Biodiversity and Conservation 6, 59±74 (1997)

Plain tales from the grasslands: extraction, value and utilization of biomass in Royal Bardia National Park,

KATRINA BROWN School of Development Studies and CSERGE, University of East Anglia, Norwich NR4 7TJ, UK

Received 11 July 1995; revised and accepted 20 November 1995

Royal Bardia National Park in Southern Nepal contains a number of internationally endangered wildlife , and a mix of habitats, including areas of grassland known as phantas. The con- servation and current use of these areas is brie¯y described, including the extraction of grass from inside the National Park by local people living in adjacent villages. The products collected ful®l a range of human needs, both practical and ceremonial. The study attempts to quantify the annual o€take in both material and monetary terms, and comments on the sustainability of the practice. Alternative sources of roo®ng materials are apparently outside the scope of poorer households. The use of these materials has particular signi®cance given the initiation of a Parks and People Project to implement a Bu€er Zone around the National Park and e€ectively to link conservation and devel- opment needs. Keywords: conservation policy; economic value; grasslands; protected areas; sustainable utilization.

Introduction This paper examines the use and management of natural resources in Royal Bardia Na- tional Park (RBNP) which lies in the southern lowlands of Nepal, the . This pro- tected area, which consists of sal () forest and areas of grassland, harbours a number of internationally important endangered wildlife species, including the Bengal tiger (Panthera tigris), Asian one-horned rhinoceros (Rhinoceros unicornis), Hispid hare (Caprolagus hispidus), and Bengal ¯orican (Houbaropis bengalensis). Under the current management regime local people are permitted to enter the park for a period of 10 days during the dry season to harvest a number of grass species from which they make di€erent products, chie¯y to meet subsistence needs. However, concern has been expressed over the extent of this utilization and its impacts on the biological diversity of the grassland ha- bitats. These concerns include the possible degradation of the grassland and changes in the ¯ora; disturbance of wildlife; increasing numbers of people entering the park for the annual cutting; and the illegal of other products, particularly wood. These put the sustainability of this practice into question. However, a contrary view is that habitat diversity is maintained, in part at least, by precisely this human intervention and that critical habitats such as the grasslands have persisted only as a result of the annual cutting. Furthermore, the annual cutting contributes signi®cantly to the livelihoods of local people. The paper describes and quanti®es the extraction of the various grasses and comments on its desirability within the context of Nepal's evolving Protected Areas policy. The next section describes that context, and then historical and ecological perspectives and current

0960-3115 Ó 1997 Chapman & Hall 60 Brown management policies of the study site are outlined. Following sections discuss the ex- traction and use of the grasses, and the implications for conservation of the grassland habitats within the RBNP. Alternative sources of roo®ng materials, their availability and cost, are discussed, and development opportunities for implementation of income gen- erating projects in the Bu€er Zone surrounding the National Park are suggested.

Conservation Policy in Nepal The organization of the state, and the ownership and control of land determined con- servation policy in Nepal in the past. Economic and cultural factors were as important in this respect as biogeography. Prior to the 1950s national security played a prominent role in Nepal's conservation strategy according to Basnet (1992). Forests, for example, were used as a means of political protection ± in the 18th century the Terai forests formed an e€ective barrier to Indian invasion, so the state had a vested interest in their preservation. These areas were also used for hunting by the Royal family, and indeed, the designation of the protected areas in the Terai was originally to provide hunting sanctuaries for use by the Nepalese Royal family and visiting dignitaries. Essentially, Nepal was a subsistence agricultural economy, with little contact with the outside world before 1950. Population growth rates were low, and relatively little pressure was put on forests for timber pro- duction, neither was there large-scale conversion of natural habitats for agricultural production. This situation has changed in recent decades, particularly in terms of demand for land, internal migration to lowlands, and technological changes in the use of resources. Legislation governing National Parks and other protected areas was introduced in 1973 by the National Parks and Wildlife Conservation Act. This Act created the Department of National Parks and Wildlife Conservation (DNPWC), which was formerly an oce of the Department of Forestry. Under the Act, the DNPWC has power to create and manage four types of protected areas: National Parks, Wildlife Reserves, Strict Nature Reserves, and Hunting Reserves. A later amendment gave the national government the power to recognise a ®fth type of protected area in response to a newer proposal, called Con- servation Areas. At present, approximately 10 % of Nepal's land area is under some form of protection, and protected areas remain a key mechanism for conservation policy. The ®ve Terai protected areas were among the earliest to be designated soon after the enact- ment of the 1973 legislation. However, even the Royal , the ®rst in Nepal, has been designated for little over two decades. Resettlement is therefore still within living memory for the local population, and indeed, one village remains within the area designated and negotiations continue to relocate inhabitants. In Royal Bardia Na- tional Park, extensions to the park in the last decade involved resettlement of numerous villages. In many of the protected areas the sites of old villages and cultivated areas are recognisable within the park by their vegetation (having a higher concentration of Im- perata). In fact, this characteristic is part of the management conundrum, whereby human intervention may be necessary to maintain some types of grassland. The dilemma appears to be how that intervention is managed, by whom, and for whose bene®t. Nepal is now considered a leader among developing nations with regard to conserva- tion programmes and legislation (Heinen and Kattel, 1992); it was among the ®rst Asian nations to develop national conservation legislation, sign CITES, and develop a national conservation strategy. The conservation policy has also followed the international tide and shifted away from a single species focus (identifying and legislating protection of en- Plain tales from the grasslands 61 dangered species, and re-location programmes such as the tiger re-location programme of 1970s), towards a more comprehensive approach to biodiversity protection, for example, in the designation of the Annapurna Conservation Area in 1988 (see Heinen and Yonzon, 1994). In a recent statement of policy, the Nepal Environmental Policy and Action Plan (HMG, 1993), HMG clearly links conservation and development, and the integration of environment and development through appropriate institutions, adequate legislation and economic incentives, and the balancing of `development e€orts and environmental con- servation for the sustainable ful®llment of basic needs of the people' (p. xi) are among its central aims. This commitment to policy integration is manifest in the development of a `Parks and People' initiative to implement Bu€er Zones around the ®ve protected areas in the Terai. The main objective of this project is to enhance the capacity of local commu- nities and the DNPWC to jointly manage the protected areas and their bu€er zones, and to improve socio-economic conditions of people living in Village Development Committees adjoining the protected areas (HMG/UNDP, 1994). [Village Development Committees (VDCs) are the form of local administration which consist of nine wards each, a ward being a village. There are 16 VDCs in the Bu€er Zone, ranging in size from under 3000 to nearly 15 000 people. Implications of this project are discussed later in this paper.]

