Kebon tatangkalan: a disappearing agroforest in the Upper CitarumWatershed, West Java, Indonesia

Parikesit1,*, K. Takeuchi2, A. Tsunekawa2 and O.S. Abdoellah3

1Institute of Ecology & Department of Biology, Padjadjaran University, Jl. Sekeloa No. 1, Bandung 40132,West Java, Indonesia; 2Graduate School of Agricultural and Life Sciences, the University of Tokyo, Yayoi 1 �1 � 1, Bunkyo-ku, Tokyo 113 � 8657, Japan; 3Institute of Ecology & Department of Anthropology,Padjadjaran University, Jl. Sekeloa No. 1, Bandung 40132, West Java, Indonesia; *Author for correspondence(fax: 62 � 22 � 2504982; e-mail: pampang@melsa.net.id or ecology@melsa.net.id)

Received 7 May 2003; accepted in revised form 2 February 2004

Key words: Biodiversity, Forest garden, Land-use change, Multi-layered system

Abstract

A comprehensive study was carried out on a typical indigenous agroforestry system, called kebon tatangkalan, inthe changing agricultural landscape of the Upper Citarum Watershed, West Java, Indonesia. The main objectivesof the study are to elucidate the structural patterns, multidimensional functions, and dynamics of this multi-lay-ered agroforest. The study has identified 12 groups of different plant assemblages indicating that this type ofagroforest contributes considerably to the heterogeneity of the agricultural landscape in the study area. Multi-variate analysis suggests that elevation and slope are the biophysical factors that correlate most significantly withthe distribution of plant species in kebon tatangkalan. The presence of this man-made vegetation is currentlydeclining from the landscape due to population growth and rapid regional economic development. Serious effortto revitalize this traditional agroforest is needed to prevent its disappearance. Introduction of highly productivespecies and establishment of integrated policies in regional land-use planning and management are measures thatshould be taken into account in preserving this traditional agroforestry system.

Introduction

Various forms of agroforestry have long been knownin the local land-use systems of the Upper CitarumWatershed, West Java, Indonesia. An agroforestrysystem locally known as kebon tatangkalan or bojongis one of the most typical ones. This agroforestry sys-tem has been evolving under the influence of variousbiophysical and socioeconomic factors. Its physiog-nomic appearance is the most distinct among the ex-isting agroecosystem components. Multi-layered veg-etation patches of kebon tatangkalan scattered withinthe matrix of agricultural land has made the landscapeof the Upper Citarum Watershed heterogeneous.

Even though the existence of talun has beenacknowledged for more than a century �Christanty et

al. 1996�, studies on its biological as well as ecologi-cal features and also its dynamics have not been re-vealed thoroughly. Previous studies �e.g., Soemar-woto and Soemarwoto 1985; Christanty et al. 1996�only revealed that, basically, there are two major cat-egories of talun, bamboo-dominated talun and mixedtalun. To date, there is no data with regard to thevariation of plant assemblages in these two catego-ries. Yet, such data are important to assess whetherthese agroforestry systems contribute to biodiversityenrichment not only at species level, but also at com-munity level.

It is presumed that kebon tatangkalan has longbeen playing a significant role in the entire produc-tion system of the agricultural landscape in the UpperCitarum Watershed. Moreover, from the standpoint of

Agroforestry Systems 63: 171–182, 2004.© 2004 Kluwer Academic Publishers. Printed in the Netherlands.

171

ecological function, the occurrence of patches of thisagroforest type is very crucial for various wildlife,birds and insects in particular �Erawan et al., 1997�.Apart from that, the multi-layered structure of veg-etation and its rooting system has made this type ofland-use very potential in the prevention of soil ero-sion.

Kebon tatangkalan in the Upper Citarum Water-shed is currently under serious pressure from popula-tion growth, expansion of capital-intensive agricul-ture and industries. However, it is regrettable that thedecline of this multi-layered agroforest has beenoverlooked by scientists, decision makers, and re-gional planners. Before 1970s, this man-made vege-tation was believed to cover the agricultural land-scape in the Upper Citarum Watershed in a veryextensive area. But in the present time, patches of thisland-use type can only be encountered in places nothigher than 1250 meter Above Sea Level �ASL�.Considering the multidimensional roles of this agro-forest, its disappearance from the agricultural land-scape of Citarum Watershed might cause significanteffects on the entire environmental setting within thewatershed.

Based on the above considerations, study that en-compasses bio-ecological and socioeconomic aspectsof kebon tatangkalan in the Upper Citarum Water-shed, West Java Indonesia was carried out. This pa-per describes some results obtained from the study.The objectives of the paper are to describe structuralas well as compositional aspects of kebon tatangkalanvegetation and its associated faunal diversity, and toelucidate the socioeconomic function and the dy-namic of this agroforest.

Materials and methods

Study sites

The present study was undertaken in two villageswith different agroclimatic zones, namely Wangisa-gara and Sukapura villages �located within 07° 04�00� � 07° 10� 30� South and 107° 40� 30� � 107°44� 45� East; 850 m � 1250 m ASL�. In general, thetopography of the study villages is flat to undulating,but in some parts of the villages, steep areas can alsobe encountered. Major land use types in the studyvillages are rice fields, upland fields �including kebontatangkalan patches�, and human settlement. The an-nual rainfall ranges from 1500 to 2500 mm. The mean

daily minimum temperature is 25 °C and the meandaily maximum temperature is 30 °C. Despite themajority of its population working in the agriculturalsector, landlessness in the two study villages is rela-tively high; not less than 40% of the households arelandless.

