This paper seeks to demonstrate the usefulness of the data held at the National Biodiversity Data Bank (NBDB) situated at Makerere University Institute of Environment and Natural Resources (MUIENR). We assess its value as a potential planning tool, based on the growing evidence that Uganda aspires to a robust Protected Area system that encompasses protection of biodiversity at the genetic, species and ecosystem levels. Analyses are presented of the coverage of 21 major vegetation types, and of species of birds and mammals. Several important vegetation types are inadequately conserved, whilst coverage of some categories of birds is also incomplete. The situation seems to be worse for mammals, although this is harder to assess because the distributions of many species are poorly known.
USING BIODIVERSITY DATA TO REVIEW COVERAGE OF UGANDA'S PROTECTED AREASDerek Pomeroy, Herbert TushabeMakerere University Institute of Environment and Natural ResourcesP.O. Box 7298, Kampala, UgandaMichael Green⟨sup⟩1⟨/sup⟩World Conservation Monitoring Centre219 Huntingdon Road, Cambridge, EnglandABSTRACTThis paper seeks to demonstrate the usefulness of the data held at the National Biodiversity Data Bank (NBDB) situated at Makerere University Institute of Environment and Natural Resources (MUIENR). We assess its value as a potential planning tool, based on the growing evidence that Uganda aspires to a robust Protected Area system that encompasses protection of biodiversity at the genetic, species and ecosystem levels. Analyses are presented of the coverage of 21 major vegetation types, and of species of birds and mammals. Several important vegetation types are inadequately conserved, whilst coverage of some categories of birds is also incomplete. The situation seems to be worse for mammals, although this is harder to assess because the distributions of many species are poorly known.INTRODUCTIONIn recent years, the term 'Protected Area' (PA) has come to embrace a wide variety of land use types, from those which are strict Nature Reserves, to others intended to meet broader needs, including those of local communities (McNeely, 1994). Countries such as Uganda, which have signed the Convention on Biological Diversity, are required to develop national conservation strategies. Inevitably, PAs will have a key role in this effort. In Uganda, the major PAs are National Parks, Wildlife Reserves and Forest Reserves (figure 1). The first two are administered by the Uganda Wildlife Authority (UWA), the last by the Forestry Department (FD). Two other categories, in practice largely notional, are Controlled Hunting Areas, and Wildlife Sanctuaries, also managed by the UWA (table 1). During the 1990s, both the FD and the UWA have undertaken detailed reviews of their estates. The FD collected extensive new data (FNCMP, 1999), whereas the UWA relied mainly upon existing data, including a draft of this article (Lamprey et al., 1998). Biodiversity data for the UWA estate remain very incomplete.⟨sup⟩1⟨/sup⟩ Present address: Broads Authority, 18 Colegate, Norwich, England.Figure 1. The major Protected Areas of Uganda showing the UWA estate and the Forest Reserves used in the analysis.In developing a robust PA system, research (including the sorts of analyses discussed here) has a major part to play. Earlier work on Uganda's PAs had already demonstrated, for example, that species richness in a PA reflects the number of ecosystems that it contains, and the effectiveness of a forest in supporting species of forest-specialist birds depends upon its size (Pomeroy, 1995). Such results are unsurprising, but until recently they were not readily accessible to PA managers, and often lacked clear recommendations (Harmon, 1994; Alexander, 1995). These are amongst the reasons why PA managers worldwide have been slow to adopt research results in conservation planning (Prendergast et al., 1999).A range of categories of PAs is recognised by IUCN (McNeely, 1994). The highest category is Nature Reserve, which in Uganda is currently found only within some Forest Reserves. At present, Uganda has ten National Parks, ten Wildlife Reserves, and more than seven hundred Forest Reserves, which range in size from a few hectares to over 100,000 hectares (FNCMP, 1999). There are 15 Controlled Hunting Areas (Lamprey et al., 1998). The PA system in Uganda, as in many African countries, originated largely to provide areas for hunting and for timber exploitation. So the fact that a reasonable proportion of Uganda's ecosystems and species are today found in Protected Areas is largely fortuitous. This is why it is now desirable to review the PA system and to plan for coverage to be as complete as possible for all elements