The Royal Bardia National Park: historical, ecological and human perspectives The RBNP covers an area of 968 km2 in Bardia and Benke Districts in the Southern Terai in the mid-Western region of Nepal (see Fig. 1). The Park extends East of the Geruwa branch of the Karnali River to the West; its Northern boundary runs along the crest of the Churiya Ridge, and its Southern boundary is demarcated by a fringe of cultivated land and government forest (Upreti, 1994). The park can be roughly divided into the Churiya Ridge which dominates the Northern portion, the Bhabar area which lies at the foothills of the Churiyas, the Terai ¯atlands and the alluvial ¯oodplain to the South and West, and the Babai Valley to the East. The South-Western ¯oodplains and ¯atlands are most easily accessible during the grass-cutting period, although the construction of a tarmac road through the middle of the RBNP has enabled access to the interior, as shown in Fig. 1. Seven major vegetation types have been identi®ed inside the RBNP (Pokharel, 1993): four types of forest and three di€erent grasslands habitats. The sal (Shorea robusta) forest covers approximately 70 % of the park area, dominating the alluvial ¯oodplain, and is also found on parts of the South-facing slopes of the Churiya. Khair-sissoo forests, composed of Dalbergia sissoo and Acacia catechu is restricted to major water courses and ¯ood plain islands. Moist riverine forest is characterized by evergreen species such as Ficus racemosa, Syzigium cuminii and Mallotus phillipinensis. Three types of grassland have been identi®ed: wooded grassland, open areas of grassland known as phanta, and ¯oodplain grassland. The ®rst two types consist of tus- sock-forming perennials such as cylindrica, , Erianthus ravennae and Vetivera zizaniodes. Wooded grassland is savannah type, often dotted with simal (Bombax ceiba), or silk cotton trees. The phantas consist of open areas; the most signi®cant of these are shown in Fig. 1. These habitats are thought to represent previously disturbed and cultivated sites. The third grassland habitat is found in the ¯oodplain and along the banks of the Geruwa, Babai and Aurai rivers in areas commonly inundated during the monsoon. These areas are characterized by tall grasses, including Saccharum 62 Brown

