Assessing deer density from trail camera survey 

On Wednesday 11 December and Thursday 12 December, a team of five Central North Island Sika Foundation volunteers went into the Kaimanawa Remote Experience Zone (KREZ) and set up 80 cameras to start a project designed to compare the effectiveness of faecal pellet counting with camera monitoring to assess deer density using a trail camera survey. Manaaki Whenua – Landcare Research, supplied aerial assistance from their helicopter which was in the area carrying out work on the deer survey project for OSPRI, to ferry the team in and out of the KREZ to enable our members to deploy the cameras.

Landcare also supplied cameras and other equipment to help us set up the project. In February 2020 a team will return and pick up the cameras and analyse the information and at the same time the DoC Tier monitoring team will also carry out faecal pellet counting so the two methods can be compared.

A big thanks to Brohn Torckler, Josh Van der valk, Ron Lenzen, Stu Emmerson and Dafydd Pettigrew for helping to get this project up and running, a big couple of days in remote and rugged country. Also, a big thank you to Ivor Yockney and his team at Manaaki Whenua – Landcare Research, without their help this project would not be possible, a good example of people working together to achieve positive outcomes.

Field plan:

Aim: To assess whether trail camera monitoring can provide indices of deer abundance that are at least equivalent to faecal pellet counts, and which could also provide estimates of actual deer density.

Proposal: To deploy trail cameras on ~20 faecal pellet count transects in the Kaimanawa Remote Experience Zone (KREZ) in early summer 2019. Pellet counts will be conducted by the DoC Tier monitoring team when cameras are removed in February 2020. The index of deer abundance based on pellet counts will then be compared with a range of activity indices based on the camera data. In addition, the camera data will be used to compare a range of new approaches for estimating actual deer density from the camera data.


Index correlation: The comparison of indices is a simple correlative study. Two pairs of cameras will be placed at each end of the 150-m long faecal pellet count transects. A major difficulty with camera trapping is that the amount of activity recorded is heavily dependent on camera placement – cameras placed on game trails detect many more deer than randomly placed cameras.

It is therefore proposed that one of each pair of cameras be placed as objectively as possible at the site of each of the first and last pellet count plots (which is likely to result in some cameras having very restricted fields of view).

The other camera in each pair will also be placed as objectively as possible nearby but in locations that meet minimum field of view criteria.

Density estimation:

A number of approaches have been developed for estimating animal density from trail camera monitoring.

These include (for deer) estimating (i) the likelihood of animal being recorded on multiple cameras, (ii) the duration of occupancy by deer of the collective fields of view (i.e.; essentially how much time deer spent on camera), and (iii) mark-recapture methods that rely on some deer having distinctive antlers that enable them to be reliably recognised every time they are recorded on camera.

The first requires a grid of about 50 cameras spaced less than one home range diameter apart, so it would only be feasible at just one or two locations with the cameras, making it infeasible to both estimate both density and to correlate with existing pellet count transect data. The main focus will therefore be on assessing the various data-modelling methods available for (ii).