Stability of Tropical Rainforest Margins

Tropical rainforests disappear at an alarming rate causing unprecedented losses in biodiversity and ecosystem services (Hughes et al. 1997, Noble & Dirzo 1997, Tilman et al. 2001, Achard et al. 2002) with Southeast Asia showing the highest rates of deforestation of any major tropical region (Sodhi et al. 2004). Despite an increased recognition of the value of these goods at national and international levels, rainforests continue to be seriously threatened by various forms of encroachments such as low-intensity harvesting of non-timber forest products by the rural poor, large-scale plantation forestry by the state or private actors, and the conversion of forested land by smallholder farmers. Transformation of ecosystems and changes in land use affect important ecosystem services and ultimatively human well-being (Robertson & Swinton 2005).

I summarize key findings from the Biological Dynamics of Forest Fragments Project, the world’s largest and longest-running experimental study of habitat fragmentation. Although initially designed to assess the influence of fragment area on Amazonian biotas, the project has yielded insights that go far beyond the original scope of the study. Results suggest that edge effects play a key role in fragment dynamics, that the matrix has a major influence on fragment connectivity and functioning, and that many Amazonian species avoid even small (<100 m wide) clearings. The effects of fragmentation are highly eclectic, altering species richness and abundances, species invasions, forest dynamics, the trophic structure of communities, and a variety of ecological and ecosystem processes. Moreover, forest fragmentation appears to interact synergistically with ecological changes such as hunting, fires, and logging, collectively posing an even greater threat to the rainforest biota.

Tropical forest moth ensembles are often extraordinarily rich in species and complex in structure. There is an increasing number of quantitative studies from all major tropical realms which now allows for a more rigorous assessment of the ways how such ensembles change along habitat transects from the natural forest towards the edge of large conservation areas, or in the course of forest recovery at such margins. Such knowledge is also essential for deciding if moths can be used for monitoring the biotic effects of forest disturbance on mega-diverse insect faunas at all, and which moth groups are the most suitable targets for this purpose. Using very large data sets from low (Mt. Kinabalu National Park, Borneo) and high elevations (Podocarpus National Park, Ecuador), we investigate how species richness and species composition of moths change at small spatial scales along gradients of disturbance at the edge of natural forest reserves. Local species diversity did not always decline, and sometimes even increased, along the gradients. Ensembles of moths from the families Arctiidae or Sphingidae were as rich as in natural forest, or were even more diverse in habitats close to the forest margin. Geometridae and Pyraloidea ensembles, in contrast, tended to be impoverished. Correlations of alpha diversity measures between moth taxa were often poor and thus did not allow for using one group as a ‘biodiversity indicator’ of others. Estimates of local diversity also depended on the temporal scale of assessment. Species composition was remarkably sensitive to habitat alterations at forest edges even at very small spatial scales, despite the high dispersal potential of many moths. Patterns of beta diversity were highly concordant across unrelated moth taxa with very different life history syndromes. Geometridae ensembles responded more sensitively to changes in canopy openness than arctiid moths and yielded more robust results with regard to sampling effects. Therefore geometrids are particularly promising candidates for environmental monitoring in tropical landscapes. Much of the faunal differentiation between forest and margin habitats was due to shifts in abundance relationships, and not to the presence or absence of ‘indicator species’. Therefore, abundance-based monitoring appears more appropriate to detect effects of environmental change on mega-diverse moth ensembles. Our results suggest that patterns of beta (rather than alpha) diversity are generally more meaningful to assess the impact of processes at forest edges relevant to nature conservation and landscape planning.

Timber harvesting is currently the most common commercial utilisation activity in tropical forests. Assessing the effects of logging on different aspects of biodiversity and general ecosystem properties is hence of prime importance if the few remaining areas of intact tropical forest are to be protected effectively and efficiently.

As we will point out in this chapter, tropical amphibian communities are an appropriate model system for studies on the impacts of human-induced environmental changes on the dynamics of complex biological systems. Here, we elaborate on patterns of diversity changes in tropical forest amphibian communities facing habitat alterations associated with selective logging in two globally important eco-regions (Côte d’Ivoire, Upper Guinea, West Africa and Guyana, the Guiana Shield, northern South America), and discuss findings from other previously conducted studies with similar focus.

A key statement that we stress on in this chapter is the fact that common measures of diversity, such as species richness and α-diversity only inadequately reflect processes of diversity change following anthropogenic disturbance. They also fail to describe actual impacts on the dynamics of complex biological systems. We argue that commonly used measures produce an incoherent and insufficient picture of diversity patterns and the underlying processes that shape these patterns. Thus, an understanding of higher levels of diversity, such as β-diversity and functional diversity (and hence compositional patterns) appears to be the key to effectively mitigating the impacts of human-induced disturbance on amphibian communities.

