Teak (Tectona grandis L.f.), the main plantation species (about 1.1 million ha) in Java, Indonesia, has been managed for over 100 years to produce high-quality timber. The plantations have provided many benefits not only for generating national income, but also for supporting the livelihoods of rural communities living in surrounding forests in Java. The plantations are managed by Perum Perhutani (PP), a state-owned forestry enterprise. Current forest management planning of teak plantations, however, has limitations in supporting sustainable forest management (SFM). It ignores the potential risk of destruction in determining annual allowable cuts, whereas forest destruction seems to be an inevitable problem in teak plantations. In addition, the existing harvest scheduling method only concerns with achieving sustained timber yields, while SFM demands for achieving sustainability of multiple forest benefits. Up to now, however, there is still lack of management planning tools for supporting SFM of teak plantations. This study therefore aimed to develop quantitative models for supporting multipurpose management planning of teak plantations at risk of destruction. The specific objectives of this study were as follows: 1) to propose an alternative method for estimating survival probability and destruction rate of teak plantations, 2) to develop models for estimating stand biomass that can be used to quantify carbon stocks of teak plantations, and 3) to develop alternative harvest scheduling models for optimizing multiple benefits (i.e., timber and carbon sequestration) of teak plantations at risk of destruction.

Chapter 2 introduces the study area, i.e., Kebonharjo forest management unit (FMU), Central Java, which is one of the 57 FMUs managed by PP. Teak is the main plantation species in this FMU, accounting for 71.2% (12,678.8 ha) of the total area (17,801.3 ha). The FMU has a dry season of 5 months (May-September) with annual rainfall of 720-1155mm, which is a favorable climatic condition for producing high-quality teak timber. The plantations are managed under a clear-cutting with artificial regeneration system. The current age-class structure of teak plantations in this FMU is imbalance in which 80.7% of the plantations are dominated by young stands (≦ 30 years old). Declining productive teak stands due to forest disturbances has becoming an emerging issue in this FMU as well as other teak plantations areas in Java.

Chapter 3 argues that the first attempt towards improving the existing harvest scheduling method is developing parametric models for estimating survival probability and destruction rate of teak plantations. There is, however, lack of reliable methods that can be used to estimate survival probability and destruction rate of forest stands. This study therefore proposed an alternative method, based on the theory of survival analysis coupled with forest register data, for estimating survival probability and destruction rate of teak plantations. The forest register data were obtained from the FMU for the period 1977-2007. Survival and destruction of the plantations were modeled using probability distribution models. Model parameters were estimated using the maximum likelihood estimation method designed for left-truncated and right-censored data. Results showed that survival probability and destruction rate of teak plantations for the period 1977-1987 were well modeled using Weibull model, while those for the periods 1987-1997 and 1997-2007 were well modeled using log logistic models. All models confirmed that destruction rates varied over stand ages and planning periods. The rates of stand destruction were relatively low (< 2% per year) in the period 1977-1987, but increased up to 3% and 14% per year in the periods 1987-1997 and 1997-2007, respectively. The highest rate of destruction mostly occurred in young stands, implying an alarming condition for the sustainability of teak plantations. Therefore, appropriate management activities to eliminate the risk of stand destruction are absolutely required to ensure the sustainability of teak plantations. The survival and destruction models developed in this study are useful for forest managers to evaluate the risk of destruction over a specific planning period and to support the development of alternative harvest scheduling methods that incorporate the risk of destruction for teak plantations in Java. The proposed method can also be applied to other regions, especially when only forest register data are available.

Chapter 4 discusses that teak plantations are not only provide financial benefits from timber production, but they also provide environmental benefits from carbon sequestration. Compared to timber benefits, quantification of carbon sequestration benefits of teak plantations is still lacking. This study therefore developed empirical biomass models that can be used to quantify carbon stocks of teak plantations. Stand biomass and other stand variables were derived from forest inventory data of the FMU. Linear and nonlinear regression models were used to develop four types of biomass models: volume-to-biomass, basal area-to-biomass, age-to-biomass, and age and basal area-to-biomass. These models were validated using an independent data set to assess their performances. Results indicated that teak stand biomass was accurately estimated using the volume-to-biomass model. This model provides a direct conversion of stand volume to stand biomass without the use of biomass expansion factors. While the accuracy of the age and basal area-to-biomass model was comparable to that of the volume-to-biomass model, the basal area-to-biomass and age-to-biomass models were less accurate. Differences in the number of trees and mean diameter, even within a certain stand age, were possible causes of the high variation in stand biomass that resulted in the weaker performances of the basal area-to-biomass and age-to-biomass models. In addition, this study confirmed that site index may not be a consistently reliable indicator for site productivity in term of forest biomass, because teak stand biomass was more strongly affected by differences in stand density (either number of trees or basal area) than by variability in the site index. Depending on the availability of input data, at least one of these models will be appropriate for estimating teak stand biomass, especially when tree-wise data are not available. These models should prove quite valuable in supporting the multipurpose management of teak plantations.