Sampling and data collection

The study encompasses ecological and socioeco-nomic aspects of kebon tatangkalan. Ecologicalaspect covers structural and compositional patterns ofvegetation in relation with faunal diversity. Floristicsurvey was carried out using sample plots; 148 plotswere randomly selected in the two study villages. Aunit of ownership was considered as one plot, andplant inventory was performed in each sampled plot.The presence and number of individuals of each spe-cies in each sample were recorded.

Avifauna and insects are considered as major taxathat have close relation with the presence of this typeof agroforest in the Citarum Watershed. Avian surveywas carried within these plots using Point Counttechnique �Bibby et al. 1992�. Data regarding insectcommunities were derived from previous survey car-ried out in the same study villages. Flying insect in-ventory was carried out using sweeping net, whereas,pit-fall trap technique was used for soil surfaceinsects �Southwood 1978�.

Socioeconomic aspect included in the present studyaimed to reveal social background of the owners andsocioeconomic function of this type of agroforest.Socioeconomic survey was undertaken using ques-tionnaires and semi-structured interviews �SSI�. Forinterviews using questionnaires, a number of ownersin the two study villages were randomly selected witha formula described in Lynch et al. �1974�. The totalnumber of owner selected as respondents was 77. Anumber of Key Informants �KIs� were purposivelyselected for SSI. The KIs were elderly people, headof the village, and villagers that depend on resourcesprovided by kebon tatangkalan.

Data analysis

To indicate species dominance in kebon tatangkalanin the study area, Importance Value Indices �IVI�were calculated for each species. Each index was cal-culated based on the relative values of species den-sity and its frequency. Multivariate analysis wasperformed to elucidate the relation between the dis-

172

tribution of trees and bamboos species in kebontatangkalan and a complex of environmental varia-bles �ter Braak et al. 1987�. The measured variableswere elevation, slope, area, light intensity, and shape.Apart from that, TWINSPAN analysis was performedto characterize species assemblages.

To elucidate the relation between complexity ofvegetation structure and diversity of birds, correlationanalysis was performed based on 31 kebontatangkalan plots that were selected randomly.Socioeconomic data were analyzed using simple sta-tistic calculations to summarize responses given bythe respondents; the results are presented in percents.

Results

Structural and Compositional Patterns of kebontatangkalan vegetation

Even though there is no official record with regard tothe exact period of the first establishment of kebontatangkalan in the Upper Citarum Watershed, someKey Informants believe that this land-use type hasexisted before 1900s. The presence of figs species�Ficus spp� with diameter at breast height more than80 cm indicate that the establishment of thisman-made vegetation is not recent. Inheritance sys-tem of kebon tatangkalan can also be used to indi-cate its long establishment in the Upper CitarumWatershed. 18.7% of respondents interviewed divulgethat they inherit the land from their parents. Some ofthem become heir to this agroforest in 1920s or be-fore. They mention that their parents inherited theland from the grandparents. Based on inventory car-ried out in the two study villages, the acreage of thisland-use type as compared to the total area of the vil-lage is very small, only 2.8% and 4.9%, respectively.Nowadays, patches of kebon tatangkalan are com-monly encountered in various sizes ranging betweenless than 0.5 ha and 10 ha.

The most dominant tree species between the twostudy villages are different. In Wangisagara village,located in the lower part, the dominant tree speciesother than bamboos are, among others, Pithecelobiumdulce Benth, Gnetum gnemon L., Arenga pinnata�Wurmb� Merrill, and Parkia speciosa Hassk. InSukapura village, the dominant species are Eucalyp-tus alba Reinw. ex Blume, Albizia montana, Callian-dra calothyrsus Meisn., and Euphorbia pulcherrimaWilld. ex Klotzch. Table 1, Table 2 summarize other

species commonly found in the two study villages. Interms of species richness, a total number of 228 spe-cies of various growth forms are encountered in the148 plots, but only 64 species belong to trees andbamboos categories. The mean densities of tree standis about 1020 mature and young individuals per ha,whereas, bamboos is about 268 clumps per ha.

Ordination analysis indicates that kebontatangkalan plots in higher elevation are usually lo-cated in sloppy areas �Table 3�. Apart from that, theshape of this multi-layered agroforest in high eleva-tion tends to be elongated following topographicalcontours. Elevation and slope are the measured bio-physical factors that correlate most significantly withthe distribution of plant species in kebon tatangkalan�Table 3�. Some woody species are only found�planted� in higher elevation, for example, Albiziamontana, Calliandra calothyrsus, Pinus merkusiiJungh., and Eucalyptus alba �in Figure 1, these spe-cies are located in the right-hand side of the ordina-tion diagram�.

In sloppy areas, some woody species commonlyfound are, among others, Mangifera foetida Lour.,Spathodea campanulata L., Acer laurinum Hassk.,Pinus merkusii Jungh., and Gigantochloa atter Hassk.Ordination diagram also indicates that in plots domi-nated by bamboos species, e.g., Gigantochloa apus�Schult & Schult�, G. verticillata Willd, and Bambusavulgaris Schrader ex Wendland, light intensity tend tobe low. In contrast, light intensity is higher in mixedkebon tatangkalan where various fruit trees can beencountered �in Figure 1, these fruit tree species canbe found in the left-hand side of the diagram�.

