In this section, biological importance is assessed primarily in terms of endemic species. Subsequently, in the Conservation Outcomes, the emphasis is on the Red Lists of threatened species that occur in the hotspot. Because of the relatively small area of this hotspot, the high degree of threat it faces (Brooks et al. 2002) and the current criteria for inclusion in the Red List (IUCN 1994), all, or at least most, of the endemics are candidate “threatened species.” This consideration is perhaps most obvious in the case of the plants where there are more than 1,500 endemic species in the hotspot, but only 236 (16 percent) are currently included in the Red List.
The global biodiversity values of the hotspot are widely recognized (Lovett 1988, 1998a, b, c; Myers 1990; Myers et al. 2000; Brooks et al. 2001; Brooks et al. 2002). This hotspot is home to at least 1,500 endemic plant species, 16 endemic mammals, 22 endemic birds, 50 endemic reptiles and 33 endemic amphibians (Lovett & Wasser, 1993; Burgess et al. 1998a; Burgess & Clarke 2000; Myers et al. 2000). It is considered as the hotspot most likely to suffer the most plant and vertebrate extinction for a given loss of habitat and as one of 11 “hyperhot” priorities for conservation investment (Brooks et al. 2002). Because of the small area of the hotspot, the densities of these endemics are among the highest in the world. At the global level, some 0.37 percent of all species (in eight major taxa) are estimated to be endemic to the Eastern Arc Mountains and 0.20 percent endemic to the Coastal Forest Mosaic (Burgess 2000).
The distribution of these endemic species within the hotspot merits special consideration. First, nearly all the forest patches have biodiversity values and most contain at least one endemic species (Burgess & Clarke 2000). Second, there are many disjunct distributions, particularly amongst the birds and the plants (Burgess & Clarke 2000). Third, there is a huge turnover of species between patches, especially in the less mobile species. Forests that are only 100 km apart can differ in 70 percent of their millipedes (Hoffman, 2000) and in 80 percent of their plants (Clarke et al. 2000). In some invertebrate taxa, 80-90 percent of species can be strictly endemic to a single site (Scharff et al. 1981; Scharff 1992, 1993; Burgess et al. 1998b).
These distribution patterns are commonly found in both the Eastern Arc Mountains and the lowland Coastal Forest Mosaic. They indicate that much of the habitat fragmentation in this area is natural and sufficiently ancient for much speciation to have taken place in isolated patches and for species to have persisted here and there due to stochastic effects. However, over a period of hundreds or perhaps thousands of years, there has also been considerable loss of habitat and habitat continuity between the natural fragments (loss of connectivity), as a result of human activities. This issue needs careful consideration when conservation interventions are planned.
In the Eastern Arc Mountains, around 40 percent (800 of more than 2000) of the plant species and 2 percent of genera (16 of about 800) are estimated to be endemic (Lovett & Wasser 1993; Lovett 1998b; GEF 2002). This area is the centre of endemism for the African violet, with 20 out of 21 species being endemic. Trees have attracted the most attention, but non-vascular plants also show significant endemism (32 of about 700 species of bryophytes) (Pocs 1998). The endemics are found in most of the forest types, as well as in intervening habitats such as rocky outcrops, heathland, montane grasslands and wetlands (Lovett 1998b).
The degree of faunal endemism in the Eastern Arc Mountains varies widely across taxa. Six percent of mammals, 3 percent of birds, 68 percent of forest-dependent reptiles, 63 percent of forest-dependent amphibians, 39 percent of butterflies and 82 percent of linyphiid spiders are endemic (GEF 2002). Some of these species have extremely limited distributions. The Kihansi spray toad, described in 1998, is found in an area of less than 1 km² (Poynton et al. 1998). Three endemic bird taxa (variously described as full species or subspecies) are restricted to the 6 km² of forest in the Taita Hills (Brooks et al. 1998). Records for the Udzungwa partridge are confined to two localities in the Udzungwas and one in Rubeho (Baker & Baker 2002). Amongst some invertebrates (linyphiid spiders, opilionids and carabid beetles), single site endemism exceeds 80 percent (Scharff et al. 1981; Scharff 1992, 1993; Burgess et al. 1998).
