Soils of Tropical Forest Ecosystems

Rational use of forest resources should be based upon accurate knowledge of land and soil properties. Selected characteristics of 149 soil profiles of the humid tropics are analyzed. General environmental characteristics, physical, chemical and mineralogical soil data have been derived from ISIS, the ISRIC Soil Information System. The humid zone includes 6 major soil groupings: Acrisols, Arenosols, Cambisols, Ferralsols, Luvisols and Podzols. For these soils the arithmetic mean and standard deviation of selected key analytical characteristics, relevant for agronomic and ecological research, are presented for a standardized topsoil and subsoil. In addition, their major agronomic constraints were assessed. About 64 % of the humid tropics is covered by Acrisols, Arenosols, Ferralsols and Podzols of low fertility, all presenting various degrees of limitations when forests are cleared for (low input) arable farming. A low plant nutrient content, a low nutrient retention capacity and a high toxic exchangeable aluminium content are major constraints. The overall data presented in this paper show that Ferralsols and Acrisols, covering 57 % of the humid tropics, have rather similar key characteristics. The taxonomic separation is principally based on a relatively small increase in clay content, which will not determine the major vegetation type or crop productivity level. When considering conservation of the natural vegetative cover and its biodiversity, several special properties of the soils have to be taken into account, such as depth of the litter layer, rootable depth and the mineral composition of the deeper subsoil and the substratum, as well as the geomorphologic-pedologic history of the terrain units and the degree of short-distance variation in soil conditions. It is concluded that land assessment studies and soil-vegetation/bio diversity research should not rely on a soil taxonomic approach alone, but should also be based on measured key pedological characteristics. The study illustrates the usefulness of the ISIS database to correlate soil characteristics and to determine agronomic-ecological constraints of major soil groupings.

The World Reference Base for Soil Resources (WRB) is the successor to the International Reference Base for Soil Classification (IRB). Its task is to apply the IRB principles of definitions and linkages to the existing classes of the Revised FAO-Unesco Soil Map of the World Legend (FAO, 1988). The main objective is to provide scientific depth and background to the Revised Legend, and to ensure that the latest knowledge relating to global soil resources and interrelationships is incorporated in a world-wide soil reference system.

At present, the WRB comprises 30 reference soil groups. Three new reference soil groups, i.e. Cryosols, Durisols and Umbrisols, are introduced compared with the Revised Legend, Greyzems are amalgamated with the Phaeozems, and Podzoluvisols renamed Albeluvisols. The concept and terminology of diagnostic horizons and properties is retained and expanded. In addition, diagnostic soil materials are defined. In the description and definition of the WRB diagnostic horizons, properties and materials emphasis is placed on field identification. Analytical criteria are given to help the identification. Some modifications are proposed to diagnostic horizons, properties and materials as defined in the Revised Legend and a number of new ones are formulated.

The basic philosophy of WRB is that the soil groups must represent a minimal geographical coverage and are the result of a major pedogenetic process. Soils are characterized by their morphological expression rather than by analytical data.

Ferralsols, Acrisols, Lixisols, Alisols and Nitisols constitute the main reference soil groups in tropical forest ecosystems occupying well drained upland positions. Other soil groups of local or regional importance are Andosols, Podzols, Umbrisols and Arenosols. The main soils occupying poorly drained positions in the tropics are Histosols, Fluvisols, Plinthosols, Gleysols and, to a lesser extent, Planosols. On one hand, the World Reference Base for Soil Resources, like the FAO-Unesco Legend, continues to build up on existing soil classification systems. Examples are the description, definition and subdivision of Anthrosols and Andosols. On the other hand, some proposals result from new ideas which are not yet reflected in published classification systems.

The final aim is a well described and defined World Reference Base for Soil Resources that is internationally accepted by the community of soil scientists. It also intends to facilitate the international use of pedological data, not only by soil scientists, but also by other users of soil and land.