of biodiversity. Such reviews require information that can only come from research.Some species are widely distributed whereas many others have very limited distributions: for example, there are two species of plant whose total world distribution, so far as is known, is Tororo Rock in eastern Uganda-an unprotected area of less than 10 ha (Okullo, 1997). Although these small distributions are easily missed in countrywide analyses, they must not be forgotten as we look for patterns.Table 1. Categories of Protected Areas in Uganda, totalling about 60,000 km⟨sup⟩2⟨/sup⟩. (Note that a few FRs overlap with UWA areas, for example in the Maramagambo Forest).CategoryAuthority⟨sup⟩a⟨/sup⟩ responsibleTotal area in Uganda (km⟨sup⟩2⟨/sup⟩)Level of protection⟨sup⟩b⟨/sup⟩National Park (NP)UWA11,680generally goodForest Reserve (FR)FD8,700variable, generally good in larger reservesWildlife Reserve⟨sup⟩c⟨/sup⟩ (WR)UWA8,630variableControlled Hunting Area⟨sup⟩d⟨/sup⟩ (CHA)UWA30,700⟨sup⟩d⟨/sup⟩variable, often slightWildlife Sanctuary (WS)UWA810usually negligible⟨sup⟩a⟨/sup⟩ UWA=Uganda Wildlife authority, formed from the Game Department and National Parks in 1995-1996; FD=Forest Department⟨sup⟩b⟨/sup⟩ Obviously there can be different opinions on this: those offered here are widespread⟨sup⟩c⟨/sup⟩ Formerly Game Reserves; both WRs and CHAs are currently undergoing re-evaluation⟨sup⟩d⟨/sup⟩ Lamprey et al., 1998: some overlap with FRsAn ideal Protected Area system would include substantial and representative areas of all the main vegetation types native to the country, and viable populations of all the native plant and animal species. Of course, ideals are hard to achieve, but they can provide guidance as to the directions in which to aim. At a later stage, as more data become available, we plan to examine the coverage of various taxa by Uganda's Protected Area system as a whole. However, at present we restrict ourselves to the following, for which there are adequate data:ovegetation types in Uganda's PAs, especially National Parks and Forest Reserves; andobirds and mammals of National Parks.METHODSAn earlier review of Uganda's PA system (Pomeroy, 1995) had depended on simple manual analyses. The electronic data used for the analyses in this paper are stored in the National Biodiversity Data Bank (NBDB) at the Makerere University Institute Environment and Natural Resources (MUIENR). This is described elsewhere (Reynolds et al., 1999).It must be stressed that in this analysis, while considerable care and effort have been put in calculating various areas covered by the different PAs, discrepancies, such as with other figures cited elsewhere, still occur. This is mainly because some PA boundaries have been changed and some remain undefined, as well as the scales at which various layers have been digitised. However we feel that this is the first step toward such analyses.VegetationThe classification of natural and semi-natural vegetation by Langdale-Brown et al. (1964) recognises 22 major categories designated by letters (A, B, C, ...)(appendix 2), and reflects major climatic and physiognomic groups. These are sub-divided into more than 80 types, distinguished mainly by their various dominant plant species. We have used only the major categories, as listed in appendix 2. The Langdale-Brown map includes many mixtures. We have treated these by identifying in each case the dominant type, as indicated on the map by the appropriate colours.In assessing the extent to which Uganda's present National Parks (NPs) and Forest Reserves (FRs) provide adequate coverage of the major vegetation types, we have taken an arbitrary total area of 300 km⟨sup⟩2⟨/sup⟩ in NPs and FRs as a measure of adequacy, comprising at least two separate blocks of 100 km⟨sup⟩2⟨/sup⟩ or more each.Birds and MammalsData for birds and mammals were compiled by MUIENR and published in 1995 by the Uganda National Parks (a precursor of the UWA). For birds, a recently-published Red Data List (Bennun & Njoroge, 1996) indicates species at risk in the East African Region, which includes Uganda.As recommended in the Global Conservation Strategy (1992), we have considered that the PA system supports individual species adequately if they occur in at least two separate parks. This criterion is weaker than stipulating two viable populations, but in Uganda such data are rarely available.RESULTSCoverage of Vegetation Types by Protected AreasThe electronically-stored information in the NBDB can be interrogated in many ways. For example, it is possible to calculate the total areas of the 21 major vegetation types recorded by Langdale-Brown et al. (1964) and to see how much of each is found within National Parks and other Protected