project updates
Ecology of Buffalo in the Okavango
March 2010
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)
For the full report, with maps showing movement patterns, click here.
December 2009
The current population of African buffalo in the Okavango Delta of Botswana is estimated to be 30 000 animals in herds varying from fewer than 50 individuals up to reported herds 3 000 strong. These numbers, together with the large body mass of individuals - up to 700kg for adult males and 500kg for adult females - suggest that buffalo play an important role in the ecosystem of the Delta.
Data pertaining to the following is being collected and analysed:
1. Population Demographics
2. Habitat Use and Home Ranges
3. Nutritional Requirements
4. Movement Patterns at Different Scales
5. Triggers for Long-Distance Movements
Discussion
The Okavango Delta is a unique system that is constantly changing, especially in terms of forage and water availability. Preliminary results suggest that these two factors are some of the most important in terms of influencing movements and behaviour of African buffalo. Buffalo require water on a daily basis to survive, but extensive flooding can reduce the area of available grazing. Understanding the mechanisms and causes of buffalo movements will broaden knowledge of the Delta ecosystem. The results of this study could therefore be applied to other species living in the Delta and help to predict the impact of future changes, such as those caused by climate change, on the species living in the Delta. The African buffalo is being used as a model species to gather information that can be applied to a broad range of species.
For the very detailed and comprehensive report click here.
October 2009
Recent reports have been receieved of large herds of buffalo from pilots in areas all across the Delta, so pilots from charter companies will be asked to record all herds that they see within a week, as a way to estimate the most frequent herd size. This was done last year, so it will be interesting to see whether there are any changes this year, given that the water levels on the Gomoti are still very high compared to last year.
GPS and habitat data for the animals that were collared for a year have been compiled. When comparing the varying proportion of habitat use across the seasons, several points stood out. Herds on the Stanley's and Gomoti side appear to use habitat types to different extents during the wet season. One main factor is the absence of the Low Mixed Woodland (LMW) habitat from the Stanley's area. This habitat type is used extensively by the Gomoti herds, whereas the Stanley's herds must turn elsewhere for their forage, likely using grassland habitats more (see Fig 6). Proportional use of Open Mopane (OMP) and Dense Mopane (DMP) seem to be relatively similar in both areas.
It is, however, difficult to ascertain how prevalent the preference for certain habitat types is without determining their availability. This is the next step for the research project. Either the vegetation types must be mapped, or thousands of randomly generated points must be plotted into Google Earth, recording the habitat type of each.
To read the full report click here
July 2009
For the majority of the early flood season (April - July), I had only two satellite collars working, B76 along the Gomoti and B78 along the Boro. They were collared in separate herds, well within my study area, but have since joined up and walked almost continuously, average 4-5 km per day. They are currently 51km from where they were collared, very close to Chief's Camp. They walked up the north-eastern edge of Chief's Island, and are now reaching Mombo Island. This is not behaviour that I have seen outside of the migration at the beginning of the wet season, although this movement is slower than the migration.
To read the full report click here
April 2009
By the end of March I have managed to sample all of the sites that I aimed for during the wet season, a total of 120 sites. I collected 180 faecal samples from 18 herds throughout the season and was able to collect population dynamics data from over 1 300 animals.
I was not able to collect as much information on population dynamics as during the last season, due to reduced visibility of the herds as a result of the denser habitats that they have been using. Most of the time I was unable to obtain an accurate estimate of the herd size, and was rarely able to see more than 50 animals during a sighting.
The fixes obtained by all of the satellite collars during the wet season have helped determine the habitat type and activity for 5 174 fixes. On average, the buffalo herds spent 38.3% of their time in dense mopane, 26.0% of their time in grassland, 18.8% of their time in low mixed woodland, and 12.0% of their time in open mopane. There are other habitat types that have been used, including open acacia and riparian woodland. Some GPS fixes were also taken in pans and on roads. The herds spent a maximum of 2% of their time in each of these habitat types, so they were left out of the sampling regime and analysis.
