Bird detector

The Soundlapse project research focusses in region that enjoys a biodiversity known worldwide for its
high endemism and which, at the same time, suffers from threats such as habitat fragmentation and
other impacts of global change. Within this context, it is one of the project’s aims to study one of
nature’s most intriguing codes: the acoustic signals that birds use to communicate and how these signals develop in temporal phases over the course of the day and over the year.

Through long-term, non-invasive acoustic monitoring, bird communities were studied in different
locations of the urban wetland network. The main challenge of this method was to analyze the huge
collection of audio recordings obtained. As a way of dealing with this overwhelming amount of acoustic
data, an artificial neural network was trained to identify songs associated with specific bird species
located in these natural environments. This approach allowed the possibility of obtaining an accurate
record of the diversity and annual activity of these bird communities and their relationship with climatic
and astronomical factors.

In collaboration with a group of researchers from the Bird Ecology Laboratory, led by the ornithologist
Jorge Ruiz an almost unknown bird species was identified. This was achieved by means of an acoustic
analysis of numerous samples of the territorial songs of this small, cryptic, and vulnerable species in a
highly urbanized area.