Behavioral classification of low‐frequency acceleration and temperature data from a free‐ranging small mammal

Abstract

The miniaturization and affordability of new technology is driving a biologging revolution in wildlife ecology with use of animal‐borne data logging devices. Among many new biologging technologies, accelerometers are emerging as key tools for continuously recording animal behavior. Yet a critical, but under‐acknowledged consideration in biologging is the trade‐off between sampling rate and sampling duration, created by battery‐ (or memory‐) related sampling constraints. This is especially acute among small animals, causing most researchers to sample at high rates for very limited durations. Here, we show that high accuracy in behavioral classification is achievable when pairing low‐frequency acceleration recordings with temperature. We conducted 84 hr of direct behavioral observations on 67 free‐ranging red squirrels (200–300 g) that were fitted with accelerometers (2 g) recording tri‐axial acceleration and temperature at 1 Hz. We then used a random forest algorithm and a manually created decision tree, with variable sampling window lengths, to associate observed behavior with logger recorded acceleration and temperature. Finally, we assessed the accuracy of these different classifications using an additional 60 hr of behavioral observations, not used in the initial classification. The accuracy of the manually created decision tree classification using observational data varied from 70.6% to 91.6% depending on the complexity of the tree, with increasing accuracy as complexity decreased. Short duration behavior like running had lower accuracy than long‐duration behavior like feeding. The random forest algorithm offered similarly high overall accuracy, but the manual decision tree afforded the flexibility to create a hierarchical tree, and to adjust sampling window length for behavioral states with varying durations. Low frequency biologging of acceleration and temperature allows accurate behavioral classification of small animals over multi‐month sampling durations. Nevertheless, low sampling rates impose several important limitations, especially related to assessing the classification accuracy of short duration behavior.

Optimisation of energetic and reproductive gains explains behavioural responses to environmental variation across seasons and years

“…This biologging research focuses on a bottom‐up regulated population in the Kluane region of southwestern Yukon (61°N, 138°W) that has been the focus of long‐term monitoring since 1987 (McAdam et al, 2007; Krebs et al, 2014). As a free‐ranging study population, red squirrels offer several advantages in documenting drivers of activity variation, including year‐round residency, large sample size, daily and seasonally variable activity patterns (Pauls, 1977; Studd et al, 2016; Studd, Landry‐Cuerrier, et al, 2019), quantifiable resources (LaMontagne et al, 2005; Fisher et al, 2019), and most importantly, the ability to accurately classify activity and inactivity on undisturbed individuals using accelerometers (Studd, Landry‐Cuerrier, et al, 2019).…”

“…Individual squirrels were captured on defended territories, weighed, assessed for reproductive condition, and fitted with an accelerometer (models Axy2/Axy3, 4 g [1.7% of body mass], Technosmart Europe) in collar form, either ventrally mounted on its own ( n = 128) or dorsally‐mounted in combination with a ventrally mounted VHF radio transmitter ( n = 361, model PD‐2C, 4 g [1.7% of body mass], Holohil Systems Limited, Carp, ON, Canada; see Studd, Landry‐Cuerrier, et al, 2019 for collar design). All accelerometers recorded acceleration between ± 8 g forces at a sampling rate of 1 Hz and temperature at a rate of 0.1 Hz, frequencies that have been shown to capture broad‐scale behaviour of small animals with high accuracy, allowing for long‐duration recordings (Tatler et al, 2018; Studd, Boudreau, et al, 2019).…”

“…Squirrels were released at site of capture and remained free‐ranging until recaptured for collar removal (3–103 days later). Accelerometer data were processed and classified into two behavioural states, active (physically moving outside a nest) and inactive (physically not moving or inside a nest) using two thresholds described in detail in Studd, Landry‐Cuerrier, et al (2019). Briefly, animals were nesting when the collar temperature was above the threshold value from a k‐means clustering of all collar temperatures recorded that day (2 clusters), and were moving when the 10 second sum of the changes in acceleration from one second to the next was greater than 1.06 g forces .…”

Open‐source, low‐cost modular GPS collars for monitoring and tracking wildlife

“…There was no evidence of drift in accelerometer data at the programmed frequency of collection however, any drift that might be occurring would have been removed during analysis as it passed through a high‐pass filter prior to analysis (Takeda et al, ). The frequency of data collection chosen in this study, at 10 Hz, is lower than many other accelerometer devices however, with the advent or more advance machine learning methods this is becoming more than sufficient for behavioural classification, with classification now being possible with as low as 1 Hz (Studd et al, ).…”

Machine learning goes wild: Using data from captive individuals to infer wildlife behaviour

“…Extrapolating behaviours from acceleration data of wild individuals is a challenge, since there is no possibility for direct output validation. Some models were trained and validated on the same wild individuals (Yoda et al, 2001;Tsuda et al, 2006;Studd et al, 2019), which requires direct observation of the studied individuals at least for a certain period of time. However, the promising advance of behaviour classification through machine learning is the ability to study the behaviour of wild animals without observing (and possibly disturbing) them.…”