TWINSPAN analysis indicates that all the speciesare clustered into 12 plant assemblages �Figure 2�.The number of species in each plant assemblage var-ies between 8 and 41 species. This indicates that thistype of land-use is not only diverse in terms of spe-cies and growth forms, but also it varies in terms ofplant assemblages. Some species usually grow in for-est area are also present in this man-made vegetation,among others, Alsophylla glauca �Blume� J.Smith,Neonauclea lanceolata Merr., Acer laurinum Hassk.,Piper aduncum L., and Macaranga tanarius �L�Muell. Arg.

Diversity of faunas in kebon tatangkalan

Kebon tatangkalan as multi-layered agroforest and itsdiverse species composition is very important habitatfor various organisms living in the agricultural land-

173

scape, particularly birds and insects. In the presentstudy villages, it was found that the majority of the62 species of birds encountered during the study usekebon tatangkalan vegetation for perching, foraging,and nesting. The present study indicates that there isa tendency of more diverse bird species in morecomplex vegetation layering �Pearson’s correlation �0.47; significant at 0.01 level�.

Furthermore, correlation analysis performed be-tween the number of vegetation layers and differentcategories of avian community indicate that there is atendency of higher diversity of insectivorous birds inmore complex vegetation structure �Pearson’s corre-lation � 0.45; significant at 0.05 level�. But weakcorrelation is exhibited between the number of layersand the diversity of frugivorous birds �Pearson’s cor-

relation � 0.13�. This result suggests that the multi-layered agroforest in the Upper Citarum Watershed isa suitable habitat for various insects that are impor-tant for birds’ prey.

The diversity of insects living in this type of agro-forest is high, sixty-six families of flying insect areidentified, and the diversity could be higher sinceparticular families consist of more than one species.While, soil surface insects are less diverse; only 23families are identified.

Socioeconomic functions of kebon tatangkalan

The ownership of kebon tatangkalan may reflect so-cioeconomic level of the owners. Inventory on landownership indicates that 96% of owners interviewed

Table 1. Some dominant species �IVI � 5%�, its main use and occurrence in kebon tatangkalan in Wangisagara village, West Java, Indo-nesia.

Botanical name Local name IVI �%� Use Occurrence

A. TreesPithecelobium dulce Benth Jengkol 23.34 Fr PGnetum gnemon L Tangkil 21.07 Fr PArenga pinnata �Wurmb� Merrill Aren 10.31 Fr/Sf/B WParkia speciosa Hassk Peuteuy 10.12 Fr/Fw PArtocarpus heterophyllus Lam Nangka 9.38 Fr PMangifera indica L Buah 7.01 Fr PCarica papaya L Gedang 6.80 Fr/Me PPersea americana Mill Alpuket 6.47 Fr PHibiscus similis Blume Tisuk 6.32 Fw PPsidium guajava L Jambu batu 5.43 Fr/Fw P

B. BamboosGigantochloa apus Schult & Schult Awi tali 72.49 C/Fw/Fe PGigantochloa atter Hassk Awi temen 51.73 C/Fw/Fe PGigantochloa verticillata Willd Awi ageung 48.51 C/Fw/Fe PBambusa vulgaris Schrader ex Wendland haur hejo 27.18 C/Fw/Fe P

C. ShrubsChromolaena odorata �L� RMKing&Hrob Kirinyuh 24.91 Fw WSida retusa Linn Sadagori 16.43 N WUrena lobata Pungpulutan 14.79 N WGlichenia linearis Paku api 14.11 N W

C. Non-graminoid herbsEupatorium riparium Regel Teklan 52.13 N WAgeratum conyzoides L Babadotan 19.91 N WOxalis corymbosa DC Calincing 14.15 N WAlternanthera sessilis �L� RBr. ex DC Gelang 7.49 N WIpomoea batatas �L� Lam. Hui boled 7.18 F/Fo PMusa paradisiaca L Cau 6.46 Fr PSynedrella nudiflora �L� Gaertn Jukut ibun 5.38 N W

C � construction material; Fw � fuelwood; Fr � fruit; Fo � fodder; O� ornamental; B�beverage; Sf � sago flour; Fe � fence; F � food;Me � medicine; N � none; P � planted; W � wild.

174

in the present study possess rice field and/or vegetablegardens, other than kebon tatangkalan land. Someowners possess one to three hectares of agriculturalland, while the average of land ownership in thepresent study area is not more than 0.2 ha. Highereconomic status can also be indicated by the fact that75.3% of owners acquire kebon tatangkalan land by

purchasing from other villagers. This suggests that inspite of elevating price of agricultural land due to itslimited availability, they still possess enough cash topurchase land.

All respondents mention that prior to the expansionof input-intensive agriculture practices in the late1970s and when the demand for bamboos and other

Table 2. Some dominant species �IVI � 5%�, its main use and occurrence in kebon tatangkalan in Sukapura village, West Java, Indonesia.

