Contents - Previous - Next

This is the old United Nations University website. Visit the new site at

Imperata - A problem or a solution?

Imperata cylindrica (L.) Raeuschel var major has been nominated as one of the world's 10 worst weeds (Holm and Herberger, 1969). Although descriptive accounts of its spread in various parts of the region had already appeared by the mid-nineteenth century, more intensive biological studies over the past 30 years have led to increased knowledge of the characteristics of this plant and have pointed to possible directions for its control. Studies of its social and cultural importance as a component of farming systems have also been carried out in recent times, while its economic possibilities have been assessed, ranging from pulp for paper through concentrated cattle feed to wine.

Although propagated by seed, one of the main characteristics of Imperata is the formation of a dense mat of rhizomes just below the surface. The apical bud of the rhizome will turn up to form a new shoot, while sub-apical buds form rhizome branches (Ayeni, 1985). Rhizomes may remain for long periods in the soil without losing their germination capacity, and they may not be directly connected with above-ground shoots (Eussen and de Groot, 1974). They are constantly capable of sending up new shoots even after destruction of the above-ground biomass after fire - a characteristic that allows this plant to persist and spread at the expense of secondary forest species, which may be destroyed by fire. Competition for space, nutrients, and water in the root zone will allow the grass to crowd out crop plants, while at the same time the Imperata secretes allelopathic substances to further inhibit the growth of competitors (Eussen, Slamet, and Soeroto, 1976; Eussen, 1977).

The main reason for this behaviour is that Imperata is light-loving and will die if tree species are able to grow tall enough to shade it out. In primary forest, it may appear briefly in natural gaps but will quickly be shaded out by tree regrowth. Once established in larger gaps or following several years of continued shifting cultivation, it will still eventually be shaded out if the regrowth is protected from fire. The establishment of what is known as "sheet alang-alang'', in which the grass is dominant over wide areas, requires human activity through regular burning, which favours the grass over other species, unless these are also fire-resistant and possess a deep root system. Introduced weeds, such as Lantana camara and Chromolaena odorata (formerly, and still sometimes in the literature, Eupatorium odoratum), may compete successfully with the Imperata (Eussen and de Groot, 1974) and enable plant species with a higher canopy, such as Vitex pubescens (labun, halabun, keleban) and types of Macaranga or Trema, to become established. Although known as a persistent weed on plantations, Imperata may fairly easily be controlled through the establishment of cover crops until shading out by the trees takes place. It is the sheet alang-alang areas that provide the biggest problems when restoration of tree cover is desired.

Although environmental and PROCEZ criticality, as examined in the Appendix, will be kept at the forefront of this analysis, it must be remembered that there is another definition of "critical land" as used by Indonesian writers. As Soerjani (1980: 9) summarized: "in Indonesia in particular and in Southeast Asia in general, there is a strong tendency for critical land to be wasteland occupied by alang-alang." He went on to state that critical land denotes unproductive land from the agricultural point of view, and may fall into three types: hydro-orologically critical (in a watershed area); physically and technically critical (in a process of degradation, such as in alang-alang fields); or socio-economically critical, where its utilization exceeds its carrying capacity. Soerjani added that "outside Java critical land is covered with alangalang and shrubs as a result of improper shifting cultivation practices." We are thus presented with a complete argu ment that at once defines critical land and provides an explanation for its origin and a solution to its spread - the elimination of shifting cultivation.

Generations of Dutch, British, and French colonial administrators throughout South-East Asia believed that country dominated by Imperata cylindrica was simply a wasteland, an idea also expounded by Gourou (1953) and for Indonesia by Geertz (1963), repeated by Soerjani (1980). Although this view has persisted among some Indonesian and Malaysian scientists, foresters, and planners, the mid 1970s saw a re-evaluation in terms of its potential use for both agriculture and grazing. Soerjatna and McIntosh (1980) emphasized the importance of the grass in providing a fast-growing cover to the soil after deforestation, thus preventing erosion and leaching. Grassland areas had been identified as preferred sites for transmigration projects, given that they were so extensive and that alternative ecological zones, such as forest or tidal swamp, had presented problems in terms of availability, accessibility, and soil quality (Soerjatna and McIntosh, 1980; Burbridge, Dixon and Soewardi, 1981).