We show that the predictability of amphibian community composition depends on the respective level of anthropogenic disturbance imposed on a particular habitat. Hence, human activities that lead to changes in the structure of a forest, such as logging, not only alter simple system descriptors, such as the number of species in a given community, but rather alter the dynamics of the entire system. In this context, functional diversity is shown to be an important aspect underlying the actual mechanism that leads to the observed change of predictability patterns. Functional differences between species, rather than number of species per se appear to be the decisive factor in sustaining desirable ecosystem states and thus in maintaining important ecosystem services.

Because biological diversity appears to play a substantial role in ecosystem resilience required to safeguard essential ecosystem functions in the face of environmental change, we call for a critical revision of common diversity assessments approaches. We advocate the reconsideration of the uncritical use of widespread measures and descriptors of biodiversity on grounds of inconsistent patterns found throughout numerous studies, including those presented herein.

The complex nature of natural communities and the multiple aspects of biodiversity at different levels, make it necessary to incorporate processes acting on different organizational and spatial scales. When investigating the impacts of human-induced environmental changes on diverse vertebrate communities in the tropics, we should hence address compositional changes, as well as β-, and functional aspects of biodiversity. Special attention should also be drawn to the particular disturbance history of a given site and large scale cross-regional comparisons should be given priority. The consideration of these approaches in future studies would likely provide deeper insight in ecosystem processes at large scales and improve the effectiveness of current ecosystem management strategies.

In most tropical countries, rainforests are encroached by the local population and converted into agricultural land or pasture at constantly high rates. An increasing portion of this new agricultural land is abandoned within a few years and left for succession by woody plants. In many regions, already a considerable fraction of the forested area is covered by secondary forests formed by rapidly growing, short-lived pioneer tree species. In the Lore Lindu National Park on Sulawesi (Indonesia), like in other tropical countries, many of the remaining patches of protected primary forest are also offended at their margins by neighbouring villagers, who practise selective logging or local planting of crops inside the forest. Thus, disturbed old-growth forest and regrowing secondary forest after clearcut represent woody vegetation types, which are of rapidly growing importance throughout the tropics. Considerable research has focused on the effects of disturbance — natural as well as human-induced — on above-ground vegetation in primary forests, and on leaf and canopy characteristics of secondary as compared to primary forest trees. Only a few studies, however, have addressed the effects of disturbance and forest conversion on the below-ground compartment in wet tropical forests.

This review surveys the literature on fine root biomass (roots < 2 mm) and its distribution and activity in disturbed old-growth and secondary tropical moist forests and compares the results with non-disturbed old-growth forests. Particular attention is being paid to the relationship between disturbance intensity, above-ground forest structure and fine root system size and structure. Secondly, we focus on the age-dependence of fine root biomass in regrowing secondary forests and the time required for fine root biomass recovery after clearcut. The results are interpreted with respect to carbon storage in the root system of tropical moist forests and its response to disturbance and conversion.

The global carbon cycle is being perturbed by changes in land-use, especially in the tropics. This chapter compares surface soil organic carbon stocks, carbon mineralization rates and soil CO₂ efflux between an undisturbed forest and a clearing at Barro Colorado Island and between a pasture and plantation at Sardinilla, Central Panama. Our results on C cycling at two study sites with contrasting parent material and soil type were compared with other studies throughout the moist tropics. Differences in soil carbon stocks in the topsoil (0–5 cm) of the clearing (15 Mg C ha⁻¹) and the undisturbed forest site (22 Mg C ha⁻¹) were statistically not significant. Our inventory revealed that the highest carbon stock (29 Mg C ha⁻¹) was found under the native tree plantation, although at least part of this high value is site-related. Thus, no carbon change could be detected two years after the conversion of the site from a pasture into a native tree plantation. Soil CO₂ efflux rates at the pasture site (8 µmol CO₂ m⁻² s⁻¹) were significantly higher than in forest, clearing and plantation (5–6 µmol CO₂ m⁻² s⁻¹). Large CO₂ flux rates in the pasture might be explained by high belowground biomass production which leads to high root respiration rates. Our incubation experiment showed that pasture and clearing soil had a higher proportion of active pool carbon than plantation and forest. Higher amounts of active pool C indicate the existence of carbon readily mineralizable by microbes. Our results demonstrate that the active pool C is a good predictor of soil respiration. Thus, active soil organic carbon is a sensitive indicator for changes in soil organic carbon following land use change.