Areas. Appendix 1 shows the total areas of each vegetation category in National Parks (NPs), as well as in the total UWA estate and in Forest Reserves (FRs). The 700 or so Forest Reserves within the Forestry Department's estate used to span an area of 15,950 km⟨sup⟩2⟨/sup⟩ (FNCMP, 1999), before some 2,500 km⟨sup⟩2⟨/sup⟩ were transferred to the management of the Uganda Wildlife Authority as National Parks. However, a good number of the FRs, most especially the smaller ones and those of little commercial value, exist only on paper, so to consider them as protective of natural vegetation would be unrealistic. In this analysis, we have used 63 of the large Forest Reserves (covering about 8,700 km⟨sup⟩2⟨/sup⟩) for which data were available in digital format from Forestry Department by then. In Uganda, a number of Forest Reserves partially overlap or are wholly within National Parks. Our analysis, as shown in appendix 1, takes this into account.The flora of Controlled Hunting Areas (CHAs) in Uganda are not accorded any legal level of protection, so the natural vegetation in these categories would not be considered as protected. As a result, the percentage figures in columns 11 and 13 of appendix 1 show both percentages of the respective vegetation types protected by UWA as a whole and what is protected excluding the CHAs. Currently, UWA is revising the extents and protection status of this category of PAs, and its hoped that those that protect unique vegetation types will be accorded some higher level of protection.With the exception of two very widespread vegetation categories (N and W), all PAs containing more than 100 km⟨sup⟩2⟨/sup⟩ of a particular vegetation category are listed individually in appendix 2.In table 2 we summarise the coverage in terms of arbitrary categories of adequacy. Despite the subjective nature of this procedure, it does enable us to identify those vegetation categories which, under present management practices, seem not to be vulnerable. Others, with less than 300 km⟨sup⟩2⟨/sup⟩ in NPs and/or FRs (or, if they exceed 300 km⟨sup⟩2⟨/sup⟩, with only one block exceeding 100 km⟨sup⟩2⟨/sup⟩) are obviously at greater risk.Table 2. A preliminary assessment of the conservation status of the main vegetation types in Uganda (A to Y) as defined by Langdale-Brown et al. (1964).Area within present National Parks and Forest Reserves⟨sup⟩a⟨/sup⟩ (km⟨sup⟩2⟨/sup⟩)Total area of each vegetation type in Uganda⟨sup⟩b⟨/sup⟩⟨1000km⟨sup⟩2⟨/sup⟩ ⟩1000 km⟨sup⟩2⟨/sup⟩Area in NPs + FRs ⟩300at least two blocksABCDG⟨sup⟩C⟨/sup⟩KNPQWof ⟩100 eachNO T VULNERABLE A T PRESENTonly one block ⟩100-HLVXFRVULNERABLEArea in NPs + FRs⟨300total PA area ⟩300SJMTSERIOUSLY ENDANGEREDENDANGEREDtotal PA area ⟨300YCRITICAL-⟨sup⟩a⟨/sup⟩ These are assumed to be the two best-protected categories of PA⟨sup⟩b⟨/sup⟩ In many cases, the actual areas will be much less, as more land is converted to agriculture and other uses.⟨sup⟩c⟨/sup⟩ Although only 83 km⟨sup⟩2⟨/sup⟩ in the appendix, part of the area is both National Park and Forest Reserve.Of greatest concern from this analysis is swamp forest, category Y, whose only significant block is amongst the group of forests collectively known as the Sango Bay Forest Reserves. They have several unique features (Davenport & Howard, 1996; Katende & Pomeroy, 1997). If, for example, the Kagera River were to be dammed, thus affecting the hydrology of the plains on which the forests lie, their nature could be changed for ever. Further, the contiguous Minziro Forest in Tanzania is being heavily felled (D. Moyer, pers. comm.), which puts more pressure on the Sango Bay forests whilst also increasing their importance.Although less clear from this analysis, because Langdale-Brown et al. (1964) only mapped the major ones, both permanent and seasonal wetlands are also at risk. Many reports have stressed their inadequate representation within Uganda's PA system (see, for example Ministry of Natural Resources, 1995; Omoding et al., 1996).Coverage of Mammals and Birds by National ParksOf the 335 species of mammals recorded from Uganda (Gathua & Vanden Berghe, 1993; Davies & Vanden Berghe, 1994), only 58% have so far been recorded within the National Park system (UNP, 1995; figure 2). Whilst many of those missing from the park lists are simply poorly known, a number of larger species are also missing, including two duikers, several squirrels, a jackal and two species of hyrax. Very few species are widespread, only four occurring in seven parks, and none in eight or more. Using the two-park criterion as a measure of adequacy, only about 40% of Uganda's native mammal. species are adequately conserved within National Parks.Birds are much more widespread, with