REPORT ON A PRELIMINARY STUDY OF THE RIVERINE FORESTS OF THE WESTERN LOWLANDS OF ERITREA
Stephen Connelly and Nicola Wilson
November 1996
EXECUTIVE SUMMARY
Introduction
Since the liberation of Eritrea in 1991 the pace of development has increased dramatically. The Western Lowlands, with their woodlands, fertile soils and sparse population, are a major focus for agricultural development and settlement. To maximise benefits and minimise conflicts from this development will require rational and careful land use planning, for which a prerequisite is reliable and up-to-date information on existing natural resources. At present, however, very little detailed information is available: this study provides an initial assessment of the status, use and management of the riverine forests, both past and present.
The area of our study was the basins of the Barka, Gash and Setit rivers. These are the three principal systems draining the Western Lowlands, an area of semi-arid hills and plains covering approximately 40% of Eritrea. Of these rivers only the Setit is perennial. The Gash and Barka are seasonal: they flow only during the rainy season, but a high water table is maintained throughout the year in their wide, sandy beds.
Most of the Western Lowlands is covered in savanna woodland of various types. However, the riverine forests are distinctive, occurring as more or less narrow bands along the main courses of the rivers and some of their tributaries. They are some of the best preserved dense forest in the country and for this reason are valued as an important national resource. The past three decades of war and drought had significant impacts on these forests and they are now at risk from over-use and clearance for commercial agriculture.
Methodology
The consultancy was divided between three months of research in Eritrea (November 1995 - January 1996) and a further two months in the UK. Field work consisted of forest visits, village interviews, discussions with government officials and gathering miscellaneous information on forest use and the related economy.
Most of the field work was done in the upper Barka and the lower and upper Gash, since the remaining areas are relatively inaccessible and the Setit valley holds very little riverine forest. As well as making qualitative assessments of the forest we collected quantitative information on dom palm stocking and growth rates. We also tested methodology for the planned GoE/SOS Sahel survey project. Semi-structured group interviews were used to gain information on the use and management of the forest by local people. Individual interviews with MoA and local administrative staff provided us with additional data on the area and on government policy and programmes. We also researched and collated secondary source material, including maps, photographs and remotely sensed data as well as documents.
The Forests of the Western Lowlands
Two basic kinds of woodland can be distinguished in the Western Lowlands: riverine forest and savanna woodland. Both formations are very variable in species composition and density across the region, but are usually easy to distinguish from each other on the ground. Riverine forest is generally denser and has bigger trees than the savanna, and usually contains species which are not found in the surrounding savanna - pre-eminent among these is the dom palm, Hyphaene thebaica.
The riverine forests in the three river systems are very different and also internally diverse. While the Barka is principally a dom forest, the Gash is much more mixed, with areas of acacia/Ziziphus woodland as well as stands of pure dom. The truly riverine forest on the Setit is merely a fringe of large trees, mainly of the same species as occur on the upper Gash. Localised variation in species and structure reflects differing site conditions, natural occurrences such as fire and flooding and the impact of humans.
These forests are very dynamic systems, with continuous changes in their extent, structure and composition taking place. Historical records suggest that dom has been an invasive species in the past, and that its range has spread dramatically in the last century. There have also been major changes in species composition in the past thirty years, of which the most significant is the removal of mature dom from the lower Gash and its replacement by mixed acacia woodland. Overall it seems that there have not been enormous changes in the total extent of the riverine forest in the recent past: in some areas there has certainly been loss of forest cover, while elsewhere there have actually been gains.
Our study indicates that the riverine forests of the Western Lowlands are not, as previously feared, on the verge of disappearance, despite the destruction wrought by war and drought and the needs of the local population. We would put the total extent of the Barka and Gash forests in the region of 30,000 hectares, of which some 20,000 are on the Barka and its tributaries, and the remainder on the Gash. The Setit supports very little riverine forest, although there are probably about 1,000-1,500 hectares of dense savanna woodland in places where the valley floor widens.
In general, forests on the Gash and Barka appear to be healthy, and are actually improving in some areas. This is particularly true of the dom thickets, some of which are showing signs of developing towards mature forest. The exceptions are the forests on the lower Gash and in the immediate vicinity of major towns. The Setit forests appear very stressed as a result of the combined impact of frequent burning and the annual influx of pastoralists seeking water and browse for their animals.
