Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

Skyla Dennise U Diaz; Jelaine L Gan; Giovanni A Tapang

doi:10.1371/journal.pone.0289001

. 2023 Jul 28;18(7):e0289001. doi: 10.1371/journal.pone.0289001

Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

Skyla Dennise U Diaz ^1,^*, Jelaine L Gan ¹, Giovanni A Tapang ²

Editor: Luca Nelli³

PMCID: PMC10381043 PMID: 37506131

Abstract

We assessed eight acoustic indices as proxies for bird species richness in the National Science Complex (NSC), University of the Philippines Diliman. The acoustic indices were the normalized Acoustic Complexity Index (nACI), Acoustic Diversity Index (ADI), inverse Acoustic Evenness Index (1-AEI), Bioacoustic Index (BI), Acoustic Entropy Index (H), Temporal Entropy Index (Ht), Spectral Entropy Index (Hf), and Acoustic Richness Index (AR). Low-cost, automated sound recorders using a Raspberry Pi were placed in three sites at the NSC to continuously collect 5-min sound samples from July 2020 to January 2022. We selected 840 5-min sound samples, equivalent to 70 hours, through stratified sampling and pre-processed them before conducting acoustic index analysis on the raw and pre-processed data. We measured Spearman’s correlation between each acoustic index and bird species richness obtained from manual spectrogram scanning and listening to recordings. We compared the correlation coefficients between the raw and pre-processed.wav files to assess the robustness of the indices using Fisher’s z-transformation. Additionally, we used GLMMs to determine how acoustic indices predict bird species richness based on season and time of day. The Spearman’s rank correlation and GLMM analysis showed significant, weak negative correlations between the nACI, 1-AEI, Ht, and AR with bird species richness. The weak correlations suggest that the performance of acoustic indices are dependent on various factors, such as the local noise conditions, bird species composition, season, and time of day. Thus, ground-truthing of the acoustic indices should be done before applying them in studies. Among the eight indices, the nACI was the best-performing index, performing consistently across sites and independently of season and time of day. We highlight the importance of pre-processing sound data from urban settings and other noisy environments before acoustic index analysis, as this strengthens the correlation between index values and bird species richness.

Introduction

The presence of most bird species can be confirmed based on their distinct vocalizations [1]. Because of this, passive acoustic monitoring is suitable for studying bird communities [2, 3], and deploying sound recorders overcomes the logistical challenges posed by active point count and line-transect surveys [4]. If enough recorders are deployed, they can cover a greater area than human observers and can be left in the field for extended periods, especially if programmed to only record at specific time intervals [5]. In areas with mobile or Wi-Fi network connections, the collected digital sound files can automatically be transmitted to a central server for long-term data storage and analysis [6]. However, the challenge with acoustic surveys is the processing effort required to analyze extensive sound data collections and translate them into biodiversity measures that can give timely information relevant to species conservation and management [7].

Several ecoacoustic indices were proposed to summarize the acoustic features of soundscapes using the sound spectrum [8]. Moreover, acoustic indices are mathematical functions that measure the distribution of acoustic energy across time and frequency in a recording [9]. These indices enable the quick processing of sound data and measure the acoustic complexity, diversity, evenness, entropy, and richness of recordings. They are based on species diversity indices and are thought to characterize the variation in sound production in animal communities across time, serving as proxies for diversity metrics such as species abundance, richness, evenness, and diversity [10]. The ease with which these indices can be calculated using R packages, such as soundecology [11] and seewave [12], makes them attractive tools for rapid biodiversity assessment and monitoring.

Despite the growing usage of acoustic indices in ecoacoustic surveys, the relationships between index values and ground-truthed measures of biodiversity remain inconclusive [13]. The reference studies wherein these indices were developed reported positive correlations with bird species richness [14–18], but some studies found the opposite [19–21]. Retamosa Izaguirre et al. [22] and Moreno-Gómez et al. [23] evaluated acoustic indices as proxies for bird species richness, and both studies reported low to moderate correlations. Alcocer et al. [24] conducted a meta-analysis of acoustic indices as proxies for biodiversity and found that acoustic indices had a moderate positive relationship with the diversity metrics, such as species abundance, species richness, species diversity. As acoustic indices measure the amplitude and frequency properties of soundscapes, they cannot discern between species and may only weakly to moderately predict bird species richness.

Acoustic index values should be significantly correlated with the number of vocalizing bird species in recordings to reflect bird species richness accurately. However, field conditions can affect the performance of the acoustic indices [25, 26]. Noise and other environmental factors can confound the index values [27]. Outdoor environments have a diverse array of biophonies, such as sounds produced by birds, mammals, amphibians, and insects. Sounds from running water or rain falling through a canopy may also be present in the forest soundscape [28]. In rural and urban areas, geophony from breezes and drizzles are sources of bias when using acoustic indices [20]. At sites near farmlands, roads, or human settlements, sounds generated by vehicles, sirens, machines, and human speech may affect acoustic index values [27]. Unavoidable ambient sounds highlight the importance of pre-processing sound data [17, 18] and ground-truthing of the acoustic indices (i.e., validation of index values with survey data) before applying them in studies.

We assessed eight acoustic indices as proxies for bird species richness in an urban green space in Metro Manila. The anthropause [29], or slowdown of human activity during the COVID-19 pandemic, was an opportune time to measure bird vocalizations with reduced human disturbance. The acoustic indices assessed in the study were the normalized Acoustic Complexity Index (nACI), Acoustic Diversity Index (ADI), inverse Acoustic Evenness Index (1-AEI), Bioacoustic Index (BI), Acoustic Entropy Index (H), Temporal Entropy Index (Ht), Spectral Entropy Index (Hf), and Acoustic Richness Index (AR). We measured the correlation between each acoustic index and the bird species richness obtained from manual spectrogram scanning and listening to recordings. The correlations using two datasets, the unprocessed (raw) and the pre-processed sound files, were also compared to assess the robustness of the acoustic indices. We used generalized linear mixed models (GLMMs) to examine how acoustic indices predict bird species richness in pre-processed 5-min sound samples dependently on season and time of day. We hypothesized that if acoustic index values are significantly correlated with the number of species heard, then the selected indices can be used as proxies for bird species richness in the field.

Materials and methods

We conducted the study at the National Science Complex (NSC) in the University of the Philippines Diliman (UPD) campus, one of the last urban green spaces in Metro Manila, Philippines. The study area is home to 42 bird species, including endemic, resident, and migrant bird species [30]. Located within the NSC is the one-hectare UP Biology—EDC BINHI Threatened Species Arboretum, home to endemic and threatened tree species [31, 32] and various bird species [33]. The anthropause [29], or slowdown of human activity during the COVID-19 pandemic, provided an opportune time to measure bird vocalizations with reduced human disturbance. With COVID-19 restrictions and suspended face-to-face classes, entering the campus was restricted to authorized personnel only. Metro Manila has a Type I climate, marked by two pronounced seasons: the dry season from November to April, and the wet season from May to October. The months of June to September receive the highest rainfall [34].

Ethics statement

No specific ethical approval was needed because automated sound recorders were used to collect data remotely.

Study area

We established three recording sites, namely the Institute of Biology (A), the College of Science Administrative Building (B), and the College of Science Library (C) (Fig 1). The recording sites were selected in consideration of Wi-Fi and electricity access. The sites are adequately spaced (>100m) from each other to avoid overlapping between recording points. Site A directly faces the UP Biology—EDC BINHI Threatened Species Arboretum, but it is disturbed by ongoing construction and maintenance work. Site B is approximately 50 meters from the closest tree line and is surrounded by a road network. Site C is enclosed by vegetation, and the recorder is less than 10 meters from the closest tree line and about 30 meters from the nearest road. Guard posts are also located at the entrance going to Site C.

