Detecting wildlife with camera traps

I am researching how we can best use camera traps to monitor foxes (Vulpes vulpes) in the mallee. I don’t know much about camera traps yet, so I decided to learn more about best practice in detecting wildlife with camera traps. Here’s what I found.

Camera traps are a useful monitoring tool that has become increasingly popular in wildlife monitoring over the past decade. They are relatively new compared to more traditional tools and techniques, but, just as with any other tool, you want to be using it the right way. What affects camera trap performance in wildlife surveys (particularly monitoring change in fox abundance) and how should they best be set up?

Camera specifications
Each different camera brand and type has different qualities that should be taken into account when designing a wildlife survey, because this defines what type of camera you will need. Some qualities are inherent to the camera (such as trigger speed), and some can be chosen when setting up the camera (such as delay).

Trigger speed
Trigger speed affects how quickly a picture is taken after detection. When a trigger speed is too slow, the animal might have left the field of view (the area captured in the photograph) before a picture is taken. However, if a trigger is too quick, it is possible that the animal has not entered the field of view of the camera, but only the detection zone (see below). Trigger speeds of 0.2–2.1 s have been found to be a good range for mammals  (Glen et al. 2013). Trigger speeds can usually not be chosen and are dependent on the camera type and brand.

Detection zone
This is zone in which the infrared sensor detects animal movement and causes the camera to be triggered and take a picture. This zone is important because it affects how likely the camera is to detect an animal. The detection zone is not always the same as the field of view of the camera and both aspects differ per brand and camera type.

Kangaroo walking into field of view

Just walking into the field of view (left).

Picture quality
The quality of the pictures taken is also important. The quality necessary depends on the type of survey. When individual recognition of animals in necessary, quality should generally be high and a white flash is needed to take pictures at night. When only species identification is necessary, infrared flash during night time is usually sufficient. When monitoring a species that is difficult to identify, it can be helpful set the camera to take multiple pictures when triggered.

Delay
A delay period can be set for a camera trap, preventing it from taking pictures for a certain time after it was triggered. This reduces the chance of detecting the same individual multiple times and saves memory.

Technical failures
Cameras can have technical failures sometimes, so make sure that enough camera traps are set up in the field to account for this. Replacement costs of camera traps and budget may influence how well this can be accounted for.

Setting up the camera trap
When setting up a camera trap in the field, it is important to have the right field of view. This means that cameras have to be set up at the right height and angle. It is important that field of view is standardised across all camera traps, so capture rates are comparable across space and time (Glen et al. 2013).

To be able to detect foxes, they will have to be set up at the right height. A rule of thumb is that the height of the camera should be similar to the core mass of the animal you want to detect (Meek et al. 2012). It is also important that the camera trap is set up the right angle. Facing downwards is preferred for small animals, but for larger ones outward facing is more often recommended.

When setting up the camera traps, vegetation that can move in the wind nearby the camera trap should be removed. Also, the camera trap should be attached to post, not a tree, as these tend to move in the wind as well. Moving vegetation and trees can cause false triggers, causing the camera to take empty pictures (no animals in it). These take up a lot of memory and increase the time it takes to sort all the pictures afterwards.

Environmental factors
Sunshine
False triggers can occur in the morning as the sun rises and starts to warm sunspots and vegetation. False triggers can also occur where the sun shining directly on the face of the camera. To reduce glare, it is a good idea to face the camera southerly.

Temperature
Most camera traps use passive infrared sensors. This means that the camera will detect heat and motion and take photos when it detects a difference between air temperature and the animal body temperature. Body heat can be masked by background heat on hot days. When ambient temperatures range between 35.7 and 41.7⁰ C, it can become difficult for the camera traps to detect foxes because the difference between body heat and ambient heat is too small. Setting the sensitivity of infrared sensor to high in summer is therefore recommended. Low sensitivity is recommended during cooler winter months.

Detection of camera traps by foxes
Camera traps emit light and sound which can be detected by a range of species. This could affect their behaviour in camera traps surveys. It has been found that foxes can easily hear some of the infra and ultra sounds emitted by camera traps as well as flash illumination and possible infrared flash lights (Meek et al. 2014).

Fox looking towards camera trap

Well hello there!

