Das Feuerregime im Queen Elizabeth Nationalpark, Uganda : Auswirkungen auf die Verbreitung und Habitatnutzung grosser Herbivoren

The conservation of biodiversity and natural resources relies to a large extent on the establishment of
protected areas. However, protected areas often fail to meet the objectives that lead to their establishment,
due to poor control over the manifold, legal and illegal human activities within protected areas,
insufficient knowledge about system characteristics, in particular with respect to disturbance regimes,
and due to inappropriate management strategies and means of implementation.
The study investigates the responses of large herbivores to anthropogenic disturbances in Queen
Elizabeth National Park (QENP), Uganda. Unique features of QENP are high human population densities
within and outside the park and a high degree of habitat fragmentation, resulting from natural (topography,
lakes, channel) and anthropogenic (villages, roads, cattle pastures) factors. Outstanding
disturbances are man-made fires that frequently devastate large parts of the park. Fire frequency and
severity have increased over the last decades, leading to widespread and long-lasting loss of herbivore
feeding grounds.
The study is based on the assumption that the present fire regime, in combination with other anthropogenic
disturbances adversely affects the population development of the abundant large grazers in
QENP under the local geo-ecological conditions. The main goals of this study are (1) to establish spatial
and temporal habitat utilisation patterns of buffalo (Syncerus caffer), Uganda kob (Kobus kob thomasi)
and hippopotamus (Hippopotamus amphibius) in relation to the prevailing fire regime and other
anthropogenic disturbances, and (2) to provide baseline data for the elaboration of a fire management
plan.
The study area is dominated by grassland in the central parts and bush-land with variable densities of
bushes towards the water. Most of the study area is surrounded by water. Point surveys were conducted
to estimate the distribution and abundance of the focal species, using Distance sampling. Environmental
and anthropogenic factors were quantified by distances from each observation point using a
Geographic Information System (GIS). All geo-referenced data were subject to multiple regression
analyses. Regression models for several stages of vegetation development before and after burning
were used to evaluate the spatial and temporal relationship between different disturbance factors and
large herbivore abundance.
The effects of increasingly frequent fires over the last decades in the study area are the spread of fire
climax plant species such as Acacia sieberiana and shrubs of Acacia hockii, the spread of Imperata
cylindrica, a widely indigestible grass species, and species-poor overgrazed grasslands in the main
distribution area of Uganda kob.
The distribution of Uganda kob is generally restricted to open grasslands and near to water in the
North and East of the study area. Their movements in postfire-periods largely depend on the locality of
burnt areas. Fresh pasture after burns is highly attractive to Uganda kob in all stages of postfire-development.
However, adult males can suffer from food shortage after burns due to their strong attachment to their mating grounds (leks). Uganda kob does not show significant avoidance with regard to
human activities, villages and roads. They are mostly affected by the decrease of suitable pasture in
parts of the study area due to the encroachment of Imperata cylindrica. The data collected show that
density of Uganda kob has significantly reduced in the western part of the study area where they were
widely distributed two decades ago. In addition, the combination of frequent fires and heavy grazing
after burns led to a loss of high quality grass species in their main distribution areas.
The distribution patterns of buffalo are characterised by an accumulation of large herds in the central
grasslands of the study area, while smaller herds occupy bush-land closer to the water. In postfireperiods
buffaloes prefer tall grass savanna in the southern part of the study area. They utilise postfire
pasture in burnt areas only in advanced stages of vegetation development. Buffaloes are affected by
all quantified anthropogenic disturbances. They clearly avoid the vicinity of villages. The main road
through the park restricts the movements of buffalo herds. Adverse effects of the habitat fragmentation
become evident when home ranges of buffalo herds are burnt to a large extent. One of the outstanding
threats to buffalo is, however, competition with life-stock. Over the last decades buffaloes have
nearly disappeared in areas with live-stock occupancy. Furthermore, inter-specific competition is probable
between buffalo and hippopotamus in their feeding grounds along the water in postfire-periods.
Hippopotamuses are still abundant in large numbers in the study area. At present, they seem not to be
affected by anthropogenic disturbances, apart from poaching.
The equilibrium of population sizes of the mega-herbivores hippopotamus and elephant (Loxodonta
africana) is important for maintaining the savanna structure in the study area, and in consequence for
the viability of other herbivores. Since elephants are not abundant in high numbers, the spread of
bushes is one of the aspects to be monitored with respect to habitat availability (open grasslands) for
other large herbivores. Appropriate fire management strategies could regulate the grassland-bushland
mosaic in the study area.
Fire management is a tool to avoid adverse and amplify advantageous effects of fire on the habitats of
herbivores. Recommendations for fire management in terms of controlled burning and protection from
fires are given on the base of habitat classification that specifies temporal, spatial and functional aspects
of the habitat use of large herbivores in time and space. The dynamic interrelations of anthropogenic
and natural factors described may additionally contribute to adaptive strategies for the protection
of wildlife in QENP.