Figure 1. Royal Bardia National Park. Plain tales from the grasslands 63 spontaneum, Saccharum bengalense and Phragmites karka. More detailed descriptions of the vegetation of the park can be found in Dinerstein (1979) and Pokharel (1993). In terms of the human population, there are a range of di€erent ethnic and caste groups inhabiting the Terai, although this has been changing over time. The Tharu are perhaps the oldest and original inhabitants of the Terai; according to Cox (1990), Tharu people have lived in the region for more than 600 years. Ghimire (1992) states that the Tharu migrated from during the 15th and 16th centuries; there is speculation that the Tharu originally came from Rajisthan; indeed, their dress and customs are similar to tribal people in that region. There are several clans of Tharu in the Terai, including Dangora, Kathariya, Dhakar and Rana Tharu. Much of the available literature alludes to these groups as living in harmony with nature, consistent with an idealized view of indigenous people as `ecosystem people' (Dasmann, 1991; see also Gadgil, 1993). For example, Upreti (1994) claims that Tharu people are part of the jungle ecosystem and that their needs and habits do not have signi®cant negative impacts on the natural environment. Tharu are traditionally dependent on the forest for timber, thatch grass and fodder, and also collect ®bres, leaves, canes and reeds, mushrooms, honey, vegetables, medicinal herbs and fruits. They snare deer, pigs, , and ®sh and aquatic snails are important additions to their diets. While the Tharu utilize a range of wildlife and other natural resources in the region, the long-term impact of these practices is not known. It seems likely that scattered groups of Tharu were the most numerous inhabitants of the area up to the 1960s. Although there had been migration and land colonization during the Rana regime, according to Ghimire (1992:50) `due to the existence of malaria, endemic disease and other dangers (wild and bandits) in the plains, the response from both landlords and peasants from the Hills to land colonisation measures remained fairly limited. The peasants showed most interest in clearing and cultivating lands only in the close adjacent plains so they could get to higher ground to sleep, and the landlords came only to the plains during the winter when malaria was less prevalent and when harvests and rents were collected.' Agricultural labourers were recruited from India in the second half of the 19th century, and there was increased immigration from India during the 1930s and 1940s, from the densely populated plains across the border. A malaria-eradication programme, and land shortages in the uplands of Nepal, led to rapid expansion of the population of the Terai from the 1960s onwards, principally as a result of migration from the Hills. Human pressures on the natural resources of the Terai have thus changed considerably over time, primarily as a result of the migration of the last two decades. This movement continues, and the areas bordering the RBNP have experienced high rates of population increase in the past 20 years. The population in Bardiya District rose from 199 000 in 1981, to 290 000 in 1991, an increase of nearly 46 % (Central Bureau of Statistics, 1993). Over the same period the population of Nepal rose by 23 %. The Village Development Committees adjoining the Park, i.e. the area de®ned by the HMG/UNDP Parks and People Project as the Bu€er Zone, consist of a total of 137 600 people in 18 600 households. The region that is now RBNP has undergone a number of shifts in property rights and changes in . Prior to the nationalization of the forests in 1956, most of the forests in Bardia were birta (private) forests (Upreti, 1994). As personal property of the Rana rulers, these areas were under strict protection. In 1969 the area was declared a Royal Hunting Reserve, although local people had free access to the forest and to graze their livestock. An area of 386 km2 was ocially gazetted as the Royal Karnali Wildlife 64 Brown Reserve in 1976 and renamed the Royal Bardia Wildlife Reserve in 1982. In 1984 it was extended to the east to include the Babai Valley. It was upgraded to national park status in 1988. Little documented evidence exists describing the human use of these areas, although we know that the area which is now the RBNP was inhabited until 20 years ago. Before the area was declared a reserve in 1976, two villages were inside the park area, one at Chi- sipani and another at Auraini. Before the establishment of the hunting reserve, the villages of Bagaura Phanta and Lamkauli Phanta were also relocated outside the reserve boundary (Upreti, 1994). Recent expansion of the Park resulted in more widespread resettlement, and Resources Nepal (1994) reports that: `Between 1979±1984 during the extension of the park a total of 1572 families, comprising over 9500 people were moved out of the Babai Valley. Some 220 families did not want to leave their homesteads, however, they were forcibly removed. The total population from the 20 villages had left 1200 ha of farmland, 1096 permanent houses and 1793 semi-permanent structures behind'. In terms of management of the grassland habitats within the area, relatively little is known of the human ecology. Pokharel (1993) describes the history of the phantas as being quite obscure. Dinerstein (1979) reported that Baghaura and Lamkauli phantas were under cultivation prior to 1975. Pokharel states that a number of people used to cultivate the phantas and are still residing outside the park, though very close to it. One of these informants is recorded by Pokharel (1993: 13): `According to Gopi, in 1965 when he was one year old, his father and friends came down from Surkhet in search of cultivable land. Obtaining land at that time was no problem. They found Baghaura a very suitable open place to live and cultivate. Later they settled there with 15 di€erent families from Surkhet. They were the ®rst inhabitants of Baghaura. Cultivation of Lamkauli could have taken place much later than Ba- ghaura, according to Gopi. He did not know if Khauraha was cultivated before. Ba- ghaura could have been a small open and ¯at grassland in 1965, which was expanded according to the need of new inhabitants by peripheral deforestation in the jungle.' Grass cutting was a regular activity in Bardia, but it is dicult to trace when this began. In RBNP the phantas were cultivated and grazed between 1965 and 1975. Records from the RBNP show authorised grass cutting since 1983, 6 years before the National Park was created. Livestock grazing in the phantas and other areas prior to 1975 was apparently very common. After 1976 the area was fenced and livestock grazing prohibited. Fire has also been used as a management tool, and in the past uncontrolled burning in the park was described. Fire is generally set at the end of the dry season when grass cutting is completed; in RBNP this is mainly carried out by park authorities. To what extent maintenance of the sward requires cutting and burning ± and therefore the necessity of human intervention to maintain the patchwork of habitats ± is relatively unknown. It can be hypothesized that grazing by domestic and wild animals, but especially the destructive browsing of elephants, was instrumental in maintaining grassland in the past when large herds were prevalent. The extent to which annual cutting and human-set ®res recreate or mimic these natural processes can only be postulated, and whether the resulting grassland communities re- semble those present in the past is uncertain (see Peet, 1994). Plain tales from the grasslands 65 The grasses and their uses A number of di€erent grass species are utilized by people from the RBNP. The three major products extracted are thatch grass, building materials and rope. Thatch grass, known as khar, consists of a mixture of the shorter grass species, predominantly and, to a lesser extent, the shortest of the tall grasses, Saccharum spontaneum. Canes and reeds are extracted for house building and a number of other uses. They are known as kharai and include taller grasses, Saccharum bengalense, Narenga and Phragmites spp. The third product is rope, made from Eulaliopsis binata, known as buncas or babiyo. Table 1 shows the most important grass species harvested from RBNP, and their main uses. Khar is collected primarily from the wooded savanna areas and the phantas, kharai from the ¯oodplains, and buncas from the foothills. Much of the analysis presented in this paper concerns the collection and use of thatch or khar. The list of grasses and products in Table 1 is by no means exhaustive and does not detail all of the di€erent species of grass and reeds used by local people, or all of the uses. There are certain species which have highly specialized and localized uses. For example, one type of grass, known locally as bus heri is used for constructing ®sh traps because of its properties of resisting rot even when repeatedly saturated in water. These di€erent re- sources are not examined here, but rather this analysis is focused on the products which are used in large amounts by a majority of people. A number of complications arise in obtaining local names for the grasses. First, a par- ticular species may be given a di€erent name, according to its use or which part of the plant is used, or its method of harvest or treatment. In the table above, for example, Narenga porphorycoma may be known by a generic term for cane or reed, kharai or khadai. The same species has a number of alternative names, according to its di€erent uses and whether it is burned before it is cut, and whether the whole, or part of the plant is used. Second, complications arise because of di€erent local languages and dialects. There are a number of di€erent Tharu clans in the area, who give di€erent names to the grasses. There are marked di€erences between the terms used in Chitwan and those used around Bardia.

