Combining data from different survey methods

questions concerning anlysis/theory using program DENSITY and R package secr. Focus on spatially-explicit analysis.

Combining data from different survey methods

Postby charlottesearle » Mon Dec 19, 2022 6:04 am

Hi all,

I’m carrying out an analysis of lion density in Tanzania using two data sets: camera trap data from a three month camera trap survey, and direct sightings data from a month-long search-encounter survey. Both the surveys were carried out in the same area and during the same period.

In addition to comparing the separate density estimates produced through these two survey types, I'd like to try and combine the different data types into a single model. This is because I expect it should improve my density estimate (by virtue of increasing the number of captures and recaptures), and more broadly because I think it could be a useful approach in certain contexts, and would like to work out how (if!) it can be done.

I’ve already modelled both these datasets separately in secr. For the camera trap data, I used proximity detectors with binary effort. For the search encounter data, I converted the study area into a 1x1km grid of cells, with effort calculated as the number of kilometres driven in that cell on that day – this means it is also proximity detector type, but with a different (non-binary) measure of effort.

I know that there is currently no function to pool capthist data from different detector arrays, and that it is recommended to merge the input files and rebuild the capthist object from scratch. However, is it possible to do this with two different survey methods with different measures of effort (binary and non-binary)?

Otherwise it would be possible to use a binary measure of effort for the search encounter data (i.e. 1 if that pixel was surveyed on that sampling occasion, 0 if it was not) and combine it with the camera trap data into a single set of input files. However, the actual effort is really quite different between the two data types, so this seems like it may not be a good way forward.

Any advice would be much appreciated!
Posts: 3
Joined: Sun Jan 05, 2020 11:22 am

Return to analysis help

Who is online

Users browsing this forum: No registered users and 0 guests