Soundscapes and artificial intelligence provide powerful tools to track biodiversity recovery in tropical forests

Tropical forest recovery is fundamental to addressing the intertwined climate and biodiversity loss crises. While regenerating trees sequester carbon relatively quickly, the pace of biodiversity recovery remains contentious. Here, we use bioacoustics and meta-barcoding to measure forest recovery post-agriculture in a global biodiversity hotspot in Ecuador. We show that the community composition, and not species richness, of vocalizing vertebrates identified by experts reflects the restoration gradient. Two automated measures – an acoustic index model and a bird community derived from an independently developed Convolutional Neural Network – correlated well with restoration (adj-R2 = 0.62 and 0.69, respectively). Importantly, both measures reflected composition of non-vocalizing nocturnal insects identified via meta-barcoding. We show that such automated monitoring tools, based on new technologies, can effectively monitor the success of forest recovery, using robust and reproducible data. Crucially, this will help ensure that forest restoration efforts result in resilient, biodiverse tropical forests and not simply ‘carbon farms’.

Methods
DNA was extracted from 200-µL aliquots using the DNEasy blood & tissue kit (Qiagen) following the manufacturer’s instructions. Multiplex PCR was performed using 5 µL of extracted genomic DNA, Plant MyTAQ (Bioline, Luckenwalde, Germany) and high-throughput sequencing (HTS)-adapted mini-barcode primers targeting the mitochondrial CO1-5P region (mlCOIintF – 5’– GWACWGGWTGAACWGTWTAYCCYCC–3’; dgHCO2198–5’-TAAACTTCAGGGTGACCAAARAAYCA–3’; following Leray et al., 2013 – also see Morinière et al.; Morinière et al.. Amplification success and fragment length were determined using gel electrophoresis. The amplified DNA was cleaned and each sample was resuspended in 50 µL of molecular water. Illumina Nextera XT (Illumina Inc., San Diego, USA) indices were ligated to the samples in a second PCR, conducted at the same annealing temperature as in the first but with only seven cycles. Ligation success was confirmed by gel electrophoresis. DNA concentrations were measured using a Qubit fluorometer (Life Technologies, Carlsbad, USA), and the samples were then combined into 40-µL pools containing equimolar concentrations of 100 ng each. The pooled DNA was purified using MagSi-NGSprep Plus beads (Steinbrenner Laborsysteme GmbH, Wiesenbach, Germany). The final elution volume was 20 µL. HTS was performed on an Illumina MiSeq using v3 chemistry (2*300bp, 600 cycles, maximum of 25 mio paired-end reads).