Literature and further reading
(Most information in this blog post is based on “An introduction to camera trapping for wildlife surveys”)

  • Glen, A. S., Cockburn, S., Nichols, M., Ekanayake, J., & Warburton, B. (2013). Optimising Camera Traps for Monitoring Small Mammals. PLoS ONE, 8(6), 1–8.
  • Meek, P., Ballard, G., & Fleming, P. J. S. (2012). An introduction to camera trapping for wildlife surveys in Australia, (p. 94). Invasive Animals Cooperative Research Centre.
  • Meek, P. D., Ballard, G.-A., Fleming, P. J. S., Schaefer, M., Williams, W., & Falzon, G. (2014). Camera Traps Can Be Heard and Seen by Animals. PLoS ONE, 9(10), e110832.
  • Meek, P., Fleming, P., Ballard, G., Banks, P., Claridge, A., Sanderson, J., & Swann, D. (2014). Camera Trapping: Wildlife Management and Research (p. 392). CSIRO.
  • O’Connel, A. F., Nichols, J. D., & Karanth, K. U. (2011). Camera Traps in Animal Ecology: Methods and Analyses. In Camera Traps in Animal Ecology: Methods and Analyses (pp. 27–43).
Advertisements

Monitoring mammal populations with camera traps

I am researching how we can best use camera traps to monitor foxes (Vulpes vulpes) in the mallee. I don’t know much about camera traps yet, so I decided to read how others have used camera traps to monitor mammal populations. Here’s what I found.

What are camera traps?
Camera traps are used to take picture of wildlife without needing presence of a human. Camera traps consist of a motion sensor linked to a camera that takes a picture (or multiple pictures, depending on the settings and camera type) when the motion sensor is triggered (e.g. Meek et al. 2012).

A typical camera trap

A typical camera trap

Benefits of camera traps vs. other monitoring tools
Because there is no presence of a human required when using camera traps, they are very effective in monitoring wildlife. Shy wildlife will not be scared away, and labour costs will be significantly reduced when monitoring with camera traps, as less people have to go into the field less often compared to other monitoring techniques (Engeman & Witmer 2000, Silveira et al. 2003).

What can we do with camera traps?
Camera traps are mainly used for three goals: hunting, wildlife photography and ecological surveys. The use of camera traps for hunting, where cameras are set up to scout prey, has led to a strong growth of the camera trap market, with variety of different camera types and qualities now available.

Use of camera traps for wildlife photography was done as early as the 1880s and nowadays still used to produce stunning pictures. National geographic wrote an interesting piece about it which you can read here. For stunning pictures see here.

Bobcats on camera trap

Bobcats in Bosque del Apache, New Mexico, US – camera trap photo (cropped) taken with a Moultrie Game Camera by J.N. Stuart via Flickr

The use of camera traps to survey wildlife has strongly increased over the past decades. These are the four main ways in which camera are used for ecological purposes:

  1. to map occupancy of a species in an area;
  2. to monitor how many different species there are in an area;
  3. to prove that a species is present;
  4. to monitor a species and estimate (changes in) their abundance.

My research will be focusing on option 4.

Using camera traps to monitor species abundance
How can we use camera traps to estimate species abundance? This depends on the type of species. Camera traps will need to be set up in different angles depending on the size of the species that will be monitored.

Depending on how the camera traps are set up, different types of abundance estimates can be obtained. Often activity or density are estimated (e.g. Bengsen et al. 2011, Rowcliffe et al. 2014), but sometimes researchers look at bait uptake (how often bait was taken from a bait station set up in front of the camera) as an indication of abundance (Hegglin et al. 2004).

Most camera traps are motion-triggered, but there is also the possiblity to set up time-triggered cameras. Hamel et al. (2014) have found that time-triggered cameras can provide smaller error rates in daily presence of species than motion-triggered cameras.

Camera traps can be set up in the field randomly, but some researchers have found that this is not the most effective way to gain abundance estimates (Guthlin et al. 2014). Other approaches are setting up the camera along known tracks or unpaved roads. Furthermore, some researchers have chosen to place lures or baits near the camera traps (Bengsen et al. 2011). All these choices affect the way the data will have to be analysed later on.

Data analysis
Most data analysis is done based on capture-mark-recapture methods, where individual recognition is necessary. This method is particularly effective when looking at species that prevail at low densities (Foster & Harmsen 2012), where each individual can be recognised, such as tigers. The main benefit of using individual recognition is that one will have certainty about whether the same animal triggered a camera several times instead of it being several animals. A disadvantage of using individual recognition is that a picture of the whole animal or even both sides of the animal are necessary, which means that camera traps tigger time should be set to trigger at exactly the right moment and should produce high quality pictures (Meek et al. 2012).

More recently researchers have started to develop abundance estimation methods for which no individual recognition is necessary. For these methods, less accurate pictures are necessary, as only species recognition is important. However, this methods also introduces uncertainty, as it is no longer possible to see if the same animal triggered a camera several times. The only methods that I have been able to find so far that allows for abudance estimation without the need for individual recognition is by Ratcliffe et al. (2008). This method is based on the assumption that animals move according to a gas model.