Much of the Western Lowlands is covered by savanna woodlands, which are still in better condition than similar areas in neighbouring countries. However, the testimony of local people suggests that the savanna woodlands have generally declined in density and area over the past few decades. At present, the savanna does not seem to be as capable as the riverine forest of withstanding the pressures on it from expanding human and animal populations. Growth rates are slow, regeneration is extremely sparse and patchy, and there is continuous attrition of mature trees through cutting for domestic and commercial poles and fuelwood.
Both the riverine forests and savanna woodlands are important habitats for wildlife, and play a significant role in maintaining biodiversity in the region. The riverine acacia forest is of particular value, and within this the uncommon pure Acacia nilotica woodlands stand out as habitats for bird life, currently threatened by agricultural clearance and invasion by mesquite. The dense savanna woodlands of south west Gash-Setit seem to us to be underrated as a habitat and worth preserving at least in part. They are currently being rapidly cleared for large-scale rainfed agriculture.
The dom palm
Because of its ecological and economic importance we singled out the dom palm for detailed investigation. Drawing together the little existing documented research on this species and supplementing it with our own observations we present a summary of dom ecology and silviculture. There is considerable disagreement between sources over the details and a marked lack of any reliable quantitative data. Clearly, however, the dom palm is the most productive of the native trees of the Western Lowlands. In addition, due to its prolific regeneration by seeding and coppicing, dom forests are capable of surviving fire and flooding, and are able to recolonise areas which have been cleared by man.
We estimate that the productivity of the dom forests is higher than previously thought. From sample plots we calculated rates of leaf production for dense dom thicket at around 60 tonnes per hectare per year (dry weight) - this figure is not statistically proven, but we believe is indicative of the scale of productivity. The fraction of this amount that could be sustainably harvested is unknown at present, but evidence from the lower Gash suggests up to 12 tonnes can be removed annually. More research is needed to establish optimal regimes to maximise sustainable leaf yields and balance this against the harvest of other products.
Our research has indicated that dom thickets are far more valuable for leaf production than mature forest. Since the leaves are the most important product in the livelihoods of the Lowland population, we stress that dom scrub should be viewed as a valuable resource in its own right, and not as degraded forest to be "improved" or removed for agriculture. (Existing regulations do not include dom thickets in their definition of "forest": this should be changed in any new legislation.)
Artificial establishment of dom has not been carried out in Eritrea, but would appear from research elsewhere to be relatively straightforward using direct seeding.
Traditional uses of the forest
The people of the Western Lowlands depend heavily on the surrounding woodland - both riverine and savanna - for many subsistence needs and for generating income. Although many species contribute, the single most important tree species is the dom palm, and the most important use is the cutting and processing of its leaves. Thousands of people in hundreds of villages do this to provide basic household necessities (including house walls and roofs) and to earn money. It cannot be overstressed that this is not a subsidiary activity using a "minor forest product", but makes a vital contribution to many people's livelihoods and food security in the Lowlands.
For some of the poorest sectors of the population - notably female-headed households - dom leaves provide the only source of income. For women in general, weaving is the means by which they can earn money which is under their control. The Tigre are especially dependent on the dom palm, and for many families dom leaves form the third strand of their farming system, along with animals and crops.
The dom palm has a dual role in food security. Since leaf cutting provides more reliable returns than agriculture, people fall back on it in order to buy their staple food in the hungry months or in years of bad harvests. In extremis, when grain is not affordable, the dom fruit becomes people's staple food. For the Kunama this is the principle reason for valuing and protecting the dom forests.
It does not seem to us that traditional users are damaging the dom forest: although leaf cutting is very widespread we saw no evidence that it was exceeding sustainable levels except in areas immediately around major settlements. There is widespread understanding among local people of dom growth rates, and they follow an informal code controlling leaf cutting which reflects their desire to preserve the resource. This includes a ban on cutting the "central bud", which it is believed would kill the tree. While productivity is sufficiently high that there is scope for further exploitation in most areas, increasing population pressure and uncontrolled harvesting could change this: the lower Gash seems particularly vulnerable at this time.
