Above-ground Biomass Recovery in Managed Tropical Forest in the Jari Valley, Eastern Amazon

Tropical forests play an important role in maintaining the regional rainfall regime and global climate, besides representing a significant stock of carbon. This study aimed at evaluate above-ground biomass (AGB) recovery, after reduced-impact logging (RIL) in a managed forest on the Jari River valley. The data were collected in 15 plots (100 m × 100 m) in the management area of the Jari Florestal Company. To estimate AGB we used a local equation adjusted for forests in the eastern Amazon. AGB before logging ranged from 157.9 Mg ha to 619.9 Mg ha, with an average of 362.5 Mg ha. AGB after logging ranged from 151.2 Mg ha to 632.8 Mg ha, with an average of 322.4 Mg ha. The time of monitoring of the plots and logging intensity were the main factors that influenced the recovery of the AGB. In 12 years after the RIL, the forest was able to recover its initial stocks of AGB, in places of low exploitation intensity.


Introduction
Forests, even to produce wood (managed), exert important role in the maintenance of the regional rainfall regime and the global climate, they emit large volumes of water into the atmosphere and represent a significant carbon stock in forest biomass (SFB & IPAM, 2011;Piponiot et al., 2016).
The recently introduced sustainable forest management (SFM) for production forests is beneficial for the maintenance of forest resources (Imai et al., 2009). The application of SFM in tropical forests increases the storage of biomass and carbon in these ecosystems. Although of the efforts of the regulatory institutions, SFM in the Brazilian Amazon continues to be performed without planning for future cutting cycles (Braz et al., 2015).
Management plans must be consistent with sustainable forest management and shall ensure the continuous production of forest products and services (Gourlet-Fleury et al., 2013). The costs associated with the planned operation are by the benefits of RIL: the reduction of the time of operation of machines and the work per cubic meter of extracted wood, reduction of wood waste and decrease of damage to the remaining trees (Jonhs et al., 1996;Sist & Ferreira, 2007;Miller et al., 2011).
The intensity of exploitation and the cutting cycle are the main determinants of extent that forest carbon stocks may be depleted (Zimmermann & Kormos, 2012). However, the effects of RIL on biomass loss are difficult to evaluate due to differences in cut intensity, a problem that is not widely recognized in the literature on managed forests (Martin et al., 2015). Reducing the intensity of exploitation could considerably reduce carbon emissions, favoring the recovery of biomass (Sist et al., 2014). For verification, whether or not there was exploration in the 15 plots field surveys we carried out at the MFA of Jari Florestal. We carried out two campaigns to know how many trees were harvested inside and (approximately 10 m away from the boundary of each plot) of the 15 plots evaluated. The first in August (8 field days) and the second in September 2015 (5 days).

Calculation of Above-ground Biomass
To estimate the AGB of each tree sampled and the forest as a whole, it was used the allometric equation of Lima (2015). The AGB calculation was performed only for live trees with DBH ≥ 10 cm, before and after the RIL. The equation (Equation 1) proposed by Lima (2015) where, AGB = Dry above-ground biomass (Mg ha -1 ); DBH = Diameter measured at height of the chest (cm) Total AGB of trees for each plot was quantified by sum of AGBi estimate of the trees for all the trees in the 1 hectare plots, according to Fox et al. (2010): We estimated the above-ground biomass (total AGB) in megagrams per hectare (Mg ha -1 ). For trees above the scope (DBH > 156 cm) of the equation of Lima (2015), the AGB was calculated by extrapolation according to the methodology of Medjibe et al. (2011) and West et al. (2014).