How to monitor foxes with camera traps?
So far I have been able to identify three ways to monitor foxes and estimate their abudance.

  1. Individual recognition has been used as a way to monitor fox abundance in foxes, but this method is rather time consuming and difficult when multiple people analyse the photos, especially when animal densities are high (e.g. Sarmento et al. 2009).
  2. Others have looked at the bait uptake by foxes by setting up camera traps near bait stations (Hegglin et al. 2004).
  3. Rowcliffe et al.’s paper on estimating abundance without the need for individual recognition might be an intersting alternative, especially if large amounts of picture need to be processed and this will be done by different people (Rowcliffe et al. 2008).

That’s it for now. If you’ve found that something I stated is incorrect, please let me know! My next step will be to sort out which will be the best method for monitoring foxes as part of the malleefowl conservation project. Stay tuned!

Literature

  • Bengsen et al. (2011) Estimating and indexing feral cat population abundances using camera traps, Wildlife Research, 38:8, 732-739
  • Engeman & Witmer (2000), IPM strategies: indexing difficult to monitor populations of pest species, In: Salmon, T. P., Crabb, A.C., (Eds.), Proceedings of the 19th Vertebrate Pest Conference. University of California, Davis, pp. 183-189.
  • Guthlin et al. (2014)Toward reliable estimates of abundance: comparing index methods to assess the abundance of a Mammalian predator, PloS one, 9:4, e94537
  • Hamel et al. (2014) Towards good practice guidance in using camera-traps in ecology: influence of sampling design on validity of ecological inferences, Methods in Ecology and Evolution, 4:2, 105-113
  • Hegglin et al. (2004) Baiting Red Foxes in an Urban Area: a Camera Trap Study, Journal of Wildlife Management, 68:4, 1010-1017
  • Meek et al. (2012) An introduction to camera trapping for wildlife surveys in Australia, Vertebrate Pest Research Unit
    NSW Department of Primary Industries
  • Rowcliffe et al. (2008) Estimating animal density using camera traps without the need for individual recognition, Journal of Applied Ecology, 45, 1228-1236
  • Rowcliffe et al. (2014) Quantifying levels of animal activity using camera-trap data, Methods in Ecology and Evolution, 5, 1170-1179
  • Sarmento et al. (2009) Evaluation of Camera Trapping for Estimating Red Fox Abundance, Journal of Wildlife Management, 2009, 73:7, 1207-1212
  • Silveira et al. (2003) Camera trap, line transect census and track surveys: a comparative evaluation, Biological Conservation, 114:3, 351-355

Fox baiting for beginners

FoxOff warning sign

A warning sign that 1080 poison is being used, at Weddin, NSW. Image courtesy: Brian Yap via Flickr

I am researching how we can best use camera traps to monitor foxes (Vulpes vulpes) in the mallee. One of the main steps in my research will involve the simulation of fox abundance under influence of different fox baiting intensities. I don’t know much about fox baiting, so I decided to do a little bit of a literature review. Here’s what I found.

Fox densities
First of all, how many foxes are there? Red foxes are highly adaptable and are found almost anywhere in the Northen Hemisphere (and Australia). Their abundance is restricted by food availablility. When food is superabundant, densities can be as high as 30 foxes per km2. In farmland one family can occupy 1 km2, whereas in suburbs this can range from 0.2-5 families per km2. In barren uplands, density can be as a low as one family per 10 km2 (IUCN, 2008). In Central Victoria (Australia), the average fox abundance is 4 foxes per km2 on farmland (DEPI).

Fox baiting
I found that a rule of thumb is to place baits at the same or a slightly higher density than fox density (information sheets on fox baiting). According to the information above, we would need 4-5 baits per km2. However, I also learned that just placing fox baits in the landscape does not automatically lead to a lasting reduction in fox abundance (Bengsen 2014). What can go wrong?

Home range
First of all, foxes need to find the bait. For that the happen, the bait needs to be in the fox’s home range at a place where it is likely to come across it. Fun fact: the average home range of a fox is around 12 km2, but can vary widely and can overlap a lot, especially in sub-adults. Each night, foxes travel 9.4 km on average within their home range (Carter et al. 2012). This variation does not make it any easier when trying to figure out where to place baits. A general guideline is to place baits at 500 – 1000 m intervals (or bait density should be slighly higher than fox density) along known tracks and has been used in intensive baiting programs (Claridge et al. 2010).