The habitat types used by the herds during the wet season are completely different to those used during the early and late flood, when the herds spent their time in areas dominated by floodplains.
In terms of activity patterns, herds spent 63.9% of their time grazing, 30.0% of their time resting and ruminating, and 6.1% of their time walking.
In contrast to last year, the herds on the eastern side of the study area spent very little time in NG43, mainly occupying NG34, with a couple of weeks spent in NG43 and in the Moremi Game Reserve.
B77 did not spend any time in NG43, but did go up into the Moremi Game Reserve. B73 and B74 both spent some weeks close to Santawani Gate but did not use areas north of South Gate. In a similar way to during the flooding seasons, the herds are using slightly different, albeit overlapping, areas. Last year, B75 spent the majority of the wet season in NG17 and the lower part of Chief's Island, but this year she stayed in NG32, between the Santantadibe and the Gomoti, for the entire season.
B78 spent all of the wet season in NG32, between the Boro and the Santantadibe. There was no overlap between B78 and B75, nor between either of these and any other satellite collared herd. However, one of the GPS collars, B7, has been in the same area as B78 for the entire wet season, although not always in the same herd.
January 2009
While several collars ceased to work over 2008, the remainder allowed some interesting information to emerge. Seasonal shifts in movement of buffalo were observed in December and in April, with herds spending the wet season mainly in mopane woodland and the rest of the year in floodplain systems. From the beginning of the late flood season, all the herds used floodplain and island systems more than other habitat types. Based on behavioural observations of herds, it has become apparent that buffalo do not spend more than one hour in a location while grazing. The grazing pressure from the rest of the herd could be pushing animals to move on to fresher patches.
October 2008
Over the last three months, the four collars that are on buffalo herds have been functioning well, sending information regularly. For the purposes of this research, seasons have been redefined so that they fit better with the movements of buffalo and the water cycles of the Okavango Delta. Seasons have been defined as Wet (December - March), Early Flood (April - July) and Late Flood (August - November). Seasonal movements of buffalo have been observed in December and in April, with herds spending the wet season in mainly mopane woodland and the rest of the year in floodplain systems. All of the herds have been using floodplain and island systems over any other habitat types since the beginning of the Late Flood season.
3 198 GPS fixes from all four collared animals have been used to determine patterns of habitat use. The habitat types that have been used most have been tertiary floodplain (34.66%) and grassland (34.07%), followed by riparian woodland (17.99%) and lastly secondary floodplain (13.28%). Google Earth has been used to identify which habitat type each GPS fix was taken in. This method has been verified by ground truthing of GPS locations in each habitat type.
After behavioural observations of herds, it has become apparent that they do not spend more than one hour in a location while grazing. The grazing pressure from the rest of the herd could be pushing animals to move on to fresher patches. In some areas, buffalo have been observed to walk straight through areas with minimal grazing, which is equivalent to avoiding grazing in those areas. The distance between each recorded GPS fix along a path has been determined using ArcView 3.3 and the Path, with Distances and Bearings v. 3.2b extension developed by Jeff Jenness. The distance between fixes has been related to the activity of the collar-bearing animal. Distances of under 50m per hour correspond to resting or ruminating, distances between 50m and 750m per hour correspond to grazing, and distances of over 750m per hour correspond to walking (avoiding).
Buffalo herds have spent the majority of their time grazing (62.13%). They have spent 33.22% of their time resting and ruminating and 4.65% of their time walking.
Although the herds have used sites within overlapping home ranges, each of these herds has spent most of its time in one or two areas, which are not overlapping. Sightings and counts of these herds have shown that they are not always the same size, ranging from fewer than 100 individuals to more than 1 000 animals.
