Performance of proximity loggers in recording intra- and inter-species interactions: a laboratory and field-based validation study

Abstract

Knowledge of the way in which animals interact through social networks can help to address questions surrounding the ecological and evolutionary consequences of social organisation, and to understand and manage the spread of infectious diseases. Automated proximity loggers are increasingly being used to record interactions between animals, but the accuracy and reliability of the collected data remain largely un-assessed. Here we use laboratory and observational field data to assess the performance of these devices fitted to a herd of 32 beef cattle (Bos taurus) and nine groups of badgers (Meles meles, n = 77) living in the surrounding woods. The distances at which loggers detected each other were found to decrease over time, potentially related to diminishing battery power that may be a function of temperature. Loggers were highly accurate in recording the identification of contacted conspecifics, but less reliable at determining contact duration. There was a tendency for extended interactions to be recorded as a series of shorter contacts. We show how data can be manipulated to correct this discrepancy and accurately reflect observed interaction patterns by combining records between any two loggers that occur within a 1 to 2 minute amalgamation window, and then removing any remaining 1 second records. We make universally applicable recommendations for the effective use of proximity loggers, to improve the validity of data arising from future studies.

Figure 1. The three types of proximity logger used in this study.

(A) Proximity logger on a collar fitted to an anaesthetised badger. (B) Cattle wearing proximity logger collars. (C) Proximity logger base station in situ near a badger latrine in a field grazed by the collared cattle.

Figure 2. Correlations between contacts recorded by interacting pairs of badger proximity collars.

Values are given for a) the duration and b) the frequency (number) of contacts for the 3 possible collar pairings based on their UHF settings (indiv. set.-indiv. set. (▴), indiv. set.-same set. (♦), same set.-same set. (▪)) and for the three data manipulation treatments to reflect the ‘real-life’ contacts (no amalgamation of broken contacts, amalgamation, amalgamation and removal of remaining 1 second contacts). The dashed line is the line of equivalence (y  =  x), along which all points would lie if collar 1 recorded exactly the same data as collar 2.