Two forest soils representative of two research sites in Leyte, Philippines, are described as to their physical, chemical and mineralogical properties. The soils, an Andosol and an Alisol, which have developed from intermediate quaternary and pleistocene to pliocene volcanoclastics, respectively, represent a chronosequence of soils. The course of weathering is explained in terms of intensity and extent of the turnover of elements, loss of basic elements, desilification and development of pedogenic minerals. The site properties are described quantitatively in terms of rootability, availability of air, water and nutrients, as well as nutrient reserves. Both soils are highly weathered and show more or less the same percentage of desilification. The loss of basic elements is slightly lower in the Andosol than in the Alisol. In addition, there is a distinctly smaller loss of P and Mn in the Andosol. By calculating the turnover and losses of elements in kg per m², the much higher extent of weathering in the Alisol can be illustrated. The clay mineral association is dominated by imogolite in the Andosol and by halloysite and kaolinite in the Alisol. The progress of mineral alteration can also be observed from the Fe-mineralogy. Both soils have excellent physical qualitites regarding rootability and aeration and water supply to plants. The limiting factor for plant nutrition is the very low amount of available P in both soils, which is aggravated further by high P-retention in the Andosol.

In order to verify the hypothesis that conversion of forest to agriculture and other secondary land uses does not necessarily cause OM and nutrient depletion in the soil, we conducted this research at two locations, one an Andosol (Andisol) area and the other an Alisol (Ultisol) in Leyte, Philippines. In each location, adjacent sites having different land uses but with the same soil type, physiography and geology, were chosen as sampling sites. These included: forest, grassland, pasture and bushland for the Andosol in Ormoc, and forest, reforestation, coconut + bushes, shifting cultivation and coconut + kudzu, for the Alisol in Baybay. Our results indicate that a change in land use or cultivation decreased the soil OM, total and available N, available K but tended to increase pH and available Ca and Mg. Available P, which was very low in both soils, showed no distinct differences measurable. The soil changes in the Andosol were greater and more distinct than in the Alisol. Land use history helps explain soil changes due to land use.

The soil chemical and hydrochemical changes accompanying tropical forest disturbances of varying intensity are reviewed. Various methodological problems are discussed with respect to the establishment of ecosystem nutrient balances and the determination of changes therein due to forest exploitation or conversion to other land uses. It is shown that amounts of nutrients removed from the forest ecosystem in harvested timber generally exceed those associated with enhanced leaching, both after selective logging and clearcutting operations. Losses of base cations through volatilization and ash dispersal upon burning residual biomass are substantial, contrary to common belief. Approximate periods required for the replenishment of lost nutrients at the ecosystem level are estimated by comparing the losses with inputs via bulk precipitation and/or mineral weathering. The results suggest typical recovery periods of 30–60 years, depending on harvesting intensity, rainfall regime and soil nutrient retention capacity. Issues requiring further research include: (i) the evapotranspiration from young regenerating and planted vegetation; (ii) nutrient losses via rapid flow through macro-pores during rainfall (as opposed to slow matrix flow); (iii) nutrient losses via erosion after mechanized tree harvesting and burning of slash (particularly in the context of short-rotation plantation forestry); and (iv) the possible contributions to the overall ecosystem nutrient budget by mineral weathering. Particular emphasis is placed on the integration of hydrological, pedological and plant ecological process research at various levels of scale within an overall catchment ecosystem context in order to facilitate the integration of different disciplines. A plea is made to concentrate future research efforts at a relatively small number of carefully selected and well-researched locations, possibly joined together in a network capturing the chief environmental variability encountered in the humid tropics.

The hydrology of Haplic Acrisols on sandstone beneath old-growth and slightly logged Dipterocarp forests, and beneath Paraserianthes and Eucalyptus plantations in the lowlands of East Kalimantan was investigated to determine possible ecological constraints for forest and plantation management. The soils are characterized by 40–60 cm thick E horizons, texture silt clay, above clayey Bt horizons. Soil moisture (θ ), measured during a period of few weeks at the end of the rain season, was typically close to field capacity. Field-saturated hydraulic conductivities (K _GP ) indicate that the E horizons are moderately permeable (medians of K _GP 2.5–42.4 cm d⁻¹ ), while the Bt horizons are virtually impermeable (medians of K _GP < 0.17 cm d⁻¹ ). Macropores generated by biota are likely the cause of the observed large within-plot variability of K _GP in the E horizons and may provide pathways for preferential water flow. Soil water budgets constructed from measurements of precipitation (245 cm y⁻¹ ), throughfall (210–230 cm y⁻¹ ), using soil hydrological constraints obtained from K _GP values measured in the Bt horizons, and assumed transpiration rates (90–110 cm y⁻¹ ) suggest that 40–140 cm y⁻¹ of the soil water is drained laterally in the E horizons. Accordingly, < 5–40 cm y⁻¹ are available for deep seepage below 40–60 cm depth. Since lateral water flow is expected to increase with distance from the top of slopes, lower slopes receive large amounts of lateral water, depending on the slope length. It is concluded that the thickness, pore size distribution and permeability of the E horizons along with slope aspect are master variables for soil hydrology in the study area and determine site quality for forestry. Degradation of the E horizons due to compaction and erosion will increase surface flow, the frequency of flooding events and the impact of droughts on Dipterocarp forests and plantations of fast growing tree species in the lowlands of East Kalimantan. The conservation of topsoil physical and hydrological characteristics is therefore a prerequisite for sustainable forestry and landuse in this region.