ten of Uganda's 1010 species known from all 10 parks (figure 3). Nevertheless, there are as many as 70 species (6.9%) that have not been recorded in a single park, and only about 70% of the species are known from two or more parks. The coverage of the more specialised species is of particular concern. The forest specialists are virtually all residents, and nearly a third of them are recorded from less than two parks. Fortunately, almost all of these occur in Forest Reserves. Most waterbirds are more mobile, and some of the larger species, such as flamingos and White Pelicans, Pelecanus onocrotalus, are present at times in thousands. Although species such as these do not breed in Uganda, wetlands are important feeding areas for them, and are thus necessary for their survival.Figure 2. The numbers of mammal species occurring in different numbers of National Paries in Uganda. Based on UNP (1995).Coverage of Threatened Birds by National ParksUganda supports a number of globally threatened species of plants and animals. For birds, the regional Red Data List enables a more detailed analysis to be made (table 3). The global categories of critical, threatened, and near-threatened, have been applied to the region (Bennun and Njoroge, 1996). There is an additional category for species for which the East African region has special responsibility. These are birds that have at least 90% of their global range within the East African region (which for this purpose includes Rwanda and Burundi, as well as Kenya, Uganda and Tanzania). Three East African habitats have a number of species endemic to them. Such species are automatically included, however common they may be. These habitats are the coastal forests, the Albertine Rift forests and papyrus swamps. Their inclusion is justified because the major representation of each of them is within the region (as is the whole of the Eastern Arc range of forests in Tanzania).Table 3. Numbers of threatened bird species in Uganda, with their representation within the National Panks (from UNP 1995 and Bennun and Njoroge, 1996). See also Table 5.National parksCategoryMgahinga Gorilla (MG)Bwindi Impenetrable (Bl)Rwenzori Mountain (RM)Queen Elizabeth (QE)Semuliki (Se)Kibale (Ki)Lake Mburu (LM)Murchison Falls (MF)Kidepo Valley (KV)Mount Elgon (ME)Not in any parkUganda Total⟨sup⟩a⟨/sup⟩⟨underline⟩GLOBALLY- LISTED⟨/underline⟩Vulnerable1415001131312Near-threatened1415513521112Restricted range1322131440003028Globally-listed total⟨sup⟩b⟨/sup⟩15301511954655443⟨sup⟩C⟨/sup⟩⟨underline⟩REGIONALLY LISTED⟨/underline⟩Critical000000000011Endangered2825422342211Vulnerable117425191812171413552Near-threatened319833171515252217669Regional responsibility17352519101616111724674Regionally-listed total⟨sup⟩b⟨/sup⟩2379398250514556575620187⟨sup⟩a⟨/sup⟩ This is the total number of species in this category recorded in Uganda; it is not the total across the table, because some species occur in more than one Park.⟨sup⟩b⟨/sup⟩ These are not numbers of species, since one species can occur in two categories, e.g. Near-threatened and Restricted Range.⟨sup⟩c⟨/sup⟩ Includes one Data Deficient species.Figure 3. The numbers of bird species occurring in different numbers of National Parks in Uganda. Based on UNP (1995).Of all the species of birds listed for Uganda, about 90% are found within at least one park (table 3). The large numbers of Albertine Rift endemics, some of which are very restricted or rare, account for the high totals of globally-listed species in the three forested parks in the south-west of Uganda. Almost all parks hold important numbers of regionally listed species, with Queen Elizabeth National Park heading the list. Forest birds, most of which have restricted distributions (Bennun et al., 1996), are also well-represented in Uganda's PAs.Broader perspectivesThere are many ways of assessing the conservation value of a PA. Using data for both birds and mammals, figure 4 illustrates this for various conservation values. Bwindi Impenetrable is especially important for Albertine Rift endemic birds, Kidepo Valley for birds and mammals found in no other park, Semuliki for forest specialist birds, and so on. Parks with a low 'score' in figure 4 are often important for other groups: Golden Monkeys Cercopithecus mitis kandti and Mountain Gorillas Gorilla g. berengei in Mgahinga Gorilla National Park, for example.Thirty-five of the 45 nationally-listed species of threatened woody plants in Uganda occur in at least one National Park or Forest Reserve (Okullo, 1997); some of these species are also globally-listed (Oldfield et al., 1998).The regional Red Data List for birds (Bennun and Njoroge, 1996) helps to put Uganda's conservation needs into a wider