Commercial Volume Equation
The volume of commercial wood (m³ ha -1 ) was estimated with the volume equation (Equation 2) developed for dense forests in the Amazon by Nogueira et al. (2008).
where, α and β are -9.008 and 2.579 for trees < 40 cm DBH and -6.860 and 1.994 for trees larger; DBH = diameter at breast height (cm). (R adjusted, R 2 aj = 0.96; Residual standard error, RSE = 0.24) for trees < 40 cm DBH and -6.860 and 1.994 for larger trees (R adjusted, R 2 aj = 0.80; Residual standard error, RSE = 0.22) The rates of recovery of AGB and commercial volume were calculated with first and last measurements for each plot. The recovery percentage biomass and commercial volume, compared to that which existed prior to which was collected, was also calculated. Changes in mortality rates were calculated according to the method proposed by Sheil et al. (1995).
A Spearman correlation matrix was generated between the variables plots (n = 15) related to the intensity of exploitation and the rates of answers. For the most significant correlations, graphs were plotted and adjusted the linear regression models. Statistical analyzes were performed with the statistical software R (R Development Core Team, 2018).
The trees with diameter ≥ 60 cm of DBH, 242 trees, with an average of 16 specimens per plot. The maximum diameter observed was 288.7 cm, as shown in Table 2. Initial values of AGB were positively correlated with trees greater than 60 cm in diameter (R 2 = 0.38, n = 15, F = 8.14, p < 0.01, Figure 2). jas.ccsenet.
The Figure  15 plots, exploitatio The AGB of recruitm significant evaluation balance, in org Figure 2. Cor e 3 shows a po denoting grea on.  In a study in the eastern Amazon, the use of RIL techniques reduced substantially the effect of the selective cut on the residual AGB, favoring the increase of AGB recovered over the 16 years, the average annual increments in AGB (recruitment + residual growth of trees minus the mortality) were 2.8 Mg ha -1 year -1 in the RIL plots (West et al., 2014). On the other hand, in French Guiana, all 36 lots operated in a conventional system (without RIL) remained as sources of above ground emissions for 10 to 12 years following (Blanc et al., 2009).
In the monitored area, the majority of the plots with 4 to 8 years are suffering from losses of AGB, besides the biomass lost with the harvest, by the AGB lost by the RIL and associated damages. In research in the Eastern Amazon, after 20 years of monitoring a managed forest, Vidal et al. (2016) reported that the volumes of wood of large trees (≥ 50 cm DBH) were recovered in 81% of the plots with RIL and the use of RIL techniques accelerated biomass recovery rates.
In the 15 plots studied in the Jari Florestal area, the plots with the longest evaluation period, approximately 12 years, showed a good development in AGB after the exploration, mainly due to the low intensity of local exploitation of these plots and the longer monitoring time in relation to other plots. According to Sist et al. (2014), to promote an operating intensity of 03 trees ha -1 would reduce CO 2 emissions during the exploration and recover the initial biomass in a 35-year cycle, established by Brazilian legislation, in addition, considering that DBH trees > 60 cm abduct on average 53 kg C year -1 due to growth, 03 trees would hijack an amount of 5.5 Mg C ha -1 over of a 35-year cutting cycle. Martin et al. (2015) suggest that further evidence is needed to assess the differences between the impacts of RIL and conventional logging. Restrict the logging intensity is essential, both from the growth point of view and survival of remaining trees (future crops) and also to the ecological sustainability of forest (Sist et al., 2003;Mazzei et al., 2010;Putz et al., 2012;Sist et al., 2005). An efficient way to improve the carbon balance in tropical rainforest production could be the preservation of large trees, individuals with DBH > 100 cm (Sist et al., 2014).

Conclusion
The analysis showed that the forest can that the forest can recover its initial stocks of AGB in 12 years after the RIL, in places with low intensity of exploitation. In with high cutting intensity, in the following years after RIL the forest continued to lose biomass, in addition to the trees removed, due to damage and mortality of the remaining trees.
Biomass studies in managed tropical forests are important and more knowledge must be produced to strengthen policies and incentives to forest management and RIL techniques, which conserve standing forests and maintain carbon stocks in forest biomass, thus contributing to the reduction of greenhouse gases emissions to the atmosphere and the balance of the climate on the planet.