Table 1. Grasses collected from Royal Bardia National Park

Grass species Local name Uses and names

Narenga khadai or kharai Cane±if burned, used for building etc. sakhata porphorycoma Cane, not burned, whole used for grain silo phank Cane, top part only silicili Saccharum narkat Cane for walls and ceiling chatati bengalense Cane for baskets kenari Phragmites kharka Cane for lamp stand for diwali festival Tifa augustifolia /Tifa pat or pater Mats, fans, mattresses, howdahs elephantina Imperata cylindrica khar Thatch, brooms; ¯ower head used for ceremonial lamp kuwar bati Eulaliopsis binata buncas, sabai, Rope used for beds, chairs, bullock carts, tying babiyo thatch; paper Saccharum spontaneum khans Reeds and canes for walls, bed, thatch, fodder 66 Brown Table 1 gives the terms most commonly reported and used in Bardia; some are Tharu names, others Nepali. A number of di€erent terms are used locally for the same grass, as in the case of Eulaliopsis binata. This may re¯ect the uses and cultural importance of the species. Martin (1995) observes that species of great cultural signi®cance are over-di€er- entiated, or split into distinct categories by local people, whereas species that are less important are usually under-di€erentiated, or lumped together in a single generic. Use values will also be re¯ected by the number of di€erent uses repeatedly cited by respondents (see Phillips et al., 1994, for a discussion of quantitative ethnobotanical techniques) and this would con®rm the importance of the three products discussed above. These ®ndings give some indication of the number of di€erent uses for the grasses, and how important these resources are for local households. This is especially true in the case of house building, where the grassland provides raw material for walling (both small and large cane) and roo®ng (cane and thatch). However, the grass products also have im- portant cultural functions. For example, they are used in ceremonial lamps, and the ¯ower head of Imperata cylindrica is used as a wick in lamps for temples or shrines. The grass cutting is an extremely important event in the annual calendar; preparations are made by the community in advance, and Tharu people celebrate the Maghi festival at the end of the grass-cutting. Saccharum spontaneum has an important role in Hindu worship and plays a role in a number of festivals (Majupuria and Joshi, 1989). Most of the discussion in this paper is con®ned to the collection and use of thatch grasses from the phantas.

The extent of harvesting in RBNP Since 1983 permits have been sold which enable local villagers to enter the protected areas in the Terai and collect grass and cane. This policy was initiated as a means of granting people partial recompense for the loss of access to these and other products as a result of protected areas implementation, and to improve relations between the local community and the Parks Authorities. In the past a nominal fee of 1 NR (75 NR = £1, in 1994) was imposed, which acted as a means of controlling the total numbers of grass collectors, and also enabled the DNPWC to keep a tally of the numbers entering the protected areas over the 15-day period, and to cover administration costs. However, as the numbers of people entering the protected areas increased over the years, the fee has been increased, and the cutting period has been shortened to 10 days (in 1994/5 season these spanned 2±11 January 1995). In RBNP in 1994/5 people reported paying 12 NR for a permit (equivalent to over 25 % of 1 day's wage of 40 NR), which would enable one person to enter the National Park each day for the 10 days. Each person can carry as much grass or cane as possible, but no carts or other means of transportation are allowed inside the protected areas. Around the entrance points, bullock carts are loaded to transport the biomass back to the village, but only people are allowed in the Park. Figure 2 shows the number of permits sold annually for grass cutting in RBNP since 1983. Permits are sold in advance of the cutting days, and are checked at each of 44 di€erent entrance points around the Park perimeter by DNPWC and army ocials during the grass cutting. Figure 2 shows a steady increase in the number of permits sold, re- presenting more than a doubling in the years 1983±1993 from 21 000 to 45 500 permits (i.e., greater than the increase in population), with 1994 showing the ®rst drop in numbers, a decrease of some 4 % of the 1993 ®gure. Over 43 000 permits were sold in 1994. Plain tales from the grasslands 67

Figure 2. Grass-cutting permits sold in Royal Bardia National Park. Source: Royal Bardia National Park.