Due to high deforestation rates, it is likely that in many regions tropical rainforests will survive in protected areas only. These refuges have to be surrounded by buffer zones where low impact forest management is permitted. In fact, increasing parts of the remaining rainforest areas are encroached by different types of low-intensity forest use. However, until now only few data exist on how different practices affect forest structure. The objectives of this study were to analyze differences in aboveground forest stand structure as caused by different types of forest utilization in the margin zone of a protected rainforest. The study was conducted in a lower montane rainforest region (800 – 1140 m) in the vicinity of the village of Toro in Central Sulawesi, Indonesia. We assessed the structure of twelve forest plots (0.15 ha each) which represented four types of forest use which are widespread in the study region. These types were natural forest without major disturbance, forest extraction of small-diameter timber, forest with selective extraction of large timbers and cacao agroforest plantations under trees remaining from the natural forest. The tree basal area ranged from relatively high values in the natural forests (52.5 m² per ha) to 19.4 m² per ha in the agroforest, and was paralleled by a decrease in tree height. Stem density peaked in stands where large timbers had been extracted and abundant tree regeneration had resumed, and it was lowest in the agroforest stands. Canopy openness, as estimated from 30 hemispheric photographs per plot, was lowest in the natural forest after small timber extraction (7 %) and highest in cacao agroforest (15 %). Leaf area index (LAI), as estimated from the same photos, averaged to 6.2 m² m⁻² in the natural forest, 5.3 in the forests with small timber extraction, 5.0 in the forests with large timber extraction, and 5.3 on the agroforest plots. The latter result confirms earlier reports of relatively high leaf areas in the cacao agroforestry system. The relatively high LAI in forest stands after major timber extraction indicates a rapid recovery of leaf area after disturbance. We also analyzed the influence of stand structural parameters on rainfall partitioning into throughfall, stemflow and interception. In all forest stands we observed a close negative correlation between mean throughfall and tree height (r ² = 0.63) indicating a higher interception loss in taller stands. We conclude that local forest management at low to moderate intensities, as it is common in the margin zones of protected forests, has a significant and quantifiable impact on several forest structural parameters which in turn may influence ecosystem functions such as rainfall partitioning in the canopy.

Land use intensification at rainforest margins increases the pressure on the primary rainforest ecosystem, known for its high biodiversity and important functional role in global climate and carbon balances. Studying this system at different levels of disturbance increases our knowledge about its stability and mechanisms of diversity changes. While most studies investigating the effects of disturbance on animal populations focus on canopy communities, the present study reviews the driving factors for soil and litter arthropod communities. In addition, results from a study along a land use gradient in Central Sulawesi (Indonesia) will be discussed.

Within natural forests, litter composition, patchiness and especially litter abundance seem to be key factors driving litter arthropod diversity and densities, by creating microhabitats for many species and individuals. With increasing levels of disturbance in tropical forests, and with forest conversion into agroforestry and agricultural systems, soil/litter temperature and moisture levels become extremely important, leading to reduced densities and species richness. Changes of soil pH, a reduction of the litter habitat and increased anthropogenic disturbance via site management practices add to degradation. In the present study, litter arthropod communities were compared between three levels of forest use (none, small and large timber extraction) and two types of cacao plantations (under natural shade trees and under a polyculture of planted shade trees). The negative effects of increasing gap fraction and thus increasing temperatures accompanied by decreasing moisture levels and litter depth could be confirmed. Abundances of Collembola, Oribatida and Symphyla in litter and soil and the number of ant species in the litter declined with increasing land use intensity. In contrast, spider density and ant activity increased when natural forest sites were compared with the two types of cacao plantations. This can be explained by the shifting community structure of these two taxa, indicating that few, but abundant species of open habitats, such as Lycosidae, were replacing forest species. Surprisingly, the decline in microarthropod abundances along the land use gradient was not linearly correlated with decreasing moisture and litter depth or increasing temperature or litter quality (C-to-N ratio). Rather low microarthropod densities were already measured in little disturbed natural forest sites with small timber extraction. A similar pattern was observed for fungal abundances and root production, indicating that even small levels of disturbance influence the litter and soil system in tropical forests. On the one hand, changing microclimate, limited litter space, reduced availability of microbial resources and increased disturbance by plot maintenance inhibit higher densities of litter and soil animal populations. On the other hand, the simplified habitat structure favours populations of open habitats, especially predatory groups adapted to higher temperatures and/or lower soil moisture. It has to be investigated how a reduction of decomposers and an increase in predators could influence pest control and thus yields in these agroforestry systems.

