilibrate the family in the area.

Growth
The periodic annual diameter increment of the whole community was 0.29 cm year -1 in the period from 1995 to 2010 (Table 2).Comparison between growth rates in tropical forests is hard, because it depends on various factors, like state of conservation (exploited or natural forest) and time passed from the last logging, intensity of the cuttings, vegetal composition, site and others.For example, Reis et al. (2010) found increments of 0.25 cm year -1 , 26 years after logging.Vidal et al. (2002), three years after logging, found an increment of 0.63 cm year -1 with application of reduced impact techniques, and of 0.37 cm year -1 with conventional logging.In the period from 1995 to 1998, PAI was higher (0.36 cm year -1 ) due to the canopy opening caused by logging, naturally decreasing because of the gradual closing of the canopy to 0.29 cm year -1 in the period from 1998 to 2010 (Table 2).Silva et al. (2001) observed that PAI is high in the first years after logging and decreases as times passes, due to the canopy's closing.
Entrance of sun radiation significantly influenced trees growth, detected by the different growth rates in diameter with relation to crown illumination intensity (Table 3), such that trees with directly illuminated crown grew more than the partially shadowed (0.37 cm year -1 ), and these also grew more than the completely shadowed by nearby trees (0.16 cm year -1 ).Clearing of crowns to receive solar radiation is, thus, a silvicultural measure to increase growth.Impact of logging helped growth by opening the canopy and decreasing competition, characterized by the analysis of illumination of crown related to growth.
Sapotaceae presented smaller PAI than the whole community, probably due to the species belonging to the ecological group of shadow tolerant species (Reis et al., 2013), characterized by slow growth rates.Increment in diameter was decreasing as time went by, with 0.29 cm year -1 from 1995 to 1998 and 0.27 cm year -1 from 1998 to 2010.In the whole period (1995 to 2010), average was 0.26 cm year -1 .Despite the family is composed by shadow tolerant species, environments with more illumination gave bigger growth rate (Table 3).
Out of the five species with biggest density, two showed significant difference in PAI (1995 to 2010), with relation to crown illumination (Table 3): M. huberi (0.44 cm year -1 ) and M. guyanensis (0.31 cm year -1 ).This total growth for M. huberi confirms results registered by Silva et al. (2001)  National Forest of Tapajós, PA, of 0.50 cm year -1 and Vitoria do Jari, AP, of 0.40 cm year -1 .Carvalho et al. (2004) pointed out that M. huberi obtained more diameter increment in the logged area than in the natural area, due to more illumination passing through canopy in the first case.PAI of M. huberi was higher in completely illuminated crown environments (0.62 cm year -1 ) and partially illuminated (0.46 cm year -1 ); with significant difference compared to environments with complete shadowing (0.17 cm year -1 ).Costa et al. (2007) in the National Forest of Tapajós also observed this behavior after logging, where plants of the species that received complete illumination grew 0.67 cm year -1 on average, while plants with partial illumination or totally shadowed had smaller growth, respectively of 0.58 cm year -1 and 0.26 cm year -1 .
M. huberi, just like other shadow tolerant species, responds positively to canopy opening in terms of growth (COSTA et al., 2007), being this an important consideration to support prescriptions of post-logging treatments, especially operations aimed to free crowns.Various authors demonstrated the post-logging treatments efficiency in growth increase and mortality decrease (OLIVEIRA et al., 2005;AZEVEDO et al., 2007;AZEVEDO et al., 2012).
Increment of trees of the whole community, without climbers, was 0.32 cm year -1 , 27% higher than in trees with climbers (Table 4).Influence of climbers on trees growth is a relatively well proved fact (VIDAL et al., 2002;COSTA et al., 2008).Species of Sapotaceae did not present significant difference in PAI between trees with and without climbers.This was due to the reduced presence of climbers totally covering Sapotaceae trees crowns.Costa et al. (2007) demonstrated that M. huberi trees presented bigger growth when without climbers (0.60 cm year -1 ) than when climbers were present (0.45 cm year -1 ), and indicated that cutting of climbers is a good silvicultural treatment for trees of this species, after logging has taken place.In this work, M. huberi did not present statistically significant difference (Table 4) between growth rates of trees without climbers (0.50 cm year -1 ) and trees with climbers (0.40 cm year -1 ).It is worth to highlight that, before logging, all climbers with diameter bigger than 2 cm had been cut, which could explain the little influence of these competitors detected when sample plots were evaluated.

CONCLUSIONS
 Reduced impact logging had influence on ingrowth and mortality of Sapotaceae species.Mortality bigger than ingrowth indicates that Sapotaceae species need more than 13 years to re-establish the original stock. RIL accelerated tree species growth, including Sapotaceae, in a short period after logging, indicating that silvicultural treatments can be applied to reduce competition for light in order to continue with accelerated growth rates.Adequate cutting of climbers before logging may reduce damages that caused mortality during RIL, as oriented trees cutting and construction of optimized paths and roads may help to reduce damages caused by logging and transportation of logs.

Figure 1 .
Figure 1.Mortality rate of the whole community (A) and of the Sapotaceae species (B) in different periods and by diameter class in a 11 ha sample of a tropical rain forest selectively logged in the Municipality of Moju, State of Pará.Figura 1. Taxa de mortalidade de toda a comunidade (A) e de Sapotaceae (B) em diferentes períodos e por classe de diâmetro em uma amostra de 11 ha em uma floresta explorada seletivamente, no município de Moju, estado do Pará.

Table 2 .
Periodic annual increment in diameter (PAIcm year -1 ), standard deviation (sd ±) and number of sampled trees of the Sapotaceae species and for the whole tree community in a 11 ha sample of a tropical rain forest selectively logged in the Municipality of Moju, State of Pará.N: number of trees.Tabela 2. Incremento periódico anual em diâmetro (IPAd -cm ano -1 ), desvio padrão (sd ±) e número de árvores amostradas das espécies de Sapotaceae e para o total da comunidade de árvores em uma amostra de 11 ha de Floresta Ombrófila Densa, explorada seletivamente, no município de Moju, estado do Pará.n: número de árvores.

Table 3 .
Periodic annual increment in diameter (PAIcm year -1 ), in the 1995 and 2010 period of the whole tree species community and of five Sapotaceae species in relation to crown illumination in an 11 ha sample of a tropical rain forest selectively logged in the Municipality of Moju, State of Pará.CE: crown emergent; CPI-Crown partially illuminated; and CCS-Crown completely shaded.Tabela 3. Incremento periódico anual em diâmetro (IPAdcm ano -1 ) no período de 1995 a 2010 do total da comunidade de árvores e de cinco espécies de Sapotaceae em relação a iluminação da copa em uma amostra de 11 ha de Floresta Ombrófila Densa, no município de Moju, estado do Pará.CE: Copa Emergente; CPI-Copa Parcialmente Iluminada; e CCS-Copa Completamente Sombreada.Anova.Means followed by the same letter in line of each species are not statistically different by the Tuckey test at 5% of probability.

Table 4 .
Periodic annual increment in diameter (PAIcm year -1 ), in the 1995 and 2010 period of the whole tree species community and of five Sapotaceae species in relation to climber infestation in an 11 ha sample of a tropical rain forest selectively logged in the Municipality of Moju, State of Pará.No climber (SC) and with climber (CC).Tabela 4. Incremento periódico anual em diâmetro (IPAdcm ano -1 ) no período de 1995 a 2010 do total da comunidade de árvores e de cinco espécies de Sapotaceae em ao grau de infestação de cipós em uma Floresta Ombrófila Densa, no município de Moju, estado do Pará.Sem Cipó (SC) e Com Cipó (CC).