Paper presented at IRRDB, 2006, Vietnam. Rubber based Agroforestry Systems (RAS) as Alternatives for Rubber Monoculture System. Gede Wibawa Lembaga Riset Perkebunan Indonesia Jalan Salak 1A, Bogor, Indonesia Laxman Joshi and Meine van Noordwijk World Agroforestry Centre/ICRAF PO Box 161 Bogor, Indonesia Eric Penot CIRAD BP 5035, 34032 Montpellier, Cedex 1, France ABSTRACT Smallholder rubber plantations in Indonesia, representing more than 80% of the total rubber areas, are very unique in the world. Most smallholder rubber areas are multi-strata in nature. Rubber is not the only perennial crop in that area, but also mixed with timber trees (forest re-growth), fruit trees, and different annual crops. Scientists identified these multistrata systems or called “Jungle Rubber” have multiple functions such as main income source for many farmers; keeping certain level of the forest biodiversity; Carbon sequestration; soil and water conservation. Many efforts have been done by the Indonesian Government to improve the productivity of jungle rubber through monocuture system such as Nucleolus Estate for Smallholding (NES); Project Management Unit (PMU); and Partial System. However, the rate of rubber replanting through those specific projects are too small compared to the total rubber areas in Indonesia. Taking into account the positive aspects of the Rubber Based Agroforestry Systems (RAS), ICRAF, CIRAD and IRRI worked jointly to implement various RAS systems in order to provide farmers better technological options for managing their farms. Results presented in this paper are collected from both on- station and on-farm research. At on-station, rubber planted at a double rows spacing (6m x 2m x 14m) with and without perennial intercrops was monitored and compared to that of the normal spacing (6mx3m). Rubber girth with double row spacing reached tappable size at five years after planting and similar to that planted with normal spacing. Rubber growth at the plot planted with Acacia mangium a very fast growing tree, at the same time with rubber, was very slow: a haft of that of the other treatments. If the fast growing trees are planted two years after rubber planting, then rubber growth is similar to that of normal spacing. Three types of RAS were tested at farmers’ plots (RAS1, RAS2, and RAS3). The total number of participants is about 150 farmers, in 100 ha plots, distributed in Jambi, West Kalimantan, West Sumatra and South Sumatra. Results of clone comparison in RAS 1 type of trial (maintenance only on rubber rows) showed that rubber growth variability was more due to the variability of farmers’ plots and frequencies of weeding. Rubber clones such as PB 260; RRIC 100 and BPM1 planted under RAS 1 can adapt the RAS conditions and can be tapped at 5-7 years after planting. These findings provide farmers alternatives to develop more environmentally friendly and divers systems in their farms, compared to that of monoculture system. This paper presents also various results related to more intensive RAS (RAS1 and RAS 3). Keywords: Agroforestry, rubber, timber trees, Acacia mangium, Imperata cylindrica, RAS, Jambi, West Kalimantan, West Sumatra, and South Sumatra. INTRODUCTION Rubber has been developed in Indonesia since more than a century and since then Indonesia has as the largest rubber area in the world (3.5 millions ha). However the productivity of smallholder rubber in this country is very low (650 kg/ha/year), or half of that of Thailand (....). This can be attributed firstly to the use of non-improved planting material (unselected seedling) and secondly to the low level of maintenance (low weeding, minimum fertilisation) practiced by most rubber smallholders (more than 80% of the total area). Due to this extensive management, smallholder rubber areas in Indonesia are mostly under “jungle rubber” forms, where rubber present as the main species grows together with other species such as timber, fruits, rattan, medicinal plants. The most important government action on the development of the commodity was started at the beginning of 1970-ies and mainly on the 1980-ies, after the foreign earning of the country, based on petroleum, starts to decrease. Various development and rehabilitation projects for smallholder tree crops have been established, which were mainly, are grouped into two schemes: Perusahaan Inti Rakyat/Nucleus Estates of smallholder (PIR/NES) and Project Management Unit (PMU), later the government has also developed partially funded projects. Except for the later, all those projects were based on monoculture and credit scheme. These schemes have had rather successful in transferring various technology innovations that implemented by tree crop smallholders and had increased farmers’ income. Apart from technical factors, the low productivity performance of the most smallholder rubbers in Indonesia is also due to economic and social factors, and weaknesses of the supporting institutions for smallholder rubber development. Traditionally smallholder rubbers in Indonesia are established after a slash and burn of secondary forest or old jungle rubber, followed by planting of annual food crops in between rubber rows for 2-3 years. The system is based