This paper discusses inter-ethnic relations and patterns of land-use and landownership, and how ethnicity is defined, used to grant preference, and to allocate socio-economic resources, by analysing two forest-margin villages in Sulawesi.

The heavy influx of migrants has altered the proportion of local people to migrants in Sintuwu and Watumaeta, villages located on the eastern border of Lore Lindu National Park. In 2001, there were over one-third households of Bugis — the largest group of ethnic migrants. Social distance, inability to adapt and unwillingness to integrate (Charras 1993, Human Rights Watch 1998) were indicated by house-clustering of the ethnic groups.

As most agricultural land inside the village has been sold to migrants, mainly to Bugis, the locals have been transformed from landlords to landless, while the migrants became landlords. In a land-based economy, land scarcity leads to declining socio-economic security. Social destabilization occurs as ethnic frictions result from this shift in resource control.

To gain back their economic power and socio-economic security, locals adopted a strategy of land expansion into ecological buffer zone areas. By outsiders, particularly the State, this is viewed as an encroachment into the national park forest, while to locals this is an attempt to re-secure their economic base. Locals justified their actions by disputing the park boundaries. Subsequently their “illegal” actions were accepted and formalized when they received usufruct rights in the form of SKPL (Letter of Land Utilization) issued by the village head and KKM (Community Conservation Agreement), as was made by the Village Council with Lore Lindu National Park administrators.

Land expansion is also an attempt to re-stabilize relationship between locals and migrants in the face of ethnic conflict. The economic buffer zone has become a socio-political buffer that prevents local tensions from becoming open conflicts. In the absence of ethnic politics at national level, this local innovation in ethnic politics is a means to neutralize destabilization of interethnic relations. Such innovation, however, cannot go on endlessly, for once formalized the locals’ new economic space will most likely be taken over by the migrants, leading to a vicious circle of forest margin destabilization.

The implication of this is the need for a land-use modeling that accounts for social constraints of inter-ethnic relations and land-use systems. Encroachment is not only perceived as an economic action, but also as a political action, as forest has not only economic value, but is also an arena of ethno-political action. Disregarding this reality can trigger, and at times has triggered, ethnic conflicts.

Recognition of the importance of economically sound conservation strategies is one of the foundations of the Convention on Biological Diversity. Because of their exceptional contribution to global biological diversity, the conservation of the Central Sulawesi (Indonesia) rainforests is a particularly important case for a successful application of such strategies. One of the obstacles to the design and implementation of economically sound conservation strategies is the lack of knowledge on the economic value of non-market benefits generated by tropical forest ecosystems and the agricultural land use systems that replace them.

Tropical forests play a key role in the world carbon cycle and in maintaining biodiversity, but agricultural activities as well as the extraction of forest products are threatening these functions. Empirical evidence from developing countries suggests that forest products play an important role as a source of income for rural households, particularly for the rural poor. There is, however, still a lack of quantitative studies on the link between poverty and forest products. The research presented in this chapter seeks to fill the gap in general knowledge on the link between poverty, livelihood systems, and extraction of forest products. Considering as an example the vicinity of the Lore-Lindu National Park (LLNP) in Central Sulawesi/Indonesia, this chapter analyses the importance of forest products, especially for the rural poor, and identifies underlying factors which drive households into the forest. Moreover, the paper investigates similarities and differences in the use of forest products in the village of Toro, where an agreement with the national park authority on the use of forest areas exists, and in the research area at large, where such agreements did not exist.

In the vicinity of the LLNP, 76% of the households collect forest products, with firewood being the most important product. The sale of forest products contributes only 7% to the total household income of all households, with 17% of the households participating in this activity. Almost three-quarters of the income from forest products originates from the sale of rattan. Differentiating forest product income by wealth groups shows the importance of forest products, especially rattan, as a source of income for the poorest households. 21% of the total household income of the poorest households originates from the selling of forest products and 30% of these households reported to have income from forest products. Participation in the sale of forest products is influenced by the wealth of the household, the area of land owned, education, ethnicity, and access to road infrastructure.