view. There are also, of course, global Red Data lists for many taxonomic groups, but consideration of those is beyond the scope of this paper. However, it is practicable to assess Uganda's importance in terms of global 'ecofloristic zones'. These are mapped and analysed by Green et al. (1996), who show that nine of the zones occur in Uganda. Two can be ignored, as only very small parts of them are found within the country, and three others, mainly in the Rwenzoris and Mt Elgon, are considered to be well-protected already. For the remaining four (table 4), three are rated as of high global priority for conservation and one is placed in the highest category of all. These assessments are based on the fact that a significant proportion of these four zones occurs within Uganda and, in the case of Transition forest, that the area currently conserved is inadequate globally, so that the PAs in this category in Uganda gain extra significance.Figure 4. The relative conservation values of Uganda's National Parks, according to four different criteria. Data mainly from UNP (1995); forest birds according to the 'FF' category of Bennun et al., 1996. For mammals, species restricted to either one or two parks are both included, the latter figure being halved to maintain the same scale. Abbreviations of Park names as in Table 3.DISCUSSIONConservation of ecosystemsBased upon the data presented in this paper, Uganda's Protected Area system provides quite good coverage for the various vegetation types, as welbas for birds and mammals. However, there are some clear gaps. Further, the data themselves have many more gaps than one would wish; this is likely to affect complementarity analyses of the type used by the FNCMP (1999), which assumes, tacitly, that all species are recorded. In reality, it is probable that only 60-80% are recorded (E. Kabesiime, T. Otim and D. Pomeroy, unpubl. data). Finally, but importantly, some PAs are less well protected than others.The analyses used by the FD and UWA used the full set of 80 Langdale-Brown categories and they defined minimum areas as 50 km⟨sup⟩2⟨/sup⟩ (FNCMP, 1999, page 40) and 100 km⟨sup⟩2⟨/sup⟩ (Lamprey et al., 1998, page 22) respectively. But even in the thoroughly-researched UK, 'large-scale habitats' such as woodland, farmland and grass/heath are useful in reaching important conservation conclusions (Gregory and Baillie, 1998). So the discussion here is limited to the use of the letter-grade categories of our analyses.Table 4. Major African habitats, according to Green et al. (1996), which are high priorities for conservation action in Uganda.Ecofloristic zone⟨sup⟩a⟨/sup⟩Global priority for conservation actionPAs in Uganda with at least 100 km⟨sup⟩2⟨/sup⟩ of the zone11 Sudanian woodland, dry evergreen forest and savannahHighMurchison Falls NP39 Transition forest, 1,000-1,800 mHighestKibale NP, Budongo, Mabira, Sango Bay and Kasyoha-Kitomi FRs43 Transition woodland and forest, 1,000-2,000 mHighKidepo Valley NP, Bokora Corridor and Pian Upe WRs and North Karamoja CHA51 Scrub forest, semi-evergreen thicketHighLake Mburo NP⟨sup⟩a⟨/sup⟩based upon a global system, as adopted by FAO: details in Green et al. (1996)The vegetation types defined by Langdale-Brown et al. (1964) can be broadly equated with ecosystems. Many of the 22 main categories are well-represented within National Parks and Forest Reserves, where they receive a reasonable level of protection (table 2; appendix 1; FNCMP, 1999). Of those whose coverage already appears inadequate, the two of greatest concern are S (Grass steppe) and Y (Swamp forest). Category S could be included in a revised PA system for Karamoja (Pomeroy and Tushabe, 1996; Lamprey et al., 1998), which would also improve the coverage of category T, which is also inadequate at present. The case for swamp forest, category Y, is more complex. Close collaboration with Tanzania, where a similar type is found in Minziro forest, would appear to be the best policy.Two other vegetation categories could be considered as threatened (table 2). Moist Acacia savannah (J) is remarkably rich for birds (T. Otim and D. Pomeroy, unpubl. data) and possibly other groups. An upgrading of part of the Katonga CHA would seem to be the best prospect here. Type M, palm savannah, occurs in both Murchison Falls and Kidepo Valley National Parks, but in both cases is subject to the depredations of fire (and, in the past, elephants Loxodonta africana) and needs careful management.Global analyses of vegetation conservation (such as that by Green et al., 1996) coincide to some extent with the analyses for Uganda given here. To assess the extent to which Uganda is meeting its global responsibilities detailed reviews are