Survey guidelines for Australia’s threatened non-flying mammals

3.3.6Nest boxes and camera traps

Use of nest boxes is a technique that requires more research, but preliminary findings suggest that it can be an excellent technique for detecting arboreal mammals, with several species being detected more readily in nest boxes in comparison to other ‘conventional’ survey techniques (for example, squirrel glider Petaurus norfolcensis). The most appropriate size of nest boxes will vary according to the species (Beyer & Goldingay 2006). Generally, larger species require larger boxes, although smaller species are not necessarily restricted to small boxes (Beyer & Goldingay 2006). Certain species have also been recorded chewing and widening the entrance of nest boxes (for example, the sugar glider Petaurus breviceps and brush-tailed phascogale Phascogale tapoatafa [Goldingay et al. 2007]). Species from 10–270 grams appear to be able to utilise nest boxes with an entrance diameter from less than 3 cm to greater than 8 cm, whereas mammals larger than 1000 grams will not use nest boxes with an entrance of less than 3 cm, and prefer them being greater than 8 cm (Beyer & Goldingay 2006). Nest box volume does not appear to greatly influence use except for mammals larger than 1000 grams which appear to prefer a size greater than 0.03 cubic metres (Beyer & Goldingay 2006).
Arboreal marsupials appear to prefer nest boxes with narrow entrances just big enough to enter (Menkhorst 1984a), and the height of the nest box can sometimes (but not always) influence the frequency of use by some species (Menkhorst 1984b; Beyer & Goldingay 2006). Large mammals (larger than 1000 grams) appear to require nest boxes higher than 6 metres above the ground, while medium-sized species (100–270 grams) require the nest box above 2 metres (Beyer & Goldingay 2006). Orientation of the nest box is unlikely to have major effect on usage, except for the potential for rain or excessive sunlight entering the box (Menkhorst 1984a).
Ward (2000) detected one feathertail glider, Acrobates pygmaeus, during 13.8 hours of spotlighting, while nest boxes captured 57 individuals within a 7 hectare area. Forty nest boxes were installed on a 5 by 8 metre grid, with 50 metres between points (5.7 next boxes per hectare). Nest boxes had narrow entrances (15–25 millimetres wide slit) and were four to five metres above the ground. However, in surveys conducted by Goldingay and Sharpe (2004), 20 feathertail gliders were detected during spotlighting and only five detected in nest boxes.
Goldingay and colleagues (2007) trialled four different types of nest boxes to investigate which would be the most suitable for the feathertail glider. The species appeared to avoid medium-sized rear-entry boxes with a 45 millimetre diameter entrance, but no clear preference was shown for three designs with a narrow entrance of less than 25 millimetres. This study deployed four nest boxes per hectare.
Any survey targeting a threatened species which may utilise nest boxes should implement the technique in conjunction with other standard techniques, particularly for species in which the nest box technique has not been proven as successful. Similarly, standard survey efforts have yet to be determined for most threatened species in relation to detecting species presence, but it is considered that at least five nest boxes per hectare should be deployed. The volume, dimensions, entrance diameter and height above the ground of the nest box should take into consideration the target species and the information provided above.
It should be noted that nest boxes have a reduced occupation rate in forests with an abundance of natural hollows. They impose additional installation and construction costs and their use requires a minimum of two visits to a site (Ward 2000).
Camera trapping is becoming a more common method of surveying (Trolle & Kery 2005). Camera traps have the advantage of potentially obtaining a wide range of significant information. Automatic camera systems are triggered by an animal passing in front of a sensor that detects movement, changes in ambient light, or a thermal differential (Moen & Lindquist 2004). The two most common types of cameras are ‘active’ and ‘passive’ triggered. Active systems consist of a transmitting unit that sends an infrared beam, and a receiving unit which is set across the target area. A picture is taken when the infrared beam is broken. Passive systems are single units that use heat and motion detectors to trigger the camera (Kelly & Holub 2008). Infrared sensors work better at cooler ambient temperatures and are less consistent in warm environments (Swann et al. 2004). Camera trapping has been found to be the most effective method of detecting species at low or moderate densities (Vine et al. 2009).
Cameras allow for the detection of species that are difficult to study due to their elusive and nocturnal habits (Mace et al. 2004). They are less time consuming, less costly, and less invasive than long-term direct observation of animals. They are also beneficial in studying animals in inaccessible or difficult to access locations such as dens and nest cavities, or in rugged terrain (Mace et al. 1994). In addition, they enable the collection of valuable information about multiple species within any given community (Rosellini et al. 2008) and provide data that is more permanent and less disputable than data gathered by direct observation. However, there is concern that camera traps affect animal behaviour, with human activity, scent and the presence of camera equipment potentially attracting or repelling animals (Major & Gowing 1994, York et al 2001). Also, positive identification of a species from camera images alone can sometimes be difficult.
Thermal infrared imagery is often used in addition to other survey methods (for example, for the Christmas Island shrew), and can be used both on the ground and in the canopy (Schulz 2004).
Recommended survey method