Most native forests in Latin America are highly fragmented. In the mid elevation areas of Northern Latin America, the agricultural matrix is frequently composed of coffee. In this region, coffee has been traditionally cultivated under the diverse canopy of shade trees, representing a high quality matrix that can contribute to the social and ecological stability of the region. This agroforestry system has been proven to be important for biodiversity conservation. Studies over the last fifteen years have shown that shaded coffee plantations maintain a high diversity of vertebrates, invertebrates and plants. These organisms play an important role in the functioning of coffee agroecosystems. Shaded coffee plantations promote a high abundance and diversity of natural enemies that help to regulate herbivores, weeds and diseases. Shaded plantations also harbor a higher diversity of native pollinators which have been shown to contribute to higher coffee yields. Likewise, the diverse shade-tree component contributes to soil fertility and soil conservation and has been shown to contribute significantly to carbon sequestration. As a matrix, coffee agroforests also contribute to the conservation of biodiversity within forest fragments by promoting migration among fragments and facilitating a metapopulation structure. Three “sustainable” coffee certification programs have been developed to help farmers cope with the vagaries of the market: organic, fair-trade and biodiversity-friendly (or shade-grown). Although certified coffees still represent a small niche market, they have the potential to promote conservation and benefit the livelihoods of small producers. Especially under conditions of low international coffee prices, as those experienced in the first years of this century, these certification programs have contributed to the ecological and socio-economic stability of the coffee growing regions of northern Latin America.

The determination and evaluation of ecosystem services provides crucial information for a comprehensive strategy to preserve near natural habitats such as rainforest margins. We show results of an economic evaluation of coffee bee pollination services and pest control in two distinct tropical regions: a low human-impact area in Indonesia with continuous near natural forests neighboring agroforestry and a high-impact landscape in Ecuador with almost no forest fragments left. We evaluate bee pollination services comparing forest destruction scenarios, where coffee yields depend on forests providing nesting sites and foraging habitats for bees and present three novel approaches: first, we show how net coffee revenues depend on pollination services of adjacent forests considering berry weight in addition to fruit set, thereby providing a more comprehensive evaluation. Second, we combine our findings on pollination with an assessment of pest management affecting coffee production. Third, we determine net welfare effects of land-use changes including the fact that former forestland is normally used for alternative crops. In both regions, crop revenues exceed coffee pollination values, generating incentives to convert forest margins even if owners would be compensated for pollination services. The promotion of certified “biodiversity-friendly” coffee is a feasible option to maintain shade-coffee systems for conservation purposes. This is of special importance in high-impact areas where only small forest fragments remain. We conclude that a comprehensive economic analysis is necessary to adequately evaluate rainforest preservation for the enhancement of ecosystem services within a mosaic of competing land-use systems.

The ongoing loss of pristine tropical rainforests increases the potential importance of agroforestry systems for the conservation of tropical arthropod diversity. Shaded agroforestry systems can still support high levels of biodiversity, even resembling those supported by undisturbed forests, but intensively managed open agroforestry systems may cause severe losses in insect diversity. In this study we evaluate the conservation value of agroforestry systems for species richness and diversity (Simpson’s index) of four insect groups at natural forest sites and three different types of cacao-dominated agroforestry systems in Central Sulawesi, Indonesia. The agroforestry systems were characterised by low, intermediate and high diversity of shade trees. Each habitat type was studied with 4 replicates, i.e. 16 study sites altogether. We compared responses of solitary bees and wasps, dung beetles and lower canopy dwelling beetles and ants. These taxa represent diverse and functionally important insect groups: solitary bees and wasps act as crop pollinators or pest predators, dung beetles as decomposers of mammalian excrements and canopy dwelling beetles and ants include abundant herbivores and predators. High percentages of forest species did not occur in agroforestry systems, but diversity and species richness in agroforests remained as high as or even higher than in the forest lower canopy. Diversity, species richness and abundance of the functionally important dung beetles and canopy ants showed strong resilience against both forest conversion and changes in agroforestry management. Diversity, species richness and abundance of solitary bees and wasps as well as canopy beetles even seemed to profit from the effects of opening the upper canopy that was related to forest conversion and changes in shade tree compositions. On the latter two groups the effects of opening the upper canopy were stronger than effects of reduced shade tree richness. Based on our results, we recommend the inclusion of agroforestry systems with a diversity of shade trees in tropical conservation plans in addition to pristine forest reserves. Furthermore, regional differences in local agroforestry management contributed to between 31% (for dung beetles) and 58% (for canopy beetles) of the total species richness, which stresses the importance of conservation policies aimed at a diversity of habitat types on a broader landscape scale.