Overall, dom forest is such a vital resource to the rural population that it should be conserved primarily for this purpose. Since the forest appears capable of sustaining this role, management and policies should be aimed at supporting the current situation, rather than trying to restrict access to the forest. Further, local use should be given priority over commercial development, unless compensatory sources of income are available to local villagers.
Local management systems for forest of all types have traditionally existed in most areas, and there is still a strong "conservationist" outlook among Lowland villagers. All these systems are based on a ban on uncontrolled cutting of live wood but vary greatly in the degree of "ownership" that villagers claim to forest areas and the extent to which they control access. These systems appear to have been unable to withstand the pressures of the recent past, and even where reinstituted after the war they will probably be unable to cope with rising populations without external support. However, there is already some partnership between villagers and government in forest management: traditional systems are now linked with official regulations, and the programme of enclosures is necessarily a joint venture. The structures through which these are organised are currently rather weak through lack of resources and experience, but form a foundation on which improvements could be based.
Enclosures could become a major tool for forest conservation and rehabilitation in the Lowlands. Although the population is not yet familiar with the concept, we found that many villagers were receptive to the idea. Any programme should, however, be treated as experimental until proven techniques are developed for the Lowland environment and patterns of land use. One problem may well be that there seems to be no recognition of the threat that grazing poses to tree regeneration, particularly in the savanna woodlands, and thus of the need for long-term enclosure.
The perception of the need for active management varies in different areas. In general, villagers perceive the savanna woodland as being under more stress than the riverine forest, and the Barka dom forests in particular are seen as an abundant resource with no need of protection or management. The only areas of riverine forest where there is an existing management system reflecting pressure on the resource are on the lower Gash.
Other uses of the forest
The settlement of large numbers of returning refugees in the Lowlands has a significant impact on the forest. This occurs in two stages: an initial peak of demand for building materials and land for agriculture followed by long-term needs for fuelwood and raw materials for income generation. The first of these can be accommodated through careful management where settlements are close to productive riverine forest, but no provision is currently being made to meet settlers' ongoing needs. This is likely to result in over-use of the forest, particularly of savanna woodlands. After food rations in the new settlements are stopped, returnees are likely to become dependent on dom for their livelihoods in the same ways as the local population. It is thus possible that areas of riverine forest may become degraded, given the concentration of large numbers of people in the new settlements.
Complete clearance of tree cover for irrigated agriculture currently poses the greatest threat to the riverine forests, since the preferred sites are exactly those occupied by the forest. At present there is tremendous interest in commercial agriculture and, given the importance attached by GoE to cash-cropping for export, there is obvious potential for conflict between national economic and environmental interests. Government regulations exist to protect the forest from destruction by agricultural concessions, but enforcement has been difficult, given both lack of resources and the absence of legal backing for any sanctions imposed by MoA. If proposed new regulations to remove agriculture to areas outside the forest are introduced and enforced, the threat to the forest from this source will be lifted. However, if these are not enforced then there will be serious consequences in terms of forest destruction and a resulting loss of crucial resources for local people.
The dom fibre factory in Agordat has recently been reopened, and could provide useful employment for local people through its demand for dom leaf cutting. Although this is potentially a sustainable use of the riverine forest, it would require a management system based on reliable silvicultural principles and data. This information does not yet exist and the current cutting cycle and harvesting areas have not been clearly defined. This may well result in degradation of the areas of dom scrub most accessible to Agordat. The factory management should prepare a written management plan, to be submitted to the MoA for authorisation and periodic review as more research is carried out.
Commercial fuelwood collection has damaged the savanna woodlands considerably in the past, since much "collection" has involved the cutting of live wood. Recent government legislation to control this has succeeded in reducing the quantities extracted, though collection - including tree cutting - continues to a very reduced extent. However, these new regulations have abruptly deprived some Lowland villagers of a significant part of their income, and any further restrictions should take this side effect into consideration.
At present gum arabic collection from Acacia senegal in the savanna woodlands is of minor importance to the Lowland population. Commercial gum collection by the local population could, however, provide a supplementary source of income and increase food security if extension efforts were made in this direction. Current government promotion is focused on large scale commercial exploitation through the allocation of concessions.