Using a subset of 239 species endemic and near-endemic to the Eastern Arc Mountains, the East Usambaras emerge as the most important site in terms of numbers of endemics, while the Ulugurus rank top for density of endemics (Burgess et al. 2001). As expected, the big forest blocks (Usambaras, Ulugurus and Udzungwas) are more species-rich than the smaller blocks (e.g., North Pare, South Pare, Ukaguru and Mahenge). Most of the endemic taxa are not only forest dependent; they are dependent on primary forest. The low-elevation forests are rich in endemics and total numbers of species, but are very limited in overall area, having suffered extensive clearance for agriculture. The uniqueness of the biodiversity in the Eastern Arc Mountains is attributable to both relictual and recently evolved species (Burgess et al. 1998c; Roy et al. 1997). Biogeographical affinities indicate ancient connections to Madagascar (45 species of bryophytes shared) (Pocs 1998), West Africa (many birds and plant genera) (Lovett 1998b; Burgess et al. 1998c) and even Southeast Asia (where close relatives of the Udzungwa forest partridge and the African tailorbird are found) (Dinesen et al. 1994).
The pattern of endemism in the Coastal Forest Mosaic is complex, reflecting the wide range of habitats and heterogeneous forest types, a high degree of turnover of local species between adjacent forest patches and many disjunct distributions (Burgess 2000; WWF-US 2003b). The ecoregion, which includes the islands of Zanzibar and Pemba, is a mosaic of forest patches, savanna woodlands, bushlands, thickets and farmland. The highest biodiversity is found in the various kinds of closed canopy forest vegetation: dry forest, scrub forest, Brachystegia (miombo) forest, riverine forest, groundwater forest, swamp forest and coastal/afromontane transition forest (Clarke 2000; WWF-US 2003b). Closed canopy forests, however, makes up only 1 percent of the total area of the Coastal Forest Mosaic.
Overall, there are more than 4,500 plant species and 1,050 plant genera (WWF-US 2003b), with around 3,000 species and 750 genera occurring in forest. At least 400 plant species are endemic to the forest patches and about another 500 are endemic to the intervening habitats that make up 99 percent of the ecoregion area (WWF-US 2003b). The majority of these species are woody but there are also endemic climbers, shrubs, herbs, grasses and sedges (Clarke et al. 2000). A substantial proportion of the endemic plants are confined to a single forest (for example, Rondo Forest, Tanzania, has 60 strict endemics and Shimba Hills, Kenya, has 12) (Clarke et al. 2000). The flora as a whole has affinities with that of West Africa, suggesting an ancient connection with the Guineo-Congolian lowland forests (Lovett 1993). Endemism is primarily relictual rather than recently evolved (Clarke et al. 2000; Burgess et al. 1998c).
Faunal endemism rates have been estimated for forest species in the Swahelian Regional Centre of Endemism (including the transition zone in Mozambique). These are highest in the invertebrate groups such as millipedes (80 percent of all the forest species), molluscs (68 percent) and forest butterflies (19 percent) (Burgess 2000). Amongst the vertebrates, 7 percent of forest mammals, 10 percent of forest birds, 57 percent of forest reptiles and 36 percent of forest amphibians are endemic (Burgess 2000). If Mozambique is excluded, endemics include 14 species of birds (including four on Pemba Island), eight mammals, 36 reptiles and five amphibians (WWF-EARPO 2002).