The large-scale exploitation and conversion of tropical forests causes growing concern about the continued existence of the rich biodiversity of these forests. In the framework of the interdisciplinary STORMA project in Lore Lindu National Park area in Central Sulawesi, we studied tree diversity in six different land use types in the margins of submontane rain forest: undisturbed forest, forest with rattan extraction, selectively logged forest, cacao forest gardens, cacao plantations with mixed canopy of planted trees, and cacao plantations with a monospecific canopy. By analyzing such a finely subdivided use gradient it was attempted to answer the question how human usage and biodiversity conservation in the study area may be reconciled. Tree species (dbh > 10 cm) were sampled in 24 plots of 0.25 ha in all six land use types (4 replicates each). In total, 251 tree species (143 genera, 59 families) were recorded. Number of tree species per 0.25 ha was 51–63 in primary forest and gradually decreased towards the studied cacao systems. However, when native and cultivated tree species were considered separately, significant differences were detected among plantation types in terms of tree diversity. Tree endemism in forest plots totalled ca. 15% and was in good accordance with endemism in woody plants of Sulawesi. The number of endemic species was strongly reduced in cacao systems, although percentage endemism did not decline significantly in cacao forest gardens. Roughly one third of tree species in the forest plots were of economic importance as commercial timber trees; timber diversity was little affected by moderate human use of the forest but was significantly reduced in cacao forest gardens and dropped to near zero in other plantation types. The mean basal area of 57 m² (36–80 m²) per ha in natural forest was lower than the previously recorded value from the study area but is still almost double as high as the mean value typical for tropical lowland forests in Southeast Asia.

The results of this study support the notion that tree diversity in the submontane forests of Central Sulawesi is unusually high and rich in large-sized timber trees, although tree size varies locally. Moderate human use of the forest ecosystems does not significantly affect tree diversity. We conclude that conservation of tropical tree diversity is compatible with human exploitation of tropical forest as long as a canopy of native trees is maintained. Future conservation policies in rain forest margin areas should therefore focus on developing measures aimed at sustainable use of the natural resources. Promotion of such activities may help to stabilize tropical rain forest margins in Central Sulawesi.

In many smallholder farming systems in the humid tropics, the slash-and-burn practice is used for land preparation. Increasing land-use intensity by shortening fallow periods often contributes to the degradation of the natural resource base of the fallow system, i.e. the fallow vegetation and soil. In the eastern Amazon region of Brazil, we therefore searched for ways to maintain the sustainability of the traditional fallow system and to adapt it to changing agro-ecological and economic conditions.

We identified two major agro-ecological constraints of the traditional fallow system with slash-and-burn: (1) high losses of nutrients and organic matter during the burn, and (2), if land-use intensity increases, fallow regeneration capacity declines.

As alternatives to slash-and-burn, we studied modifications to those practices recognized to be harmful to the ecological sustainability of the fallow system, i.e. mulching for the management of soil organic matter and fire-free land clearing with bush choppers to transform fallow vegetation into mulch. Mulching allows extending the cropping period, planting crops off-season, and modifying crop rotations. Additionally, biomass and nutrient accumulation of degraded fallow vegetation can be improved by enrichment plantings using fast-growing leguminous tree species.

Our socio-economic analysis focused on the implications of technology change on income and land-cover change at farm and field levels. Based on farm-household data collected from 270 randomly selected households, a farm-household level bio-economic model was developed to analyze the consequences of improved access to mechanized plowing and mechanical mulching for typical smallholdings in the study area.

Model simulations suggest that the costs of mulching are still very high compared to other mechanized land preparation technologies that provide similar economic benefits from the farmers’ point of view. Among the recommendations to reduce mulching costs is the use of simpler mulching equipment on areas with young fallows.

Technology scenarios indicate that countervailing policy measures are necessary if mechanized chopping and mulching is to bring about the desired ecological benefits.

Taxes on ecologically undesirable forms of land preparation, e.g. slash and burn, are promising policy options to promote chopping-and-mulching or other fire-free land preparation techniques given that they can be provided at costs that range between 60 to 110 Euros ha⁻¹. Tax revenues could be used for financing environmental conservation payments and/or a crop yield insurance that applies to crops that are produced using environmentally friendly production technologies.

This paper analyzes the comparative advantage of alternative forest management policies taking into account the subjective welfare judgment of individual rural households. The methodological approach chosen for this study is a financial farm household analysis combining both, economic data at the household level and natural science data of forest resources obtained in a research area in the Northwest of Madagascar. This interdisciplinary approach takes account of the interdependence between ecosystem dynamics and economic decision-making processes. By applying a poverty index, it is possible to make a further distinction between groups of rural households that are inherently different with regard to their economic potential and activities. The results of this study demonstrate that, due to their livelihood strategies, it is especially the poorest households that suffer most from a strict conservation approach, while better-off households would benefit more due to an improved provision of indirect forest services, particularly watershed protection. We also analyze the potential for compensating opportunity costs of restricted forest utilization through international transfer payments and the potential of implementing economic incentives for afforestation at the household level. In this context, we extend our analysis beyond mere economic aspects of resource management and provide further information how alternative management strategies can meet the increasing direct demand on forest resources, while also achieving the objective of biodiversity conservation. Lastly, the potential of forest resources to contribute to local and regional economic development in developing countries is discussed.