In tropical forests growing on nutrient poor and acid soils nutrient cycling is tight and ecosystem nutrient losses are low. In this chapter the effects of selective logging on leaf decomposition rates and soil nutrient leaching are evaluated. Two logging gaps of different size (730 m² and 3440 m²) were each divided into three zones: skidder (open, disturbed soil), crown (debris, no soil disturbance) and undisturbed (small trees still present, no soil disturbance).

The soils in the gaps were wetter than in closed forest, especially during dry periods. This led to about 50% more water percolation through the soil in the skidder zone of the gaps during the first two years. The concentrations of NO₃-N, Al, Ca, K, Mg, Mn, Na and Cl in percolating water in the gaps were significantly higher, and the pH lower than in closed forest, particularly during the first 15 months after logging. In the large gap, nutrient losses through leaching were greater and extended for a longer period than in the medium gap. The logging induced nutrient losses decreased in the order: skidder zone, crown zone, undisturbed zone.

The absence of living roots through soil disturbance and by removing a group of trees is a more important cause for increased leaching than the sudden addition of large amounts of fresh, rapidly decomposable organic matter to the ground surface. After logging, nitrification is a key process since it enhances cation leaching and induces soil acidification and increased Al solubility.

Leaching losses may contribute up to 50% of the total nutrient losses associated with logging in case of a large gap, but significantly less in smaller gaps. Therefore, sustainable forest management systems should aim to reduce leaching as much as possible. Limitation of skidder activity to prevent excessive soil disturbance and avoiding the creation of large canopy gaps are crucial to minimise leaching losses.

Chemical analysis of total calcium in soil samples from Mendolong research area revealed that they had a very low calcium content. The content was so low that the harvesting of stemwood and stembark from all trees larger than 20 cm DBH should result in the loss of about 19% of all calcium in the ecosystem down to 50 cm depth. This finding was not expected based on earlier results of analyses of foliar contents and contents of exchangeable calcium in the soil.

Comparable figures for total calcium contents in the soil are not available from other tropical rainforest ecosystems. They have therefore been calculated by multiplying literature values for concentrations of total calcium in areas with tropical rainforest climate by soil weights from corresponding depths in Mendolong. The total calcium contents in the mineral soil to 50 cm depth varied from more than 20 000 kg ha^{− 1} in Chile to less than 200 kg ha^{− 1} in Brazil. This great variation indicates that not all tropical rainforests have low contents of calcium in their mineral soil. However, low contents can be found in certain areas which can be rather widespread in the humid tropics. In such areas, sustainable forestry is not possible without replacing the calcium lost from forest harvesting.

The soils associated with most tropical forests are exposed to significant leaching. This leaching has weathered the soils’ mineral suite and reduced their capacity to bind nutrients, exacerbating the potential for nutrient loss from the forest or any derived ecosystem such as slash-burn agriculture. Phosphorus is often the most limiting nutrient in ecosystems on weathered tropical soils. Examples are given for P budgets of South American forests and for an estimate of P loss from a highly dystrophic rain forest. Despite a biological turnover of 6% of the forests P every year, only about 0.001% of the total P stocks are lost annually, indicating that biogeochemical P recycling is highly efficient. The mechanisms for such efficient cycling are internal retranslocation of P, and a litter and root mat with abundant mycotrophism which result in an effective separation of the biological P cycle from a potentially P-fixing soil. In the examples of less dystrophic forests from NE Brazil, P retention in the litter mat is less efficient and P uptake from the soil is more important. Therefore, the potential for non-forest land use is greater in NE Brazil, although P tranformations to less available forms limit the length of cultivation cycles. Few data exist on alternative land uses, such as tree crops or agroforestry, and the complex methodology of studying P has limited the number of experiments aiming at an understanding of the complete set of P transformations which determine an ecosystem’s resilience to disturbance and limit net P losses.