Fig 1 — This map was generated with data from OpenStreetMap [35] and Wikimedia Commons [36]. (A) Institute of Biology, (B) College of Science Administrative Building, and (C) College of Science Library. The inset shows the recording system (USB microphone attached to Raspberry Pi 3 Model B+).

Acoustic data collection

At each recording point, we placed a low-cost sound recorder built using an omnidirectional USB microphone attached to a Raspberry Pi 3 Model B+ (Fig 1) [37]. All three systems, running with Raspberry Pi OS, were programmed to record 5-min sound clips every 6-min, using a sampling rate of 44.1 kHz. This resulted in a 1-min gap between recording bouts. The systems were recording continuously from July 2020 to January 2022 to cover the wet and dry seasons and account for migratory birds. These were done using the program arecord [38], called in a script that also uploads the WAV files to Google Drive for long-term data storage and analysis.

Low-cost recorders may have varying sensitivities, resulting in different responses in terms of frequency range and signal intensity [39]. Thus, we tested the response of each low-cost recorder across a range of frequencies. Using version 3.0.0 of Audacity® recording and editing software [40], we generated a 30-sec tone from 1 Hz to 10,000 Hz and played it through a loudspeaker held at breast level, gradually increasing the distance from the recorder. We plotted the spectrum data of the 30-sec generated tone against the average response of each recorder (S1 Fig). The frequency response was not flat over the entire frequency range, and higher frequencies were increasingly attenuated, except for recorder A (S1 Fig).

Data pre-processing

We randomly selected 840.wav files from the database based on the YYYY-MM-DD-HH-mm timestamp, totaling 4,200 minutes (≈70 hours). Through stratified sampling, we selected 20 sound samples from each hour between 5:00 AM and 6:00 PM to cover the daily activity period, including the dawn and dusk chorus. This resulted in a total of 280 samples for each recording point. We did not remove recordings with heavy rain or strong wind from the sample.

We performed pre-processing on the 840.wav files by adding a high-pass filter (400.0 Hz) and performed subsequent noise reduction and signal amplification in Audacity® macros [40] to reduce anthropogenic and geophonic background noise [18]. We used the default settings to streamline the processing of multiple.wav files, and upon visual inspection, the noise reduction setting (12 dB) was enough to reduce the unwanted background noise.

Manual spectrogram scanning and listening to recordings

We used the 840 pre-processed sound samples for manual spectrogram scanning and listening to recordings. We viewed each 5-min sound sample in Audacity®, using Spectrogram View set to Linear Scale, with the Minimum Frequency set to 0 kHz and the Max Frequency set to 8 kHz (Fig 2). S. Diaz was the primary observer who analyzed all the recordings to reduce variability between observers. We listened to the selected recordings and identified all bird vocalizations up to the species level; thus, each WAV file had a list of species heard. Sound files from online repositories of recorded avian vocalizations, such as Xeno-canto [41] and eBird [42], were used as references in identifying bird species in the recordings. In addition, expert consultation from experienced birders (J. Gan, C. Española, and A. Constantino) helped in further species identification. Unidentified vocalizations with distinct time-frequency characteristics were still included as sonospecies.

Fig 2 — The x-axis represents the time expressed in seconds, while the y-axis represents the frequency expressed in kHz. Spectrogram parameters: linear scale, 8 kHz maximum frequency, and window size = 4096. Amplitude ranges from -60 dB to -44 dB (raw) and -60 dB to -29 dB (pre-processed).

The limitation of our study is that we did not measure bird species abundance because of the difficulty of telling apart individuals based simply on their calls. Hence, we focused on the relationship of acoustic indices with bird species richness.

Acoustic index analysis

We used two data sets for the acoustic index analysis: raw (direct from recording) and pre-processed.wav files. We calculated eight acoustic indices: Acoustic Complexity Index (ACI) [14], Acoustic Diversity Index (ADI) [15], Acoustic Evenness Index (AEI) [15], Bioacoustic Index (BI) [16], Acoustic Entropy Index (H) [17], Temporal Entropy Index (Ht) [17], Spectral Entropy Index (Hf) [17], and Acoustic Richness Index (AR) [18] using the soundecology [11] and seewave [12] R packages [43]. We used a sampling frequency of 44,100 Hz and a fast Fourier transform (FFT) window of 512 points (S1 Table). This corresponds to a frequency resolution of FR = 86.133 Hz and a time resolution of TR = 0.0116 s.

All acoustic indices, except H and AR, were calculated using the multiple_sounds function in soundecology. We used the default settings for the ACI, ADI, AEI, and BI (S1 Table). The maximum frequency of ADI and AEI was set to 10 kHz, and the frequency step was set to 1 kHz. The minimum frequency of BI was set to 2 kHz, and the maximum frequency was set to 8 kHz. We calculated the H and AR in seewave using the default settings (S1 Table). We extracted the Ht from the results of AR and each H value by its corresponding Ht value to obtain Hf.

Since the results given by ACI are accumulative (i.e., very long samples will return very large values) [11], we divided each ACI value by the ACI value of a silent recording (= ACI/8843.141) to get a range of values easier to compare, herein referred to as normalized Acoustic Complexity Index (nACI). We also calculated the inverse Acoustic Evenness Index (1-AEI) to facilitate comparability with the ADI [21]. Other indices had no further calculations.

Normalized Acoustic Complexity Index (nACI)

The ACI is calculated from a spectrogram divided into frequency bins and temporal steps [14]. This index calculates the absolute difference between adjacent sound intensity values in each frequency bin. The absolute differences for all frequency bins are summed to obtain the ACI for the entire recording [14]. Based on this formula, high ACI values correspond to intensity modulations with the assumption that bird vocalizations dominated the recording. In contrast, low ACI values correspond to constant intensities, increased insect buzz, and anthrophony (i.e., human-generated sounds). The ACI correlated positively with the number of bird vocalizations in a beech forest in Italy [14].

Acoustic Diversity Index (ADI) and Inverse Acoustic Evenness Index (1-AEI)

The ADI and AEI are calculated on a matrix of amplitude values extrapolated from a spectrogram divided into frequency bins. Both indices first obtain the proportion of amplitude values in each bin above a dB threshold [15]. The ADI applies the Shannon diversity index to these values, whereas the AEI applies the Gini index. A higher Shannon’s index indicates higher diversity, whereas a higher Gini index indicates lower evenness and diversity. Consequently, ADI values should correlate positively and AEI values negatively with bird species richness. In Indiana, USA, the ADI values were higher, and AEI values were lower during the dawn and dusk choruses, indicating the presence of more vocalizing bird species during these periods [15]. We calculated the inverse of AEI, herein referred to as the inverse Acoustic Evenness Index (1-AEI) to facilitate comparability between the two indices [21]. Higher ADI and 1-AEI should indicate higher frequency occupation of the spectrum and higher sound diversity.

Bioacoustic Index (BI)

The BI is calculated based on a spectrogram generated from a fast Fourier transform (FFT). This index obtains the area of the FFT (i.e., total sound energy) between 2 and 8 kHz, which is the typical frequency range of bird vocalizations [16]. The BI was strongly and positively correlated with avian abundance in Hawaii [16].

Acoustic Entropy Index (H): Temporal Entropy Index (Ht) and Spectral Entropy Index (Hf)

The H is calculated as the product of Ht and Hf, where Ht represents the Shannon entropy of the amplitude envelope, and Hf represents the Shannon entropy of the frequency spectrum [17]. A flatter amplitude envelope, or low entropy, indicates constant intensities, whereas higher H values indicate more intensity modulations [17]. In Tanzanian lowland coastal forests, H increased with bird species richness [17].

Acoustic Richness Index (AR)

The AR is based on Ht and the median of the amplitude envelope (M) [18]. Lower AR values (i.e., flatter amplitude envelope and lower entropy) indicate lower acoustic richness and the presence of fewer bird species. In contrast, higher AR values correspond to increased amplitude modulations, higher acoustic richness, and more bird species. In France, the AR correlated positively with bird species richness [18].