needed. These should become possible in the next few years.Conservation of speciesAnalysis for coverage by species is only possible for birds in National Parks, where the data are comparatively complete at a presence-or-absence level, and to a lesser extent for mammals. At least for the more serious threat categories, the few species of birds that are not found within any National Park are unlikely to be helped by additions to the park system (table 5).Many species of birds and mammals are recorded from only one park or none at all (figures 2, 3). There is no satisfactory way of defining an acceptable minimum number of sites, since the way in which metapopulations function varies between taxa, but in any case is poorly known (Harrison, 1993). In the future, many of these species are likely to be found in other PAs, although not necessarily in viable numbers. But doubts must remain until more information is available. Perhaps wisely, most practical conservationists are wary of putting minimum numbers on ecosystems, or individuals (Groombridge and Jenkins, 1996).One would of course like to know, not simply that a species is represented in two or more secure reserves, but that it has viable populations in each of them. Similarly, adequate representation of vegetation types depends amongst other things on deciding what is meant by adequate. Our criteria are somewhat different from those of WCMC (Green et al., 1996). There is obviously a need for closer collaboration with other countries, with whom some responsibilities may be shared. For example, the Maccoa Duck and the Cape Grass Owl, rarely recorded in Uganda (table 5) can be better conserved in neighbouring Kenya and Tanzania where they are more common (Britton, 1980).Uganda's Protected Areas as a systemUganda has three major groups of PAs, which include a number of National Parks (figure 1):• EastKidepoValley National Park to Mount Elgon National Park, possibly with some new gazettments;• South-westMgahinga Gorilla National Park via Queen Elizabeth to Rwenzori Mountains, Kibale and Semuliki National Parks. Some of these, such as Mgahinga Gorilla, Rwenzori Mountains and Queen Elizabeth National Parks are contiguous with other large parks in the Democratic Republic of Congo and in Rwanda.• WestThe Murchison group, including the important Budongo Forest Reserve.Table 5. Bird species in the highest IUCN threat categories which are not represented in any of Uganda's National ParfcsCategoriesDistribution and StatusPAs where present in UgandaGlobalRegionalGlobalRegionalGreat Crested GrebePodiceps nigricollis-CRWidespreadDeclining rapidlyNoneMaccoa DuckOxyura maccoa-ENEast and South AfricaKenya, Tanzania, local[No record since 1930]Taita FalconFalco fasciinuchaVUVULocal in East and South AfricaScarce, few localitiesNone⟨sup⟩a⟨/sup⟩CorncrakeCrex crexvuVUBreeds PalearcticWidespread, rareNoneLesser Spotted CrakePorzana pusilla-VUE and S AfricaLocal, rareNoneCape Grass OwlTyto capensis-VULocal in E and S AfricaLocal in highlandsNoneTurner's EremomelaEremomela tumerivuEN/RREast AfricaVery localNyondo FRHeuglin's Masked WeaverPloceus heuglin-VUEastern DRC⟨sup⟩b⟨/sup⟩ to KenyaLocal in Uganda, very rare in KenyaNone recorded⟨sup⟩a⟨/sup⟩ This species has been recorded breeding near Sipi Falls, some 3 km from Mt. Elgon NP; but there is evidence of a recent decline (K. Otte, pers.comm.).⟨sup⟩b⟨/sup⟩ Democratic Republic of CongoLong-term planning might aim to consolidate these groups of National Parks with adjacent WRs and FRs; data analysis could be designed to test the effectiveness of such a system. We would then be in a much better position to assess how much of the country's biodiversity is conserved through the Protected Area system. With more data, and an increasing capacity in MUIENR to undertake GIS analyses, it will become possible to employ a more extensive gap analysis (Scott et al., 1992) and further improve the design of the Protected Area system for the country.CONCLUSIONSThe use of electronic databasesoThe examples given in this article and by Reynolds et al. (1999) are sufficient to show the potential of electronic databases in planning conservation strategies. Establishment of the databases, coupled with well-planned field work, can go a long way towards providing a ready source of information on the status of a particular resource, whether it is within a Protected Area or not.oAttempts to produce more sophisticated analyses for Uganda have been hampered by the lack of earlier environmental data. Even where such data are available, they are often in formats unsuited to the objectives of setting up a database for conservation planning.Nevertheless, we hope to have demonstrated that methods already exist