Ecology of Buffalo in the Okavango

Region: Okavango Delta, Botswana

The Okavango Delta is a unique ecosystem that supports a high diversity of wildlife, including many herbivorous species. The African buffalo (Syncerus caffer) is one of the largest herbivores present in the Okavango Delta, with males weighing up to 700 kg and females weighing up to 500 kg. Herbivorous species with a large body mass are more tolerant of low quality foods than those with a smaller body mass, and are consequently able to forage in a wider range of habitats.

However, the region is under threat, especially from water loss and the encroachment of human activity on its boundaries. There are plans to build dams in Angola and Namibia, where the water from the Okavango River would be diverted to towns and used for irrigation. This would diminish the amount of water flowing into the Delta system, causing a reduction in the area of floodplains and all the wildlife that they support.

The buffalo’s ability to consume low quality vegetation, together with the trampling effect that would ensue from a large herd of buffalo passing through a habitat, could provide a basis for the facilitation of other species of herbivore. The feeding habits of the buffalo may remove the taller grasses, as well as those of lower nutritional value, thus ‘opening’ the habitat for smaller herbivores by increasing access to shorter, higher quality forage. This phenomenon which has been observed in the Ngorongoro Crater has buffalo and zebra entering the drying grasslands before wildebeest and gazelles, which follow once the taller grasses have been grazed and trampled to a more convenient height.

The project aims to increase understanding of the ecology of the African buffalo in the Okavango Delta so that a conservation and management plan for the species can be developed.

Very little information is available on the behaviour and ecology of the buffalo in this specific region, yet they are likely to play a key role within the ecosystem. Defining this role will help to predict the impact of any potential changes to the delta, not only on the buffalo themselves, but also on the habitats that they use and on the sympatric herbivores that may benefit from their presence.

Objectives

• To monitor the demographics of a population of buffalo, including age and sex ratios, recruitment rates and body condition scores, in order to identify any trends and apply these to the entire population of the Okavango Delta.
• To map the monthly and seasonal home ranges of the study population, including any habitat preferences, and identify causes for these preferences.
• To determine the foraging patterns of the population in terms of the plant species consumed and the factors influencing their choice of forage species.
• To quantify effects that the feeding patterns shown by the buffalo may have on the quality and composition of the vegetation species in each habitat type, and to determine whether these effects could provide the basis for a facilitation interaction with sympatric herbivorous species.
• To investigate the problems with the buffalo cordon fence (details below), as seen by the local farmers, and to suggest feasible solutions for those problems.

Increasing our understanding of the ecology of the Okavango Delta can only help to conserve this unique and fragile ecosystem so that it can be enjoyed for many more years to come.

The first step in determining whether this type of interaction is taking place is to identify the ways in which buffalo modify the habitats that they use.

African buffalo have been studied in a few other sites across Africa; however information with regards to the buffalo population of the Okavango Delta is sparse. Based on previous observations, one would expect the buffalo to consume a reasonably wide range of grass species, concentrating on the most nutritionally viable ones whenever possible. However, when those species with high nutritional value become less abundant, for example during the dry season, buffalo tend to reduce their selectivity and consume higher quantities of lower quality plants to compensate.

Using various methods, including fitting nine individuals with GPS collars and observational information, the project will gather information on population demographics, home ranges, diet, and habitat modification.

Buffalo cordon fence
The study will also investigate the relationship between buffalo and the buffalo cordon fence that surrounds much of the Okavango Delta, which prevents the interaction between wildlife and domestic animals. Buffalo are carriers of numerous diseases that are also transmissible to cattle, such as foot and mouth disease and bovine tuberculosis. They graze similar areas and could present themselves as competitors for the forage available to cattle. The maintenance of the cordon fences themselves is apparently not satisfactory, with the result that elephants could damage them, allowing wildlife and cattle to mingle. This would cause problems with cattle depredation and the destruction of property by wildlife.

The GPS data obtained from the collared cows will be used to establish whether the herds approach the cordon fence. Areas where this is the case will be investigated to uncover any distinguishing features, such as weak points at the fence, or water or shade, which farmers may have provided for cattle. The species and nutritional quality of the vegetation found in such areas, together with the availability of such resources as water and shade, will be compared to those of their preferred habitats away from the cordon fence. This will also be useful in determining how buffalo would react to an increased human presence in the areas surrounding the Delta.

The farmers will be interviewed, with the help of a local to translate if necessary, to record their grievances with the fence, and what they suggest could be done to solve the problems. The farmers will be asked how they currently deal with any problems, whilst being aware that some measures may not be in accordance with protection laws. Stretches of the cordon fence will be examined at random to determine the average level of maintenance. This maintenance should be kept highest in areas where farmland and villages are located in close proximity to the fence.

