Economic analysis of replacement regeneration and coppice regeneration in eucalyptus stands under risk conditions

,


INTRODUCTION
The use of Eucalyptus species is very popular in Brazilian forest enterprises, and knowledge of the optimal harvest age and optimal stand regeneration season is key to ensuring project success (REZENDE et al., 2005).According to Rodrigues et al. (1999), these forest enterprises seek to maximize the return on capital invested, regardless of the management being adopted, whether replacement regeneration or coppice regeneration.
The decision on whether to adopt replacement regeneration or coppice regeneration of a forest stand is infl uenced by several factors which include land availability for new forest crops, changes in project end use, oscillations in demand and technological advancements.
In some cases, coppice regeneration might be the most profitable alternative, while in other cases replacement regeneration might be the answer, due to increased yield potential through genetic improvement.Use of genetic improvement allows selecting progenies and productivity gains.In studies about genetic gain for Eucalyptus clones in Minas Gerais state, Botrel et al. (2007) found 6.98% gains in density and 6.89% gains in dry matter.According to Volker et al. (1990), estimates of volume gain for Eucalyptus globulus in Tasmania are in the range of 7% to 17%.Borralho et al. (1992) found a 20% to 47% gain in dry matter, in a study about expected genetic gain for Eucalyptus globulus in Portugal.Rocha et al. (2006) found a 13% to 22% gain in diameter for Eucalyptus urophylla originated in Indonesia and planted in Minas Gerais.According to Trugilho et al. (2001), the greater the variability in the trait of interest among clones or progenies, the greater the estimated gains resulting from genetic improvement.
Conducting a viability analysis is the natural course of business in every venture, looking to fi nd the best alternative for capital investment.Studies involving economic assessment of forest management regimes usually address optimal harvest or rotation age, optimal stand regeneration season, ideal number of stump sprouts to be left in each planting hole when coppicing, number of coppice operations and economic rotation of the stand (NOBRE; RODRIGUEZ, 2001).
That said, projects are nonetheless subject to uncertainties that obstruct the decision-making process.Uncertainties involving forestry projects are even greater because they are combined with time of return on capital invested, being medium to long term.For successful forest planning, it is necessary to minimize uncertainties by converting them into risks.
Risk is the variability of returns associated with a given asset and risk acceptance is divided into three categories: indifference to risk, aversion to risk and preference for risk (GITMAN, 2002).It is thus related to the creation of scenarios and simulation of factors influencing project viability, providing input to help decision-making.
Traditional methods of uncertainty assessment in the forestry sector typically resort to sensitivity analysis, which is used where information available about each variable is scarce or nonexistent.This analysis only assesses the impact of one independent variable at a time.However, when more than one variable infl uencing the decision-making process are analyzed, it is necessary to try all possible combinations between them, which makes this analysis more complex.
The Monte Carlo method is a robust analysis method known to be a powerful and useful tool, as it involves a set of random variables inherent in the project, in order to verify existing risk prior to decisionmaking.This methodology analyzes the distribution of relevant variables, as captured by a given probabilistic representation (COELHO JUNIOR et al., 2008).
The objective of this study is to analyze the economic viability of replacement regeneration and coppice regeneration of eucalyptus stands, under deterministic and risk conditions.

Development of the economic model
Table 1 provides most likely costs for establishing eucalyptus stands, according to forestry companies in northern Minas Gerais state engaged in timber production for energy purposes.Costs are divided into: implementation, maintenance, harvest, transportation, carbonization, replacement regeneration, coppice regeneration and cost of land.
Cost of harvest was represented by the use of a feller buncher along with a grapple saw and a skidder, for timber harvest and log stacking at the edge of the stand.Cost of coppice regeneration comprised fertilization (35.4%), soil conservation (23.9%), ant control (9.95%) and stump sprout thinning (30.75%).The annual cost of land computed interest rate on land value.
Based on the above costs and considering the revenues from charcoal sale, cash fl ow was structured for the economic analysis (Figure 1).The fi rst situation considered replacement regeneration after harvesting the high-tree planted stand, assuming that the future cloned stand can be more yielding than the preceding stand, in other words, stand replacement will ensure gain due to genetic improvement.In the second situation, cash fl ow considered that, after harvesting the high-tree planted stand, sprout regeneration, also known as coppice regeneration, will be adopted in the stand.There is no genetic gain in this case.
For the situation of replacement regeneration, the following costs and revenues were considered: C 0 is cost of implementation, C 1 , C 2 , C 3 and C 4 are costs of maintenance, C 5 is cost of land, C 6 and C 11 are costs of harvest, C 7 and C 12 are costs of carbonization, C 8 and C 13 are costs of transportation, C 9 is cost of replacement regeneration, while R 1 and R 2 are revenues from timber sale at age 7 and 14 respectively.Figura 1 -Fluxo de caixa para o povoamento de alto fuste (0 aos 7 anos) e de condução da brotação (7 aos 14 anos).