To eat or to cache
There are different types of bait. The main distinction that is made is between fresh or dry bait. Fresh bait is often some type of meat injected with 1080 poison. Dry bait can be commercially manufactured cubes (FoxOff) that contain 1080 poison. Australian foxes prefer fresh bait (deep fried beef liver) over commercially manufactured bait (Petel et al. 2001). Chances are much, much higher that a fox will immediately eat fresh bait than dry, commercially manufactured bait. The fox will likely cache the latter and dig it up later when food abundance is low. However, when trying to choose the bait for a baiting campaign there is a trade-off, because dry bait has a much longer shelf-life than fresh bait. Instruction flyers on fox baiting state that what bait should be used is highly dependend on the area where baiting will be done.

What if another species eats the bait?
There is a small chance that other species will eat the bait before a fox finds it. To decrease the chance that this happens, baits have to be buried at certain depth defined by the state where the baiting happens (usually about 10cm). The poison in baits, sodium fluoracetate or 1080, is found in Australian plants. Therefore, native species have a higher tolerance to 1080, especially on the west coast where plants with 1080 are more abundant. Most baits contain 3 mg 1080, which is highly lethal to an adult fox but usually not to native species (see table below for information on lethal dose for different species). Dogs, however, are extremely sensitive and therefore signs should always be put up in areas where 1080 is used to warn dog owners. I have not been able to find any indications of numbers on how many baits are taken by other species than foxes, so if you’ve got any tips please let me know.

LD1080

Lethal dose of 1080 for different native and introduced species. 1080 is not cumulative. Source: Animal Control Technologies FoxOff brochure

Territories
When a fox dies, his territory becomes available to nearby foxes. This is why a once-off baiting campaign doesn’t work: neighbouring foxes will quickly move into the newly available territory. To avoid this, fox baiting has to be done at high intensity for a long period of time in an extensive, coherent area (Bengsen 2014). Fox control is most effective during late winter and spring, because foxes are rearing young. Therefore they don’t change territories much and at the same time have high food demands. Delaying reinfestation is much more difficult in autumn, when young foxes have left their den and are on the move to find their own territory (DAFWA).

How much does fox abundance decrease?
DAFWA states on their website that 5 baits/km2 will kill at least 80% of the foxes and increased baiting density does not increase this percentage further. Thompson and Fleming (1994) found a mean population reduction of 69.5% when they used a mean bait density of 12 per km2. Thomson et al. 2000 found an estimated population reduction of >95% with a baiting density of 5 per km2. As you can see, results of baiting campaigns vary widely and I have not yet been able to find a clear overview of the relationship between fox baiting and fox density. This may be difficult to find, as every baiting campaign is different and population reduction is not always studied.

Fox eating squirrel

Red Fox with an Arctic Ground Squirrel, Yukon Territory, Canada. Image courtesy: Keith Williams via Flickr

It looks like fox baiting is not as straightforward as I thought. The effect of baiting is influenced by density of baits (should be about 5 km2), the scale of the campaign (a large scale connected area, the larger the better to reduce reinfestation), the season in which is baited (during breading season is best), what type of bait is used and so on. I did not get the impression that there a ‘one size fits’ all baiting campaign, nor are there any clear indications on how each aspect of the baiting program will affect the outcome. If I am going to simulate fox densities under different baiting intensities, I may just have to assume stuff without knowing exactly what is going on. But who knows, maybe I missed something. Let me know if you’ve got any useful insights!

Literature

  • Bengsen 2014, Effects of coordinated poison-baiting programs on survival and abundance in two red fox populations, Wildlife Research, 41, 194-202
  • Carter et al. (2012) Fox-baiting in agricultural landscapes in south-eastern Australia: a case-study appraisal and suggestions for improvement, Ecological Managament & Restoration, 12:3, 214-223
  • Claridge et al. (2010), Trends in the activity levels of forest-dwelling vertebrate fauna against a background of intensive baiting for foxes, Forest Ecology and Management, 260, 822-832
  • Petel et al. 2001, Bait palatability influences the caching behaviour of the red fox (Vulpes vulpes), Wildlife Research, 2001, 28, 395–401
  • Thompson and Fleming (1994) Evaluation of the Efficacy of 1080 Poisoning of Red Foxes using Visitation to Non-toxic Baits as an Index of Fox Abundance, Wildlife Research 21, 27-39
  • Thomson et al. (2000). The effectiveness of a large-scale baiting campaign and an evaluation of a buffer zone strategy for fox control. Wildlife Research 27, 465–472

 

The fox, the malleefowl and the camera trap: my research plans

In March 2014 I started with my masters in the qaeco research group at the University of Melbourne. Since it is almost June now and my research proposal has been accepted, I thought it was about time I give a bit of an explanation about what I will be doing in the next two years.