A series of household interviews were conducted in sample villages in VDCs adjacent to the RBNP and inside the proposed Bu€er Zone. (Structured interviews were undertaken in 26 households in six villages by the researcher, an assistant and two Nepali collaborators. In addition, interviews and discussions were held with a range of key informants, including VDC chairmen, local teachers, DNPWC personnel, and information obtained from group interviews carried out in the sampled villages. Data on numbers of permits and bundles of grass collected were obtained by direct observation, supplemented by records of the DNPWC, during the 10-day cutting period.) Interviews were undertaken in the weeks prior to the cutting period, and direct observations and enumeration of collection of grasses during the 10-day cutting period carried out. All of the households interviewed, with one exception, were sending adult members into the RBNP to collect grass. Many households committed all their adult labour to this task, and although both women and men participated, some households send just adult males. From the survey, the average number of permits per household was 3.4, and the household size ranged from 5 to 40 (n + 26). The total numbers of permits sold indicate that the average for households adjacent to the Park (in the Bu€er Zone) is 2.34 per household. Total numbers indicate that 32 % of the total population of the de®ned Bu€er Zone (i.e. the 16 VDCs adjacent to the RBNP) participate in the grass cutting. This is con®rmed by the ®ndings of the household survey: much of the available adult labour is deployed to the task of grass cutting and collection. 68 Brown How does the demand for the grass by local communities compare with the supply of biomass inside the RBNP? There is an assumption by conservationists that over-ex- ploitation, caused by increased numbers of cutters entering the park, is leading to changes in the composition of the grassland, and there is some anecdotal evidence of a decline in the quality of the thatch grass, perhaps caused by a change in the species mix. (The change in quality may be a result of changes in the species mix, in the growth habit or quality of the plants themselves, or because cutters are utilizing more marginal areas of grassland.) However, cutting and/or burning is required to discourage the colonization by taller grass species, shrubs and small trees, and possibly therefore slow the succession to sal forest. On- going ecological experiments aim to examine the impact of di€erent cutting and burning regimes on the grassland diversity. There are likely to be other ecological impacts of the regular removal of such large quantities of biomass, which will depend in part on local environmental conditions, including, for example, the removal of nutrients. Using the observations and measurements from the survey, and data from Pokharel's study (1993) on the area of the phantas, the total amount of thatch grass available from the southern phantas (Upper and Lower Baghaura, Upper and Lower Khauraha, and Lam- kauli) is calculated, and this is shown in Table 2. This gives the estimated production of biomass from the ®ve southern phantas, using measurements from the 1994 survey. (Bundles of grass harvested from Lower Baghaura phanta were sampled and weighed just prior to the cutting period. The mean weight of a bundle was 2.52 kg. The same pro- portion of yield/productivity, i.e. the same rate of o€take as Pokerel's study (1993) were assumed, and con®rmed by ®eld observations.) These calculations indicate that a total of 841 000 bundles of khar were available from the ®ve southern phantas in RBNP in January 1995. Observations of the grass collection, based on a sample of cutters entering and leaving the Park at Thakadwara Gate in the south west, indicate that the possible total harvest of khar may be in the region of 1 050 000 bundles, or approximately 2650 tonnes. This ®gure is larger than the estimates for the phantas alone, as small patches of grassland occur throughout the RBNP and will be utilized by local people. This estimate is likely to be an overestimate for a number of reasons. First, the o€take rate may well be lower than estimated, as the observations may have been biased to the higher quality parts of the phanta. Second, the observations from Thakadwara gate were used, and then extrapolated to give a ®gure for the whole Park according to the number of

Table 2. Estimates of biomass production in ®ve southern phantas in RBNP

Area Production Yield Number of bundles (ha) (t/ha)a (t/ha) available U Baghaura 41.6 13.98 8.5 140 673 L Baghaura 58.7 8.06 4.9 114 442 U Khauraha 94.7 11 6.7 251 973 L Khauraha 34.5 8.57 5.2 71 517 Lamkauli 111.3 9.74 5.9 262 220 840 825 Total bundles

aProduction is net primary production. Plain tales from the grasslands 69 permits sold each gate. However, Thakadwara is situated close to the phantas so a higher proportion of khar may be removed through this gate. Third, assumptions have been made about the amount of time spent collecting khar and other products, and (as indicated by observations) that khar collection was concentrated in the ®rst 3 days of harvest. It has therefore been assumed that each of the 43 580 people with permits will have collected khar for 3 days out 10 days. Such an assumption is con®rmed by household interviews, al- though again this may be biased due to the geographical proximity of the surveyed villages to the southern phantas. How do these estimates of harvest relate to the predicted demand? According to the village surveys, many households have at least one thatched roof, and, in response to household interviews, a typical roof might require 250 bundles of khar for thatching every 2±3 years. Given the estimates of the yield, this implies that only 56 bundles are available per household in the Bu€er Zone each year, or enough to thatch one roof per household every 4±5 years. These ®gures indicate a short fall in thatch available, which is con®rmed to a certain extent by the interviews and information obtained from key informants. It may also go some way to explain the drop in the number of permits sold. A perception of scarcity of the grass means that people are not willing to invest the permit fee when they are not sure of collecting enough grass for their needs.