Guedes, I. C. de L. et al.
For the situation of coppice regeneration, the following costs and revenues were adopted: C 0 is cost of implementation, C 1 , C 2 , C 3 and C 4 are costs of maintenance, C 5 is cost of land, C 6 and C 11 are costs of harvest, C 7 and C 12 are costs of carbonization, C 8 and C 13 are costs of transportation, C 10 is cost of coppice regeneration, while R 1 and R 2 are revenues from timber sales in year 7 and 14 respectively.
To enable a comparison of the two situations, a 14-year planning horizon was considered, which is the time required for one replacement regeneration and one coppice regeneration (Figure 1).
Revenue was based on the per ton price of dry matter (Pms), calculated on the basis of per cubic meter price of charcoal (mdc), in such way as to allow changes in basic density (Db), due to genetic improvement or coppice regeneration (WHITTOCK et al.;2004).The equation used for the calculation is given as follows: where Fc is the conversion factor from m³ of timber to m³ of charcoal and Pc is the price per m³ of charcoal.Matthews (1992) and Underdown and Bush (2002) argue that, as a result of an established root system, volume yield from coppice regeneration can be up to 25% higher than that of the high-tree planted stand.
Loss from mortality, on the other hand, and/or postharvest damage to stumps, can dramatically reduce yield in the coppice method in comparison to the high-tree planted stand (FARIA et al. 2002;MATTHEWS, 1992).
Yield of the stand adopting replacement regeneration, and yield of the stand adopting coppice regeneration, were calculated as follows: where: Pr melhorado is yield of the stand adopting replacement regeneration, Pr is yield of the original stand, gg is percentage of genetic gain, gv is percentage of genetic gain relative to volume gain, Pr talhadia is yield of the stand adopting coppice regeneration and %Pr is variation in yield of coppice regeneration relative to yield of original stand.
According to Albertsen et al. (2000), the cost of harvest in coppiced stands is higher than in original hightree stands.To adjust this situation, cost of harvest for the coppiced stand was increased by 5% in relation to cost of harvest for the original stand.Table 2 provides a summary of additional information used in the economic assessment.

Identifi cation of input variables
Opportunities and threats infl uencing the yield in each management regime were identifi ed, and input variables (inputs) included the yield of the high-tree planted stand, or original stand, and basic density.
A triangular distribution was used, the most likely value of basic density being 530 kg.m -3 for timber from the original stand (SCOLFORO et al., 2008).For timber from the coppiced stand, a 5% drop in basic density was assumed in relation to the original stand.For the stand subjected to replacement regeneration, it was assumed that timber density can be higher in the new clone, due to genetic improvement (SESBOU e NEPVEU, 1991).Economic analysis of replacement regeneration ... As regards productivity, the most likely volume value for the high-tree planted stand was assumed as 300 m³/ha, while maximum and minimum values were assumed as 20% above and below that respectively.

Identifi cation of the output variable
Both the deterministic analysis (risks are not taken into account) and the risk analysis adopted the Net Present Value (VPL) method as output variable, computed for an annual interest rate of 6%.According to Rezende e Oliveira (2008), VPL is given by the following equation: where Rj is value of revenues; Cj is value of costs; i is interest rate; j is period in which revenues or costs occur; n is project duration.

Risk analysis
The risk analysis was conducted running the Monte Carlo method so as to simulate values for the input variables and, as a result of the random values generated, to obtain values for the output variable.
Cash fl ows were structured, and 10,000 simulations were run for the output variable using pseudorandom numbers, in other words, a series of values was generated for the variable and, consequently, its simple and cumulative frequency distribution was obtained.

Deterministic economic analysis of managemen t regimes
Table 3 provides VPL results according to deterministic analysis (risks are not taken into account), indicating that all tested alternatives are economically viable.In the situation of replacement regeneration it was noted that, with improved genetic gain, volume yield increases and so does VPL.Likewise, for different yields of the coppiced stand in relation to the original stand, as yield increases, so does expected return.Rezende et al. (2005) found the same VPL pattern relating to yield in studies of optimum cutting age for Eucalyptus.
In the harvest season, should there be a clone available capable of producing the same yield as the hightree planted stand (zero genetic gain = G0), the VPL will be R$ 3,696.39.Assuming the yield of the coppiced stand is the same as that of the high-tree planted stand (T100), the VPL will be R$ 3,792.40,therefore coppice regeneration should be preferred over replacement regeneration.On the other hand, if the yield of the coppiced stand is only 90% of that produced by the original stand (T90), the VPL then drops to R$ 3,401.54, in which case replacement is a better option than coppice regeneration.
Should there be a clone yielding 10% more than the high-tree planted stand (10% genetic gain = G10), then replacement should be more profi table than coppice regeneration, providing yield from coppicing is not more than 10% the yield of the high-tree planted stand (T110).For genetic gain of 30% (G30) or over, in no event should coppice regeneration be more viable economically than replacement regeneration.This analysis is interesting for decision makers, as it provides return possibilities according to different management options prior to implementing them in the harvest season.For that, it is necessary to have prior information on expected yields of the coppiced stand and also on clones available for new crop planting.Rocha et al. (2006) reported 13% to 22% gains in diameter for Eucalyptus urophylla.Botrel et al. (2007), in a study with Eucalyptus clones in Minas Gerais, found a 6.98% gain in basic density and a 6.89% gain Guedes, I. C. de L. et al.
in dry matter.These results are strong indicators of the difficulty encountered today in attaining significant volume yield gains of wood or dry matter resulting from genetic improvement or from updated management and silvicultural techniques.