The negative impact of fast-growing tree plantations, especially when grown in short rotations, on nutrient reserves of poor tropical soils is often described. The intensity of soil and forest degradation may be assessed by different approaches, including detailed observation of deficiency symptoms and yield development, foliar and soil analyses, as well as diagnostic fertilizer experiments. Input/output analyses, taking into consideration water and nutrient fluxes between the different plant and soil compartments are means to evaluate the sustainability of tropical tree plantations. Only a small amount of data is available up to now, indicating poor sustainability. Soil and tree nutrient status can be improved by the application of mineral fertilizer, which is usually rapidly effective. In contrast, amelioration by management practices is only slowly effective, nevertheless important in view of sustainability, environmental protection and biodiversity. A big gap exists in reliable data on water and element dynamics in forest plantations with different tree species under different ecological conditions, and under different management practices.

Declining supplies of cabinet-timber from natural stands of tropical rainforest combined with an increasing global demand for all classes of timber has given rise to a significant interest in developing plantations of high-value cabinet timber species in many parts of the tropics. In many cases the plantations are being established on land with low inherent fertility or sites that have been degraded as a result of poor landuse practices. The use of fertilisers has the potential to increase the production on these sites, but our understanding of the rates, timing and types of fertiliser to use with different species on different soil or site types is poor. This paper describes results of experiments at three sites in the wet tropics of north Queensland with similar soil types to sites being used for plantation establishment in South East Asia. The experiments aimed to determine the response to varying rates of P fertiliser in three high-value rainforest cabinet timber species, the Australian natives: black bean (Castanospermum australe Cunn. ex C. Fraser), and kauri pine (Agathis robusta (C. Moore ex F. Muell.) Bailey) and the exotic West Indian cedar (Cedrela odorata L.). Responses to fertilisation varied between sites and between species. There was a more pronounced response to fertilisation on coarse-textured soils derived from granite parent materials than on finer extured soils derived from metamorphic or basalt substrates. We discuss these results in light of their implications for plantation management and for the design and implementation of further nutritional research.

Tropical pine plantations are an important resource in helping to meet the future timber needs of southeast Asian countries. Information on the soil resource and nutrient management of these plantations is scant and given that a large proportion of these industrial plantations are on degraded soils, there is a huge potential for increasing sustainable productivity through effective site specific use fertilisers. An Australian example of tropical pine management is detailed to illustrate intensive plantation management based on understanding of sites and specific fertiliser requirements of these sites.

The Queensland Department of Primary Industries (QDPI) Forestry manages in excess of 130 000 hectares of exotic pines, predominantly slash pine (Pinus elliottii Engelm. var. elliottii Little and Dorman) and Honduras Caribbean pine (P. caribaea Mor. var. hondurensis Bar. and Golf.). Since 1991 the hybrid between these two species has been planted extensively in the south of the State.The plantations are established essentially on the coastal lowlands on infertile, light textured, acid soils derived from granitic outwash or sedimentary materials. Drainage varies from good on the low ridges to poor in the lower lying areas. The success of the plantation program is dependant upon the widescale use of fertilisers especially phosphorus. Lesser responses to nitrogen applied at planting occur on a range of sites under particular conditions and severe deficiencies of potassium and copper are common on the podzols. Fertiliser prescriptions based on soils and site history are described. Current fertiliser prescriptions have evolved over a long period as research information has become available, species and sites planted changed and fertiliser cost and availability altered. Major nutrient problems have been effectively solved by the efficient use of fertiliser. More sophisticated nutrient management on a site specific basis is being pursued to maximise productivity at minimum cost and to maintain or improve site productivity in the long term while minimising any adverse consequences of fertiliser addition or nutrient manipulation.