Statistical analysis

Spearman’s correlation

We calculated Spearman’s rank correlation between each acoustic index and the bird species richness obtained from manual spectrogram scanning and listening to recordings. We used two data sets: 840 raw and 840 pre-processed.wav files, herein referred to as pooled raw data and pooled pre-processed data. Additionally, we calculated Spearman’s rank correlation separately for each recording point. We calculated Fisher’s z-transformation to compare the correlations between the raw and pre-processed datasets.

Generalized linear mixed models (GLMMs)

We used generalized linear mixed models (GLMMs) to examine how acoustic indices predict bird species richness in pre-processed 5-min sound samples depending on season and time of day. For each sound sample, we specified the acoustic index as the dependent variable, bird species richness, season, and time of day as fixed effects, and the recording point as the random effect. We fitted a set of three models for each acoustic index, which included the full model and fixed effects models in R using the lme4 [44] and glmmTMB [45] packages (S2 Table). We used GLMMs to fit nACI, ADI, and BI with Gaussian distribution and identity link function [46]; AR with Gaussian distribution and log link function; and 1-AEI, H, Ht, and Hf with beta distribution [47]. We computed the corrected Akaike Information Criterion (AIC_c) using the AICcmodavg [48] R package and chose the model with the lowest AIC_c value as the best model for each acoustic index [49]. We considered all models with Δ AIC_c < 4 as equally plausible [49]. Lastly, we compared the plausible models using likelihood-ratio tests (LRT).

Results

Summary of bird species heard

From the 840 pre-processed 5-min sound samples analyzed, we heard 21 bird species from 20 families in the recordings from the NSC (Table 1). This represented only 52.5% of the 40 known species sighted at the NSC based on recent eBird checklists [50] submitted from July 2020 to January 2022. However, two species, namely the Philippine Nightjar and Pygmy Flowerpecker, were heard in the recordings but not included in the list because there is insufficient eBird data available for both species. Of the 21 detected bird species, seven (33%) are Philippine endemics (Table 1). Thirteen are resident species. The Brown Shrike is the only migrant species. Seventeen (81.0%) bird species were found at all three sites. Given the proximity of the three sites, the bird community across them is most likely shared. We assume that the bird abundance and bird richness are similar across the three sites.

Table 1. Summary of bird species heard and number of sound samples each species was detected out of 280 samples per recording point.

Family	Common Name	Scientific Name	Site A	Site B	Site C
Acanthizidae	Golden-bellied Gerygone	Gerygone sulphurea	32	127	143
Alcedinidae	Collared Kingfisher	Todiramphus chloris	46	76	105
Ardeidae	Black-crowned Night Heron	Nycticorax nycticorax	2	0	2
Artamidae	White-breasted Woodswallow	Artamus leucorynchus	3	2	0
Campephagidae	Pied Triller	Lalage nigra	2	5	3
Caprimulgidae	Philippine Nightjar*	Caprimulgus manillensis	2	1	2
Columbidae	Zebra Dove	Geopelia striata	1	2	11
Corvidae	Large-billed Crow	Corvus macrorhynchos	32	127	143
Dicaeidae	Red-keeled Flowerpecker*	Dicaeum australe	24	4	35
	Pygmy Flowerpecker*	Dicaeum pygmaeum	0	0	1
Laniidae	Brown Shrike	Lanius cristatus	13	6	1
Megalaimidae	Coppersmith Barbet	Psilopogon haemacephalus	0	7	31
Muscicapidae	Philippine Magpie-robin*	Copsychus mindanensis	5	1	1
Nectariniidae	Olive-backed Sunbird	Cinnyris jugularis	69	47	28
Oriolidae	Black-naped Oriole	Oriolus chinensis	130	121	86
Passeridae	Eurasian Tree Sparrow	Passer montanus	84	49	86
Picidae	Philippine Pygmy Woodpecker*	Yungipicus maculatus	15	3	4
Psittaculidae	Philippine Hanging Parrot*	Loriculus philippensis	21	16	33
Pycnonotidae	Yellow-vented Bulbul	Pycnonotus goiavier	115	155	97
Rhipiduridae	Philippine Pied Fantail*	Rhipidura nigritorquis	36	21	23
Zosteropidae	Lowland White-eye	Zosterops meyeni	18	66	8

Open in a new tab

From July 2020 to January 2022, 21 species from 20 families were heard in the National Science Complex. Asterisk (*) indicates Philippine endemic species.

An uneven distribution of species was heard in the recordings (Fig 3). The highest number of bird species heard in a single recording was seven. However, this only comprised twelve (1.43%) of the 840 pre-processed.wav files. Seventy percent (588.wav files) had one to four species recorded, and 15% (129.wav files) had zero bird species heard.

In 840 pre-processed 5-min sound samples, the highest mean richness was observed from recordings between 6:00 AM and 8:00 AM, indicating the dawn chorus (Fig 4). The dusk chorus was less pronounced.

In addition to bird sounds, we also detected various other biotic sounds, including those produced by mammals, amphibians, reptiles, and insects. These include the sounds of chickens, dogs, geckos, frogs, sheep, and crickets. We also identified anthrophonies, such as human speech, construction noise, and road traffic, as well as geophonies, such as rain and wind. Although we detected other non-avian sounds in our recordings, we did not take their frequencies into account.

Spearman’s correlation between acoustic indices and bird species richness

The Spearman’s correlation and Fisher’s z-transformation analyses revealed that bird species richness correlated more strongly with the pre-processed data than with the raw data (Tables 2 and 3). Regarding the pooled pre-processed data, five acoustic indices, namely the nACI, ADI, 1-AEI, Ht, and AR, had significant but weak negative correlations with the number of bird species heard in the recordings (Table 2). The three other acoustic indices, BI, H, and Hf, had no significant correlation with bird species richness.

Table 2. The Spearman’s rank correlation coefficients between acoustic indices and bird species richness.

Site	nACI	ADI	1-AEI	BI	H	Ht	Hf	AR
Raw (unprocessed)
Pooled	-0.10 ^**	-0.02	-0.02	-0.01	-0.07 ^*	0.02	-0.07*	-0.001
A	-0.25 ^***	-0.14 ^*	-0.15 ^*	-0.01	-0.19 ^**	0.02	-0.19 ^**	-0.09
B	-0.05	-0.01	-0.01	-0.16 ^**	-0.01	0.08	-0.02	0.01
C	-0.04	0.15 ^*	0.17 ^**	0.05	0.01	0.12	0.004	0.09
Pre-processed
Pooled	-0.15 ^****	-0.09 ^**	-0.09 ^**	-0.01	0.02	-0.10 ^**	0.05	-0.11 ^**
A	-0.25 ^****	-0.20 ^**	-0.20 ^***	0.01	-0.08	-0.05	-0.05	-0.13 ^*
B	-0.05	-0.10	-0.10	-0.18 ^**	0.15 ^*	-0.03	0.16 ^**	-0.04
C	-0.20 ^***	-0.05	-0.05	0.06	0.05	-0.20 ^***	0.10	-0.16 ^**

Open in a new tab

Based on the pooled pre-processed data, the nACI, ADI, 1-AEI, Ht, and AR had significant but weak negative correlations with bird species richness. Asterisks and bold indicate significance levels

* p < 0.05

** p < 0.01

*** p < 0.001.

Table 3. The results of Fisher’s z-transformation comparing Spearman’s rank correlation coefficients for the raw and pre-processed data sets.

