Identification of New Zealand bats (Chalinolobus tuberculatus and Mystacina tuberculata) in flight from analysis of echolocation calls by artificial neural networks

Abstract

Time-expanded and heterodyned echolocation calls of the New Zealand long-tailed Chalinolobus tuberculatus and lesser short-tailed bat Mystacina tuberculata were recorded and digitally analysed. Temporal and spectral parameters were measured from time-expanded calls and power spectra generated for both timeexpanded and heterodyned calls. Arti®cial neural networks were trained to classify the calls of both species using temporal and spectral parameters and power spectra as input data. Networks were then tested using data not previously seen. Calls could be unambiguously identi®ed using parameters and power spectra from time-expanded calls. A neural network, trained and tested using power spectra of calls from both species recorded using a heterodyne detector set to 40 kHz (the frequency with the most energy of the fundamental of C. tuberculatus call), could identify 99% and 84% of calls of C. tuberculatus and M. tuberculata, respectively. A second network, trained and tested using power spectra of calls from both species recorded using a heterodyne detector set to 27 kHz (the frequency with the most energy of the fundamental of M. tuberculata call), could identify 34% and 100% of calls of C. tuberculatus and M. tuberculata, respectively. This study represents the ®rst use of neural networks for the identi®cation of bats from their echolocation calls. It is also the ®rst study to use power spectra of time-expanded and heterodyned calls for identi®cation of chiropteran species. The ability of neural networks to identify bats from their echolocation calls is discussed, as is the ecology of both species in relation to the design of their echolocation calls.

The importance of exotic plantation forest for the New Zealand long-tailed bat (Chalinolobus tuberculatus)

“…Other research into long-tailed bats in Kinleith Forest found they were detected more commonly along roads among P. radiata which were near harvest age, and that bat activity along a road in one area of P. radiata exceeded that along the same road where it passed through a nearby indigenous remnant (Moore 2001). Nevertheless, rates of detection of bats in Kinleith Forest were far lower than in comparable bat detector-based research in the native Nothofagus forest of Fiordland's Eglinton Valley (O'Donnell 2000a;Borkin & Parsons 2009), and whilst no direct relationship has yet been established between detection rates and actual numbers of bats present (Parsons 2001), we presume that this indicates a smaller population in Kinleith Forest.…”

Variability in echolocation call design of 26 Swiss bat species: consequences, limits and options for automated field identification with a synergetic pattern recognition approach

“…PARAMETRIC APPROACHES Parametric measurements taken e.g., from oscilloscopes, powerspectra or spectrograms allow for later analysis with a variety of approaches ranging from discriminant function analysis or DFA (Zingg 1990;Vaughan et al 1997;Jones et al 2000;Russo & Jones 2002) to decision trees (Herr et al 1997) and artificial neural networks or ANN (Stocker 1998;Parsons 2001). For measurements taken from data-reduced samples high recognition rates above 90% were reported for some species, others failed completely to be separated (Herr et al 1997).…”

Habitat use and nocturnal activity of lesser short‐tailed bats(Mystacina tuberculata)in comparison with long‐tailed bats(Chalinolobus tuberculatus)in temperate rainforest

“…One frequency dial on the Bat Box III detectors was set at c. 27 kHz (±4 kHz; Waters & Walsh 1994) corresponding with the peak amplitude of lesser short-tailed bat calls and the other on c. 40 kHz to correspond with peak amplitude of long-tailed bats (Parsons 2001). Total bat passes per hour per night were counted from the tapes.…”

“…Passes were generally classified as lesser short-tailed bats if they were recorded at 27 kHz and as longtailed bats if they were at 40 kHz. Although there is overlap between the fundamentals and harmonics of calls of both species (Parsons 1997), overlapping calls are distinguishable because lesser short-tailed bat call rates were twice as fast as those of longtailed bats (O'Donnell et al 1999;Parsons 2001). Verified reference tapes of both bat calls recorded in a variety of different habitat types were used as an additional aid to identification.…”

“…Nevertheless, this seems largely to be the case, with both long-tailed bats and lesser short-tailed bats demonstrating little overlap, and little flexibility, in their habitat use. Differences between lesser short-tailed and long-tailed bats in wing ecomorphology (Norberg & Rayner 1987;Jones et al 2003;O'Donnell 2005), echolocation (Parsons 2001), nocturnal activity patterns (O'Donnell 2000b(O'Donnell , 2005; this study), roost site selection (Sedgeley & O'Donnell 1999a,b;Sedgeley 2003), diet (Arkins et al 1999;O'Donnell 2005) and morphology (O'Donnell et al 1999;Lloyd 2005;O'Donnell 2005) suggest that differences between the species range across the full gambit of behaviours. Some caution should still be used when making definitive statements about habitat selection in New Zealand bats.…”