Abstract Sampling methods able to capture information about various taxa, over broad time and spatial scales are essential to assess the successes of protected areas (PAs) and pest control programmes. Passive acoustic monitoring (PAM) coupled with acoustic indices and automated identification are promising tools for biodiversity monitoring. However, two technical bottlenecks are still important limitations for their wide use. Automatic identification commonly presents high false positive rates and there are no standardised protocols for the use of acoustic indices for monitoring. In this thesis I approach these gaps and test the use of PAM associated with automated identification and acoustic indices for monitoring PAs and conservation management. In chapters 2 and 3, instead of using acoustic indices as biodiversity indicators, I use them as filters that allow the identification of the acoustic region that differs most between sites. I define the acoustic regions as units of analysis bounded by a specific time period and frequency range adjusted to capture the main groups of biologically relevant acoustic events within a soundscape. By splitting indices data into acoustic regions, I facilitate statistical analysis of indices results and simplify the identification of sounds that are driving the indices results. In chapter 2 I test if acoustic indices are sensitive enough to measure significant differences in the soundscapes for the Serra do Cipó National Park, Brazil, and a surrounding farmland area. The soundscapes differ significantly for all the 12 indices tested during autumn from 05:30 – 09:00am and within the range of 0.988-3.609 kHz. Sonotype results show that the soundscape outside the park is strongly influenced by domestic animals (present in 63% of the sound files aurally analysed). In chapter 3 I propose and test a workflow for the monitoring of two sites within the Waitakere Ranges Regional Park, New Zealand, that have different pest mammal management levels. The analysis of variance and pairwise comparisons indicated the acoustic region encompassed within 21:00 to 23:59 and a range of 0.988-3.609 kHz in autumn as the one that differs most between sites. The sounds responsible for the main differences on indices measurements are emitted by the activity of invasive mammals in the site with no pest control. In chapter 4 I present and test the Assemblage of Focal Species Recognizers - AFSR, for decreasing false positives of automated acoustic identification for 5 seabird species from Burgess Island, New Zealand. I used MatlabHTK, a hidden Markov models interface for bioacoustics analyses, for illustrating AFSR technique by comparing two approaches, 1) a multispecies recognizer where all species are identified simultaneously, and 2) an assemblage of focal species recognizers (AFSR), where several recognizers that each prioritise a single focal species are then summarised into a single output, according to a set of rules designed to exclude unreliable segments. False positive rate improved for all the five species when using AFSR achieving a remarkable 0% false positives and 100% precision for three of five seabird species. Instead of attempting to withdraw useful information from every fragment in a sound recording, AFSR prioritises more trustworthy information from sections with better quality data. AFSR can be applied to automated species identification from multispecies PAM recordings worldwide. These results confirm that PAM sampling associated with automated identification and acoustic indices are able to represent condition and detect trends in acoustic communities, which are the main focus of monitoring programmes. PAM is able to provide information on acoustic community composition and dynamics, affording useful information for PAs management and conservation programmes. |