Assessing the economic importance and value of grass collection The grass and cane collected from the protected areas of the Terai undoubtedly have importance, especially, as the foregoing discussion has shown, in terms of their uses in house building. Information on the relative importance of these products, and how households would fare if access were stopped, is essential for future management strate- gies. A number of di€erent approaches can be made to assess the economic value of extracted biomass (see for example, Godoy et al., 1993; Martin, 1995). Three approaches are most commonly used; based on the marketed value of products, the cost of time or labour involved in their extraction, or the cost of substitutes. For the purposes of this paper some indication of the relative value of these products and their importance in the livelihood strategies of local people is required. Limited data on the quantities of materials extracted, their local prices, and the costs of substitutes, primarily for khar, were collected and are now brie¯y discussed. This represents only a cursory survey of the economic values rather than a full analysis, and concerns only the direct use value of these resources. For a discussion of the range of economic values of non-timber forest products see Chopra (1995). In an earlier study by Lehmkuhl et al. (1988) the gross value of cane and grass collected from Chitwan National Park (RCNP) in January, 1986, was calculated to be in the region of US$451 836. The net value, with labour and permit costs subtracted, amounted to US$252 103, as shown in Table 3. The study also examined the costs of alternative building materials, and concluded that most villagers close to the park do not have enough capital to invest in other supplies, but they do have the time to cut their own grasses. The standard of living of these people would therefore be lower, without access to park re- sources. Similar calculations for the amount of khar harvested from RBNP in January, 1995, can be made using the estimates outlined in the preceding section. If it is assumed that 1 050 000 bundles were harvested at a cost of 1.5 NR per bundle (prices indicated in our 70 Brown Table 3. The value of grasses and cane collected from Royal Chitwan National Park, 1986

Product Number Weight Value Value of loads (`000 kg) (million Nrupees)a (`000 US$)b Khar 162 592 6406 4.55 206.9 Kharai 134 265 4726 5.39 244.9 Total gross value 9.94 451.8 Net valuec 5.46 252.1

Source: data from survey reported by Lehmkuhl et al. (1988). aLocal prices for khar 0.71 NR/kg, kharai 1.14 NR/kg. bExchange rate of US$1 = 22 NR (1986). cPermit (1 NR pp/day) and labour costs (20 NR pp/day) subtracted. household interview) then the gross value of production is 1 575 000 NR. If a labour cost of 40 NR per day (the agricultural wage rate) is assumed, and a permit cost of 12 NR per person, but that grass cutters spent only 3 days collecting khar, then the net value of khar from RBNP is in the region of 819 000 NR, or £10 900 or US$16 400 (75 NR = £1, 50 NR = US$1 in 1994). This ®gure is considerably lower than the estimate of Lehmkuhl et al. for RCNP in 1986. This is for a number of reasons, not only the vagaries of international exchange rates. The biomass production and harvest, and the actual area of grassland, are considerably lower in RBNP than RCNP. Fewer people enter RBNP to cut grass than they do in RCNP, where some 80 000 permits were sold in 1994. The harvest period is now restricted to 10 days, whereas it was 15 days at the time of the study of Lehmkuhl et al. The cost of permits and labour are now higher than in 1986. The prices of the grasses di€er, and around Bardia a relatively smaller proportion of the grasses are sold. In ad- dition there are a number of problems associated with using the price of labour to impute the value of extracted materials. Although the agricultural wage rate of 40 NR per day is used, this does not accurately re¯ect the opportunity cost of time, as the 2 weeks when grass collection takes place are the slack season for farm labour, and it is unlikely that there are many employment opportunities at that time. These are just some of the factors which make attaching a monetary value to the collection of these materials problematic. In addition, the economic value may not re¯ect or be indicative of the cultural signi®cance of these resources (see Brown, 1994 for a discussion). Despite the apparently low economic value of the grass using this type of analysis, these ®ndings indicate that the grass is a very important resource, especially for the Tharu people. This is con®rmed by the wide range of uses ± most important in housing ± grain silos, beds, mats, bullock carts and other uses. There are few ± or no ± alternative sources of grass outside the protected areas. Perhaps surprisingly, the survey found only limited local markets for thatch and cane, although there are many reciprocal arrangements between friends and neighbours. The grass cutting has cultural signi®cance ± in terms of the uses (for example, in ceremonial lamps in religious activities) ± and as a social event, and is linked to the Maghi festival. So the direct use value as estimated here indicates only part of the social and livelihood signi®cance of these resources. Cultural values of the grasses and products, and the full range of ecological and wildlife ben®ts of the grasslands should also play important roles in the consideration of management of Plain tales from the grasslands 71 these resources. Continuing to con®ne the analysis to direct use values, the costs and use of tiles as an alternative to thatch is now brie¯y discussed.