Chapter 5 further discusses the limitation of harvest scheduling method that currently used by PP in the management planning of teak plantations. The existing harvest scheduling method is not only unable to incorporate the potential risk of destruction when PP determine annual allowable cuts, but it also less support for multipurpose management planning of teak plantations. This study proposes an alternative harvest scheduling model for optimizing multiple benefits (i.e., timber and carbon sequestration) of teak plantations at risk of destruction. The proposed model, which incorporated the destruction and biomass models (developed in Chapter 3 and 4, respectively), simulated the dynamic of age-class structures from one (a 5-year) period to another and then optimized harvest levels throughout a planning horizon using linear programming models. The optimal harvest levels were sought for three management scenarios: optimizing harvest volumes, optimizing net present values (NPV), and optimizing NPV while increasing carbon stocks. Each scenario was evaluated using four destruction rates (i.e., zero, low: 4.4-6.3% per period, medium: 9.4-15.9% per period, and high: 18.5-49.6% per period) and three cutting-age limits (i.e., 51, 61, and 71 years). The results showed that increasing destruction rates resulted in the reduction of harvest levels, financial benefits, carbon stocks, and led to imbalance forest structures. Meanwhile, reducing cutting-age limits increased harvest levels and financial benefits, but resulted in greater removal of carbon stocks and mature stands (especially when destruction rates were high). These findings suggest that instead of reducing cutting-age limits (implying shortening rotation lengths), eliminating or minimizing forest destruction seems to be an appropriate management strategy for ensuring the sustainability of teak plantations. In addition, incorporating carbon sequestration benefits (i.e., increasing carbon stocks by 10% from period to period) into the management objectives resulted in the reduction of financial benefits, but such multipurpose management scenario produced better forest structures than timber management. This study suggested that the proposed harvest scheduling model provides greater flexibility for forest managers to develop appropriate management scenarios at risk of destruction; hence, it can be used as an alternative harvest scheduling model for supporting multipurpose management planning of teak plantations.

Chapter 6 concludes that this study provided scientific contributions for improving the existing management planning of teak or other plantation forests in Java. Specifically, this study proposed an alternative method for estimating destruction rate, developed empirical models for estimating stand biomass, and proposed an alternative harvest scheduling model for integrating multiple benefits (i.e., timber and carbon sequestration) of teak plantations at risk of destruction. Although the results of this study might be specific for the study area, the proposed methods are general and applicable for other areas because plantation forests management in Java has some similarities.

インドネシア・ジャワ島においてチーク(Tectona grandis L.f.)は100年以上の造林の歴史があり、国営の林業公社によって大規模な経営が行われている。高品質の木材を生産するばかりでなく、多様な公益的機能を発揮するため、その人工林を適切に管理することは産業としても地域住民の生活の視点からも極めて重要となっている。ところが近年、同国の政治経済が不安定化したのを契機に、チーク人工林の持続的管理に懸念が生じている。本研究は、長期間にわたるチーク人工林経営に外的要因による森林破壊等のリスクを織り込んで、森林の多面的機能を高める管理計画を策定し、その評価を定量的に行ったものである。本論文は全6章からなる。

第1章では、研究の背景と目的が記されている。森林管理における木材生産と公益的機能はいわゆる伐ること・伐らないことの間の便益のトレードオフとしてしばしば議論されてきたことだが、それは少なくとも森林が順当に成長することを前提としたものであった。ジャワ島では違法伐採や森林火災、家畜の林内放牧などの影響により、チークの若齢林が重大な被害を受け、復旧再造林されるケースが多発し、森林管理が計画通りできない状況にある。森林破壊が防げない現状では、それを織り込んで計画を策定すべきことを指摘した。これは多くの途上国に共通する事情であるが、研究で考慮された例はほとんどない。

第2章では、本研究の対象地と用いたデータの概要が述べられている。対象地はジャワ島中央部北側のKebonharjo経営区約1万8千haである。ここは他に比べてデータが蓄積されており、1977~2006年の30年間の齢級別森林面積の移り変わりや年度別収穫量などの長期間にわたるデータが利用可能である。チークは陽樹で苗の植栽時に間隔を大きくとるため、始めの2,3年間は苗間でツンパンサリと呼ばれる間作農業が営まれている。

第3章では上記のデータをもとに、チーク人工林植栽後の経過年数ごとの森林残存率・破壊率が寿命関数の理論を援用して考察されている。データの存在する30年間では、最初の10年間で破壊率は低かったものの、次の10年間では破壊率が中程度に高まり、最近の10年間では著しく高まったことが分かった。この結果として経営計画では60~70年の伐期を想定しているにも関わらず、森林面積の80%以上が30年生以下に集中することとなった。破壊率を低く抑えなければ高齢級の森林面積を増やすことはできないことが明らかとなった。

第4章ではチーク人工林の環境機能のひとつとして二酸化炭素吸収機能に着目し、森林の総バイオマスを林分レベルの統計量から推定する方法を探求した。対象地域は森林破壊を受けやすいため現存する林分でも密度水準が様々で、従って統計量に密度の情報を含む林分蓄積や断面積合計を用いた推定式では精度が高かったが、林齢などは変数として有効でなかった。これらの結果は同国内の他の地域や樹種にも当てはまると考えられた。

第5章では、林分が順当に成長した場合の木材生産と二酸化炭素吸収量に森林破壊のリスクを組み込んで、多様な条件下で価値の最大化を試みた。すなわち3通りの破壊率、3通りの伐期齢、4通りの経営目標を設定し、それぞれの組み合わせで森林全体を2回循環させる多目的計画を解いて現在価値を比較した。その結果、低い破壊率の下では長い伐期齢が相対的に有利で、木材と炭素の相対的価値の大小によって最適な経営目標も異なることが定量的に明らかとなった。

最後の第6章ではこれらの結果をまとめて考察し、特に森林の破壊率が高い水準ではいかなる経営目標も大きく損なわれることを強調して示した。本研究では木材生産以外の環境価値として二酸化炭素吸収機能を取り上げたが、これ以外の公益的機能にも本論の手法が適用可能であることが述べられている。

以上,本研究はインドネシア・ジャワ島のチーク人工林を対象に、森林破壊のリスクを考慮に入れた多目的管理計画の策定手法について探求したもので、森林の多面的価値を様々な条件下で定量的に評価することに大いに寄与するものである。本研究で開発された評価手法は応用可能性が極めて高く、学術上応用上貢献するところが少なくない。よって審査委員一同は本論文が博士(農学)の学位論文として価値あるものと認めた。