Index	z_raw	z_{pre-processed}	z_diff	t	p
nACI	-0.104	-0.148	-0.044	-1.817	0.069
ADI	-0.023	-0.092	-0.069	-2.839	<0.01
1-AEI	-0.015	-0.093	-0.077	-3.165	<0.01
BI	-0.007	-0.006	0.001	0.040	0.968
H	-0.072	0.024	0.096	3.927	<0.001
Ht	0.022	-0.099	-0.121	-4.942	<0.001
Hf	-0.074	0.052	0.125	5.133	<0.001
AR	-0.001	-0.111	-0.110	-4.514	<0.001

Open in a new tab

Pre-processing significantly strengthened the ADI, 1-AEI, Ht, and AR correlations.

Fisher’s z-transformation results showed no significant improvement in the correlation between the nACI and bird species richness (Table 3). Pre-processing significantly strengthened the correlations for ADI, 1-AEI, Ht, and AR, whereas the correlations for H and Hf became significantly weaker (Table 3).

Generalized linear mixed models (GLMMs)

The 1-AEI, Ht, and AR predicted bird species richness in pre-processed 5-min sound samples, and they showed significant differences with season and time of day (Table 4). Specifically, the 1-AEI, Ht, and AR were significantly higher during the wet season than the dry season. The nACI also predicted bird species richness; however, we found no significant differences associated with season and time of day (Table 4). Conversely, the ADI, BI, H, and Hf did not significantly estimate bird species richness (Table 4). We found that season and time of day were stronger predictors for these indices.

Table 4. The results of GLMM examining how eight acoustic indices reflect bird species richness, season, and time of day, using the 840 pre-processed 5-min sound samples dataset.

	Coefficient	SE	t	p
	Normalized Acoustic Complexity Index (nACI)
Intercept	1.039	0.011	94.470	<0.001
Richness	-0.003	0.001	-4.698	<0.001
Season [wet]	-0.003	0.002	-1.422	0.156
Hour	-0.0004	0.0002	-1.651	0.099
	Acoustic Diversity Index (ADI)
Intercept	1.787	0.233	7.672	<0.01
Richness	-0.013	0.010	-1.348	0.178
Season [wet]	0.253	0.033	7.652	<0.001
Hour	-0.014	0.004	-3.285	<0.01
	Inverse Acoustic Evenness Index (1-AEI)
Intercept	0.574	0.447	1.283	0.200
Richness	-0.068	0.020	-3.417	<0.001
Season [wet]	0.369	0.067	5.529	<0.001
Hour	-0.021	0.009	-2.419	<0.05
	Bioacoustic Index (BI)
Intercept	7.735	2.552	3.031	0.087
Richness	-0.086	0.065	-1.335	0.182
Season [wet]	1.065	0.214	4.966	<0.001
Hour	-0.079	0.028	-2.822	<0.01
	Acoustic Entropy Index (H)
Intercept	1.714	0.152	11.282	<0.001
Richness	0.001	0.008	0.072	0.942
Season [wet]	0.075	0.028	2.693	<0.01
Hour	-0.008	0.004	-2.230	<0.05
	Temporal Entropy Index (Ht)
Intercept	3.915	0.162	24.201	<0.001
Richness	-0.030	0.010	-2.991	<0.01
Season [wet]	0.125	0.034	3.667	<0.001
Hour	-0.009	0.004	-2.103	<0.05
	Spectral Entropy Index (Hf)
Intercept	1.854	0.150	12.340	<0.001
Richness	0.006	0.009	0.684	0.494
Season [wet]	0.064	0.029	2.197	<0.05
Hour	-0.008	0.004	-2.137	<0.05
	Acoustic Richness Index (AR)
Intercept	-1.066	0.135	-7.880	<0.001
Richness	-0.072	0.018	-4.021	<0.001
Season [wet]	0.580	0.068	8.588	<0.001
Hour	-0.038	0.008	-5.028	<0.001

Open in a new tab

We specified the acoustic index as the dependent variable, bird species richness, season, and time of day as fixed effects, and the recording point as the random effect.

Discussion

Acoustic indices and bird species richness

Our study showed that the nACI, 1-AEI, Ht, and AR estimated bird species richness in pre-processed 5-min sound samples, showing a significant negative relationship between index values and bird richness. We found nACI to be the most robust, as it predicted bird species richness independently of season and time of day and consistently across sites. The negative relationship between nACI and bird species richness may be attributed to the overlapping vocalizations of multiple birds in the recordings. This leads to reduced differences in sound intensities and lower nACI values. The ACI also showed a significant negative correlation with bird species richness in a grassland ecosystem in the U.S.A [51] and a tropical rainforest in Costa Rica [52]. The negative correlation was explained by less complex bird calls at the study sites and the presence of dominant species, which resulted in less variable sound intensities [51, 52].

While the nACI was a more robust index, the 1-AEI, Ht, and AR were reliable predictors of bird species richness in urban settings when pre-processing was applied to the.wav files. After pre-processing, the 1-AEI, Ht, and AR showed stronger correlations with bird species richness (Table 3). Therefore, using the 1-AEI, Ht, and AR with pre-processed data can lead to better predictions of bird species richness in green spaces and other urban environments.

The negative correlation between 1-AEI and bird species richness could be due to bird species occupying a wide frequency range, resulting in high index values. Since recordings with heavy rain and intense anthrophony were also not excluded, these recordings yielded high 1-AEI values, despite having little to no audible bird species.

The Ht also had a significant negative relationship with bird species richness in a rainforest in Chile [23]. The low diversity of birds in the recordings and the presence of one or few dominant species may explain the negative correlation of the Ht with bird species richness. If only a few bird species vocalize at a time, the amplitude modulations become more distinct, and the envelope becomes less flat, returning higher Ht values.

In contrast, the BI, H, and Hf showed inconsistent and poor performance in predicting bird richness. The correlations remained inconsistent across three sites, even after pre-processing. Therefore, we do not recommend using the BI, H, and Hf as proxies for bird species richness in urban settings.

Given the performance of the eight acoustic indices assessed, we recommend using only the nACI, 1-AEI, Ht, and AR as proxies for bird species richness in urban settings. However, ground-truthing of the acoustic indices should be conducted before applying them to studies in different environments as their performance depends on various factors including, the local noise conditions, bird species composition of the site, season, and time of day.

Although most bird species in the NSC were heard at all three sites, different local noise conditions due to the construction and maintenance work at site A, the proximity of site B to the road, or site C to the tree line, resulted in inconsistent correlations between index values and bird species richness. However, pre-processing strengthened the correlation between select acoustic indices (ADI, 1-AEI, Ht, AR) and bird richness. Our study highlights the importance of pre-processing sound data from urban and noisy environments before using acoustic indices. If background noise has been preliminarily removed from recordings by applying amplitude threshold cut-off filters (e.g., high-pass filter), noise reduction, and signal amplification, then acoustic indices can perform consistently across diverse acoustic conditions and accurately reflect bird species richness [17, 18].

Conclusions

We demonstrate passive acoustic monitoring using low-cost sound recorders to study urban bird communities in the Philippines and attempt to address the challenge of processing and analyzing extensive sound data collections. To our knowledge, this is the first study assessing acoustic indices as proxies for bird species richness in an urban green space in the country. Among the eight acoustic indices assessed in the study, the nACI was the best and most robust, while the 1-AEI, Ht, and AR were also good predictors of bird species richness. We emphasize the importance of ground-truthing the indices before applying them to studies. Research exploring the acoustic indices and their relation to bird species richness is needed to understand the inconsistent correlations fully.

As the frequency response of our recorders was not flat over the entire frequency range, which impacts the calculation of acoustic indices, low-cost recorders should be calibrated before use in ecoacoustic studies to offset this limitation. We also recognize the importance of pre-processing sound data, such as applying high-pass filters, noise reduction, and signal amplification, to account for differences among recorders, remove unwanted ambient sounds, and improve the detection of bird species.