for preliminary assessments of the extent to which biodiversity is being conserved by some of Uganda's Protected Areas.Conservation of ecosystems and speciesoThe present coverage of terrestrial ecosystems is inadequate, according to analyses of the classification of both Langdale-Brown et al. (1964) and Green et al. (1996). However, possible strategies exist to conserve all of the less-well-represented categories.oWetlands are seriously under-represented, and plans to address this require vigorous action (Ministry of Natural Resources, 1995; Omoding, 1996).oConsolidating scattered PAs into groups would provide better insurance and larger areas of key habitats, notably wetlands and forests, would be linked.oThe still extensive gaps in the data can only be filled by well-co-ordinated research.ACKNOWLEDGEMENTSWe greatly appreciate the inputs made by the Global Environment Facility (GEF) project- "Institutional Support for the Protection of East African Biodiversity"-towards the development and maintenance of the database. We also wish to acknowledge the contribution of Dr. Jake Reynolds of WCMC, who initially developed the application, and whose continued technical inputs have helped the NBDB to realise its initial objectives. Analyses of the maps of Langdale-Brown et al. (1964) were made by the late James Omoding II, and more recently by Fiona Mutekanga, to whom we are most grateful.We wish to express our indebtedness to Dr Alan Rodgers and Mr Robert Nabanyumya of the GEF Project, and to an anonymous referee, for having read and commented on earlier drafts of this paper.Lastly, we would like to acknowledge the efforts made by the staff of the MUIENR NBDB to make the Data Bank what it is.REFERENCESAlexander, M. (1995). Management planning in relation to protected areas. Parks 5: 2-11.Bennun, L., C. Dranzoa & D. Pomeroy (1996). The forest birds of Kenya and Uganda. Journal of East African Natural History 85: 23-48.Bennun, L. & P. Njoroge (eds) (1996). Birds to Watch in East Africa: a preliminary Red Data List. Research Reports of the Centre for Biodiversity, National Museums of Kenya. Ornithology 23. Nairobi, KenyaBritton, P.L. (ed.). (1980). Birds of East Africa: Their Habitat, Status and Distribution. East Africa Natural History Society, Nairobi.Davenport, T. & P. Howard (eds). (1996). Sango Bay Forest Reserves: Biodiversity Report No. 20. Forest Department, Kampala.Davies, G. & E. Vanden Berghe (1994). Checklist of the Mammals of East Africa. East Africa Natural History Society, Nairobi.FCNMP (Forestry Nature Conservation Master Plan) (1999) Forest Department, Kampala.Gathua, M. & E. Vanden Berghe (1993). Mammals of East Africa: preliminary statistics. EANHS Bulletin 23: 72-74.Global Conservation Strategy (1992). Washington, USA, WRI/IUCN/UNEP.Green, M.J.B., M.G. Murray, G.C. Bunting & J.R. Paine (1996). Priorities for biodiversity conservation in the tropics. WCMC Biodiversity Bulletin 1.Gregory, R.D. and S.R. Baillie (1998). Large-scale habitat use of some declining British birds. Journal of Applied Ecology 35: 785-799.Groombridge, B. and M.D. Jenkins (eds) (1996). Assessing Biodiversity Status and Sustainability. WCMC, Cambridge.Harmon, D. (ed.) (1994). Co-Ordinating Research and Management to Enhance Protected Areas. IUCN, Gland, Switzerland.Harrison, S. (1993). Metapopulations and conservation. In: P.J. Edwards, R.M. May & N.R. Webb (eds.), Large Scale Ecology and Conservation Biology, Blackwell Scientific Press, Oxford. Pp. 111-128.Katende, A.B. & D.E. Pomeroy (1997). Was Sango Bay a Pleistocene refugium? EANHS Bulletin 27: 7-8.Lamprey, R., A. Muhimbura-Atukunda, E. Buhanga, F. Egunyu, J. Okecha, J. Omoding & F. Michelmore (1998). Strategic Plan for Biodiversity Conservation in the Former Controlled Hunting Areas of Uganda. Uganda Wildlife Authority, Kampala.Langdale-Brown, I., H. A. Osmasto, & J. G. Wilson (1964). The Vegetation of Uganda and Its Bearing on Land-Use. Government Printer, Entebbe.McNeely, J.A. (1994). Protected areas for the 21st century: working to provide benefits to society. Biodiversity and Conservation 3: 390-405.Ministry of Natural Resources (1995). National Policy for the Conservation and Management of Wetland Resources. Kampala.Okullo, J.B.L. (1997). A draft Red Data List of woody plants of Uganda. Kampala, MUIENR; WCMC, Cambridge.Oldfield, S., C. Lusty & A. MacKinven (1998). The World of Threatened Trees. World Conservation Press, Cambridge.Omoding, J., T. Otim, P.E. Etyang & N.M. Mutekanga (1996). Inventory of Wetland Biodiversity in Uganda: Activities, Methodologies and Results. Institutional Support for the Protection of East African Biodiversity. Food and Agricultural Organisation of the United