Assessing Raspberry Shake and Boom sensors for recording African elephant acoustic vocalizations

Abstract

In this work we assess the performance of the low-cost Raspberry Shake and Boom sensor package for detecting and monitoring African elephants (Loxodonta africana). This is the first documented test of this particular sensor package for recording animal behaviour; the unit was originally designed for detecting tectonic earthquakes and low frequency (<50 Hz) atmospheric acoustics. During a four day deployment in October 2019 in South Africa, we used five Raspberry Shake and Boom units to record acoustic and seismic vocalisations generated by a group of African elephants. Our results highlight a varied degree of success for the RS&B units in detecting the signals of interest. The acoustic microphone recorded intricate details of low-frequency (<50 Hz) vocalisations that were not clearly recorded by more sensitive instruments, but was not able to record higher frequencies due to the low sampling rate (100 Hz). The geophone was not able to consistently record clear seismic waves generated by vocalisations, but was able to record footsteps within a 50 m range. We discuss future directions and deployment strategies to improve the sensitivity of the sensor package. Nevertheless, we conclude that the RS&B unit shows great promise as a low-cost tool for monitoring African elephants.

Airborne Infrasound Makes a Splash

“…Other activity such as colliding ocean waves (the “ocean microbarom”) (Waxler & Gilbert, 2006), wind blowing over mountains (Bedard, 1978; Walterscheid & Hickey, 2005), and severe storms (Goerke & Woodward, 1966) can create long duration signals that comprise much of the Earth's infrasonic background at low frequencies. Regional to local infrasound (and occasionally audible emissions) have been reported from small earthquakes (Johnson et al., 2020; Sylvander et al., 2007), iceberg calving (Richardson et al., 2010), tornadoes (Elbing et al., 2019), thunder (Arechiga et al., 2014; Dessler, 1973), the aurora (Pasko, 2012), and even elephants (O. D. Lamb et al., 2021), see Campus and Christie (2010) for a comprehensive list. Human‐caused events can make infrasound as well.…”

Seismic savanna: machine learning for classifying wildlife and behaviours using ground‐based vibration field recordings

“…Sugumar and Jayaparvathy (2013) also investigated using seismic data for elephant classification, but only tested their method on a few dozen simulated datapoints. Lamb et al (2021) recently deployed Raspberry Shake sensors in South Africa to assess their viability for monitoring seismic vocalizations and locomotion of elephants, but not designed algorithms to detect and classify the animals. All these results are encouraging, and in this study, we take this idea to the next step, with large amounts of field data consisting of seismic and cameratrap recordings, and a fully data-driven automated deep learning approach.…”