As acoustic indices were created to measure different characteristics of soundscapes quickly, they are not species-specific and lack sound recognition capabilities. For this reason, machine learning via convolutional neural networks (CNN) is increasingly being used in bioacoustic studies for birdsong classification [53–55]. Neural networks trained on Xeno-canto and Macaulay Library recordings have successfully classified North American and European bird species [53]. This remains to be explored in Philippine bird species. The use of other analysis, such as principal component analysis (PCA) [56] and cluster analysis [57], also help discriminate among different soundscapes in urban areas based on ecoacoustic indices.

Supporting information

S1 Fig. The frequency response of recorders A-C with respect to the generated tone (black line) in Audacity® [40].

(PDF)

Click here for additional data file.^{(109.4KB, pdf)}

S1 Table. Summary of parameters used in the calculation of acoustic indices.

(PDF)

Click here for additional data file.^{(130.4KB, pdf)}

S2 Table. The results of GLMM examining how eight acoustic indices reflect bird species richness, season, and time of day, using the dataset of 840 pre-processed 5-min sound samples.

(PDF)

Click here for additional data file.^{(149.7KB, pdf)}

S3 Table. The results of the acoustic index analysis for 840 raw 5-min sound samples.

(PDF)

Click here for additional data file.^{(524.1KB, pdf)}

S4 Table. The results of the acoustic index analysis for 840 pre-processed 5-min sound samples.

(PDF)

Click here for additional data file.^{(509.5KB, pdf)}

Acknowledgments

The authors appreciate the help of expert birders, Dr. Carmela P. Española and Adrian Constantino, in species identification.

Data Availability

All relevant data are within the manuscript and its Supporting Information files.

Funding Statement

The author(s) received no specific funding for this work.