Growth, biomass and nutrient accumulation by 14-year-old stands of Caribbean pine (Pious caribaea Morelet var. hondurensis Barr. and Golf.) on three subhumid savanna sites in Nigeria were investigated. Soil organic carbon, total N, extractable P and exchangeable K, Ca and Mg showed significant decline in the upper 10 cm of the mineral soil as a result of reafforestation with the pine.

Estimates of the mean residence time of organic matter and nutrients in the litter layer of the ecosystems were 2 to 4 years thus indicating a slower rate of litter decomposition compared to litter of mixed natural savanna vegetation and broad-leaved plantations in which the residence time of organic matter rarely exceeds 6 months. Nutrient inventory in the different compartments of the pine ecosystem showed that a sizeable proportion of P and K was held in the tree biomass. Nutrient budget analysis showed that total biomass removal in above-ground tree harvest could lead to problem of P and K nutrition in future rotations on some sites. Estimates of nutrient fluxes and their relative contributions to tree nutrition were made. It was concluded based on the relative contributions of the fluxes that the soil remains the dominant source of the nutrients taken up by the tree and consequently tree nutrition as well as tree growth were soil fertility-dependent. Sites with low inherent nutrient quality were considered to be at a risk of nutrient depletion.

Retranslocation of nutrients before leaf abscission is an important pathway of nutrient conservation. This article reports the extent of nutrient withdrawal by Acacia mangium Willd. and Khaya ivorensis A. Chev., before leaf fall. Measurements were carried out based on the nutrient content in intact green leaves and freshly fallen litter, collected at the plantation age of 24 months. Results showed that A. mangium consficated back about 78% of P compared to only 26% by K. ivorensis. A. mangium could probably need some of these P for the N₂ fixation process beside for surviving on a soil with poor P availability. The withdrawal of nitrogen seems less critical but the trend relates to the availability of nutrients in the soil. The fact that A. mangium took back about 22% of N, which was far more higher than what has been taken back by K. ivorensis, is not fully understood as the former species has the capability to fix the atmospheric nitrogen. The withdrawal of K was about 13% for both species and this value shows K is more limiting than N as far as K. ivorensis is concerned. Magnesium, on the other hand, was not retranslocated.

Effects of fertilizer combinations on early growth of fast-growing plantation tree species were tested in two separate experiments on Imperata cylindrica dominated grasslands in South Kalimantan, Indonesia. Effects of commercial NPK and locally made NPK mixtures on the growth of Acacia mangium, Gmelina arborea, Paraserianthes falcataria and Swietenia macrophylla were tested by applying the fertilizer one year after planting. In the second experiment the effects of N, P, K, and their combinations with trace elements on the growth of A. mangium, G. arborea and Eucalyptus deglupta were tested by applying the fertilizer in two doses, immediately after planting and one year after planting. Two years after planting different NPK-mixtures did not have significant effect on the growth of A. mangium or S. macrophylla. On the contrary, all the NPK treatments roughly doubled the basal area of P. falcataria. For G. arborea the fertilizing effect was somewhat similar, but not statistically significant at p < 0.05. In the second experiment significant differences in growth of A. mangium, G arborea and E. deglupta between sites and fertilizer treatments were detected. Responses were strongest in the treatments including phosphorous. Generally A. mangium had steady performance compared to G. arborea, P. falcataria or E. deglupta, which were more sensitive to fertilization treatments and site effects. Fertilization is important on forest plantations on grasslands as it clearly speeds up the early growth that is important for gaining site control over the grass. Phosphorous must be included in the fertilization treatments on grasslands. NPK-fertilization is recommended as a standard solution if detailed information on the site is not available. Studies are needed on the long term effects of fertilizers, rotation-length fertilization requirements and the economic profitability of fertilizing, as well as the interactions between tree species, fertilizers and soil.

Multipurpose woody species can play an important role in the improvement and maintenance of soil productivity in traditional and planted fallow systems. This study examined the longterm effects of woody species grown in hedgerows in alley cropping and of soil tillage, on soil properties, runoff and erosion, and crop performance. Observations conducted from 1988 to 1993 on a trial established in 1982 on a sloping land is reported in this paper. Six land use systems were compared: Leucaena leucocephala and Gliricidia sepium hedgerows planted at 2 and 4 m interhedgerow spacings and tilled, and no-till and tilled controls without hedgerows.