Botanical name Local name IVI �%� Use Occurrence

A. TreesEucalyptus alba Reinw ex Blume Kalices 29.83 C/Fw PAlbizzia montana Selong montana 28.43 C/Fw PCalliandra calothyrsus Meisn. Kalianra 25.37 Fw/Fo PEuphorbia pulcherrima Willd ex Klotzch Kastuba 11.81 O PCoffea arabica L Kopi 10.89 B PArenga pinnata �Wurmb� Merrill Aren 9.53 Fr/Fw/Sf/B WArtocarpus heterophyllus Lam Nangka 8.80 Fr/Fw PSyzigum densifolia Jambu 8.68 Fw WToona sureni Merr. Suren 7.68 C/Fw/Fo PPersea americana Mill. Alpuket 6.83 Fr PMangifera indica L Buah 5.44 Fr P

B. BamboosGigantochloa atter Hassk. awi temen 74.90 C/Fw/Fe PGigantochloa apus Schult&Schult awi tali 67.26 C/Fw/Fe PGigantochloa verticillata Willd awi ageung 43.83 C/Fw/Fe P

C. ShrubsChromolaena odorata �L� RMKing&H.Rob Kirinyuh 52.67 Fw WSida retusa Linn. Sadagori 51.14 N WUrena lobata Pungpulutan 16.85 N WCestrum nocturnum L. Kembang dayang 14.46 Fe P

D. Non-graminoid herbsEupatorium riparium Regel Teklan 63.27 N WOxalis corymbosa DC Calingcing 29.17 N WAgeratum conyzoides L Babadotan 17.08 N WAlternanthera sessilis �L�RBr. Ex DC Gelang 8.28 N WIpomoea batatas �L� Lam Hui boled 6.91 F PMusa paradisiaca L Cau 6.87 Fr P

C � construction material; Fw � fuelwood; Fr � fruit; Fo � fodder; O� ornamental; B�beverage; Sf � sago flour; Fe � fence; F � food;N � none; P � planted; W � wild.

Table 3. Canonical coefficients, intraset correlation coefficients for axes 1 and 2, and weighted correlation among variables in kebontatangkalan in the Upper Citarum Watershed �only for woody species�.

Variable Canonical Coeff. Intraset Corr. Matrix of correlation among variables

Axis 1 Axis 2 Axis 1 Axis 2 Elevation Slope Light Area Shape

Elevation 0.979 0.602 0.990 0.121 – 0.401 � 0.485 0.067 0.489Slope 0.141 -0.997 0.408 -0.880 – � 0.206 0.168 0.313Light intensity 0.006 0.075 � 0.511 0.064 – 0.026 � 0.229Area � 0.038 0.111 0.033 0.024 – 0.126Shape � 0.043 � 0.297 0.521 � 0.143 –

175

timber used for various purposes were substantial,kebon tatangkalan was an important source ofincome. But this is no longer the case in the presenttime; none of the owners rely on this traditional sys-tem as their main source of income. All respondentspoint out that they rely more on rice and/or vegetablecultivation as family income.

Despite decreased production, all owners admitthat in the present time the price of particular kebontatangkalan product, particularly bamboos and tim-ber, is higher than in five years ago. The current priceof one bamboo varies between Indonesian Rupiah�IDR� 5,000 and IDR 8,000, or approximately US$0.6 – US$ 0.9 �US$1 � IDR 8,900, January 2003�,depending on the kind �species� of bamboo. While,the price of one mahogany �Swietenia mahagoni �L�Jacq.� tree, with trunk diameter about 50 cm, is ap-proximately US$11.23. The price of a 6-year-old tree

of Albizia falcata �one of the commonest introducedtree species� with trunk diameter about 20 cm is be-tween US$ 4.50 and US$ 5.60. In the last five years,Albizia woodlots tend to expand gradually due to itscharacteristic as fast growing species, therefore, theowners are more attracted to grow this species andalso due to its higher market demand.

In the present time, the productive function of ke-bon tatangkalan is only to support subsistence needin fruit and vegetable consumption. Inventory carriedout in the present study indicate that there are about16 fruit-producing tree species found in all sampledplots. Fruit tree species such as Artocarpus hetero-phyllus Lam. �jackfruit�, Psidium guajava L. �guava�,Mangifera indica L. �mango�, and Persea americanaMill. �avocado� were among the most commonlyplanted in this agroforest. But nowadays, their sub-sistence role seems to be diminishing. The owners,

Figure 1. Biplot of CCA ordination diagram of woody species and some environmental variables in kebon tatangkalan in the Upper CitarumWatershed, West Java, Indonesia; species names are abbreviated �see Appendix for full scientific names�.

176

for various reasons, infrequently carry out replantingnew fruit trees. The study reveals that no more than20% of respondents still perform tree planting in thelast five years. More than 60% of the respondents re-veal that constraint in marketing the product andthievery are among the main reasons that make themreluctant in replanting new seeds.

Almost all of the owners reveal that kebontatangkalan in the present time is a valuable assetmerely from the point of view of the land rather thanthe products it produces. They admit that very lim-ited land available in the Upper Citarum region makesthe price of land very expensive. The price of onetumbak �local term which is equivalent with 14 m2�of land in Sukapura village, for example, could be ashigh as IDR 400,000 �approximately US$ 44.90�.Therefore, most of the owners consider the land as akind of saving that can be sold in case they need somecash. Some owners rent their kebon tatangkalan landin order to improve its economic function.

Despite its decreasing productive function forhousehold economy, kebon tatangkalan still has so-cial function. This multi-layered agroforest has beenan important part in inheritance system for genera-tions. The parents usually inherit the land to their de-

scendants. Almost all respondents admit that ownerwill not sell this asset unless emergency situation oc-curs, e.g., big cash is needed for ‘Hajj’ pilgrimage.

Another social function of this land-use type is infuelwood fulfillment, particularly for poor villagers.There are various plants growing in this multi-layeredagroforest that can be used as fuelwood �see Table 1,Table 2�. All the interviewed owners state that theyusually allow other villagers to use their land assource of fuelwood as long as no tree is cut. Fieldobservation confirms that in villages located far fromforested lands, this multi-layered agroforest is themost important source of fuelwood for poor villagers.

