26 October 2016

Koeberg Data Summary - The Patterns

One of the benefits to using camera traps to monitor wildlife is that it can be active for 24 hours per day, day in and day out. This makes it possible to start examining the day-night cycles and even moon phase activity patterns of animals. Below are some of the theories I came up with. (As always please take all of these "statistics" with the appropriate grain of salt.)

The "dune cliff" at Koeberg Nature Reserve

One of my favorite things to compare between species is the difference in their day-night activity patterns.

The closer you get to the poles, the greater is the seasonal difference between sunrise and sunset times. As a result when working with more than a year of camera trapping data you can't simply compare the observations based on the time alone. Instead you have to compensate for the difference in the length of the day.


I like to group the observations into categories, for example sunrise, early morning, mid day, etc. The general distribution still looks the same, but in this way I can more reliably compare observations over a longer period of time.


The Eland (Eland - Tragelaphus oryx), Common Duiker (Duiker - Sylvicapra grimmia) and Steenbok (Steenbok - Raphicerus campestris) all showed more or less the same activity pattern. Showing peaks of activity during early morning and late afternoon as well as an increase in activity around midnight. But there are small differences between the species, for example the Duiker seem much more active at night and the Steenbok seem to be more active at mid day. The Eland was also the only one to be more active in the late afternoon compared to the early morning.


However, the Springbok (Springbok - Antidorcas marsupialis) showed a different pattern compared to the other herbivores mentioned above. It lacked the spikes in activity around late afternoon and at mid night. Why this is I'm not sure. Maybe during the later part of the day the herd moved out of the areas where the cameras were placed, or maybe it is a reflection of seasonal changes in behavior. Most of my camera trapping in the area where the Springbok hang out happened in the autumn months.


The above species all showed both diurnal and nocturnal activity, but some species are very strict about their preference. As alwys the Porcupine (Ysterverk - Hystrix africaeaustralis) at Koeberg are strictly nocturnal.


The Porcupine even seem to go as far as to avoid the moon. The number of observations are heavily skewed towards times when the moon is not shining. This trend at Koeberg is consistent with my observations at Tygerberg were the Porcupine showed an almost identical pattern, being highly active during new moon and less active during full moon. Both locations also show far more observations during periods the no moon visible, but this latter effect is less pronounced in the Koeberg data compared to Tygerberg. My guess would be that because Koeberg is more remote and has less human activity at night, compared to Tygerberg, the animals are a little more comfortable to go about their nightly routine.


The polar opposite of the Porcupine must be the Small Grey Mongoose (Kleingrysmuishond - Herpestes pulverulentus) which is only active during the day.


The Small Grey Mongoose and Caracal (Rooikat - Caracal caracal) both prey on rodents, but it seems fairly safe to assume that they prey on different species, for the most part. The chart below is fairly busy, but I wanted to show how the activity periods of the predators at Koeberg overlap with their prey. The two Gerbil species are mostly exposed to predation from Caracal and Small-Spotted Genet (Kleinkolmuskeljaatkat - Genetta genetta), while the Four-Striped Grass Mouse (Streepmuis - Rhabdomys pumilio) primarily has to deal with the Small Grey Mongoose and a small amount of pressure in the late afternoon from the Caracal.


By far the majority of observations were of single animals, but there are some interesting exceptions. Porcupine families tend to travel together fairly regularly and as a result they have a much larger percentage of observations with two or more individuals.


For some species it is also possible to differentiate between the sexes. For both Steenbok and Duiker the males seem to get photographed more frequently. Maybe they patrol their territories more actively, or walk around more while looking for females? Interestingly the male-female sex ratio seems to be even more exaggerated for Duiker compared to the Steenbok.


The last theory I want to touch upon is one that originated from way back in March 2014 when I did a blog post about it, over here. I had a theory that the Duiker at Koeberg seem to adjust their activity pattern based on the season. The data seems to support the idea, showing that the Duiker are very active late at night during the warm summer months and much more active early in the morning during the cool winter months.


This past couple of posts analyzing the camera trap data from Koeberg is just the tip of the iceberg. I'm sure there are many more discoveries to be made in the future, but in the next couple of posts it is time to share some more camera trap photos and a new nature reserve.

19 October 2016

Koeberg Data Summary - The Maps

When I started camera trapping at Koeberg Nature Reserve I decided to try and cover as much area with my few (4-5) camera traps as possible.

Some scenery at Koeberg Nature Reserve

I decided to try and move each camera trap to a new location after 2 weeks.



With over a year of camera trapping data it is time to look at the maps.

First up is a heat map showing all observations plotted on the map. Just looking at this map the data may seem a little disappointing, with some areas clearly dominating the others in term of the number of observation records.


The map above can be misleading, especially when some camera trap locations were spaced close together or one camera trap recorded very high animal activity.

What if instead of just plotting each observation on the map, we try to compensate for the length of time the camera was active. The map below shows the areas on the map with the highest frequency of observations during each camera trap period. (The number of observations divided by the number of days the camera trap was active.)


I was still not happy with the two maps above and decided to make another map that to more accurately show the amount of effort (time) I put into camera trapping at a specific area. The map below ignores the observations, and instead just uses the number of days the camera trap was placed at a given location. (Using just the number of days the camera trap was active.)