on extensive management both for rubber and intercrops, either during the first two-three years of intercrop establishment, or afterward. After completion of annual intercrops, farmers abandon the land to seek other portions of land to be planted with similar intercropping system. Weeding or slashing of the forest re-growths was done once to twice a year in the first three years after intercrop and maximum once a year before rubber starts to be tapped. Even in a rural development context these systems are important sources of income for farmers; require only low capital, labour and cash inputs and produce a diverse ranges of foods, fruits and timber; however there is still needs to improve the productivity of the these systems with moderate changes in farmer's management in order to improve land productivity. Various consequences of this low farm management are identified such as a) slow and heterogeneous rubber growth and long immature period or late reaching tappable size (8 to 12 years after rubber planting) and; b) rapid growth of forest re-growth. Some useful tree species re- growth may be kept by farmers due to its important economic value such as fruits, rattans, medicinal and timber trees, in the future. These extensive and low management systems develop toward a complex agroforests based on rubber trees. De Foresta and Michon (1996) defined complex agroforests as forest structures managed by farmers for the production of various forest and agriculture products on the same 2 piece of land, mimic natural forest structures, with a complex structure and a closed or almost closed canopy dominated by few tree species. In other hand, simple agroforestry refers to associations involving a small number of components arranged with obvious, usually well ordered pattern: one or a couple of tree species, either as a continuous canopy, in equally distant lines or in edges, and some annual species for ground cover (Michon and de Foresta, 1998). It was known that outside of those government project areas, most smallholders could not implement recommended technologies (Supriadi, 1997; Supriadi et al., 1999; Supriadi and Nancy, 2001), mainly because of it is costly, the technologies are not always adapted to their various circumstances, lack of institutional and capacity buildings, and the technologies are sometimes not available for farmers. These constraints need to be taken into account in developing new technologies for smallholder farmers. Rational of Rubber Agroforestry Systems (RAS) The main challenge for researchers is to search and to test new models for improving smallholder rubber production systems, based on the current farmer practice ones rather than replacing them with estate-like or monoculture, conserving the biodiversity and environmental benefits of agroforestry practices. The important characteristics of rubber clonal planting materials are well known such as high productivity (2-3 times of that seedling originated plants), homogenous growth, and response to production inputs (fertilizers, weeding). These clonal materials are costly for farmers (between 0.25–0.30 USD/polybag in 2005). However, due to it capacity to double or even triple latex production, compared to unselected rubber trees, it is worth to use clonal rubber in RAS. Clonal planting material has been historically selected for estate monoculture management and optimized for the highest level of maintenance. Testing clonal rubber in agroforest environments with a certain level of extensive practices means that clone will be selected for other environments where competition is far higher than that of monoculture and based on reduced inputs and labour. Testing clonal materials includes also assessment of rubber production with and without fertilizer, growth under different levels of weeding and clones comparison under different levels of weeding. The research aims to identify the key components (use of improved planting material, fertilization, and combination of both) to improve productivity of the rubber agroforestry systems, which can be established in pioneer and buffer zones, as well as in zones where replanting is required. Since1994 and planned up to 2007, World Agroforestry Centre (ICRAF) in association with CIRAD-France and Indonesian Rubber Research Institute (Sembawa Research Station) established a network of trials to study rubber agroforestry systems and test different approaches suitable for different conditions under SRAP (Smallholder Rubber Agroforestry Project) and SRAS (Smallholder Rubber Agroforestry System) project. The project was funded by various funding agencies such as: USAID, French Embassy, Gapkindo, and CFC (the Common Fund for Commodities). This paper presents brief results of different Rubber Agroforestry Systems 3 (RAS), both simple and complex RAS, as alternatives to intensive rubber monoculture systems (Penot, 2001). MATERIALS AND METHODS The network trials were developed since the last 10 years either at controlled environments (on- station) or at farmers’ circumstances (on-farm). Increase of productivity of jungle rubber in Indonesia may be attained by providing improved planting materials of the tree components to the farmers and