We are collecting data across the three ecological seasons (table 1; comprising cold / dry, hot / dry, and hot / wet) as opposed to the four traditional or meteorological seasons (namely spring, summer, autumn and winter). Ecologically speaking, a season is a period of the year in which only certain types of floral and animal events happen as opposed to meteorological seasons which are determined by temperature, with summer being the hottest quarter of the year and winter the coldest quarter of the year.
*(Supporters of the Roadkill Project who reside in the Northern Hemisphere must remember that the seasons in South Africa are in reverse ….. in case you are shuddering at the thought of January being called hot, when in the ‘North’ it most certainly is not!)
Table 1: Ecological seasons and the month to be sampled (source: Zambatis, 1987)
|ECOLOGICAL SEASON||RANGE||SAMPLING MONTH|
|Cold / Dry||May – August||July|
|Hot / Dry||September – December||October|
|Hot / Wet||January – April||February|
The most cost and time effective method of detecting roadkill for all species involved:
- Driving at a speed of ~40 km/h
- Starting the transect one-and-a-half hours after dawn
- Driving a transect distance of 100 km, over a 40-day period. This distance and time frame was selected based on the study counting multi-species and consequently the spread of diversity within these taxon. Birds were by far the biggest number of roadkill detected, and therefore need further distances to be driven and longer days spent on the road in order to represent the taxon as fully as possible
We also sub-sampled within the transect and drove an additional 20 km stretch of gravel road each day in addition to the 100 km stretch of tar road. This was to compare possible differences in roadkill numbers detected on gravel and tar roads. Further, night-time surveys were conducted starting one-and-a-half hours after sunset and for a distance of 20 km, as well as pre-dawn surveys starting one-and-a-half hours before sunrise. These additional transects were to determine when wildlife was being ‘hit’ and therefore how long the carcasses remained on the road (figures 1, 2 & 3).
Figures 1, 2 & 3: Black-backed Jackal roadkill carcass showing decomposition rates over a 3-week period (March 2011)
Having completed two of the three ecological seasons of sampling, I am now busy analysing the data. In figure 4 and table 2, a comparison of the two seasons shows birds being by far the most impacted species for both seasons. On the other hand, mammal roadkill was greater during the hot / dry season than during the hot /wet, with the opposite being found for reptile roadkill which was higher during the hot / wet season (figure 4 and table 2).
Figure 4: Roadkill detected over a 40-day period on a 100 km transect across two ecological seasons
Table 2: Number of roadkill detected across two ecological seasons
I am currently looking at the rainfall data for the last two seasons to see if there is a relationship between rainfall and roadkill. Unfortunately, despite this last season being the ‘hot /wet’ season, we had little rain compared to the rest of the country. However, I do have data collected over the last two years which I am able to compare ‘hot /wet’ data from 2012 against 2010 and 2011 (figure 5 and table 3).
Figure5: Percentage of roadkill detected over a three-year period during the hot /wet season (January – April)
Table 3: Number of roadkill detected over a three-year period
There are many other variables that may cause roadkill besides rainfall which I shall also be examining, such as; animal movement on roads in relation to the cycle of the moon, behaviour of key species during certain seasons, for example, the Scrub Hare was a noticeable victim of roadkill during the hot / dry season and less so during the hot / wet. This is possibly due to the hot /dry season also being the breeding season of Scrub Hares (figure 6) and are therefore more active.
Figure 6: Scrub Hare
Additionally, Flap-neck Chameleon have been noticeable more visible on the roads during the 2011 and 2012 hot / wet season, again, possibly due to it being their breeding season (figures 7 & 8).
Figures 7 & 8: Flap-neck Chameleon
I am also recording data on fence type where the roadkill was detected, such as, was it a game fence, electric fence or cattle fence, as well as the type of habitat adjacent to the road (figures 9, 10 & 11) and the height and thickness of the grass (figures 12, 13 & 14).
Figures 9, 10 &11: Photographic examples of different habitats along the transect; dense mixed bushveld, open grassland and koppie outcrops
Figures 12, 13 & 14: Photographic examples of different grass heights and thickness, from n0-grass to partial-grass to very thick grass.
These variables are all probably contributors to wildlife roadkill, but whether it is just one of these variables that is the main cause or a combination will be determined once the cold / dry season is completed in July.
In addition to all the surrounding variables being recorded at each roadkill site, a GPS reading is also taken. Figure 15 shows a simple scatter plot of each GPS roadkill point; one can already start to notice certain hotspot areas, areas that are more sparse for roadkill, and others where no roadkill was detected over the 40 days. The non-roadkill points are of particular interest as it raises the question of why no roadkill was found at that site during that particular season?
Figure 15: Roadkill points detected along transect (100 km tar road / 20 km dirt road)
Another interesting question is the number of roadkill that disappears by the next day. Many roadkill will be naturally scavenged by roadside predators such as Brown Hyena, Chacma Baboon, Birds of Prey, often at risk of becoming roadkill themselves. Whilst driving my transect one morning, I watched a Martial Eagle swop down from his telephone-pole-perch and snatch up a Scrub Hare roadkill carcass from the road (figure 16).
Figure 16: Martial Eagle
Of the 470 roadkill detected during the hot / wet season, 383 of them had disappeared by the next morning (figure 17). From a questionnaire conducted in February 2012 at a local petrol station, drivers were asked if they would ever remove roadkill from the road and for what reason.
Figure 17: Number of roadkill detected for each taxon showing number disappeared by next day (n = 383/470)
Of the 129 respondents, 72% said they would remove a dead animal from the road for the safety of other drivers, 15% admitted removing roadkill for food whilst only 1 of the 129 respondents admitted to removing roadkill for muthi (traditional medicine). 61% of the respondents said they had considered taking roadkill but were afraid of the repercussions from the law.
Interestingly, many of the recipients were unable to identify the roadkill to species level, unless it was a large mammal, and didn’t consider snakes and birds to be roadkill.
There is a vast amount of data that has been collected and whilst some of the results have been shocking, I am incredibly excited about trying to understand why roadkill happens and what can be done.
If you’re not already shocked by the what the data has shown so far, then table 4 gives a breakdown of each species and how many were found over the 40 days!
And to finish ……. In addition to measuring roadkill rates, we are also monitoring traffic volume, speed and weight. Table 5 gives a brief summary of traffic data during the hot / wet season. What is most concerning is the number of vehicles exceeding the national speed limit in an area that is rich in wildlife!
Table 5: Summary of traffic volume, speed and class over a 2-day period of the transects.
Table 6: classification scheme for traffic data
Thanks to the following for assistance with species identification: Andrew Rae, Duncan MacFadyen, Dan Parker, Ric Bernard, Jeanne Tarrant, Andre de Kock, Dave Pugh and Ian Little.
Thank you to Mike Rae, Adrian Stander and Engen Garage, Musina for assistance with the roadkill questionnaires.