The dynamics of kebon tatangkalan and its affectingfactors

All Key Informants argue that expansion of intensiveagriculture affects the existence of kebon tatangkalan.Nowadays, intensive agricultural land, small- as wellas large-scale industrial areas, and human settlementcover no less than 80% of the total area of the upperpart of Citarum Watershed �approximately 2,000km2�. All respondents also point out that continuingpopulation growth is another factor that causes con-

Figure 2. Floristic classification �TWINSPAN� for 148 plots of kebon tatangkalan in the Upper Citarum Watershed, West Java, Indonesia;numbers in the dendogram indicate the number of plant species belonging to each group �only woody and bamboos species are listed; speciesnames are abbreviated�.

177

version of the existing agricultural land, includingkebon tatangkalan.

According to respondents, the occurring land-usechange is aimed to accommodate economic improve-ment efforts. This has caused less productive land-usetype to diminish from the landscape of the UpperCitarum Watershed. Kebon tatangkalan is among themost affected in the dynamics of land-use change inthis area. Unfortunately, none of the respondents canestimate the annual conversion rate of this agroforestto other land-use types. However, the tendency of de-creasing number of owners can be used as indicatorof the decreasing area of kebon tatangkalan.

The present study reveals that the number of own-ers that have this traditional agroforest land withacreage larger than 400 m2 is lessening. The numberof owners decreases slightly compared to inventorycarried out in 1995. In Sukapura village, there were58 households owned kebon tatangkalan in 1995, andit decreases to 47 households in year 2000. Similarly,the number of owners also decreases in Wangisagara,from 105 households in 1995 to 100 households inyear 2000. Based on the inventory, the present totalnumber of households in these two villages �about2600 households� that own kebon tatangkalan is onlyabout 2.5%.

Field observations carried out between 1995 and2000 in higher elevation area �above 1000m ASL� in-dicate that many kebon tatangkalan patches havebeen converted to a more productive agricultural landlike cash crop fields. In this area, climatic and edaphicconditions are more suitable for farmers to grow an-nual crops such as potatoes, cabbage, green onion,and carrots. Interviews with some owners suggest thatif farmer converts one-patok �approximately 400 m2�kebon tatangkalan to be planted with carrots, hecould earn at least IDR 75,000 �about US$ 8.80�within 70 days. In comparison, he has to wait for atleast five years in order to log the planted timber �Al-bizia or Eucalyptus� to gain the same amount of cash.

The ongoing decline of kebon tatangkalan has re-sulted in greater effort carried out by fuelwood usersto collect adequate amount of fuelwood. Nearly 50%of fuelwood user respondents reveal that nowadaysthey have to spend more time to gather fuelwoodfrom this agroforest. They argue that to collectadequate amount of fuelwood for daily need theyhave to go from one place to another as patches ofthis agroforest are becoming more fragmented in theagricultural landscape of Citarum Watershed.

Discussions and conclusions

Structural patterns of kebon tatangkalan andcomparison with other forest gardens

The biological features of kebon tatangkalan exhibitssimilarities with an agroforest type, called talun,found in Soreang district� also located in theCitarum Watershed �Christanty et al. 1996�. Varioustree and bamboo species are common to both areas,among others, Arenga pinnata, Parkia speciosa,Mangifera indica, Albizia falcata. The four bamboospecies found in the present study sites are also verycommon in talun vegetation. Bamboos appear to bean important component because they are alwayspresent in every patch of this agroforest in both ar-eas. Specific term ‘talun awi’ �meaning bamboo ta-lun� is popular among farmers in Soreang district, butnot in the present study area.

Despite its similarity, there is to some extent a dif-ference between the two areas in terms of manage-ment of this man-made vegetation. A variation oftalun system, called talun-kebun rotation cycle, iscommonly practiced by farmers in Soreang district.As described by Mailly et al. �1997�, the cycle con-sists of one year of mixed species vegetable cropping,followed by a year of cassava cropping, and fouryears of bamboo fallow, this makes the total cyclelength of six years. In the present study area, a mix-ture of perennial plants and annual crops can also beencountered in some kebon tatangkalan patches. Butthis mixture is not a talun-kebun system because norotation system is intended by the farmers.

Compositional feature of kebon tatangkalan alsodemonstrates similarity with multistoried village gar-den found in Bogor area, also located in the Provinceof West Java. Various trees occupying main canopiesof kebon tatangkalan such as Pithecelobium, Parkia,Mangifera, Artocarpus, Arenga, Gnetum, and bam-boos are also common in this type of village garden�Michon and Mary, 1994�. Compositional similaritybetween the two systems also occurs in undergrowthshrubs such as Sauropus androgynus, Claoxylon po-lot, Phylllanthus niruri, Glochidion borneense, andsome species belonging to Asteraceae and Verben-aceae.

Nevertheless, kebon tatangkalan has less numberof species compared to multistoried village garden inBogor. Michon and Mary �1994� point out that com-mon garden plots, 300 to 500 m2 in area, may con-tain more than 50 different tree and herb species. In

178

kebon tatangkalan with approximately the same areaas that in the village garden in Bogor, the number ofspecies is not more than 40 �the range is between 12and 40; with average 23.9 � 7.1�. Similar conditionis also true for tree density, in which village gardenin Bogor has higher mean density of mature andyoung trees than kebon tatangkalan.