In terms of species richness, there are at least 158 species of mammals (17 percent of all Afrotropical species), 94 reptiles and 1200 molluscs (WWF-US 2003b). As with the plants, endemism is primarily relictual (Burgess et al. 1998c) and single site endemism and disjunct distributions are common. This makes it extremely difficult to prioritise the forests in terms of their biodiversity. Burgess (2000) made a preliminary analysis on the basis of species richness and endemism, using vascular plants, birds, mammals, reptiles and amphibians. This showed that different forests are important for different groups. For example, while Arabuko-Sokoke is top for endemic birds and for mammal species richness, it barely makes it into the top ten for plants. Overall, the five most important forests are Rondo (plants and birds), lowland East Usambaras and Arabuko-Sokoke (birds, mammals and reptiles), Shimba (plants and birds) and Pugu Hills (birds and mammals). Pemba Island, with an area of only 101 400 ha, is extraordinarily important for birds with four endemic species (Baker & Baker, 2002) while Zanzibar has six endemic mammals and three endemic birds (Siex, pers. comm.).
Forests in this hotspot are located in two countries and fall under multiple management regimes. In Kenya, the protected area network at national level consists of national parks, national reserves, forest reserves, nature reserves and national monuments (Bennun & Njoroge 1999). Many of the national monuments on the coast are sacred forests called Kaya Forests. At a lower level, many forests are located on trust lands and fall under the control of County and Municipal councils. In Tanzania, the protected area network at national level consists of national parks, game reserves, government catchment forests, game controlled areas, forest reserves and nature reserves (Baker & Baker 2002). Below the national level a large number of forests, particularly in the coastal forest belt, fall under local authorities, owned and managed by the villagers. In both countries, no exploitation is allowed in national parks and protection levels are generally high (but see below for an exception in Kenya). In both countries, confusing and overlapping legislation on the environment and natural resources is being rationalized through the enactment of new polices.
Within the Kenyan area of the hotspot, there is one national park, a 6 km² area to the northwest of Arabuko-Sokoke Forest. This park is, however, somewhat of an anomaly, as it contains no closed forest and exists only on paper. There are four national reserves (Shimba, Tana River, Boni and Dodori) (WWF-EARPO 2002). These fall under the jurisdiction of the Kenya Wildlife Service (KWS). The Shimba Hills were gazetted as National Forest in 1903 and then double-gazetted (with the exception of two small areas that remained as forest reserves under the control of the Forest Department) in 1968 as the Shimba Hills National Reserve (Bennun & Njoroge 1999). Protection levels are higher in the area controlled by KWS, as they have armed rangers and a clearer institutional mandate for conservation. The Tana River Primate National Reserve contains 16 out of the 70 patches of riverine forest found along the lower Tana River (Butynski & Mwangi 1994). There forests have suffered severe damage during the past three decades from farmers clearing land for agriculture and possibly from the construction of several dams up-river that have reduced the incidence of flooding (Butynski & Mwangi 1994, Wieczkowski & Mbora 1999-2000). The biodiversity in Boni and Dodori is poorly known because security problems have prevented biological surveys.
The largest of the Kenyan forest reserves is Arabuko Sokoke (417 km²). For the last 10 years this forest has been under multi-institutional management (KWS, the Forest Department, Kenya Forestry Research Institute (KEFRI) and the National Museums of Kenya, (NMK)) (Arabuko-Sokoke Forest Management Team 2002). This arrangement has been taken as a model for other indigenous forests in Kenya but has been rarely implemented. Protection levels suffer from the proximity of the tourist resorts of Malindi and Watamu and the resultant demand for carving wood and timber. The effectiveness of management has been variable over time, being subject to the commitment of the personnel on the ground, the working relationships between KWS and the Forest Department and the level of resources available. Generally, however, management has been more effective than in the other 17 forest reserves (WWF-EARPO 2002) within the Kenyan coastal forest belt. In the fragmented forests of the Kenyan portion of the Eastern Arc Mountains (Taita Hills), some patches, including plantation, have been gazetted as forest reserve. Others are on trust land administered by the local county council, some of which have been recommended for gazettement as forest reserves (Bennun & Njoroge 1999).