Nations, Kampala.Pomeroy, D. (1995). Using distributional data for conservation: how well will Uganda's protected areas conserve the country's wildlife? In L.A. Bennun, R.A. Aman and S.A. Crafter (eds), Conservation of Biodiversity in Africa, National Museums of Kenya, Nairobi. Pp. 363-376.Pomeroy, D. & H. Tushabe (1996). Biodiversity of Karamoja. A report for the Uganda Wildlife Authority. MUIENR, Kampala.Prendergast, J.R., R.M. Quinn & J.H. Lawton (1999). The gaps between theory and practice in selecting nature reserves. Conservation Biology 13(3): 484-492.Reynolds, J., H. Tushabe & P. Kasoma (1999). The establishment of a National Biodiversity Data Bank for Uganda. EANHS Bulletin 29(3): 6-9.Scott, J.M., F. Davis, B. Csuti, R. Noss, C. Groves, H. Anderson, S. Caicco, F. D' Erchia, T.C.J. Edwards, J. Ulliman & R.G. Wright (1992). Gap analysis: a geographic approach to the analysis of biodiversity. Wildlife Monographs 123: 1-41.UNP (Uganda National Parks) (1995). Bird and Mammal Checklists for Ten National Parks in Uganda. Kampala.Appendix 1: Approximate coverage of vegetation types by Uganda's PA system. Columns 2-10 show the areas of each Langdale-Brown category (in km⟨sup⟩2⟨/sup⟩) in Uganda as a whole and in various categories of PA. Where an area is jointly administered, (eg Maramagambo FR, which is partly a Forest Reserve overlapping a National Park), this is indicated by Column 9. Column 6 is the result of the difference between Columns 3 and 5; and Column 7 of 6 and 9. Column 10 includes areas in Column 9.12345678910111213Category⟨sup⟩1⟨/sup⟩TotalIn UWA (all)⟨sup⟩2⟨/sup⟩In UWA excl. FRs⟨sup⟩3⟨/sup⟩InCHAsIn UWA excl CHAs. UWA excl. CHAs & FRsInNPsFRs in UWA⟨sup⟩4⟨/sup⟩Total in FD% ProtectedBy UWA (excl. CHAs & FRs)ByFDBy UWA (incl. CHAsA683674662136616556616696199B309526001800102815727871476785886252984C3235469435046943545534491131514D527983145526805430606375190383616F2461462657320442236928153418123G2599100774058841929411312520011839H4189155815582501308130888505031137J62326776714742031971807703111K150891250114851573567956655346528L25874307324312606467-5805256770312M26646476325111361218315155124N38020115029792897725252431308228222751630P14424143513158136225024181201983110Q138954371429245639153549256636526226231R15801580151311474333662786868234100S801801801332469469000590100T427742393986354869144818124324310699V46153398353723921006986336202621174w188257763759151332630242316020723213141X8626677681229448448374190518Y260342303423321110193913z692433438315817617325252315Open Water3584420020151149149116017001Totals24164549747452212945120296149981110252988625⟨sup⟩1⟨/sup⟩See Appendix 2 for details⟨sup⟩2⟨/sup⟩includes area of Forest Reserves contained partially or wholly within NP boundaries.⟨sup⟩3⟨/sup⟩Excludes area of Forest Reserves (FRs) contained partially or wholly within Uganda Wildlife Authority (UWA) boundaries.⟨sup⟩4⟨/sup⟩Areas of Forest Reserves which overlap with UWA categories.Appendix 2: The different vegetation categories according to Langdale-Brown et al. (1964) within Uganda's current PAs.Vegetation CategoryNPswith ⟩100km⟨sup⟩2⟨/sup⟩Other UWA designations with ⟩ 100 km⟨sup⟩2⟨/sup⟩FRswith⟩100km⟨sup⟩2⟨/sup⟩AHigh altitude heath and moorlandME, RM--BHigh altitude forestBl, ME, RMN.Karamoja, Sebei, S.Karamoja CHAsKadam, Moroto, Napak, TimuCMedium altitude moist evergreen forestBl, KFKatonga CHAKalinzu, Kasyoha-KitomiDMedium altitude moist semi-deciduous forestKF, QE, SeKigezi WRBudongo, Bugoma, Mabira, MaramagamboFForest/savannah mosaicKFKaruma WR, Katonga CHA-GMoist thicketQEN.Karamoja, S.Karamoja CHAs, Otzi Forest WSOtziHWoodlandMFE.Madi CHA, Karuma WR-JMoist Acacia savannah-Katonga CHA-KMoist Combretum savannhaMFE.Madi, Karuma Falls CHAs, Karuma WRBudongoLButryospermum savannahE.Madi, Lipan, N.Karamoja, N.Teso CHAs, Mt.Kei WSMt.KeiMPalm savannah-Semuliki CHA-NDry Combretum savannahKV, MF3 WRs and 8 CHAs7FRsPDry Acacia savannahLM, QEPian Upe WR, Katonga, Sebei, S.Karamoja CHAs-QGrassland savannahLM, MF, QEBokora Corridor, Kyambura, Pian Upe WRs, S.Karamoja CHARwoho Plantation ForestRTree and shrub steppeKVMatheniko WR, N.Karamoja, S.Karamoja CHAsSGrass steppeMatheniko WR, N.Karamoja, S.Karamoja CHAsTBushlandKVMatheniko WR, N.Karamoja, S.Karamoja CHAs ZuliaVDry thicketMFBugungu, Matheniko WRs, Kaiso-Tonya, N.Karamoja, S.Karamoja CHAsWCommunities on sites with impeded drainageKV2 WRs and 7 CHAsNapakXSwampQEKatonga CHA-YSwamp forest---UWA - Uganda Wildlife Authority, NP - National Park, WR - Wildlife Reserve, CHA - Controlled Hunting Area, WS - Wildlife SanctuaryKEY to NPs: Bl - Bwindi Impenetrable, KF - Kibale Forest, KV - Kidepo Valley, LM - Lake Mburo, ME - MtElgon, MF - Murchison Falls, QE - Queen Elizabeth, RM - Rwenzori Mountains, Se - Semuliki.