evaluating which are the most appropriate and affordable for smallholders. This research program is based on four major components: a) the characterization of selected areas to achieve a “situation typology” covering a wide range of conditions, b) a network of on-farm trials using participatory approach, c) a farmer typology reflecting all strategies and constraints encountered in the rubber growing areas of Kalimantan and Sumatra, and d) in-depth studies on particular relevant agronomic and ecological topics. Data analysed in this paper present only on agronomic study of the RAS. Sources of Data Data analysed in this paper are explored from two types of trial, on-station controlled plots (Sembawa Research Station, South Sumatra province) and on-farm participatory plots established in Jambi, West-Sumatra, and West Kalimantan provinces of Indonesia. On-station trials A series of testing of simple RAS, in which rubber (BPM 24 clone) was mixed with Paraserianthes falcataria, a fast growing timber tree, planted at the same planting date with rubber was established in Sembawa Research Station. Rubber was planted in a double row spacing (4m x 3m x 16m) and three densities of Paraserianthes falcataria (3, 4, and 5 rows in between wide rubber rows, or comparable to 450, 600 and 750 trees/ha respectively). These treatments are compared to rubber monoculture planted in a double row spacing and in a normal spacing of 6.7m x 3m (500 trees/ha). Apart from the above trial, another trial was established to test the effect of different planting densities of Paraserianthes falcataria (950, 475 and 240 trees/ha) on rubber growth planted under a two year old rubber (BPM 24 clone) with a normal spacing of 7m x 3m. The third trial consist of testing various tree species under rubber planted in double rows spacing distance (6m x 2m x 14m) or of density of 500 rubber trees/ha. The non-rubber tree species tested are Tectona gradis, Acacia mangium, Eucalyptus sp. of a density of 450 trees/ha (3 rows). A normal density of rubber (500 trees/ha) of 6m x 3.3m spacing distance, planted without intercrop, was established as control plot. Rubber and intercrops were planted at the same planting time in 2000. On-farm participatory network and RAS Types tested A network of farmer-managed trials is carried out in Jambi, West-Sumatra and in West Kalimantan provinces of Indonesia. The trials were established in three different planting year 4 phases. The first phase: planting in 1995-1996 consist of 10 types of trial; the second phase: planting in 2001 consist of 3 types of trial; and the third phase: planting in 2004 consist of 7 types of trial. Results considered in this paper will be explored from the selected first phase of the trials only. The trials, with an average of 3 to 5 farms or replications per trial, covering 100 hectares and involving about 150 farmers have been established. Each farmer’s field is considered as a replication with 1 or 2 simple treatments such as: rubber weeding levels, rubber fertilization, rice variety x fertilization, type of associated trees, and types of cover crops (Multi Purpose Trees (MPT)/Fast Growing Trees (FGT)) combination. These experiments take into account the limited resources of smallholders. Labour is one the main factors being considered in assessment of a system’s suitability. The first, RAS 1, is similar to the current jungle rubber system, in which unselected rubber seedlings are replaced by adapted clones. The main objectives are to determine if clonal rubber germplasm succeed to grow well under jungle rubber environment, to increase yields significantly, and to assess the minimum required management level of RAS. A secondary objective is to assess the level of biodiversity conservation in the jungle rubber system. It is expected that the rubber clones be able to compete with the natural secondary forest growth. Various planting densities, clones, and weeding protocols are tested. This will identify the minimum amount of management needed for the systems. RAS 1 requires a certain level of existing biodiversity (old jungle rubber, tembawang or other type of timber/fruits agroforestry systems, home gardens, secondary or primary forest) for establishment. In other words, RAS 1 is aimed not to be established in Imperata grassland areas (Penot, 1994; 1997). The second, RAS 2, is a complex agroforestry system in which rubber and perennial timber and fruit trees are established after slashing and burning, at a density of 550 rubber and a range of 90/250 other perennial trees per hectare. It is very intensive, with annual crops being intercropped during the first 2-3 years, with emphasis on improved upland varieties of rice, with various levels of rice fertilization. RAS 2 is aimed to answer the following questions: how is total productivity and income affected by intercrops? what are the dynamics of species interactions? And what are the crop alternatives during rubber immature period? Intercrops are annual (predominantly upland rice or rotation rice/leguminous such as groundnut) or perennial (cinnamon), during the first years of establishment. Previous experimentation has shown the positive effect of annual intercropping on rubber growth (Wibawa, 1996, 1997). The range of trees that can be grown in association with rubber in agroforestry associations and the market potential of their products are being examined: tekam (Endospermum malaccensis), meranti (Shorea spp), tembesu (Fragraea fragrans), and sungkai (Pheronema canessens) trees for timber trees; durian (Durio zibetinus), rambutan (Nephelium lappaceum), duku (Lancium domesticum), langsat (Lancium sp), cempedak (Artocarpus sp), petai (Parkia speciosa), jengkol (Archidendron pauciflorum), and tengkawang (Shorea macrophylla) for fruit trees. The third system, RAS 3, planted only in West Kalimantan, is also a complex agroforestry system with rubber and other trees planted at the same density as that as in RAS 2, but with no 5 intercrops except in the first year, followed by a combination of leguminous cover crops, and Fast Growing Trees (FGT). It is established on degraded lands covered by alang-alang grass (Imperata cylindrica) (Penot, 1995). The grass bounds the growth of annual crops so selected cover crops (Mucuna, Flemingia, Crotalaria) or MPTs (Calliandra, Wingbean, Gliricidia) and FGTs ( G. arborea, P. falcataria., A. mangium) are established with various density between 50- 110 trees/ha. It had been assumed that the FGT could be harvested in 7 or 8 years to provide timber and wood for the existing pulp industry. The objective of RAS 3 is to reduce the weeding requirement by providing a favourable environment for rubber and the associated trees to grow, cover the soil as soon as possible to bound imperata growth. The first of the RAS3 trials were planted in 1996 in three farmers’ fields in Kopar village in West Kalimantan. High yielding clonal rubber plant (PB260), grew in polybags, were planted in the field after land clearing through slashing and burning. Mucuna, Pueraria, Flemingia were planted four rows between rubber rows planted at a density of 13-5.000 holes per hectare; the naturally occurring Imperata and Chromolaena were also retained for comparison. At Trimulya village, in a Javanese transmigration zone, FGT were planted between rows of rubber trees planted at a density of 110/ha. All plots, both with cover crops and FGT, were weeded (manually or using herbicides) only along rubber rows every three months; limited fertilizers (rock phosphate and urea) were applied only in the first two years. Regular measurements of rubber trees, presence and dominance of ground vegetation were taken and form the basis for analysis of different treatments. RESULTS AND DISCUSSION Results derived from on-station trials Effects of P. falcataria on rubber growth (double row spacing). Rubber (BPM24 clone) growth planted in double row spacing (4mx3mx16m) up to 18 months is comparable to that planted in normal spacing 6.7m x 3m (Fig. 1). The gap of rubber girth between those two treatments increased afterward and started to be significant at 24 months. The presence of P. falcataria at different densities reduced rubber growth significantly since 24 months. At 51 months, rubber girth at intercropped plot was 30% and 15% less than that at monoculture normal spacing and at monoculture double row spacing respectively. The monthly increment of girth was 0.6cm; 0.8cm and 0.9cm respectively. The slowest increment was observed during dry season where intercrop reduced increment up to 70% compared to control and 50% if compared to monoculture double row spacing. The canopy of rubber in double row spacing started to shade the soil since 30 months. Only about 60% of light penetrated in monoculture double row spacing, and far below that point in intercrop treatments (36-52%) and about 70% in normal spacing. At 54 months, in all double row plots, the intensity of light was below 35%, however in normal density the light intensity was 50% (Table 1). These data indicated that the intra-plant competition for light may be started 6 earlier in plots with intercrop and in plot without intercrop with double row spacing, compared to normal spacing. 45.0 40.0 m) 35.0 c H ( 30.0 T S450 R 25.0 S600 GI ER 20.0 SNo75 I0ntrc B Control B 15.0 U R 10.0 5.0 0.0 18 21 24 27 30 33 36 39 42 45 48 51 AGE (MONTHS) Fig. 1 Rubber growth planted at different treatments: double row vs. normal spacing and monoculture vs. P. falcataria intercrop Similar to the above data, the light condition in between rubber row and in intercrop row (location B), was varied due to treatments. P. falcataria of a density of 750 trees/ha reduced light intensity since 18 months, and lesser reduction of light intensity due to lower P. falcataria density was observed. P. falcataria shaded the soil more than 50% since 18 months (Table 1). Table 1. Light intensity (%) in different locations within double row plots (4mx3mx16m) and normal spacing plot (6.7mx3m) at different rubber ages. Rubber age (months) Location Treatment 18 30 54 A S450 99 30 16 A S600 94 52 26 A S750 99 26 31 A No Intercrop 100 62 32 A Control 99 69 50 B S450 68 35 31 B S600 52 41 31 B S750 37 31 41 B No Intercrop 98 82 70 B Control 99 90 31 C S450 49 36 34 C S600 40 38 25 C S750 20 35 37 C No Intercrop 100 99 100 C Control - - - A = Measurement in the middle of rubber rows 7 B = Measurement in between rubber row and