Note: I am going to be saying a lot of ‘we ‘in this article. Since I am not the queen I am not referring to myself, but mainly to a whole group of people who have been involved with research on the malleefowl for much longer than I have. Read more here and here.

Threatened bird
Have you ever heard of the malleefowl? It looks like this:

The Malleefowl or Leipoa ocellata. Source: Wikipedia

The Malleefowl or Leipoa ocellata. Source: Wikipedia

It is a cool bird for three reasons:

  1. It build nests that function as an incubator:
    Schematic malleefowl mound

    Cross section of a Malleefowl mound, showing a layer of sand (up to 1 m thick) used for insulation; egg chamber; and a layer of rotting compost. The egg chamber is kept at a constant 33°C by opening and closing air vents in the insulation layer, while heat comes from the compost below. Source: Wikipeda (text and figure)

  2. It’s is only found in Australia, is part of aboriginal culture and has a lot of different names in different aboriginal languages such as Nganarmara
  3. It’s name in Dutch (my native language) is thermotervogel, which translates back to English as “thermometerbird”

If you have ever seen one, you are lucky, because the malleefowl is rare. In fact, this funny mound-building bird is listed as Vulnerable on the IUCN red list.

Does the fox eat the bird?
The malleefowl is threatened for a whole lot of reasons and no one knows exactly why. It is thought that the red fox, an invasive species in Australia, plays a role in the decrease of malleefowl (Benshemsh 2007). However, there is no scientific consensus on this particular problem (Walsh et al. 2012).

Red fox or Vulpes vulpes

The red fox was introduced in Australia in the 1830s and is now considered a major pest species. This particular fox is a Vulpes vulpes crucigera, I am not sure if these are the ones in Australia.  (Honestly, I haven’t been able to figure out what particular subspecies of foxes is dwelling around in Australia.)  Source: Wikipedia

Fox baiting: 1080 poison

To figure out if foxes play an important role in the threatening of malleefowl, fox baiting is done in some areas in Australia.

Fox bait logo and warning sign

Sodium fluoroacetate (‘1080’), a.k.a. fox bait or FOXOFF is poisonous to canids. Sources: Foxoff and Bubs Smith

The reduction in fox population size is expected to lead to an increase in malleefowl abundance. If this is actually true is what we are trying to figure out.

Monitoring the fox population
To figure out if foxes affect malleefowl populations, we need to monitor fox and malleefowl populations to see if there is any increase in malleefowl populations after fox populations decline.

Also, we need to check if the fox baiting is actually working and killing enough foxes (it doesn’t always work in the long term (Thomson et al. 2000)).

To answer the question of whether or not fox baiting is working, we will need to collect data by monitoring fox populations.

That’s where I come in. I will focusing on the effective monitoring of fox populations.

Monitoring is generally expensive, you need people with skills, you need time, you need transport, you need do to it more than once. The newest trend in wonderful world of species monitoring is the camera trap.

This is a camera trap

A camera trap is basically a camera that is triggered by motion and takes pictures of animals. It sits in a robust casing so it doesn’t break easily. Source: Wikipedia

Okay. Now we get to what I will be doing:

  1. I will look at pilot data that was collected when monitoring a fox population and answer superpractical questions like:
    1. how high should you position the camera,
    2. how long do you need to monitor to need a good estimate,
    3. how far do you want to space the cameras apart?
  1. Using this data and using findings of others who have used camera traps for monitoring, I will design a way to set up camera traps in the field that should be most effective in monitoring fox populations and detecting changes in population size.
  1. Then I will do simulations of fox baiting at different intensities to analyse how good my design is at detecting the effects of fox baiting at these different intensities. This way I find out what the minimum number of camera traps that is necessary to still detect a difference between fox baiting and no fox baiting.

That’s it!

————————————————————–

References:

Benshemesh, J., 2007. National Recovery Plan for Malleefowl. Department for Environment and Heritage, South Australia.

Thomson, P. C., Marlow, N. J., Rose, K., Kok, N. E., 2000. The effectiveness of a large-scale baiting campaign and an evaluation of a buffer zone strategy for fox control. Wildlife Research 27 , 465–472.

Walsh, J.C., Wilson, K.A., Benshemesh, J., Possingham, H.P., 2012. Unexpected outcomes of invasive predator control: the importance of evaluating conservation management actions. Animal Conservation 15, 319-328.