It is quite interesting that kebon tatangkalan sharessome common species with another typical forestgarden, called dusun, in Central Maluku �Kaya et al.2002�. Yet, the two sites are separated by a distanceof not less than 2000 km. Among these common spe-cies are Mangifera indica Blume, Cocos nuciferaLinn, Syzygium aromaticum (L.) Merr, Durio zibethi-nus Moon, Toona sureni Merrill, and Arenga pinnataMerrill. This suggests that these species are importantcomponents of multi-layered agroforest in Indonesiasince a long time ago. Despite this, both systems havedifferent origin of initial establishment. The dusunwas initially established with the cutting of a patch ofpristine forest, but fruit trees are usually left uncut�Kaya et al. 2002�. Furthermore, the starting point ofdusun is the establishment of a clove or coconutplantation in open areas. The initial establishment ofkebon tatangkalan was undertaken by planting anyuseful species of trees and bamboos in privatelyowned land.

The existence of kebon tatangkalan in the UpperCitarum Watershed represents the evolutionary devel-opment of local knowledge in the formation of artifi-cial vegetation. The owners are able to combine theirpreference and the abiotic conditions, which can beseen from the result of ordination analysis. The own-ers living in different agroclimatic zones are able tocharacterize species suitable to grow in that area. Onthe other hand, they also acknowledge the importanceof tree planting in sloppy area.

General classification described in Christanty et al.�1996� only mentioned two different types of kebontatangkalan �talun� commonly found in West Java,i.e., mixed talun and bamboo-dominated talun. Thepresent study has identified twelve different plant as-semblages in kebon tatangkalan. The occurrence ofthese variations suggests that this agroforest typecontributes to the heterogeneity of the agriculturallandscape in the Upper Citarum Watershed.

Apart from high variation in plant assemblages�communities�, kebon tatangkalan patches in thestudy area also exhibit high species variability as in-dicated by low values of Sorensen’s Similarity Index�Parikesit et al. 1997� between paired samples. This

suggests that variation of the overall species compo-sition among patches is apparent. High species vari-ability also suggests that resemblance on speciespreference among owners only occurs for particularspecies such as the four species of bamboos, Artocar-pus heterophyllus, Pithecelobium dulce, and Gnetumgnemon. On the other hand, variation of microcli-matic condition under the main canopy in differentpatches is likely key factor resulted in high speciesvariability in the agroforest.

Vandermeer et al. �1998� emphasized two distinctcomponents of biodiversity in agroecosystem per-spective, called ‘planned diversity’ and ‘associateddiversity’, both are relevant in the context of kebontatangkalan. Owners’ decision with regard to speciesplanted in their land is based on their preference anddifferent purpose of uses such as for timber and con-struction materials, fuelwood, subsistence foods, andprevention of soil erosion. This has resulted in highplanned diversity in this agroforest. On the otherhand, unintensive maintenance of this agroforest typehas resulted in the presence of various unplanted plantspecies with different life forms that eventually cre-ates more complex vegetation structure. This kind ofstructure generates suitable habitat for numerous fau-nas like insects and birds. The presence of numerousunplanted plants, birds, and insects is the associateddiversity in kebon tatangkalan.

At larger scale, the presence of the multi-layeredagroforest with its variation of plant assemblages inthe agricultural landscape of the Upper Citarum Wa-tershed makes the landscape structure heterogeneous.Overall, this is very important in maintaining biodi-versity condition and soil conservation in the courseof continuous changes of the landscape due to humandisturbance. Previous study indicated that the rate ofrelative erosion situation in this agroforest was muchless, only a half than that in upland fields planted withcassava or rice �Institute of Ecology, 1981�.

Socioeconomic perspectives of kebon tatangkalanand the impacts of its conversion

It is likely that the current economic function of theproducts from kebon tatangkalan is not as importantas that in some other forest garden systems. In ‘tem-bawang’ system in West Kalimantan, for example,selective weedings is performed during harvest sea-son to facilitate fruit collection and also to maintainseedlings and saplings of valuable species. Theincome earned by many families throughout the

179

month-long season was substantial �Padoch and Pe-ters 1994�. No such management practice is commonin kebon tatangkalan system; none of the ownersperform weeding or other agricultural practices aimedto obtain good fruit harvest.

The productive function of kebon tatangkalan thatstill can be found relates to vegetable consumption.Leaves and inflorescence of some species supplyvegetable to villagers as part of their daily diet.Among these species are Gnetum gnemon, Sauropusandrogynus, Claoxylon polot, Xanthosoma sp., andManihot esculenta. These same species are alsopresent in village gardens in Bogor area and theymake up about 70% of vegetation supply for localvillagers �Michon ands Mary 1994�. But, differentwith that in village gardens, villagers in the UpperCitarum Watershed do not consume shoots of somebamboo species �Gigantochloa apus, G. verticillata,and G. ater�. Similarly, use of bamboo shoots forcommercial purpose is also absent. All respondentsare not familiar with bamboo shoots market, althoughthey acknowledge that there is a potent demand forthis product. Bamboos are grown mainly for buildingmaterials such as for house floor and wall, animalshelter, irrigation pipes, bridges, and to make poles toattach legume vines.

For many villagers, the presence of multi-layeredagroforest in the Upper Citarum Watershed cannot beneglected. Recent study �Parikesit et al. 2001�revealed that this multi-layered agroforest is the mostimportant source of fuelwood for poor people livingin villages far from forested area. About 51% of therespondents used fuelwood for daily cooking, inwhich 88% of the fuelwood users used kebontatangkalan as their source of fuelwood. Therefore, ifthis main source totally disappears from the agricul-tural landscape of the Upper Citarum Watershed, thesupply of ‘free energy’ from agroecosystem that hasbeen taking place for decades will come to an end.