The establishment of tree crop plantations such as oil palms, coffee, or cocoa considerably contributes to the loss of tropical forests. Taking the case of cocoa production in Central Sulawesi as an example, this chapter investigates whether there is a potential for reducing deforestation by improving the productivity of tree crop plantations in rainforest margin areas by better crop management. Increased productivity would enable farmers to earn a living from a smaller area of land; thus, the expansion of low-productivity perennial cropping systems into forest land can be viewed as a waste of forest resources. In order to assess whether the productivity of the existing cocoa plantations in the research area could potentially be improved, the level of technical efficiency attained is estimated. A farmer is technically inefficient if he fails to produce the maximum output attainable for the level of inputs he uses. Technical inefficiency is caused by lacking know-how, for example with respect to the timeliness of agricultural operations such as weed control or crop hygienic measures.

After introducing the concept of technical efficiency (TE) and describing the method of Stochastic Frontier analysis to empirically estimate the level of TE, we show how this method was applied to our Sulawesi case. Using data collected in a sample of 202 farm households, we estimate separate Stochastic Frontier production functions for the two most important crops in the research area: cocoa, the primary cash crop, the cultivation of which has become widespread only in the 1990s, and irrigated rice (paddy) that has been grown for generations. Apart from estimating the levels of TE attained, we also investigate the influencing factors of efficiency in one single statistical procedure.

The estimated average TE in rice production is 77% as opposed to only 37% in cocoa cultivation; thus, the potential to increase production by improving crop management (not by increasing the level of input use!) is particularly large in the case of cocoa. The analysis of the factors influencing TE shows that poverty and illiteracy have an efficiency reducing effect in both rice and cocoa production. Agricultural extension services significantly increase efficiency in rice cultivation while this effect is not observed in cocoa production.

Acknowledging that increasing the productivity of perennial cropping systems in forest margin areas may also create incentives for deforestation, we then discuss the conditions under which enhanced productivity can be expected to have a forest saving or a forest clearing effect. We finally assess the likely outcome of improved TE in the case of Central Sulawesi and conclude that the comparatively low efficiency level currently found indicates a considerable potential for reducing deforestation by increasing farm incomes on already converted forest land, thus meeting both environmental and economic objectives. Policy interventions aimed at realizing this potential should include improved agricultural extension focusing on technical advice on the proper management of cocoa, but, at the same time, they need to control the influx of migrants attracted by the profitability of cocoa cultivation.

Human activity endangers tropical forests in different parts of the world. The conflicting interests of nature conservation on the one hand, and the livelihood of farmers living at the forest margins, on the other, clash noticeably in so-called hotspots of biodiversity, such as the Lore Lindu region of Central Sulawesi, Indonesia. Biodiversity generally decreases along a land use gradient from natural forest to agroforestry and annual crop systems. Thus, before solutions for a sustainable balance between conservation and the needs of people living at the forest margins can be sought, changing land use strategies and the factors that influence them must be analyzed. While similar studies often concentrate on economic indicators of land use change only, this chapter highlights the importance of two cultural realities, namely migration and ethnicity. We will demonstrate the great influence of these two factors on land use decisions and on the accessibility of land in the Lore Lindu region.

Based on results from previous qualitative studies, we selected three upland villages, which represent a continuum that leads, in terms of migration history, from a pre-transition village, to a transitional village, to a post-transition village. We used quantitative and qualitative methods for our analysis.

The results show a general land use change in the region, ranging from a strategy that places food first (e.g., rice) to one that places cash first, especially by the cultivation of cacao. This change cannot be observed to the same extent in the study region as a whole. It varies from village to village, depending on the share of households belonging to the Bugis migrants. Bugis as an ethnic group and as migrants have an enormous effect on the land use decisions of local ethnic groups in their respective villages. Strong ethnic networks among Bugis migrants play an important role not only with respect to their village preferences, but also with respect to the transfer of knowledge regarding agricultural management. A lack of interaction between local and migrant ethnic groups results in differences in the efficiency of cacao cultivation. While the Bugis migrants mainly buy their agricultural land from local farmers or, depending on the village’s migration history, from other Bugis households, members of the autochthonous ethnic group usually clear the primary forest. The findings of the present study reveal that an analysis of migration and ethnicity is a crucial precondition to finding sustainable solutions for ensuring that the margins of the rainforest in the Lore Lindu region remain stable. This chapter will emphasize that, in addition to economic considerations, an analysis of the cultural forces that influence land use changes cannot be neglected.