In recent times the choice of forest sites for transmigration has come under pressure from the international conservation movement, giving even more impetus for the selection of locations in the grasslands. Inputs such as fertilizer have been suggested, however, as essential for permanent farming. Tjitrosoedirdjo and Wiroatmodjo (1982) recommended herbicide with minimum or zero tillage for the preparation of grass-based transmigration sites, but more common has been the integration of cattle as work animals to eliminate the alang-alang by ploughing (Soewardi and Sastradipradja, 1980). Government agencies opposed to shifting cultivation saw the development of permanent grassland farming as the preferred alternative for local farmers who worked swidden plots in forested areas near grassland sites.

Soewardi (1976) suggested that development of animal husbandry in the alang-alang dominated areas was feasible, since its qualities as forage were not unfavourable when compared with other grasses, although some feed supplements would usually be required. He considered that in large areas of Kalimantan, under conditions of low land capability and lack of socioeconomic capacity for extensive pasture improvement, alang-alang should be considered not a weed but an opportunity (Soewardi, 1976; Soewardi and Sastradipradja, 1980).

Meanwhile anthropologists and geographers were also urging a closer examination of the activities of indigenous populations in creating and maintaining areas of alang-alang. Seavoy (1975: 49) stated that grasslands were an integral part of the economy and social structure of all villages of shifting cultivators that he had visited in Kalimantan, whether Dayak or Malay. "Every village has at least one small field of Imperata cylindrica (alang-alang) nearby. Kalimantan shifting cultivators purposely create this field by burning and reburning one or more abandoned ladang [dry fields] until Imperata becomes a pure stand." Seavoy claimed that grasslands were created for hunting purposes only and were not used by shifting cultivators to graze cattle. A second group, "part-time grazers," according to his description grew little rice and exchanged cattle for their rice needs. They would burn Imperata to secure fresh feed for their animals.

In his description of Batak grassland farming in the Lake Toba region of Sumatra, Sherman (1980) challenged Geertz's (1963) assumption that alangalang areas were useless "green desert." Bataks would turn the heavy grass sod with hoes (in earlier times digging sticks) and eventually break up the rhizomes by hand, on fields carefully terraced to prevent erosion. Both cattle and buffalo were reared, the latter being used to plough nearby wet-rice fields. They and the cattle, which were raised for sale, were stall-fed or pastured on young Imperata shoots. Their manure was collected and used on cash crops of onions. Complementarity thus existed between raising livestock and farming grassland.

Dove (1981) claimed that the existence of large areas of alang-alang in the Riam Kanan valley of South Kalimantan was attributable to the swidden practices of the local Banjarese population. According to Dove's sources, prior to World War II the people had, over a long period, deliberately planted newly opened land in dry rice for three out of the succeeding five years, after which only Imperata would grow on the site. As the forest diminished they had developed techniques for grassland cultivation, involving slashing, burning, and ploughing the sod using cattle. The latter, a recent innovation, replaced a former hoe-based system (canghulan from the Indonesian term for the common work-hoe, cangkul). Grassland cropping could be carried out for up to seven consecutive years, after which three years of fallow would be undertaken to prevent the Imperata from being replaced by short grasses. Dove stated that the Banjarese regarded the grasslands not as a problem, but as one of the solutions to their struggle for existence. This explanation of Banjarese behaviour differs in significant detail from both historical accounts and Potter's observations in the neighbouring valley, the Riam Kiwa. It is none the less illustrative of the kinds of differences that may well exist within quite small areas and forms part of the re-evaluations of grassland-based systems that took place in Indonesia during the 1970s and 1980s.

Indonesian government interest in agriculture in the alang-alang has now declined following the advent of large-scale reforestation projects using the fast-growing tree Acacia mangium. As described in chapter 5, extensive plans were drawn up in the 1980s for the creation of industrial forest estates (HTI), many of them in sheet alang-alang areas, destined for pulp and paper production. Major problems now beginning to surface relate to disputes with local villagers over claims to land in what were officially considered unoccupied grasslands, and the labour force available to work the estates, leading to the use of Javanese workers as also described above. Examining the social impacts of Imperata reforestation is the focus of new research efforts by a number of international agencies.

The occurrence of Imperata cylindrica in Borneo

There is little precise cartographic information about the extent of lalang in Sarawak, but it is not believed to be a major problem. Shim (1993: 23) attributes the absence of large tracts of sheet lalang to the lack of a pronounced dry season: "nearly year-round rains do not allow lalang to reach a fire climax, and it nearly always loses in competition with other weeds." Freeman, in his well-known Iban Agriculture (1955), documented cases where too-frequent cropping by shifting cultivators had led to lalang infestation, which was apparently common in the Kapit district. Padoch (1982b) has questioned the universality of this behaviour among Iban pioneers. As noted in chapter 6, the phase of pioneering shifting cultivation by Iban groups is now over. The only figures available for the extent of grassland 72,000 ha - date from 1974 and include scrub (Hatch, 1982). Shim (1993) believes the actual area of grass to be much smaller.