Conversion or simplification of vegetation struc-ture �e.g., through woodlot establishment� of kebontatangkalan is believed to provide the owners withhigher economic gain. Cultivation of annual cashcrops through input-intensive agricultural practices toreplace traditionally managed agroforest has resultedin considerable increase of economic output. None-theless, whether the gain from cash crop cultivationis sustainable in a broader sense should be ques-tioned. The detrimental effect of modern agriculturein various parts of the world’s ecosystems, includingthe extensively practiced conventional monoculture

systems, is widely recognized �e.g., Giller et al. 1997;Sotherton 1998; Kammerbauer and Ardon 1999�.

Overall, conversion of this multi-strata agroforestat the landscape level will reduce spatial heterogene-ity. It is also important to point out that the conver-sion might cause habitat loss for various faunas livingin this complex man-made vegetation. This ultimatelywould affect the integrity of the entire ecosystemwithin the landscape of Citarum Watershed.

It is obvious that the decline of this long timeknown agroforest is due to its decreasing role in gen-erating significant income to the owners. Therefore,maintenance of this traditional agroforest should beundertaken through revitalization effort. To encouragethe owners to maintain their kebon tatangkalan, anincentive in the form of free seedling assistanceshould be considered. This is in concordance withtheir expectation that government or any developmentagency working in rural development should providethem with high quality fruit trees that have good mar-ket demand. However, it should be pointed out thatrevitalization should not put too much emphasis onthe perspective of the economic function. In concor-dance with the goal of agroforestry development, ithas to take into account any relevant ecological con-sideration.

Above all, the decline of kebon tatangkalan is aconsequence of lack of policy and institutional sup-port to maintain this complex man-made vegetation.The present policy in land-use planning and conser-vation established by the West Java Provincial Boardof Planning and Development does not address spe-cifically for the maintenance of this ecologically im-portant land-use. On the other hand, conservationeffort is likely to be confined in more natural ecosys-tem such as the remnant forest in mountainous areas.

Constraints and prospects of maintaining kebontatangkalan in the changing agricultural landscape

Vandermeer et al. �1998� point out that in multi-spe-cies systems, human dimension makes the problemmore complicated because its indicators have moredynamic character than biophysical ones. Apart fromthat, future sociopolitical aspect has to be taken intoconsideration as well in carrying out evaluation of thecurrent land-use options. The above statement sug-gests that effort to halt the decline of kebontatangkalan will certainly have considerable con-straints. To convince farmers that complex agroforestand its economic benefit are more advantageous in the

180

long-term than agricultural practice that has maxi-mum but short-term profit is not a simple task. It isalso a great challenge for agroforestry advocates topersuade farmers that continuing expansion of capi-tal-intensive agriculture that leads to the homogeni-zation of land-use type will ultimately cause ecosys-tem instability.

The custom of tree growing practiced for more thana century by the local people is in fact a crucial‘capital’ for the preservation of this multi-layeredagroforest. Apart from that, there is an old traditionamong the owners that this traditional agroforest is afamily asset and it is a part of their inheritance sys-tem. On the other hand, climatic and edaphic condi-tions of the area are suitable for various tree speciesto grow. The introduction of new and high productivespecies should be encouraged since the present studyrevealed that in the study area the so-called ‘Cinder-ella’ species �Leakey and Newton 1994; cited in Nair1998� is almost non-existent. There are various spe-cies from other regions that can be introduced suchas Artocarpus champedens L, A. communis Forst.,Anacardium occidentale L, Achras sapota �Gilly�Lundell, and Nephelium lapaceum L.

Acknowledgements

Financial support for this study was obtained fromThe John D. and Catherine T. MacArthur Foundation,Japan Society for Promotion of Science, and BasicResearch Project – Indonesian Ministry of Educationand Culture. The authors would like to thank Ms. T.Sodjakusumah, Mrs. Arifianto and A. Susilo for theirassistance during the fieldworks.

Appendix

Table A1. List abbreviation and full scientific names of tree, bam-boos, and other woody species in kebon tatangkalan in the studyarea.

Per ame Persea AmericanaMag ind Mangifera indicaCar pap Carica papayaErt cars Erythrina carssifoliaArt ela Artocarpus elasticusArt het Artocarpus heterophyllusGig ver Gigantchloa verticilataGig apus Gigantchloa apusGig ater Gigantchloa atterCal cal Calliandra calothyrsusEuc alba Eucalyptus alba