Economic analysis of management regimes under risk conditions
Results of Monte Carlo simulation for replacement regeneration and coppice regeneration are provided in Tables 4 and 5 respectively.No probability of negative VPL values occurring were noted in any of the simulations, in other words, there is no risk that the options being tested are economically inviable.
The values of standard deviation help confi rm that risks of economic inviability are virtually nonexistent.In Table 4, for instance, for the no genetic gain (G0) situation, a standard deviation of R$ 836.63 indicates that more than 5.6 standard deviations would be required for a zero VPL.For the 50% genetic gain (G50) situation, risks of negative VPLs are even slighter, requiring for that a drop of 7.6 standard deviations in VPL.
As is illustrated in Table 5, for the situation where the yield of the coppiced stand equals that of the original stand (T100), the percentile analysis indicated a 10% probability of VPL having minimum values of R$ 3,780.37 and a 90% probability of it having maximum values of R$ 5,867.67.
Mode values are lower than median values and these in turn are lower than mean values, indicating that the probability distribution of VPL values is approximately normal with positive skewness.
Figure 2 provides several genetic gain situations for the high-tree planted stand, and yields from coppicing relative to the original planted stand.The analysis considered the point at which the cumulative distribution function intersects the theoretical curve of frequency distribution.VPL occurrence probabilities were noted to range between 70% and 85% according to each different situation.Guedes, I. C. de L. et al.
The option providing the best risk-return combination is the 20% genetic gain situation (G20), as it combines a high probability of return occurrence (82.52%) with a high return (VPL = R$ 6,765.36).It should be noted, however, that attaining a 20% genetic gain with eucalyptus stands at present is a diffi cult task, due to the current technological stage of Brazilian silviculture.
The no genetic gain (G0) situation provides the second best risk-return combination.Here, the risk is lower (probability of return occurrence = 82.61%)than in G20, but then return too is lower (VPL = R$ 5,532.09).
The no genetic gain (G0) situation is really what happens in practice with several forestry companies, since replacement regeneration uses the same clone as that of the original stand.Where the yield of the coppiced stand being 10% higher than the yield of the original stand (T110) is considered a plausible alternative, one should be aware of the risk of that not occurring in practice as a result of technology defi ciency (operational risk).
In accepting the above risk, assuming that technology will enable yield from coppicing to be 10% higher than yield from the original stand (T110), coppice regeneration is a better option than replacement regeneration if there is no genetic gain (G0), that is, if the yield of the future stand equals that of the original stand.Considering that the yield of the coppiced stand equals that of the original stand (T100), the best option is replacement regeneration, if there is no genetic gain (G0).

CONCLUSIONS
Deterministic and risk analyses for the various options of replacement regeneration and coppice regeneration of Eucalyptus stands led to the following conclusions: -replacement regeneration is economically viable, even if the yield of the new stand is the same as that of the original stand; -coppice regeneration is economically viable, even if the yield is only 70% of the original stand (high-tree planted stand); -the best risk-return ratio option is substituting (replacement regeneration) the original stand with another yielding 20% more; -the probabilistic analysis running the Monte Carlo method revealed that invariably there is economic viability for the various replacement and coppice regeneration options being studied, minimizing uncertainties and consequently increasing confi dence in decision-making.

Table 1 -
Costs of charcoal production from eucalyptus wood in northern Minas Gerais.Custos de produção de carvão vegetal de madeira de eucalipto na região norte de Minas Gerais.Where C 0 is cost of implementation, C 1 , C 2 , C 3 and C 4 are costs of maintenance, C 5 is cost of land, C 6 and C 11 are costs of harvest, C 7 and C 12 are costs of carbonization, C 8 and C 13 are costs of transportation, C 9 is cost of replacement regeneration, C 10 is cost of coppice regeneration, while R 1 and R 2 are revenues from timber sales at age 7 and 14 respectively.
Source: Companies in the forestry sector producing timber for energy purposes, with stands in northern Minas Gerais.mdc* = cubic meter of charcoal

Table 2 -
Data used in the economic assessment.Dados utilizados na avaliação econômica.

Table 3 -
VPL for different percentages of genetic gain in replacement regeneration and different yields of coppice regeneration in relation to the high-tree planted stand (original).-VPL para as diferentes porcentagens de ganho genético do povoamento reformado e as diferentes produtividades da talhadia em relação ao povoamento do alto fuste (ou original).

Table 4 -
Descriptive statistics of VPL (R$/ha), considering different percentages of genetic gain of the high-tree planted stand.