The presence of hedgerows and no-till treatments maintained higher soil organic carbon (Org. C) and exchangeable cations contents. Highest Org. C and exchangeable cations were observed in plots with 4 m Leucaena hedgerows. Treatments with hedgerows also showed lower soil bulk density than the controls. The no-till treatment showed the highest soil bulk density. No relationships were observed between soil bulk density, aggregate stability and infiltration rate. Mean maize yields for the five years showed the following order: 4 m Leucaena > 4 m Gliricidia > 2 m Leucaena > 2 m Gliricidia > no-till > control tilled control plots. Intercropped cassava root yields showed the highest stability in the three cropping seasons in the tilled control and Gliricidia alley cropped plots. Lower cassava root yields were attributed to soil compaction in no-till treatment and to shading with 2 m Leucaena alley cropping

At the “Experimento Central” of CATIE (Centro Agronomico Tropical de Enseñanza, Turrialba, Costa Rica) long term studies have been carried out to determine productivity and the soil fertility (organic matter and nutrients) in agroforestry systems. In this paper results are presented and discussed as examples of sustainable agroforestry, using the systems of cacao (Theobroma cacao) under Laurel (Cordia alliodora) or Poro (Erythrina poeppigiana) as a case of study.

The following productivity and soil fertility parameters were found:

Therefore it is concluded that the agroforestry systems analysed maintain their soil organic matter status and promote efficient nutrient cycling. They are also sustainable and appropiate for soil fertility conservation.

Soil characteristics were examined in relation to tree species diversity of a tropical rain forest on a 100 m transect plot covering from ridge to valley with 0.10 ha and a long-term observation plot, 1.0 ha, in Mt. Gadut area near Padang, West Sumatra, Indonesia. The area experienced an annual rainfall of more than 6000 mm with no real dry season. In the 0.10 ha transect plot, detailed soil characterization was performed. From each sub-plot having the measurements of 5 × 5 m, 5 samples were taken from each 0–5 and 5–15 cm depth. In this transect plot, the subplots positioned on the ridge showed a greater diversity of tree species and higher biomas than those positioned in the valley. Soils found on the ridge showed less fertility and greater variation of fertility characteristics than those found in the valley. The examination of vertical soil profile distribution of ridge, middle and valley positions in this transect also revealed that exchangeable Calcium, total carbon and total nitrogen of ridge soil tended to accumulate of these nutrients at the upper horizons and of dissipation at lower horizons more than those of valley soil profile. Detailed characterization of 115 subplots of 1.0 ha permanent plot showed that horizontal variations of soil fertility characteristics were also important in relating the tree species diversity of tropical rain forest. These results suggest that both vertical and horizontal variations of soil characteristics are imperative in supporting the tree species diversity.

The distribution of mineral elements in trees and soils in a 1 ha permanent observation plot on mixed dipterocarp forest in West Sumatra, Indonesia was studied. Seven elements, including S, K, AI, CI, Si, G and Sr, revealed significant correlation between leaves and barks, suggesting that barks can also be used to study the nutritional characteristics of trees instead of leaves as the first approximation. The concentrations of elements in 457 different samples of bark showed quite a wide range, indicating the enormous diversity of nutritional characteristics of mineral elements among these tree species. The concentration ranges in mg kg⁻¹ were Mn (0.1–3800), Sr (0.1–1070), Si (2–1700), Fe (0.6–4100), Al (10–46000), Mg (10–7200), Zn (0.4–290), K (60–22000), Cu (0.8–150), P (19–2540), CI (110–9200), S (170–13600), Ca (2300–100000), B (4–70) and Na (130–700).

The accumulators and excluders were defined for each element as the trees which were the top 5 % and the bottom 5 % respectively (i.e. top and bottom 23 in elemental concentrations in their barks among the 457 tree samples). The soil edaphic factor appears to regulate the distribution patterns of some accumulators and excluders. The accumulators of K, Ca, B and Sr showed a tendency to distribute in the areas rich in those extractable elements in soils. On the other hand, the excluders of K, Al, Na, Cu, Mn and Sr showed a tendency to distribute in the areas poor in those extractable elements in soils. The accumulators and excluders of the other elements, however, did not show such a relationship to the soil edaphic characteristics. Genetic character of trees and/or other environmental factors including complex biological interaction may superimpose in edaphic factors in the distribution of these trees.

