Plant physiologists have known for several decades that plants emit sounds (Milburn and Johnson 1966). A bigger part of these ‘crackling’ or ‘whispering’ sounds are of transpiratory/hydraulic origin and are therefore related to the circulation of water and air within the plant as part of the transpiration process (Sandford and Grace 1985). The frequencies of the loudest acoustic emissions (the so-called cavitation pulses) lie mostly in the ultrasonic range, depending on the species-specific characteristics of plant tissues (Mayr and Rosner 2011).
Cavitation pulses (audible in our recording at the right side as single, loud clicks) are indications for embolism in the water transport system, which occurs when a plant is subjected to drought stress and desiccation (Jackson and Grace 1996). The excessive water tension in the water-conducting system leads to the rupture of water columns in plant vessels (caused by cavitations). Plants therefore ‘whisper’ louder as it gets drier in their environment.
In the audification process of the ultrasonic cavitation pulses and all other water transport-induced noises, we transposed the recorded signals in the audible domain using a resampling algorithm that decimates the input signal and converts inaudible high-frequency signals into audible low-frequency signals (applied frequency shift ratio in the recording on the right side: 10-1 [Specht 1998–2002]). Applying this method allowed us to keep the recording’s original time scale by just changing the frequency of the signals.
Each plant species – in fact, each individual plant, has its own acoustic signature, related to its structure and local climatic conditions (figure 1). Investigating the acoustic emissions of a tree (figure 2) in response to dynamically changing climatic conditions might reveal biological or physical properties that place them in a broader ecophysiological context and make processes explainable that have only recently become well understood (Zweifel and Zeugin 2008; Zweifel, Rigling, and Dobbertin 2009).
As you may hear in our audio example here, there exist many further constantly present sounds in plants; controversially, it has been discussed whether these noises have a communicative function in the root area, using the soil as transmission medium (Gagliano, Mancuso, and Robert 2012).
Various artistic projects have subjected plant sounds to artistic investigation, intending to reveal a normally inaudible world. We mention here only a few examples: Justin Bennett’s Hoor de Bomen, 2009; or Tree Listening by Alex Metcalf, 2007/2010; Data Tree by Christa Sommerer and Laurent Mignonneau; as well as Forest Symphony by Ryuichi Sakamoto, 2013. Artists have used our recordings of plant sounds for their own works, including Annea Lockwood, Kontraerklang (Chiara Kramer), Stephanie Wehowski, and others.
Recording and measurement of acoustic emissions
To be able to record and measure the acoustic emissions of a tree in the field, we had to adapt and ‘hack’ technical equipment normally used in non-destructive material testing (NDT) and bioacoustic research. The NDT acoustic emission sensors had to be made waterproof, and the bioacoustic preamplifiers, normally used to amplify the signals of hydrophones, had to be altered to provide a higher amplification rate due to the very low acoustic energy of the tree sounds.
We also built our own acoustic sensors, based on DIY technology like those used in many artistic experiments and performances: a copper wire pin was soldered onto cheap piezo elements to allow them to be coupled optimally into the tissue of a plant (figure 3). Piezo elements, so-called buzzers, are normally used in pocket calculators and watches as acoustic signallers (producing the beeps), but they may also be used as electro-acoustic transducers. Piezo buzzers are even more sensitive than expensive NDT transducers, because the piezoelectric ceramic layer is mounted on a very thin brass plate that needs much less sound energy to vibrate than a thicker ceramic element in an industrial transducer (Rossing 2007). Our frequency response tests showed that the piezo buzzers have resonances at 1 and 140 KHz, exactly the frequency domains where we expected most of the acoustic emissions (cavitation pulses and water transport noises).
The incoming signals were digitised and automatically measured with bioacoustic ultrasonic audio interfaces and recording software from Avisoft (with some adjustments to the software to enable recording and correct measurement of ultra-short pulses) and then logged together with the ecophysiological data. These data sets were then used for the data sonification.