Acoustic assessment of experimental reforestation in a Costa Rican rainforest

Effective forest restoration requires tools for evaluating and comparing restoration approaches. Nevertheless, measuring restoration progress can be difficult and expensive. Passive acoustic monitoring (PAM) can be an inexpensive assessment strategy to collect large amounts of biodiversity informati...

Descripción completa

Guardado en:

Detalles Bibliográficos
Autores principales:	Álvaro Vega-Hidalgo, Eleanor Flatt, Andrew Whitworth, Laurel Symes
Formato:	article
Lenguaje:	EN
Publicado:	Elsevier 2021
Materias:	Bioacoustics Ecological restoration Reforestation Ecoacoustics Soundscape Katydids Ecology QH540-549.5
Acceso en línea:	https://doaj.org/article/382479775ad14fd29d6828293e4fb66d
Etiquetas:	Agregar Etiqueta Sin Etiquetas, Sea el primero en etiquetar este registro!

id	oai:doaj.org-article:382479775ad14fd29d6828293e4fb66d
record_format	dspace
spelling	oai:doaj.org-article:382479775ad14fd29d6828293e4fb66d2021-12-04T04:33:24ZAcoustic assessment of experimental reforestation in a Costa Rican rainforest1470-160X10.1016/j.ecolind.2021.108413https://doaj.org/article/382479775ad14fd29d6828293e4fb66d2021-12-01T00:00:00Zhttp://www.sciencedirect.com/science/article/pii/S1470160X21010785https://doaj.org/toc/1470-160XEffective forest restoration requires tools for evaluating and comparing restoration approaches. Nevertheless, measuring restoration progress can be difficult and expensive. Passive acoustic monitoring (PAM) can be an inexpensive assessment strategy to collect large amounts of biodiversity information at scale. Nevertheless, analyzing and interpreting this information remains a difficult challenge. In this study we applied and compared three approaches to assess restoration treatments using recordings collected from PAM. We tested the hypothesis that variation in forest structure translates into differences in the species composition and acoustic signature of sites. For this purpose we used a reforestation experiment on the Osa Peninsula of Costa Rica, where we compared the mature forest to four restoration treatments. The treatments included natural regeneration and three treatments that varied the ratio of balsa, a pioneer tree species, and other native species. Our first approach consisted of visual and acoustic review of recordings to describe taxonomic groups found in each location. Our second approach consisted of measuring the acoustic energy present in the 10–30 kHz frequency band, an acoustic range primarily occupied by the mating signals of katydids and other insects, important elements of the food web and are often less mobile than birds and mammals. In our third approach we created 24-hour spectrograms that represented sites and treatments. Using the 24-hour spectrograms, we calculated a PCA and used a tSNE to evaluate the differences in acoustic signature and visualize clusters of treatments. The first approach revealed that relying on visual and acoustic review would fail to find the diel acoustic patterns that were captured in the other two approaches. The approaches varied substantially in the amount of acoustic data incorporated and the amount of human processing time. Subsampling recordings demonstrated that using only 10 sec instead of 40 sec per recording generated comparable results. The failure to differentiate among restoration treatments could reflect insensitivity in the approaches, but more likely represents the fact that the restoration plots are newly established and that substantial differentiation is more likely to arise during the time course of restoration.Álvaro Vega-HidalgoEleanor FlattAndrew WhitworthLaurel SymesElsevierarticleBioacousticsEcological restorationReforestationEcoacousticsSoundscapeKatydidsEcologyQH540-549.5ENEcological Indicators, Vol 133, Iss , Pp 108413- (2021)
institution	DOAJ
collection	DOAJ
language	EN
topic	Bioacoustics Ecological restoration Reforestation Ecoacoustics Soundscape Katydids Ecology QH540-549.5
spellingShingle	Bioacoustics Ecological restoration Reforestation Ecoacoustics Soundscape Katydids Ecology QH540-549.5 Álvaro Vega-Hidalgo Eleanor Flatt Andrew Whitworth Laurel Symes Acoustic assessment of experimental reforestation in a Costa Rican rainforest
description	Effective forest restoration requires tools for evaluating and comparing restoration approaches. Nevertheless, measuring restoration progress can be difficult and expensive. Passive acoustic monitoring (PAM) can be an inexpensive assessment strategy to collect large amounts of biodiversity information at scale. Nevertheless, analyzing and interpreting this information remains a difficult challenge. In this study we applied and compared three approaches to assess restoration treatments using recordings collected from PAM. We tested the hypothesis that variation in forest structure translates into differences in the species composition and acoustic signature of sites. For this purpose we used a reforestation experiment on the Osa Peninsula of Costa Rica, where we compared the mature forest to four restoration treatments. The treatments included natural regeneration and three treatments that varied the ratio of balsa, a pioneer tree species, and other native species. Our first approach consisted of visual and acoustic review of recordings to describe taxonomic groups found in each location. Our second approach consisted of measuring the acoustic energy present in the 10–30 kHz frequency band, an acoustic range primarily occupied by the mating signals of katydids and other insects, important elements of the food web and are often less mobile than birds and mammals. In our third approach we created 24-hour spectrograms that represented sites and treatments. Using the 24-hour spectrograms, we calculated a PCA and used a tSNE to evaluate the differences in acoustic signature and visualize clusters of treatments. The first approach revealed that relying on visual and acoustic review would fail to find the diel acoustic patterns that were captured in the other two approaches. The approaches varied substantially in the amount of acoustic data incorporated and the amount of human processing time. Subsampling recordings demonstrated that using only 10 sec instead of 40 sec per recording generated comparable results. The failure to differentiate among restoration treatments could reflect insensitivity in the approaches, but more likely represents the fact that the restoration plots are newly established and that substantial differentiation is more likely to arise during the time course of restoration.
format	article
author	Álvaro Vega-Hidalgo Eleanor Flatt Andrew Whitworth Laurel Symes
author_facet	Álvaro Vega-Hidalgo Eleanor Flatt Andrew Whitworth Laurel Symes
author_sort	Álvaro Vega-Hidalgo
title	Acoustic assessment of experimental reforestation in a Costa Rican rainforest
title_short	Acoustic assessment of experimental reforestation in a Costa Rican rainforest
title_full	Acoustic assessment of experimental reforestation in a Costa Rican rainforest
title_fullStr	Acoustic assessment of experimental reforestation in a Costa Rican rainforest
title_full_unstemmed	Acoustic assessment of experimental reforestation in a Costa Rican rainforest
title_sort	acoustic assessment of experimental reforestation in a costa rican rainforest
publisher	Elsevier
publishDate	2021
url	https://doaj.org/article/382479775ad14fd29d6828293e4fb66d
work_keys_str_mv	AT alvarovegahidalgo acousticassessmentofexperimentalreforestationinacostaricanrainforest AT eleanorflatt acousticassessmentofexperimentalreforestationinacostaricanrainforest AT andrewwhitworth acousticassessmentofexperimentalreforestationinacostaricanrainforest AT laurelsymes acousticassessmentofexperimentalreforestationinacostaricanrainforest
_version_	1718372961927299072

Acoustic assessment of experimental reforestation in a Costa Rican rainforest

Ejemplares similares