intercrop row (1.5m from rubber) C = Measurement in the middle of intercrop’s row Note: S450: means that density of P. falcataria is 450 trees/ha The competitive effects of P. falcataria on rubber can be reduced by planting it, two years after rubber planting, with a density of about 250 trees/ha (planted 1 row/rubber inter-row and 6m within row). The trial in Sembawa showed (Fig. 2) that rubber planted without intercrop (monoculture) can reach a tappable size (girth at 1m above union is 45cm) 60 months after planting, and with intercrop (P. falcataria) 67 months after planting. This data is very useful for farmers who interest to develop such perennial intercropping with rubber. Providing good growth conditions to rubber during the first years of establishment phase, especially when the intercrops are the fast growing trees is a wise strategy. This strategy enables rubber to explore better factors of growth and compete with other perennials. 45 ) 40 m c 35 ( h 30 Rubber+ P.falcataria t Gir 25 Rubber Monoculture r 20 e b 15 b u 10 R 5 0 950 475 240 P. falcataria density (trees/ha) Fig. 2 Rubber growth, 60 months after planting, intercropped with P. falcataria with different densities which was planted two years after rubber planting. Double row spacing for rubber as potential spacing for RAS In order to search better rubber spacing for perennial intercropping under rubber, a new double row spacing was tested by widening the space within rubber (RRIC 100 clone) to 6mx 2mx14m (same density of 500 rubber trees/ha). The trial was carried out in Sembawa since April 2000. Results showed (Fig. 3) that rubber growth in double row spacing plots can be comparable with that grew in normal spacing (control) plot. These results were better than that mentioned above. This may be due to the wider rubber row spacing within rubber and rubber clone used, RRIC 100, a fast growing clone. This trial also indicated that planting perennial intercrop after rubber (almost 2 years) is a good strategy to avoid high competition to rubber. Eucayptus sp planted under rubber none significantly reduced rubber growth. However, Acacia mangium a fast growing timber tree which planted at the same time with rubber, competes rubber growth very significantly since the second year of the establishment. Rubber reach the tappable size between 56 to 63 months after planting, and no significant difference was observed between rubber girth 8 in double row with or without intercrops to that planted in monoculture normal spacing up to 62 months. The findings from this trial may be very useful as basis of the recommendation of planting rubber in double row spacing (6mx2mx14m), as alternative to normal recommended ones (6mx3m or 7mx3m). The land in the wider space in between double rows (14m) can be used by farmers to grow food crops in a longer period (more than three years) and for perennial tree crops (timber or fruits). The light intensity is expected still more than 70% up to 54 months. The longer the intercrops are practiced the saver the plantation from the external factor pressure (fire, pests). 55 50 45 RD+Eucalyptus ) 40 m RD monoculture c ( 35 RD+Tectona h t r R Control i G 30 RD+Acacia 25 20 15 20 30 40 50 60 70 Age of Rubber (months) Fig. 3 Rubber growth planted in double row spacing (6mx2mx14m) and at different perennial intercrop treatments, compared to rubber planted in a normal spacing. Note: R: Rubber with 6mx3.3m spacing and monoculture; RD Rubber with 6mx2mx14m spacing. Tectona grandis planted 21 months after rubber. The timber trees planted under double row spacing of rubber grew well as in monoculture conditions. Stem growth (measured at 1m from soil) of timber trees planted under rubber (Fig. 4) indicated that Acacia has the fastest growth (twice of that of rubber), while Eucalyptus’ slow growth was due to the virus leaf disease attacking the trees since two years old. The growth of Tectona is good at the current South Sumatra conditions. 9 80 70 60 50 m) Eucalyptus c h ( 40 Acacia t Gir 30 Tectona 20 10 0 26 30 33 36 40 45 51 56 Age of Rubber (months) Fig. 4 Trees growth planted under double row spacing of rubber Note: Acacia and Eucalyptus planted at the same time as rubber, while Tectona grandis planted 21 months after rubber. Eucayptus was cut after 45 months due to virus leaf disease. Results from On-Farm participatory experiments Effects of levels of weeding on rubber growth under RAS 1 In West Kalimantan, farmers do not follow the protocol of the trial targeting to test the effects of frequencies of weeding within rubber row on rubber growth, since the first year of establishment (Table 2). The level of weeding implemented by farmers is less than that expected by the protocol of the trial. Farmers tend to weed rarely their plantation. Even, the lowest frequency (Low) is not respected by farmers. As results, rubber growth was slower than the expected one. The consequences were that we observed a comparable rubber growth (no significant difference) in plots “medium”, “intensive” or “intensive with LCC” due to the similar frequencies of weeding from the first up to 5th year. Rubber growth in this group of treatments was better than that in low weeding plot (Fig. 5) 10