Remote cameras allow for a relatively inexpensive tool to gather data on a broad range of species. The main limitation to camera systems is their sensitivity and false triggering, which often results in the memory card being filled with empty images and depletion of the battery. Another limitation is the failure to photograph a target animal (Swann et al. 2004). These false triggers are known to be more common for camera systems with wider detection zones (Swann et al 2004). Passive infrared-triggered camera systems have a number of set-up options for resolution, time delay between photos and day/night modes which if adjusted properly can minimise these false triggers.

The following guidelines can increase the success rate of cameras:

• 
  choose a camera based on target species. Some systems are better at detecting smaller species than others, whilst all will detect larger species (for example, use white flash rather than infra-red to increase detail in images for identification of small species)
  
• 
  choose a suitable camera for the size of the target area (for example, systems with a narrow detection zone are more appropriate for areas with a narrow entrance, whilst a wider detection zone is more suitable for recording activity within an larger area)
  
• 
  use a very firm support to avoid false triggers
  
• 
  for passive units set sensor height according to target species
  
• 
  place the sensor in a position where there is no vegetation in the foreground
  
• 
  set the camera 2–5 metres from the target area to avoid out of focus pictures
  
• 
  ensure all cords are reinforced with duct tape or similar materials to help reduce cord loss due to animals chewing them (Swann et al 2004).
  

Some surveys use a food based attractant, although this is not always practical as an animal may stay in an area for a longer time, essentially modifying an animal’s behaviour. This is acceptable for a detection type camera, but for a mark recapture camera survey project behaviour modification would not be acceptable (Moen & Lindquist 2004). Kelly and Holub (2008) found that different studies required different camera systems, depending on the species in question and the aims of the study. Camera traps are increasingly being used as they can reduce adverse effects that may be caused by more invasive methods such as physical capture (Kelly & Holub 2008; Silveira et al. 2003)

Baited remote cameras have been successfully used to detect small mammals (Nelson et al 2009). Ten cameras were deployed for 18 nights at sites where the smoky mouse was known to occur in Victoria, in conjunction with Elliot A traps and hair tubes (one hectare area). Each camera station consisted of a Sony 7.2 megapixel digital camera and a passive infrared sensor (using a white flash), attached to either a tree trunk or a wooden stake, with the sensor positioned approximately 25 centimetres above the ground. Five out of six cameras detected the species after 10 nights.
Despite the recent development of camera traps for use in wildlife surveys, the survey effort required for species detection has only been documented for a limited range of species groups. Further research is necessary to document the full range of threatened species described in this document. Based on known information the following survey design and effort is recommended:

• 
  cameras should be deployed for at least 14 nights, and
  
• 
  approximately 10 cameras should be deployed per hectare.
  

Camera traps should not be used as the only survey method and should always be used in conjunction with other standard survey techniques (for example, spotlighting, hair tubes, Elliot traps, cage traps etc). This is particularly important for species in which the technique has not been proven successful. Furthermore, as with other established survey techniques, the failure to detect a species following a survey using a camera trap does not mean the species is absent from the study area.