References

1.Byers BE, Kroodsma DE. Avian vocal behavior. Cornell Lab of Ornithology. Handbook of Bird Biology. 2016:355–405. [Google Scholar]
2.Bardeli R, Wolff D, Kurth F, Koch M, Tauchert KH, Frommolt KH. Detecting bird sounds in a complex acoustic environment and application to bioacoustic monitoring. Pattern Recognition Letters. 2010. Sep 1;31(12):1524–34. doi: 10.1016/j.patrec.2009.09.014 [DOI] [Google Scholar]
3.Penar W, Magiera A, Klocek C. Applications of bioacoustics in animal ecology. Ecological complexity. 2020. Aug 1;43:100847. doi: 10.1016/j.ecocom.2020.100847 [DOI] [Google Scholar]
4.Abrahams C. Bird bioacoustic surveys-developing a standard protocol. In Practice. 2018(102):20–3. [Google Scholar]
5.Zwart MC, Baker A, McGowan PJ, Whittingham MJ. The use of automated bioacoustic recorders to replace human wildlife surveys: an example using nightjars. PloS one. 2014. Jul 16;9(7):e102770. doi: 10.1371/journal.pone.0102770 [DOI] [PMC free article] [PubMed] [Google Scholar]
6.Aide TM, Corrada-Bravo C, Campos-Cerqueira M, Milan C, Vega G, Alvarez R. Real-time bioacoustics monitoring and automated species identification. PeerJ. 2013. Jul 16;1:e103. doi: 10.7717/peerj.103 [DOI] [PMC free article] [PubMed] [Google Scholar]
7.Gibb R, Browning E, Glover‐Kapfer P, Jones KE. Emerging opportunities and challenges for passive acoustics in ecological assessment and monitoring. Methods in Ecology and Evolution. 2019. Feb;10(2):169–85. doi: 10.1111/2041-210X.13101 [DOI] [Google Scholar]
8.Eldridge A, Casey M, Moscoso P, Peck M. A new method for ecoacoustics? Toward the extraction and evaluation of ecologically-meaningful soundscape components using sparse coding methods. PeerJ. 2016. Jun 30;4:e2108. doi: 10.7717/peerj.2108 [DOI] [PMC free article] [PubMed] [Google Scholar]
9.Sueur J. Sound analysis and synthesis with R. Cham: Springer; 2018. Jun 6. doi: 10.1007/978-3-319-77647-7 [DOI] [Google Scholar]
10.Sueur J, Farina A, Gasc A, Pieretti N, Pavoine S. Acoustic indices for biodiversity assessment and landscape investigation. Acta Acustica united with Acustica. 2014. Jul 1;100(4):772–81. doi: 10.3813/AAA.918757 [DOI] [Google Scholar]
11.Villanueva-Rivera LJ, Pijanowski BC. Soundecology: soundscape ecology. R package version 1.3.3. 2018. Available from: https://cran.r-project.org/package=soundecology [Google Scholar]
12.Sueur J, Aubin T, Simonis C. Seewave: sound analysis and synthesis. R package version 2.2.0. 2022. Available from: https://CRAN.R-project.org/package=seewave [Google Scholar]
13.Bateman J, Uzal A. The relationship between the Acoustic Complexity Index and avian species richness and diversity: a review. Bioacoustics. 2022. Sep 3;31(5):614–27. doi: 10.1080/09524622.2021.2010598 [DOI] [Google Scholar]
14.Pieretti N, Farina A, Morri D. A new methodology to infer the singing activity of an avian community: The Acoustic Complexity Index (ACI). Ecological indicators. 2011. May 1;11(3):868–73. doi: 10.1016/j.ecolind.2010.11.005 [DOI] [Google Scholar]
15.Villanueva-Rivera LJ, Pijanowski BC, Doucette J, Pekin B. A primer of acoustic analysis for landscape ecologists. Landscape ecology. 2011. Nov;26(9):1233–46. doi: 10.1007/s10980-011-9636-9 [DOI] [Google Scholar]
16.Boelman NT, Asner GP, Hart PJ, Martin RE. Multi‐trophic invasion resistance in Hawaii: bioacoustics, field surveys, and airborne remote sensing. Ecological Applications. 2007. Dec;17(8):2137–44. doi: 10.1890/07-0004.1 [DOI] [PubMed] [Google Scholar]
17.Sueur J, Pavoine S, Hamerlynck O, Duvail S. Rapid acoustic survey for biodiversity appraisal. PloS one. 2008. Dec 30;3(12):e4065. doi: 10.1371/journal.pone.0004065 [DOI] [PMC free article] [PubMed] [Google Scholar]
18.Depraetere M, Pavoine S, Jiguet F, Gasc A, Duvail S, Sueur J. Monitoring animal diversity using acoustic indices: implementation in a temperate woodland. Ecological Indicators. 2012. Feb 1;13(1):46–54. doi: 10.1016/j.ecolind.2011.05.006 [DOI] [Google Scholar]
19.Machado RB, Aguiar L, Jones G. Do acoustic indices reflect the characteristics of bird communities in the savannas of Central Brazil?. Landscape and Urban Planning. 2017. Jun 1;162:36–43. doi: 10.1016/j.landurbplan.2017.01.014 [DOI] [Google Scholar]
20.Mammides C, Goodale E, Dayananda SK, Kang L, Chen J. Do acoustic indices correlate with bird diversity? Insights from two biodiverse regions in Yunnan Province, south China. Ecological Indicators. 2017. Nov 1;82:470–7. doi: 10.1016/j.ecolind.2017.07.017 [DOI] [Google Scholar]
21.Dröge S, Martin DA, Andriafanomezantsoa R, Burivalova Z, Fulgence TR, Osen K, et al. Listening to a changing landscape: Acoustic indices reflect bird species richness and plot-scale vegetation structure across different land-use types in north-eastern Madagascar. Ecological Indicators. 2021. Jan 1;120:106929. doi: 10.1016/j.ecolind.2020.106929 [DOI] [Google Scholar]
22.Retamosa Izaguirre M, Barrantes-Madrigal J, Segura Sequeira D, Spínola-Parallada M, Ramírez-Alán O. It is not just about birds: what do acoustic indices reveal about a Costa Rican tropical rainforest?. Neotropical Biodiversity. 2021. Jan 1;7(1):431–42. doi: 10.1080/23766808.2021.1971042 [DOI] [Google Scholar]
23.Moreno-Gómez FN, Bartheld J, Silva-Escobar AA, Briones R, Márquez R, Penna M. Evaluating acoustic indices in the Valdivian rainforest, a biodiversity hotspot in South America. Ecological Indicators. 2019. Aug 1;103:1–8. doi: 10.1016/j.ecolind.2019.03.024 [DOI] [Google Scholar]
24.Alcocer I, Lima H, Sugai LS, Llusia D. Acoustic indices as proxies for biodiversity: a meta‐analysis. Biological Reviews. 2022. Dec;97(6):2209–36. doi: 10.1111/brv.12890 [DOI] [PMC free article] [PubMed] [Google Scholar]
25.Gasc A, Pavoine S, Lellouch L, Grandcolas P, Sueur J. Acoustic indices for biodiversity assessments: Analyses of bias based on simulated bird assemblages and recommendations for field surveys. Biological Conservation. 2015. Nov 1;191:306–12. doi: 10.1016/j.biocon.2015.06.018 [DOI] [Google Scholar]
26.Eldridge A, Guyot P, Moscoso P, Johnston A, Eyre-Walker Y, Peck M. Sounding out ecoacoustic metrics: Avian species richness is predicted by acoustic indices in temperate but not tropical habitats. Ecological Indicators. 2018. Dec 1;95:939–52. doi: 10.1016/j.ecolind.2018.06.012 [DOI] [Google Scholar]
27.Fairbrass AJ, Rennert P, Williams C, Titheridge H, Jones KE. Biases of acoustic indices measuring biodiversity in urban areas. Ecological Indicators. 2017. Dec 1;83:169–77. doi: 10.1016/j.ecolind.2017.07.064 [DOI] [Google Scholar]
28.Pijanowski BC, Farina A, Gage SH, Dumyahn SL, Krause BL. What is soundscape ecology? An introduction and overview of an emerging new science. Landscape ecology. 2011. Nov;26(9):1213–32. doi: 10.1007/s10980-011-9600-8 [DOI] [Google Scholar]
29.Rutz C, Loretto MC, Bates AE, Davidson SC, Duarte CM, Jetz W, et al. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife. Nature Ecology & Evolution. 2020. Sep;4(9):1156–9. doi: 10.1038/s41559-020-1237-z [DOI] [PubMed] [Google Scholar]
30.Vallejo BM Jr, Aloy AB, Ong PS. The distribution, abundance and diversity of birds in Manila’s last greenspaces. Landscape and Urban Planning. 2009. Feb 15;89(3–4):75–85. doi: 10.1016/j.landurbplan.2008.10.013 [DOI] [Google Scholar]
31.Energy Development Corporation. EDC establishes threatened species arboretum in UP. 2014. December 11. Available from: https://www.energy.com.ph/2014/12/11/edc-establishes-threatened-species-arboretum-in-up/ [Google Scholar]
32.Cabel IC. Walk and learn to save PH native tree species. Philippine Daily Inquirer. 2019. April 8. Available from: https://newsinfo.inquirer.net/1104179/walk-and-learn-to-save-ph-native-tree-species [Google Scholar]
33.Villasper J. Destinations: The University of the Philippines Diliman Campus–Choose Your Own Adventure. Birdwatch. 2020. March 2. Available from: http://birdwatch.ph/2020/03/02/destinations-the-university-of-the-philippines-diliman-campus-choose-your-own-adventure-2/ [Google Scholar]
34.Philippine Atmospheric, Geophysical and Astronomical Services Administration (PAGASA). Climate of the Philippines. Available from: https://www.pagasa.dost.gov.ph/information/climate-philippines [Google Scholar]
35.OpenStreetMap Contributors, CC BY-SA 2.0. 2015. Available from: https://openstreetmap.org [Google Scholar]
36.HueMan1—Own work, CC BY-SA 4.0. Blank map of the Philippines. 2021. Available from: https://commons.wikimedia.org/w/index.php?curid=108050493 [Google Scholar]
37.Fletcher AC, Mura C. Ten quick tips for using a Raspberry Pi. PLoS computational biology. 2019. May 2;15(5):e1006959. doi: 10.1371/journal.pcbi.1006959 [DOI] [PMC free article] [PubMed] [Google Scholar]
38.Kysela J. arecord (1): command-line sound recorder and player for ALSA. [Cited 2021 August 31]. Available from: http://manpages.org/arecord [Google Scholar]
39.Benocci R, Potenza A, Bisceglie A, Roman HE, Zambon G. Mapping of the Acoustic Environment at an Urban Park in the City Area of Milan, Italy, Using Very Low-Cost Sensors. Sensors. 2022. May 6;22(9):3528. doi: 10.3390/s22093528 [DOI] [PMC free article] [PubMed] [Google Scholar]
40.Team Audacity. Audacity®: free audio editor and recorder [Computer application]. Version 3.0.0. 2021. Audacity® software is copyright © 1999–2021 Audacity Team. Available from: https://www.audacityteam.org/. It is free software distributed under the terms of the GNU General Public License. The name Audacity® is a registered trademark. [Google Scholar]
41.Planqué B, Vellinga WP, Pieterse S, Jongsma J, de By RA. Xeno-canto. 2005. Available from: https://www.xeno-canto.org [Google Scholar]
42.Sullivan BL, Wood CL, Iliff MJ, Bonney RE, Fink D, Kelling S. eBird: A citizen-based bird observation network in the biological sciences. Biological conservation. 2009. Oct 1;142(10):2282–92. Available from: https://www.ebird.org/ [Google Scholar]
43.R Development Core Team. The R Project for Statistical Computing. R Version 4.0.0. 2020. Available from: https://www.r-project.org/ [Google Scholar]
44.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear Mixed-Effects Models using ’Eigen’ and S4. R package version 1.1–32. 2023. Available from: https://cran.r-project.org/package=lme4 [Google Scholar]
45.Magnusson A, Skaug H, Nielsen A, Berg C, Kristensen K, Maechler M, et al. glmmTMB: Generalized Linear Mixed Models using Template Model Builder. R package version 1.1.6. 2023. Available from: https://CRAN.R-project.org/package=glmmTMB [Google Scholar]
46.Budka M, Sokołowska E, Muszyńska A, Staniewicz A. Acoustic indices estimate breeding bird species richness with daily and seasonally variable effectiveness in lowland temperate Białowieża forest. Ecological Indicators. 2023. Apr 1;148:110027. doi: 10.1016/j.ecolind.2023.110027 [DOI] [Google Scholar]
47.Sánchez‐Giraldo C, Bedoya CL, Morán‐Vásquez RA, Isaza CV, Daza JM. Ecoacoustics in the rain: understanding acoustic indices under the most common geophonic source in tropical rainforests. Remote Sensing in Ecology and Conservation. 2020. Sep;6(3):248–61. doi: doi.org/10.1002/rse2.162 [Google Scholar]
48.Mazerolle MJ. AICcmodavg: Model Selection and Multimodel Inference Based on (Q)AIC(c). R package version 2.3–2. 2023. Available from: https://CRAN.R-project.org/package=AICcmodavg [Google Scholar]
49.Burnham KP, Anderson DR. Model selection and multimodel inference: A practical information-theoretic approach. New York, NY: Springer New York; 2002. doi: 10.1007/b97636 [DOI] [Google Scholar]
50.Ebird. eBird: An online database of bird distribution and abundance. eBird, Cornell Lab of Ornithology. UP Diliman—MSI Hotspot. 2022. Available from: https://ebird.org/hotspot/L10877450?yr=all&m=&rank=mrec [Google Scholar]
51.Shamon H, Paraskevopoulou Z, Kitzes J, Card E, Deichmann JL, Boyce AJ, et al. Using ecoacoustics metrices to track grassland bird richness across landscape gradients. Ecological Indicators. 2021. Jan 1;120:106928. doi: 10.1016/j.ecolind.2020.106928 [DOI] [Google Scholar]
52.Retamosa Izaguirre MI, Ramírez-Alán O, De la O Castro J. Acoustic indices applied to biodiversity monitoring in a Costa Rica dry tropical forest. Journal of Ecoacoustics. 2018. Feb 26;2. doi: 10.22261/JEA.TNW2NP [DOI] [Google Scholar]
53.Kahl S, Wood CM, Eibl M, Klinck H. BirdNET: A deep learning solution for avian diversity monitoring. Ecological Informatics. 2021. Mar 1;61:101236. doi: 10.1016/j.ecoinf.2021.101236 [DOI] [Google Scholar]
54.Marchal J, Fabianek F, Aubry Y. Software performance for the automated identification of bird vocalisations: the case of two closely related species. Bioacoustics. 2022. Jul 4;31(4):397–413. doi: 10.1080/09524622.2021.1945952 [DOI] [Google Scholar]
55.Liu J, Zhang Y, Lv D, Lu J, Xie S, Zi J, et al. Birdsong classification based on ensemble multi-scale convolutional neural network. Scientific Reports. 2022. May 23;12(1):1–4. doi: 10.1038/s41598-022-12121-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
56.Benocci R, Roman HE, Bisceglie A, Angelini F, Brambilla G, Zambon G. Eco-acoustic assessment of an urban park by statistical analysis. Sustainability. 2021. Jul 14;13(14):7857. doi: 10.3390/su13147857 [DOI] [Google Scholar]
57.Benocci R, Brambilla G, Bisceglie A, Zambon G. Eco-acoustic indices to evaluate soundscape degradation due to human intrusion. Sustainability. 2020. Dec 14;12(24):10455. doi: 10.3390/su122410455 [DOI] [Google Scholar]

