Page 48 - WINTER2019

Plant Bioacoustics
to detect and measure sound and the increasing sophistica- tion of computational signal-processing abilities. Studies using bioacoustic tools to detect plant-based pests will continue to expand, given that over 218 species across 12 different insect orders have been identified as using sounds or vibration for communication, with the true number of species certainly being much higher (Hill, 2008).
Investigators across the United States study a wide variety of pests. Richard Mankin at the US Department of Agricul- ture (USDA) has conducted research on how to spot invasive pests like the devastating Asian citrus “jumping plant lice” (Diaphorina citri) and Asian longhorn beetle (Anoplophora glabripennis) in living trees, whose larvae break fibers while feeding or moving through woody tunnels (Mankin et al., 2011, 2018). Alexander Sutin at the Stevens Institute of Technology (Hoboken, NJ) has focused on insect detection in agricultural shipments and wood packing materials (Sutin et al., 2017). Richard Hofstetter at Arizona State University (Tempe) has conducted detailed investigations into the sounds of multiple species of piñon bark beetles, which naturally live in native piñon trees throughout the American West (Hofstetter et al., 2014). However, a drying climate has stressed these trees, leading to outbreaks of bark beetle infestations that have been estimated to kill nearly a third of piñon trees in the United States.
Despite the wide variety of species examined, there are con- sistent patterns and challenges faced by all these researchers. Many insect larvae in woody substrates produce broadband, even ultrasonic, pulse trains of short, 1- to 10-ms impulses. The pulses often occur in short bursts, with interpulse inter- vals of 200 ms or less. Adults of other species generate tonal signals with substantial harmonics. The signals are normally detectable to only several centimeter ranges when propagat- ing through the air, but whenever the vibrations of a single animal travel through plant substrates, they have been detected by accelerometers up to 4 meters distant before fading into the background noise spectrum. Ultrasonic sounds (>20 kHz) are particularly effective for long-range detection because back- ground noise levels are generally low.
The increasing scrutiny of these sounds has made plant researchers appreciate how crucial the structure of plants is in transmitting and modifying sound. The anisotropic and heterogeneous nature of plants causes filtering, waveguide dis- persion, and even the resonant enhancement of insect signals
in a manner that a seismologist or ocean acoustician would find familiar.
Lujo et al. (2016) have presented in detail one example of such research in plant lice. Males vibrate their wings to generate 0.2-kHz tonal signals with harmonics up to 3 kHz, which are transmitted down their legs into tree branches. The relative strength of these harmonics changes with distance along a branch, but females tend to respond as long as a few harmon- ics are still present. Waveguide dispersion effects are apparent, with different frequency components traveling at different speeds along the branch, and multiple researchers have spec- ulated that insects might exploit these frequency-dependent absorption and dispersion effects to estimate the direction and distance to a potential mate.
Even more intriguing is the limited evidence that suggests that insect sounds can be transmitted between different plants, even those that aren’t physically touching. Shira Gordon and her col- leagues at the USDA (2019) have investigated how mating calls from the glassy-winged sharpshooter (Homalodisca vitripen- nis) are transmitted through grapevines by using a transducer to mechanically vibrate a single plant stem and then using a laser Doppler vibrometer (LDV) to characterize the resulting transmission along the plant. Like other researchers, Gordon et al. found evidence of frequency-dependent dispersion, in that the higher harmonics of the call had higher group sound veloci- ties than the lower harmonics, consistent with Bernoulli-Euler beam flexure theory. What was more surprising was that they found evidence that these artificially generated calls could be transferred between nonconnected plants by sound radiation from the broad grape leaves that act like crude diaphragms in speakers. The LDV could still detect 100-Hz vibrations on plants separated by up to 10 cm from another plant agitated by the transducer, albeit attenuated by 60 dB, with higher frequency components detectable at shorter distances.
The practical importance of plant propagation effects on insect signals is that it complicates efforts to automatically detect and distinguish insect sounds from other acoustic sources, a prob- lem familiar to many animal bioacousticians. Furthermore, researchers are investigating whether acoustic playbacks of insectsoundsorvariationsthereofcouldbeusedtopreventor even expel existing pests from trees using various sounds (e.g., Hofstetter et al., 2014), and knowledge of the transmission path through plants is an obvious requirement for reproducing con- vincing fidelity.
48 | Acoustics Today | Winter 2019