A global project on ‘Alternatives to Slash and Burn’ agriculture was initiated by a consortium of international and national research institutes to facilitate intensification of the use of converted forest land, in order to help alleviate poverty and protect the remaining forest areas for their biodiversity values and their role in mitigating greenhouse gas emissions.

Data for the Indonesian benchmark areas in the lowland peneplain, piedmont and mountain zone of Sumatra are presented. A significant amount of forest land, especially in the lowland peneplain, has been converted in the last ten years into agricultural use, usually following logging concessions. Soils on the peneplain are poor (oxi- and ultisols) and current intensive crop based production systems are not sustainable. In the piedmont zone on better soils (inceptisols), rubber agroforests (still) characterize the area. Agroforests have emerged during the 20’th century as the major alternative to slash-and-burn agriculture, based on a shift of emphasis from food crops to cash-earning tree crops. Emphasis on food crops, however, continues in government resettlement schemes.

Differences in organic C content of the topsoil between forests and crop land are about 0.5% C, with agroforests and tree crop plantations in an intermediate position. A new size-density fractionation scheme for soil organic matter demonstrated larger changes in light and intermediate fractions. Forest soils can be significant sinks for methane and thus partly compensate for the methane emissions in lowland rice production.

Overall, the Sumatra benchmark areas demonstrate the need to combine intensification of land use at the field/household level with effective protection of remaining forest areas at the community level and reducing other driving forces of deforestation at the national level.

Results of fine root inventories in Dipterocarp forests and plantations of Eucalyptus deglupta Bl. and Paraserianthes falcataria (L.) Niel. in East Kalimantan, Indonesia are presented. The influence of different forest land use on the amount and vertical distribution of fine roots in the soil is examined. Possible consequences for physical soil parameters related to root growth and morphology of roots are discussed. Differences in nutrient acquisition between tree species are indicated by the chemical compositon of fine roots. Fine root development and turnover in the Dipterocarp forests and the plantations is assessed by the comparison of samplings taken during dry and wet season.

We have found that young roots of D. zibethinus (durian) from a village forest garden (complex agroforest) in West Kalimantan, Indonesia, form extensive VA mycorrhizas. Cells of the root cortex contained many coils of hyphae, arbuscules and infrequent vesicles. Intercellular hyphae were not seen. VA mycorrhizal fungi were less frequently found in older roots but other fungi were present, some of which colonized the root surface. Nephelium lappaceum (rambutan) and Artocarpus integer (cempedak), cultivated in association with durian, formed similar VA mycorrhizas. Roots of Lansium domesticum (langsat) only showed initial stages of colonization — i.e. entry-points and coils of hyphae, but this may be a function of the age of the roots that were sampled. Although we did not survey the many other species in the forest gardens, work elsewhere by others on the same species or their relatives suggests that most plants will also be VA mycorrhizal.

The possible nutritional significance of the VA mycorrhizal symbiosis in this forest garden agroecosystem is briefly discussed and suggestions are made for further study.

Physical properties of mineral top soil were investigated on and outside 6 and 13 year old tractor tracks (and outside tractor tracks) on Haplic Acrisol (Ultisol) under selectively logged forest, secondary forest after forest fire and plantation forest. The analysis of possible rehabilitation to pre disturbance soil properties also included earlier published data for new tractor tracks and control forest in the same experimental area. Dry bulk density (DBD), loss on ignition (LOI) and steady state infiltrability (SSI), as well as visual descriptions of vegetation, top soil and erosion, were recorded for 3 positions in slopes as well as for 3 degrees of tractor disturbance to try to resolve the process of top soil rehabilitation.

13 year old tracks under secondary forest grouped significantly together with controls for DBD and LOI, in contrast to nearby 13 year old ones under the canopy of selectively logged forest. Differences in amount and quality of litterfall are discussed to explain this difference. For SSI there was a significant increasing gradient from new to older tracks, but no tracks were as high as control.

From these and other data, it may be argued that DBD decreases and possibly total porosity increases over time with increasing organic content (LOI). Soil permeability is rehabilitated more slowly than DBD and total porosity, as macro porosity is further developed in a later stage with further increased organic content. This later development of macroporosity may also be attributed to changes in the quality of organic material as DBD may not be further lowered.