PLoS One. doi: 10.1371/journal.pone.0289001.r001

Decision Letter 0

Cihun-Siyong Alex Gong

2 Feb 2023

PONE-D-22-34645Acoustic indices as proxies for bird species richness in an urban green space in Metro ManilaPLOS ONE

Dear Dr. Diaz,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please submit your revised manuscript by Mar 19 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: https://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols. Additionally, PLOS ONE offers an option for publishing peer-reviewed Lab Protocol articles, which describe protocols hosted on protocols.io. Read more information on sharing protocols at https://plos.org/protocols?utm_medium=editorial-email&utm_source=authorletters&utm_campaign=protocols.

We look forward to receiving your revised manuscript.

Kind regards,

Cihun-Siyong Alex Gong, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. We note that Figure 1 in your submission contain [map/satellite] images which may be copyrighted. All PLOS content is published under the Creative Commons Attribution License (CC BY 4.0), which means that the manuscript, images, and Supporting Information files will be freely available online, and any third party is permitted to access, download, copy, distribute, and use these materials in any way, even commercially, with proper attribution. For these reasons, we cannot publish previously copyrighted maps or satellite images created using proprietary data, such as Google software (Google Maps, Street View, and Earth). For more information, see our copyright guidelines: http://journals.plos.org/plosone/s/licenses-and-copyright.

We require you to either (1) present written permission from the copyright holder to publish these figures specifically under the CC BY 4.0 license, or (2) remove the figures from your submission:

a. You may seek permission from the original copyright holder of Figure 1 to publish the content specifically under the CC BY 4.0 license.

We recommend that you contact the original copyright holder with the Content Permission Form (http://journals.plos.org/plosone/s/file?id=7c09/content-permission-form.pdf) and the following text:

“I request permission for the open-access journal PLOS ONE to publish XXX under the Creative Commons Attribution License (CCAL) CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). Please be aware that this license allows unrestricted use and distribution, even commercially, by third parties. Please reply and provide explicit written permission to publish XXX under a CC BY license and complete the attached form.”

Please upload the completed Content Permission Form or other proof of granted permissions as an "Other" file with your submission.

In the figure caption of the copyrighted figure, please include the following text: “Reprinted from [ref] under a CC BY license, with permission from [name of publisher], original copyright [original copyright year].”

b. If you are unable to obtain permission from the original copyright holder to publish these figures under the CC BY 4.0 license or if the copyright holder’s requirements are incompatible with the CC BY 4.0 license, please either i) remove the figure or ii) supply a replacement figure that complies with the CC BY 4.0 license. Please check copyright information on all replacement figures and update the figure caption with source information. If applicable, please specify in the figure caption text when a figure is similar but not identical to the original image and is therefore for illustrative purposes only.

The following resources for replacing copyrighted map figures may be helpful:

USGS National Map Viewer (public domain): http://viewer.nationalmap.gov/viewer/

The Gateway to Astronaut Photography of Earth (public domain): http://eol.jsc.nasa.gov/sseop/clickmap/

Maps at the CIA (public domain): https://www.cia.gov/library/publications/the-world-factbook/index.html and https://www.cia.gov/library/publications/cia-maps-publications/index.html

NASA Earth Observatory (public domain): http://earthobservatory.nasa.gov/

Landsat: http://landsat.visibleearth.nasa.gov/

USGS EROS (Earth Resources Observatory and Science (EROS) Center) (public domain): http://eros.usgs.gov/#

Natural Earth (public domain): http://www.naturalearthdata.com/

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

Reviewer #3: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: No

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: No

Reviewer #3: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have performed an interesting study on natural sound from bird species in an urban environment, using well-known and standard acoustic indices. They obtain robust results using pre-processing data by analysing cross-correlations between indices and bird species, thus discerning more accurately their effective applications to classify such natural sounds. While I can recommend the publication of the paper in PLOS ONE, I have few suggestions which the authors may consider for improvement of the presentation.

(1) Fig.3: The data for ADI and 1-AEI look identical. Is there a reason for that coincidence? Similar observation holds for H and Hf. Can you give us a hint for this?

(2) The different correlations observed at different sites should also reflect the different local noise conditions. Can you perhaps estimate such effects in order to understand the variations of correlation values found?

(3) There is quite recent literature about birds sounds in urban environments that the authors seem not to be aware of. I recommend to check the following references:

(3a) Benocci, et al. Mapping of the Acoustic Environment at an Urban Park...

Sensors 22, 3528 (2022).

(3b) Benocci, R., et al. Auto-correlations and long time memory of environment sound..

Ecological Indicators 134, 108492 (2022).

Reviewer #2: The study examines effectiveness of acoustic indices in bird species richness estimation in urban area. Similar studies have been conducted in various environments, so it is not novel approach. However inconsistent results reported in various studies point out that we still need more examples of such comparisons to understand where and when acoustic indices can be used as an approximation of biodiversity. From this perspective the study is valuable.

Authors identified bird species in 840 5-min sound samples (collected across more than one year) by manual spectrogram scanning and listening to recordings, calculated acoustic indices, and compared it by using Spearman’s correlation. This is very simple approach and, in my opinion, does not fully explore data. Authors reported weak or insignificant correlations between bird species richness and acoustic indices. The limitation of the study is sample size (only 3 recording points). The strong side of the study is that authors sampled soundscape across whole year. There is not a lot of studies analysing effectiveness of acoustic indices across a day and season. I think that daily and seasonally variable effectiveness of acoustic indices may be the main reason of found weak correlations. Did you check how well work acoustic indices when you compare separately dawn chorus, day and dusk chorus, or breeding vs non-breeding season? I suggest re-analysing data that way. I do not know how sound samples are distributed across day and season, maybe it will be necessary to analyse more sound samples to get optimal distribution, but authors can examine how each acoustic index estimate bird species richness dependently on time in a day and season (at least breeding vs non-breeding season). They can do this in simple way applying GLMMs in which they will specify the index for each sound sample as a dependent variable, number of detected bird species, time in a season, time in a day and one-way interactions as fixed effects, point id as a random effect. At the current stage of analysis, in my opinion, it will be difficult to publish results of the study. Also description of the methods should be more precise. I suggest considering reorganization of discussion and describing all indices together instead each of them separately. I suggest adding raw data as a supplementary materials.