A great number of studies have been dealing with land-cover mapping of tropical regions using earth remote sensing technology recently. This is partly due to a growing number of operational sensor systems for both scientific and commercial use and also because of an increasing demand for land-cover information relevant to global environmental issues and international policy instruments (e.g. the Kyoto protocol). Within this context, the present article discusses the state of the art of data processing and analysis for the assessment of broad scale land-cover and land-cover change in tropical regions. Current global scale land-cover maps are compared with regional satellite mapping products (Landsat/ETM+) for a test region in the humid tropics of Central Sulawesi, Indonesia.

The paper suggests the land cover classification system (LCCS) to be used as the conceptual basis for future land-cover analysis in Sulawesi because it delivers a consistent and comparable scale-independent class structure for satellite image-based land-cover mapping and monitoring. The results of the comparative analysis of land-cover and land-cover change document the inhomogeneity, inconsistency and hence high uncertainty of existing estimates. The outcome of the harmonized and generalized land-cover products for two base years (1992 and 2000) indicates considerable disagreements in area estimates and spatial distributions of land-cover classes for a single date that in some cases exceed the detectable changes between years.

Future work aiming at a long-scale operational land-cover mapping of tropical environments has to account for (a) a further harmonization of existing and planned land-cover definitions and products, (b) the regional validation of products and (c) the implementation of a multi-level standardized technical and conceptual classification workflow for ecosystem mapping and monitoring in tropical regions.

Greenhouse gas and energy fluxes between the land surface and the atmosphere are important aspects for the evaluation of land-use options in tropical areas. Changes in vegetation cover alter the capacity to absorb carbon dioxide and solar radiation from the atmosphere and influence the magnitudes of latent and sensible heat flows to the atmosphere. If happening at a larger spatial scale, land-use change can lead to significant local feedbacks like drought, flooding, soil erosion or shifts in local climate.

Up to now only little was known about how typical ecosystems of the rain forest margin areas interact with the atmosphere. We present here results from a sub-program of STORMA (Stability of rain forest margins in Indonesia, SFB-552-B1), to describe the energy, CO₂ and water fluxes between a larger tropical area and the atmosphere.

Field measurements with the eddy correlation method showed that a Cacao agroforestry system (AFS) was a small source of CO₂, whereas a tropical rain forest was a strong CO₂ sink during the one year observation period. In addition, the rain forest evaporated much more water compared to the Cacao AFS.

We applied a series of models to extrapolate the results from the field investigations to the region, including so-called SVAT models of differing complexity, and a simple PAR efficiency model to predict net primary production from regional meteorological and remote sensing data. In all cases, we discovered a large spatial variability in CO₂ and water vapour fluxes. We investigated the sensitivity of the models to changes in climatic drivers or land-use parameters from the current to a more intensified agriculture. The selected land-use scenarios reduced carbon dioxide sequestration and total evaporation and increased sensible heat fluxes and thus surface temperatures.

We conclude that there is still a lack of field observations to better understand and simulate the behaviour of tropical land-use systems. If the necessary detail on model parameter values, which characterise the differences between the land-use systems, is known, the existing model approaches are suitable to investigate consequences of land-use change to regional biogeochemical cycling, water utilisation and climate change.

Deforestation in the Brazilian Amazon increased from 10 million hectares in the 1970s to more than 60 million hectares at the turn of the century, resulting in growing awareness about deforestation impacts like greenhouse-gas emissions, loss of biodiversity, and motivating a number of initiatives involving the science and technology (S&T) field to address the issues of deforestation and sustainable development in the Amazon. The present work summarizes part of the large-scale land cover-use changes that occurred in the region and then analyses the organization of four S&T programs carried out in the Amazon in the context of alarming deforestation rates and rapidly changing land use in the Amazonian frontier. The four programs include two major research programs - the Large-Scale Biosphere-Atmosphere Experiment in the Amazon (LBA) and the Science & Technology Subprogram (S&T) of the Pilot Program to Conserve the Brazilian Rain Forest (PPG7) - and two examples of environmental monitoring and management - the Ecological-Economic Zoning (ZEE) and deforestation monitoring programs. In the context of high rates of forest loss and generally very weak institutions, the initiatives organized with the concourse of the S&T field may have significantly contributed to advance the discussions of sustainable development and sustainable land use in the frontier, and to mature some ideas about the participation of civil society, national environmental policy and, also, international cooperation. These exercises also suggest that the tasks of reducing and mitigating deforestation impacts and fostering sustainable land use are not to be engineered but, rather, negotiated, and that understanding how to contribute to such negotiations seems to be a major challenge for the science and technology field in Brazil.