Availability of alternative roo®ng materials A sizable number of households in the vicinity of RBNP use clay and, to a lesser extent, cement tiles as an alternative to thatch grass as roo®ng for houses and other buildings. The household survey found almost exactly half of the buildings of the households sampled were tiled, and that nearly 75 % of the households had at least one tiled building (most households had between two and ®ve buildings which consisted of a dwelling and cattle shed). These ®gures were con®rmed by direct observation in the villages visited. Tiles provide a number of advantages over thatch; they last longer, needing to be replaced only every 20±25 years compared with 5 years or less for thatch; they reduce ®re hazards, a characteristic noted by several informants; and, if of reasonable quality, they will keep a house drier than thatch during the monsoon. In general, tiles are generally perceived to be desirable, as `modern', and re¯ect the high status of a household. The disadvantages of tiles primarily concern their cost, the lack of skills available locally to produce them, and access to materials such as clay and sand for their construction, and fuel for ®ring. The reported costs of tiles varied widely. This is for a number of reasons. First, households had installed tiles at di€erent times through the years, the earliest of those reported being manufactured almost 20 years ago. Second, there were di€erent sizes and types of tiles, some made from di€erent materials. Lastly, the prices varied, as some people bought ready-made tiles, whereas others paid for tiles to be made and installed. The di€erent costs reported are not directly comparable, and are only indicative of the mag- nitude of costs. Households reported that the cost of tiling a house ranged from 4000 to 20 000 NR. The cost of tiles ranged from less than 0.5 NR per tile when hiring labourers to make tiles using clay from a houshold's own land, to 5 NR per tile to buy the ®nished product. One household reported the cost of tiles as 500 NR per 1000 to buy, or 350 NR per 1000 to make. Most households who had tiled buildings had hired labourers to make tiles using clay collected from their own land. Sand was reported to be freely available locally, and tiles were dried in the sun for a few days and then ®red using cattle dung. Four or ®ve labourers, working for 4 or 5 days, could produce enough tiles (10 ±20 thousand) to roof an average size building. Lehmkuhl et al. (1988) report a cost of between 2500±3700 NR for ceramic tiles to roof a typical 30 m2 house in 1986. Most of the households which had tiled buildings had hired labourers to make the tiles. The `tile season' was de®ned as December±February, as would be expected, the dry sea- son, but also the slack time of the year in the agricultural calendar. Many of the house- holds hired itinerant labourers from India, not local casual labour, which implies a lack of skills in tile manufacture, and this is con®rmed by the observations of Lehmkuhl et al. (1988) around Chitwan. This is despite the fact that tile use is not a particularly recent phenomenon, that it is relatively widespread, and that materials appear to be available locally. Use of tiles is undoubtedly increasing, and whether this is spurred by a perception of increasing scarcity and/or higher costs of thatch is uncertain. However, the loss of access to grasslands has in the past provided an incentive to change from thatch to tiling. A few households reported that they had ®rst thatched their buildings about 18 years ago, and when asked about the reasons for their changes, said that when the Wildlife Reserve 72 Brown was ®rst designated they had lost access to the grasslands and had not expected that it would be regained. One man told how his father, the leader of the Panchayat, had been the ®rst person to use tiles in his village, and had actively encouraged his neighbours to use tiles for fear of losing access to source of thatch grass. The most important feature of tiles as an alternative to thatching, and one which has particular implications for poorer households in the community, is that installation of tiles involves, in every case reported, a shift from a mainly subsistence activity to a cash transaction. In most cases this cash transaction involved a large lump sum investment, so even though the net costs of tiles annually may be competitive with thatch, they are unlikely to be readily available or attractive to poorer households. More investigation into the viability of providing training in tile-making, and in setting up tile works as income generating projects as part of the Bu€er Zone Management Plan is required. In addition, the wider environmental implications of more widespread tile making need to be considered; the large-scale removal of clay and sand, and the avail- ability of fuel associated with tile manufacture could result in a new set of ecological, environmental and socio-economic problems. Encouraging tile production may appear to be a logical step to decreasing pressure on grassland resources, but it requires careful examination.

Conclusions The grassland habitats inside the protected areas of the Terai are a most important re- source for both wildlife and local villagers. Grasses are utilized by local people, particu- larly Tharu groups, for a wide range of uses. The case study of RBNP has shown that although the direct use value of the main grass products, thatch and cane, may not be especially high in ®nancial terms, they are important subsistence products which make signi®cant contributions to the welfare of local villagers. Without access to the grassland products, traditional methods of house building could no longer be maintained and there are virtually no sources of thatch and cane outside the protected areas. The grassland products play an important role in livelihood strategies and have cultural value in terms of ceremonial uses. The ecological impacts and sustainability of biomass extraction are currently uncertain, although on-going ecological experiments intend to shed light on any changes occurring in the grasslands. On the one hand, biologists argue that the disturbance associated with cutting and burning will have detrimental impacts on small animals, insects and birds, while, on the other, the ¯ush of short grass resulting from burning may encourage un- gulate species. These in turn attract large predators such a tigers, which are most highly prized by tourists. There is therefore much uncertainty, further clouded by the vested interests of the parties concerned, of the relative bene®ts and costs of current management regimes. It may well be that certain types of cutting and burning, at particular times of the year, produce the mosaic of habitats which optimise multiple uses by both animals and humans. The diversity of the grasslands, where a patchwork of diverse habitats is provided by micro-scale di€erences of site, means that management of the protected area needs to take localized features into account, rather than attempting to implement one regime across all the Terai protected areas. In summary, the rich biological diversity of the Terai protected areas is maintained in part by human intervention, and current cutting practices bring bene®ts to both local Plain tales from the grasslands 73 communities and to wildlife. Whether the practices are sustainable, and in what way they can be made so, is open to question, but the current focus of conservation and develop- ment policy is to relieve perceived pressure on resources inside the protected areas by encouraging the more intensive use of those in the designated Bu€er Zone surrounding the National Park (HMG 1993; HMG/UNDP 1994). The ®ndings of this study have high- lighted the importance of grass collection for local people, and also indicate that alter- natives for one of the main uses, thatch, are available in the form of ceramic or concrete tiles. However, the cost of tiles may well be prohibitive for poorer households, and as yet, the skills for their production are not prevalent locally, and in addition the environmental and economic impacts of more widespread production are unknown. This is one area which could usefully be explored as part of the community development activities of the Parks and People Project. A critical issue for implementation of this project in the Bu€er Zone is how e€ective participation of local people can be facilitated and how bene®ts, especially for poorer households, can be maintained. Progress in understanding the ecological and socio-economic dimensions of the con- servation±development conundrum, their dynamics and potential trade-o€s, demands extensive research which transcends traditional disciplinary boundaries. This research takes a step towards this aim. Such work is necessary to inform policy if sustainable utilization and long-term management is to succeed in conserving areas such as the RBNP.