Nine more GPS-enabled collars have been refurbished and are now ready to be fitted to buffalo cows. Two of these collars are the same as the ones that are already on, with satellite capabilities. The other seven will be used as store-on-board collars, taking fixes to the same schedule as the others (one GPS fix per hour). These data will be retrieved at the end of the study for analysis. These collars will provide a greater insight into the movements of herds in the study area. The VHF component will enable the collars to be tracked regularly from the air and from the ground so that herd size and composition can be monitored.
July 2008
Two out of the six collars have stopped working, one having dropped off the animal's neck, and the other a technical malfunction. Both have been sent back to Sweden for repair. The other four are functioning well.
Throughout the wet season, all of the herds spent most of their time in areas dominated by mopane woodland. There they were feeding on the grasses that came through during the rains. These species included Urochloa trichopus, Digitaria eriantha, Eragrostis rigidior, Panacum maximum and Cenchrus ciliaris. Although the majority of their time was spent in mopane dominated areas, the buffalo herds grazed mainly in small open areas with the mopane. They rested under the trees during the heat of the day. They walked straight through certain grassland areas with tall grasses densely packed together. This could be because they could not select the highest quality species in those habitat types and preferred areas with more spread out grasses. The majority of their time was spent within 1km of water, usually in the form of pans within the mopane areas.
At the end of the wet season, the herds moved out of mopane habitats and into areas dominated by floodplains close to channels. They are spending a lot of time on the edges of islands in semi-flooded areas. These locations are still close to water, although the water is in channels and floodplains as opposed to pans. The main habitat types that are being used now are secondary floodplain, tertiary floodplains, riparian woodlands and grasslands in the interior of islands.
Herd sizes were assessed several times and do not appear to be constant. The fission-fusion society observed during the wet season is still valid now. A variety of herd sizes have been observed and recorded, ranging from fewer than 50 to more than 1 000. In other systems, it has been suggested that the population comes together at certain times of the year into a few large herds. This could happen in a different season, but has not yet been observed. Most sightings have been of herds between 100 and 400 animals.
Age/sex ratios have been estimated for these herds. There are approximately 0.74 males for every female, which is a similar figure to ratios found in other ecosystems. The ratio of young (juveniles and calves) to adult females is of 0.33 calves to every adult female. This may change, and will be reassessed at the end of this season.
Over the past three months, research has been somewhat hampered by the unusually high rainfall, particularly in January. However, by February some areas had dried out sufficiently. The herd in which each collared individual was found was followed for 48 hours, and observed visually during the day and using GPS location at night. Age, sex and body condition of each member of the herd was recorded.
In NG34, most of the vegetation consisted of open grasslands dominated by tall grasses and Kalahari apple leaf shrubs, tall open mopane and very dense stands of shrub mopane. Some of these dense stands were quite small and separated by small open areas. This type of habitat seemed to be favoured by the herds for resting in the shade during the hottest hours of the day. The buffalo appeared to merely walk through the grasslands, occasionally stopping to eat a mouthful or two of grass, but they spent more time grazing in the open mopane areas, which was also the habitat type where the probability of finding water was highest.
The buffalo were expected to spend some time predominantly grazing, with little directional movement. However, from the data collected, the herds seem to graze as they walk. Although the dense vegetation prevented the entire herd from being visible so that it was not possible to obtain a head count, the high density of tracks and scat indicated that the herds being followed were at least several hundred strong.
In the next few months, more collars will be put onto buffalo cows; there are a further eight GPS-enabled collars, these with UHF remote download capability, that will be put onto eight more cows in these two areas to increase the sample size.
December 2007
Six adult female buffalo were darted and collared, providing updated information about the animals' movements. Based on this, it is possible to see that there is a marked difference in the movement patterns of the cows collared in different areas. It appears from the GPS data that the cows collared in one area are approaching the veterinary fence, which could be preventing them from following migration routes. Alternatively, the forage available in this area could be of a much higher quality than that in the Moremi area, therefore sustaining the herds for longer.