National monument status has been given to 39 out of nearly 50 of the sacred Kaya forests (WWF-EARPO 2002), but the level of protection gained from this status is below that of the forest reserves. An additional national monument at Gede Ruins is not a Kaya, but it includes a fenced 350 ha coral rag forest that is in good condition and very well protected. There are numerous Local Government or County Council Forests. Unfortunately, protection of these forests is virtually non-existent, to the point where local councillors have sold forest plots for agricultural settlement (e.g., at Madunguni and Mangea Hill). A large proportion (nearly 40 percent) of the Kenyan coastal forests fall into this category or is totally unprotected (data from WWF-EARPO 2002).
In the Tanzanian portion of the Eastern Arc Mountains, there are two national parks (Udzungwa Mountains National Park, gazetted in 1992, 1,960 km²; and Mikumi National Park, 3,230 km²), two game reserves (Selous and Mkomazi) and a nature reserve (Amani Nature Reserve, gazetted in 1997, 83.8 km²) (GEF 2002; Roe et al. 2002). However, more than 90 percent of the total forest area in the Tanzanian portion of the Eastern Arc Mountains and almost 75 percent of the total forests are gazetted as government catchment forest reserves (Burgess pers. com.). These range in area from more than 557,000 ha (Ngindo) to less than 10 ha and include all the larger forests in the Kilimanajaro (e.g., Chome), Tanga (e.g., Nguru North, Shume Magambe) and Morogoro (e.g., Uluguru, Nguru South) regions. Most of the remainder are local authority forests, ranging in size from 57,300 ha (Mbalwe/Mfukulembe) to less than 10 ha, although there are a few private forests, mainly on tea estates (e.g. Ambangulu Tea Estate) and some of which have been covenanted for conservation. In the national park, protection levels are high, but elsewhere they are highly variable. The important catchment forest reserves are, in general, better protected than the local authority forests (Burgess et al. 1998).
In the Tanzanian coastal forests, management regimes are more complicated. Most are either forest reserves (80) or are on public land (20) with no protection status (WWF-EARPO 2002). Four are private forest reserves (Magotwe, Kichi Hills, Mlungui and Magoroto). Only three are entirely managed by the district government as local authority forest reserves, although some have double status (two overlapping with forest reserves and two more with private forest reserves ). There are two Catchment forest reserves (Mselezi, Ziwani) (Burgess and Clarke 2000; WWF-EARPO 2002) managed by the Central Government Forest and Beekeeping Division. Two others, Zaraninge and the former Mkwaja ranch, are being incorporated into the new Sadaani National Park (WWF-EARPO 2002). Some patches are also found in the Selous Game Reserve and others in Mafia Island Marine Park. Offshore protected areas are also found in Zanzibar (Jozani Forest Reserve) and Pemba (Ngezi Forest Reserve). There are also smaller areas in Zanzibar that are important for water catchment (e.g. Masingi) and for endemic species (e.g. Unguja Ukuu Forest Plantation). There is a proposal to upgrade the Jozani Reserve in Zanzibar (now known as the Jozani-Chakwa Bay Conservation Area) to a national park.
Management and protection of most of the forests throughout the hotspot have suffered from inadequate stakeholder involvement, conflicts of interest and corruption. Where forests are gazetted, the boundaries tend to be respected but the forests themselves suffer steady degradation. The levels of protection achieved on the ground are strongly dependent on local factors such as proximity to urban areas, pressure for land, ease of access, presence of valuable timber and the capacity and morale of the local forestry officers (WWF-US 2003a). There is a general move toward various forms of participatory forest management (PFM), in the hope that an exchange of forest user rights for community management responsibilities and ownership (where appropriate) will lead to better protection by the people who often know best what is going on in the forests. Although this hope is widely held, it has not yet been scientifically tested within the Eastern Arc Mountains and Coastal Forests hotspot. The alternative strategies of direct payments and easements are being explored, but have not yet been implemented.
Previous: Background / Next: Conservation Outcomes