May 2011
August 2010 was the last month in terms of fieldwork for this PhD study. By end August, all collars that could be located were removed. Two collars that had abruptly ceased to send GPS information or emit a VHF signal could not be found despite numerous attempts.

Water levels were extremely high for the majority of the year, which had a major impact on the study in terms of access, as well as on the buffalo themselves. None of the collared animals crossed the Gomoti Channel on their return from the wet season home ranges. They avoided secondary floodplains almost completely and spent very little time in tertiary floodplain. Presumably this was because most of the forage was under water.

I left Botswana in November 2010 to return to Bristol in order to process information, analyse data and write up the thesis, due to be submitted at the beginning of 2012. My last few months in Botswana were spent at the Harry Oppenheimer Okavango Research Centre (HOORC), now the Okavango Research Institute (ORI). Of 1 800 grass samples collected and separated, 540 were pseudo-randomly selected for nitrogen analysis. Between August and November, most of these were prepared for analysis but a variety of factors, including faulty equipment, lack of supplies and strikes, meant that no analysis was carried out. A lab assistant has been tasked with completing the analysis and sending through the results, which should be available by February 2011.

August 2010
Since the last report, all field data collection has been completed. The last vegetation sampling trip was at the beginning of August, and although some vegetation types were represented to a lesser extent than last year, I succeeded in meeting the quota of samples that I had set out for myself. Given that I have completed the fieldwork component of my work, I have been removing collars from buffalo cows.

The return migration from the wet season home ranges to the flood season home ranges took place late and slowly this year. The late rains soon before the cold winter months meant that the majority of the pans in mopane woodland still held water and a lot of the grasses still had a green flush. This could have reduced the urgency of triggers to return to floodplains, although there was still a definite shift between the seasons. The water held in the pans would be relatively stagnant and could be less attractive to the buffalo than the fresh running water available in floodplains and channels, and this could have been a factor in causing the move.

The high flood this year seems to have had an effect on the movements of the collared animals. There were several large herds of buffalo seen on a regular basis in the Santawani area, to the east of the Gomoti. None of the collared animals crossed the Gomoti, and I received reports of low numbers of sightings in the Chitabe area, as well as on the Gomoti itself. This could be explained by the lack of dry floodplain available to be grazed. I have not yet determined the habitat type for each of the fixes that were taken from the collars, but from the data that I have, secondary floodplain was almost not used at all, and tertiary floodplain was not used significantly. Indeed, I could not find sufficient locations in tertiary floodplain to warrant sampling that habitat type, as opposed to previous years, when buffalo herds spent approximately a third of their time in tertiary floodplain.

Body condition scores did not differ remarkably from previous years; this indicates that the floodplain grasses are not essential to maintaining condition during the early flooding season. They may play more of a role during the late flooding season, but the floodwaters should start to drop by then, exposing fresh grasses when those in the grasslands are at their least productive. However, given that buffalo previously selected tertiary floodplain to a similar extent to grasslands during the early flooding season, it is conceivable that there are some nutrients or minerals present in the floodplain grasses that are beneficial to buffalo. It would be very interesting to see whether predicted further rises in the water level in coming years could have a negative impact on the buffalo if floodplain grazing areas remain inaccessible to them.

For the next few months, until November 2010, the project will be moving into the laboratory phase at the Harry Oppenheimer Okavango Research Centre, where I will be analysing several hundred grass samples that have been separated into leaf and stem portions, as well as approximately 60 faecal samples to determine nitrogen content. This will give me an indication of the quality of grasses that buffalo are selecting for. This will be compared to the body condition scores from herds to identify any correlation between the two, as well as identifying any changes in diet quality according to season.

The rainy season was supposed to end in March, but clearly it didn’t know that. This has prevented the buffalo from returning to the floodplains, which is necessary in order for sampling to begin for this, the last season of fieldwork for this project. The next step is data analysis and thesis write-up.

Of the collars that are currently deployed, four are working well. Those that aren’t will be tracked down and replaced, if possible. All the collared cows have been in different areas, with no overlap between then. All information from satellite fixes has been compiled in order to estimate distances travelled by buffalo herds:

• Average hourly distance travelled was 322m (1 056 ft)
• Average distance travelled per 24 hours was 7724m (4.8 miles)
• Average daily distance travelled was 4402m (2.7 miles)
• Average nightly distance travelled was 3312m (2 miles)