Methods - Soundscape Ecology: Principles, Patterns, Methods and Applications - page 246

Fig. 9.18 Schematic view of the method to collect and process audio data. In in situ sampling , the
animal acoustic community is recorded, then in ex situ processing the audio files are processed and
signal analysis is performed applying metrics (Modified with permission from Sueur et al. 2012 )
sampling/collection of field data using a single or a combination/array of
microphones. The ex situ phase encompasses all the processing of audio files,
their labeling, and insertion in a database. Then, indices are applied to extract the
required information. This procedure in the future could be done in a portable
all-in-one system where the entire process could be obtained from a specific
application.
9.9 Functional Diversity Indices
Biodiversity is a well-known concept but despite its universal use remains quite
difficult to quantify appropriately (Kaennel 1998 ; Petchey and Gaston 2002 ; Duelli
and Obrist 2003 ). According to Wagner and Danchin ( 2010 ) the information that is
contained in the acoustic repertoire of vocal animals can be considered a social
information with two subrepartitions: public information (accessible to others) and
private data (inaccessible to others).
Acoustic complexity of biophonies is a relevant part of biodiversity, although
the contribution that every vocal species can hear remains undetermined without an
aural species-specific determination (Obrist et al. 2010 ).
The complexity that emerges from a spectrogram of a community of vocalizing
animals can be measured by applying functional diversity indices in which at every
frequency category is assigned a value of functionality (Schleuter et al. 2010 ). This
procedure can be used by transferring the algorithms designed for calculating the
functional richness (FR) to the spectral domain where each frequency category
could be considered a “vocal trait.” An example of application of this approach to
acoustic complexity has been posed by Pekin et al. ( 2012 ) when modeling acoustic
diversity for a comparison with LIDAR methodology and by Gasc et al. ( 2013 ).
These last authors have considered the acoustic distance between species and
used four different dissimilarity indices: spectral dissimilarity index (Df) (Sueur
et al. 2008 ), Kolmorov-Smirb ` nov distance (KS) (Rachev 1991 ), symmetric
Kullback-Leibler (KL) distance (Kullback and Leibler 1951 ), and the similarity
RV correlation coefficient (Josse et al. 2008 ).