In Sabah the figures are also old, giving an area of 155,000 ha in 1970; more than half of this was in the West Coast Residency around Kudat and Kota Belud (Shim, 1993). An examination of recent topographical maps in Sabah revealed much apparently successful reforestation with Acacia mangium in this region. Reforestation is also being tried in another famous lalang area, the Sook Plain, in the Interior Residency (see chap. 8). The formation of this grassland has been well documented as taking place after a fire in 1914, in which all the existing swamp forest was burnt and did not regenerate, the exposed white sand soil being colonized by a mixture of lalang and bracken (Matthews, 1917; Cockburn, 1974; Shim, 1993).1

In contrast to the quite moderate expanses in Malaysian Borneo, the four Indonesian provinces of Kalimantan have been estimated to contain 1.4 million ha of grassland, out of 10.3 million ha in the entire country (Tjitrosoedirdjo, 1993). These figures are drawn from the maps compiled by the Regional Physical Planning Programme for Transmigration (RePPProT) and represent the situation around 1982. Although not all this grass is Imperata, especially in eastern Indonesia, in Kalimantan the grasslands are dominated by that species. South and West Kalimantan are the two provinces in which the human impress upon the land has been most marked. They contain the highest population densities and, perhaps coincidentally, the largest expanses of Imperata cylindrica, although these usually occur in rolling or hilly areas marginal to the main population centres. Alang-alang is estimated by the RePPProT study (1987b) to cover about 17 per cent of the small province of South Kalimantan - about 623,000 ha. This constitutes a grassland area almost twice that of West Kalimantan (340,000 ha), which makes up only 2 per cent of that province (RePPProT, 1987a).

The western foothills of the Meratus mountains (including the Riam Kiwa) and the middle Melawi basin (including the Ela Hulu) stand out prominently for their vast fields of alang-alang (fig. 9.1). In South Kalimantan the grasslands extend almost unbroken in a swathe of more than 300 km from north to south. Many of these areas occur in a low-rainfall belt in the lee of the Meratus mountains, but climate is not a sufficient explanation because the grass also exists in higherrainfall areas. It becomes prominent again east of the mountains, and occurs in patches up the east coast into East Kalimantan, for example near Samarinda. In West Kalimantan the patches tend to be more scattered, but notable areas include an extensive and compact block in the middle Melawi (about 150,000 ha) and a dense mosaic of grassland interspersed with swamp forest at Kendawangan in the far south. In Central Kalimantan there are also important stretches of sheet alang-alang in the south near Pangkalan Bun.

Lest the occurrence of such grassland becomes exaggerated, there are many areas throughout Kalimantan where the map sheets (at a scale of 1:250,000) indicate either no Imperata or very small areas only. Most of the mountainous interior and the swampy coast are unaffected. The Land Resources of Indonesia: A National Overview (RePPProT, 1990) does not separate grassland from scrub and shifting cultivation in its vegetation map, thus creating the impression that it is all the same, which is far from being the case. Although the south-east stands out as by far the most prominent grassland district in Kalimantan, it proved impossible to map the alang-alang at a scale small enough to represent the four provinces on one page; in most areas the patches would be too tiny to be visible.

Fig. 9.1 Two grassland areas of Kalimantan (Source: RePPProT, 1987a, 1987b. Map series 1: 250,000, Land Use and Current Forest Status)

On the ground their visibility is not in doubt. The occurrence of such large expanses of sheet alang-alang leads to the question of

when and how these landscapes were created. Human activity has certainly been the trigger, but what kind of activity, in what particular cultural or economic circumstances, and over what period? What has been the relevance of particular site features such as climate or soils? Has population pressure, or lack of it, been part of the explanation? Simplistic explanations such as "improper shifting-cultivation practices" will obviously not suffice. Shifting cultivation, in particular circumstances, might be one causal factor among several, but is unlikely on its own to have created the really large expanses of grass. In-depth probing of both contemporary informants and historical sources is necessary to try to tease out the picture. Studies of current usage of the grasslands and the long-term sustainability of present livelihood systems form the other essential component to any study of "criticality."2

Contents - Previous - Next