I think it gives a much better indication of what areas were covered and what areas weren't.

I then wanted to see in what areas I recorded the highest species richness. The map below shows the number of species observed during the camera trap's active period. (The number of species divided by number of days the camera trap was active.)


I'm really pleased with how each of these maps highlight different areas of the reserve and tells a different story.

Another thing I wanted to try at Koeberg was to compare the maps of different species to determine whether some sort of rudimentary distribution map for the reserve can be created. Below are the distribution maps of common herbivores on the reserve. I'm very pleased with the results.

The Steenbok (Steenbok - Raphicerus campestris) were mostly found in the sandy dunes to the North.



The Common Duiker (Duiker - Sylvicapra grimmia) seldom ventured onto the open dunes and preferred the somewhat denser vegetation to the East and South.



The Eland (Eland - Tragelaphus oryx) ventured throughout the reserve, but had a preference for drinking at the dam in the North-West and the grassy area around the offices (which also has a small dam and is where they are fed once a week).



The Porcupine (Ystervark - Hystrix africaeaustralis) avoided the open sandy dunes all together and focused its attention on the thicker vegetation towards the East and South.



The other species listed above are natural to the area, but the Plains Zebra (Bontsebra - Equus quagga) are usually found much further to the North and East of the country. Interestingly they also have a very small distribution inside the reserve, preferring to stay on the grassy area around the offices. The extinct Quagga was closely related to the Plains Zebra and was historically found throughout the greater Western Cape province. The Plains Zebra at the reserve don't seem to move around much. Based on the behavior of these animals at Koeberg I think that either the Quagga would have specifically adapted to the region's conditions or it didn't occur in the reserve, instead preferring more grassy areas outside the reserve.



It really is amazing what can be done with only a hand full of camera traps on a small nature reserve if you are willing to persevere, keep the cameras in the field as much as possible and record the data in a structured manor. (I use WildLog, but any structured database or spreadsheet can do.)

Note: All these maps use Google Earth as the base layer. I cropped the images to make it easier to compare the different maps. Hopefully I won't get into too much trouble. :)

12 October 2016

Koeberg Data Summary - The Numbers

In the next three posts I'll be summarizing my camera trapping results form my time at Koeberg Nature Reserve. (You can click on the images to enlarge the charts.)

The white sand dunes at Koeberg Nature Reserve with Table Mountain in the background

I started camera trapping at the reserve during September 2013 until around November 2014. Earlier this year I returned to try and photograph the Leopard from May 2016 to June 2016.



A good balance of mammal and bird species showed up at the camera traps. The reptile and amphibian species were a pleasant surprise. I saw more snakes at Koeberg than any previous reserve I camera trapped at, but it seems that snakes are not easily camera trapped. The reptile species that showed up on the photographs were of Red-Sided Skink, Cape Skink and Angulate Tortoise.


Looking at the number of observations the mammals have slightly more total observations than the birds. Breaking it down and looking at the number of observations per species, we see that the mammals have a fairly good spread, but for the birds almost half of the camera trap observations were of Cape Francolin, and another quarter was of Cape Robin-Chat.


The species accumulation curve shows the usual trend: An initial ramp as new species are camera trapped for the first time, and then a steady decline as more time passes before encountering the next new species. The smaller "jumps" on the curve are usually due to one or more cameras being moved into a new "micro-habitat" such as a small pond or rodent nest. I'm surprised to see the strong leap in new species during 2016, after my long absence from the reserve. It seems like the reserve still has a lot to offer and new discoveries to be made. A new university study is currently underway and it would be very interesting to compare our results.


Not all camera trap locations are equal and some have a much higher species richness than others. Yet both the number of species per camera trap location and the total number of observations per camera trap location show the expected pattern: A few locations with lots of species / observations and many locations with fairly low amount of species / observations.


The duration of camera trap observations also reflects the expected trend. With a very high number of observations shorter than 10 seconds and then a steady decline. I suspect the small bump around the 1 minute mark is mostly due to the long delay between photographs on the Cuddeback cameras (also the odd time when I set the interval to more than 30 seconds on the Bushnell cameras). At the end of the chart I group all very long (7 minutes or more) observations together, which all add up to more or less the same amount of observations as the 10-20 seconds group.


I always enjoy interpreting the charts below. When a species visits a camera trap, what is the chance of that same species showing up again during the same day-night cycle?



These charts also nicely show the danger of interpreting data in this manner. From the charts the Honey Badger seems to be almost guaranteed to visit a camera trap more than once per day-night cycle, but this was only true for the 3 observations I recorded, and that is far too small a sample size to establish any meaningful trend. Another thing to note is that the Honey Badger observations were in fact "baited" by the presence of a carcass and does not reflect normal foraging behavior. In contrast a species such as the Duiker or Steenbok, with much larger sample sizes, will provide a much more accurate reflection of reality.

The rodent species usually top these charts, but there can be meaningful differences even for larger species like the Steenbok and Duiker.


The Bush Karoo Rat construct and live in large wooden lodges and thus its data pattern is very different to most other species.

One of the Bush Karoo Rat lodges at Koeberg Nature Reserve

As you'll see in the posts to come Steenbok and Duiker, in particular, provided a very good comparison - as both species have decently large sample sizes, but showcase very different behavior.