Pin mer Pinus merkusiiAlb mon Albizia montanaToo sur Toona sureniHib mac Hibiscus macrophyllusHib sim Hibiscus similisCam sin Camelia sinensisCof arb Coffea arabicaDur ere Duranta erectaEup pul Euphorbia pulcherrimaMal arb Malvaviscus arboreusAre pin Arenga pinnataFic amp Ficus ampelasAcr lau Acer laurinumSyz aqu Syzygium aqueumPsi gua Psidium guajavaAla chi Alangium chinensisMac tan Macaranga tanariusRic com Ricinus communisFic brev Ficus brevicuspisFic spp Ficus spJat cur Jatropha curcasBam vul Bambusa vulgarisMic alb Michelia albaMan foe Mangifera foetidaSyz den Syzigium densifoliaLit sp Litsea spAlb fal Albizia falcataCit max Citrus maximusMim pud Mimosa pudicaMim pig Mimosa pigraPan edu Pangium eduleMor cit Morinda citrifoliaMel umb Melochia umbellataMel aze Melia azedarachHib mac Hibiscus macrophyllusCla pol Claoxylon polotLan cam Lantana camaraChr odo Chromolaena odorataCoc nuc Cocos nuciferaEla pie Elaeocarpus pierreiSam java Sambucus javanicusGlo bor Glochidion borneenseBix ole Bixa oleranaSau and Sauropus androgynusSpa cam Spathodea campanulataAca sia Acasia siameaLea ind Lea indicaGal dul Garcinia dulcisBel chi Belamcamda chinensisDys nut Dysoxylum nutansHom pul Homalanthus pulponeusFic sep Ficus septicaFic ben Ficus benyaminaAlb chi Albizia chinensisDis chi Disporium chinensisGne gnm Gnetum gnemon

181

References

Bibby J.C., Burgess N.D. and Hill D.A. 1992. Bird census tech-niques. Academic Press, London, UK.

Christanty L., Mailly D. and Kimmins J.P. 1996. ‘Without Bam-boo, the Land Dies’: Biomass, Litterfall, and Soil Organic Mat-ter Dynamics of a Javanese Bamboo-Talun-kebun System.Forest Ecology and Management 87: 75–88.

Erawan T.S., Djuangsih N., Muchtar M., Setiana H. and Istanti L.S.1997. Community structure and diversity of fauna in UpperCitarum River Basin, West java, Indonesia.. In: Dove M.R., Sa-jise P.E. �eds�, The Conditions of Biodiversity Maintenance inAsia. East – West Center. Program on Environment, Honolulu,Hawaii, pp. 73- 111.

Giller K.E., Beare M.H., Lavelle P., Izac A.M.N. and Swift M.J.1997. Agricultural intensification, soil biodiversity, and agroeco-system function. Applied Soil Ecology 6: 3–16.

Institute of Ecology 1981. Land-use and vegetation changes anderosion situation in Citarum Watershed Research report. Padjad-jaran University, Bandung, Indonesia. �In Indonesian�

Kammerbauer J. and Ardon C. 1999. Land-use dynamics and land-scape change pattern in a typical watershed in the hillside regionof central Honduras. Agriculture, Ecosystems, and Environment75: 93–100.

Kaya M., Kammesheidt A. and Weidelt H.- J 2002. The forest gar-den system of Saparua Island, Central Maluku, Indonesia, andits role in maintaining tree species diversity. Agroforestry Sys-tems 54�3�: 225–234.

Lynch F., Hollnsteiner M.R. and Covar L.C. 1974. Data gatheringby social survey. Trial edition. Social Science Council, Inc. Qu-ezon City, the Philippines.

Mailly D., Christanty L., Kimmins J.P. 1997. ‘Without bamboo, theland dies’: nutrient cycling and biogeochemistry of a Javanesebamboo talun-kebun system. Forest Ecology and Management91: 155–173.

Michon D. and Mary F. 1994. Conversion of traditional villagegardens and new economic strategies of rural households in thearea of Bogor, Indonesia. Agroforestry Systems 25: 31–58.

Nair P.K.R. 1998. Directions in tropical agroforestry research: past,present, and future. Agroforestry Systems 38: 223–245.

Padoch C. and Peters C. 1994. Managed forest gardens in WestKalimantan, Indonesia.. In: Potter C.S., Cohen J.J. and JanezwkyD. �eds�, Perspectives on biodiversity: case studies of genetic re-source conservation and development. AAAS Press, Washington,DC, USA, pp. 167- 176.

Parikesit , Djuniwarti and Hadikusumah H.Y. 1997. Spatial Struc-ture and Floristic Diversity of Man-made Ecosystems in UpperCitarum River Basin.. In: Dove M.R., Sajise P.E. �eds�, TheConditions of Biodiversity Maintenance in Asia. East – WestCenter. Program on Environment, Honolulu, Hawaii, pp. 17- 43.

Parikesit , Takeuchi K., Tsunekawa A. and Abdoellah O.S. 2001.Non-forest Fuelwood Acquisition and Transition in Type of En-ergy for Domestic Uses in the Changing Agricultural Landscapeof the Upper Citarum Watershed, Indonesia. Agriculture,Ecosystems, and Environment 84: 245–258.

Soemarwoto O. and Soemarwoto I. 1985. The Javanese Rural Eco-system.. In: Rambo T.A., Sajise P.E. �eds�, An Introduction toHuman Ecology Research on Agricultural System in SoutheastAsia. University of the Philippines at Los Banos, ThePhilippines.

Sotherton N.W. 1998. Land-use changes and the decline of farm-land wildlife: an appraisal of the set-aside approach. BiologicalConservation 3: 259–268.

Southwood T.R.E. 1978. Ecological methods with particular refer-ence to the study of insect populations. Chapman and Hall,Wiley, London-New York, 524 pp.

ter Braak C.J.F., Jongman R.H.G. and van Tongeren O.F.R. 1987.Data Analysis in Community and Landscape Ecology. Cam-bridge Press, Cambridge, UK.

Vandermeer J., Noordwijk M., Anderson J., Ong C. and Perfecto I.1998. Global change and multi-species agroecosystems: con-cepts and issues. Agriculture, Ecosystems, and Environment 67:1–22.

182