Acknowledgements Fieldwork was conducted under the auspices of the UEA Tall Grasslands Research Project which is funded by the UK Department of Environment Darwin Initiative for Survival of Species, in collaboration with the Department of National Parks and Wildlife Conservation in Nepal. The author would like to thank the following people for their assistance and advice: Dr Uday Sharma, Dr Bijaya Kattel, Mr Karki, and Mr B. Pathak at DNPWC; Mr T. Khatri at the King Mahendra Trust for Nature Conservation; Mr Gopal Sharma, Ram Din Mahato, Nic Peet and Neil Adger. Views expressed here are the authors alone, and any omissions or errors remain sole responsibility of the author. CSERGE is a designated research centre of the Economic and Social Research Council. An earlier version of this paper was presented at the International Association for the Study of Common Property Conference, Norway, May 1995.

References Basnet, K. (1992), Conservation practice in Nepal: past and present. Ambio, 21.6, 390±3. Brown, K. (1994) Approaches to valuing plant medicines: the economics of culture or the culture of economics? Biodiv. Conserv. 3, 734±50. Central Bureau of Statistics (1993). Statistical Yearbook of Nepal. Kathmandu: CBS, HMG. Chopra, K. (1995) The Valuation and Pricing of Non-timber Forest Products: a study in Raipur District of Central India. Delhi: Mimeo, Institute of Economic Growth. Cox, T. (1990) Land rights and ethnic con¯ict in Nepal. Econ. Polit. Weekly June 16±23, 1318±20. Dasmann, R.F. (1991) The importance of cultural and biological diversity. In Biodiversity: Culture, Conservation and Ecodevelopment (M. Old®eld and J.B. Alcorn, eds) pp. 7±16. Boulder: Westview Press. Dinerstein, E. (1979) An ecological survey of the Royal Karnali-Bardia Wildlife Reserve, Nepal. Part 1: vegetation, modifying factors and successional relationships Biol. Conserv. 15, 127±50. 74 Brown Gadgil, M. (1993) Biodiversity and India's degraded lands. Ambio 22, 167±72. Ghimire, K. (1992) Forest or Farm? The Politics of Poverty and Land Hunger in Nepal. Delhi: Oxford University Press. Godoy, R., Lubowski, R. and Markandya A. (1993) A method for the economic valuation of non- timber tropical forest products. Econ. Bot. 47, 220±33. Heinen, J.T. and Kattel, B. (1992) A review of conservation legislation in Nepal: past progress and future needs. Environ. Manage. 16, 723±33. Heinen, J.T. and Yonzon, P.B. (1994) A review of conservation issues and programs in Nepal: from a single species focus toward biodiversity protection. Mountain Res. Dev. 14, 61±76. HMG (1993) Nepal Environmental Policy and Action Plan: Integration Environment and Develop- ment. Kathmandu: Environment Protection Council. HMG/UNDP (1994) Parks and People Project Document. Kathamandu: DNPWC. Lehmkuhl, J.F., Upretti, R.K. and Sharma, U.R. (1988) National parks and local development: grasses and people in Royal Chitwan National Park, Nepal. Environ. Conserv. 15, 143±8. Majupuria, T.C. and Joshi, D.P. (1989) Religious and Useful Plants of Nepal and India. Lashkar (India): M. Gupta. Martin, G.J. (1995) Ethnobotany. London: Chapman and Hall. Peet, N. (1994) Fire and Biodiversity Conservation in the Tall Grasslands of Nepal and Northern India. Norwich: Mimeo, School of Biological Sciences, University of East Anglia. Phillips, O., Gentry, A.H., Reynel, C., Wilkin, P. and Galvez-Durand, C.B. (1994) Quantitative ethnobotany and Amazonian conservation. Conserv. Biol. 8, 225±49. Pokharel, S.K. (1993) Floristic Composition, Biomass Production and Biomass Harvest in the Grassland of the Royal Bardia National Park, Nepal. MSc thesis, Agricultural University of Norway, As, Norway. Resources Nepal (1994) Habitat Himalaya Resource Nepal News®le, January±February, p. 2. Kathmandu: Resources Nepal. Upreti, B.N. (1994) Royal Bardia National Park. National Planning Commission, HMG, Nepal, in Collaboration with IUCN.