More detailed comments:

L36: …840, 5-min sound samples.

L36-37: You used raw and pre-processed data.

L37: Spearman’s correlation

L38: …manual spectrogram scanning and listening to recordings.

L47: Please remove solely.

L52-53: Mobile network also serves such possibility and seems to be more useful than wi-fi.

L60-61: Not only these packages and not only R, but I agree that calculation is easy.

L64-65: Please remember that some indices correlate negatively with soundscape complexity.

L66-67: Please look at this review https://doi.org/10.1111/brv.12890, it can be useful.

L71-80: Consider mentioning about complexity and variation of birds vocalization – two different species (or even individuals within a species) will generate different values of acoustic indices.

L78-80: Great point. But also specific species composition may be one of the main factors responsible for effectiveness of acoustic indices.

L91: Not many papers reported strong correlation, it is rather moderate or weak.

L96: ...is home to 42 bird species, including endemic...

L96-98: How many of them are breeding species, how many are sedentary? Please write how long is a breeding season and when it has place.

L108-109: The distance between recorders was ca 100 m what means that you probably recorded the same individuals from different recording points (individuals between point A, B and C probably were recorded by three recorders). Most studies report detection distance for most of songbirds ranging between 100 and 200 m. Do you have information about detection distance of used by you equipment? Recording of the same individuals from various points do not limits your study, because soundscape recorded in each point is different.

L121-122: Please write about frequency response, signal to noise ratio and detection distance of recorders. Did you calibrate recorders before using?

L124:125: Impressive. Do you use these data in silent cities project?

L129-130: 840 five minutes sound samples. Please describe how did you select recording? Was it random choose? How was the distribution of sound samples across the year (because you recorded soundscape more than year, does it mean that months from July to June were sampled twice?)? Was it the same number of sound samples analysed per recording point? Did you validate recordings before analyzing (quality, noise, species richness)?

L130: …daily activity period, including dawn and dusk chorus.

L132: Which software and settings of spectrogram did you use? How many observers analysed recordings and if they were experienced in birds vocalisations? Did you recognise the species based on songs, calls or both? This is important to point out, because in the breeding season the same species richness will generate different complexity of soundscape than in non-breeding season.

L136-137: How commonly vocalizations of animals other than birds were recorded? The information about other biological sources of sound in your study area is missing.

L140: Similar effect you can get modifying setting of acoustic indices calculation.

L150-154: Please add a table (e.g., as a supply mat) in which you will show settings for each index.

L155-159: This point is not clear. Did you calculate indices for 5-min sound samples? If yes, you had constant sound-sample duration and you can use raw or standarised indices values for further analyses. Please be precise, describing the dataset used for manual spectrogram scanning and listening to recordings (840 sound samples) and calculating indices (840 sound samples or whole dataset).

L138: Please look at this review: https://doi.org/10.1080/09524622.2021.2010598

L175-176: So these two indices should correlate negatively and positively with bird species richness.

L197-199: It is very simple method. I think that you could use more sophisticated analyses to show how acoustic indices estimate bird species richness dependently on time in a day and time in a season. For example you can use GLMM for each index to check how the index predict bird species richness dependently on time in a day, season and recording point. Such analysis will considerably improve sounds of your study. When you use Spearman’s rho you should demonstrate evidence of a monotonic relationship and lack of severe skewness in raw data. If I good understood you compared the number of bird species detected manually in 840 sound samples with many acoustic indices calculated on the same 840 sound samples. I do not see test comparing effectiveness of indices on raw and pre-processing data, similarly like between recording points, but you write about it in discussion.

L202-203: 840 5-min sound samples? I suggest adding histogram (or modifying Table 1) to show how many sound samples each bird species was detected.

L205-206: Why?

L206-209: You do not need mention them here, please add marker in Table 1 which species are endemic.

L213: I suggest to show in how many sound samples each species was detected in each recording point instead markers.

L241: You showed the same in the table and on fig. It would be better to see raw data on the fig., e.g., number of bird species vs index value for each sound sample and the mean/median +/- 95%cu. If you would like to check difference in bird species richness estimation between recording points I suggest using GLMM with point ID as an effect. Additionally in such model you probably would be able to test the effect of time in a day and season on effectiveness of acoustic indices. Please consider such approach.

L247: In my opinion discussion needs reorganization. I suggest writing it more holistically (like in conclusions) instead describing each index separately. When you write whether pre-processing data increase effectiveness of the index in bird species richness estimation you need to support it by statistical test. More references to effectiveness of acoustic indices in estimation of bird species richness in urban areas are needed.

L249-255: In your study you did not test the effectiveness of soundscape recorders in comparison to human observers. Therefore this part of discussion is not necessary. You estimated bird species richness based on soundscape recording and compared it with acoustic indices, what is appropriate approach.

L343-344: There was a few studies examining effectiveness of acoustic indices in urban areas from other regions. Please look at them.

Reviewer #3: The paper “Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila” describes the correlation results between bird species richness and eco-acoustic indices.

The paper is clear but requires supplementary analysis before being accepted. It focuses just on a specific aspect of the soundscape recorded at three sites linked to bird species richness. Generally, eco acoustic indices represent summaries of the whole soundscape and highlight specific features of the frequency spectrum. This concept should be included in the introduction.

More in detail in the Material and Methods section (Data collection), you should include the sampling rate of the recorder and the frequency response. The latter is important especially for low-cost sensors.

Sect. Manual aural identification: Here, report also all other sound sources identified and their frequencies. This is important especially in urban green areas

Sect. Acoustic index analysis: when you introduce the pre-processing, please report an example of a spectrum for a raw and pre-processed audio file.

All indices are based upon a FFT computation of audio recordings. Please specify the number of FFT points use, the frequency and time resolution.

In Sect ADI and IAEI , it is reported that each frequency bin represents a particular bird species. This is NOT true!!

At line 176, I would say that higher ADI and 1-AEI indicate higher frequency occupation of the spectrum.

Fig 2 is not clear: As ordinate axes use a clearer title. Here you should report also differences among the sites in terms of richness and abundance of bird species. This is important because the correlation between indices depends also on the species abundance.

If possible include a temporal trend of the mean bird richness as well as for the eco-acoustic indices.

Lines 264-266: this part needs to be improved

Line 272-273. This is why abundance of birds should be included in the analysis.

Line 303-304: this is why you need to include a more detailed aural survey.

In the Conclusions, you should also stress that, besides CNN and machine learning, an approach which implies the use of statistics is also able to help discriminate among different soundscapes in urban areas based upon eco-acoustic indices, as reported in:

Benocci, R.; Roman, H.E.; Bisceglie, A.; Angelini, F.; Brambilla, G.; Zambon, G. Eco-Acoustic Assessment of an Urban Park by Statistical Analysis. Sustainability 2021, 13, 7857.

Benocci, R.; Brambilla, G.; Bisceglie, A.; Zambon, G. Eco-Acoustic Indices to Evaluate Soundscape Degradation Due to Human Intrusion. Sustainability 2020, 12, 10455.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: No

Reviewer #3: No

**********

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

PLoS One. 2023 Jul 28;18(7):e0289001. doi: 10.1371/journal.pone.0289001.r002

Author response to Decision Letter 0

2 Apr 2023

Author Rebuttals to Initial Comments:

Reviewer #1

The authors have performed an interesting study on natural sound from bird species in an urban environment, using well-known and standard acoustic indices. They obtain robust results using pre-processing data by analysing cross-correlations between indices and bird species, thus discerning more accurately their effective applications to classify such natural sounds. While I can recommend the publication of the paper in PLOS ONE, I have few suggestions which the authors may consider for improvement of the presentation.

(1) Fig.3: The data for ADI and 1-AEI look identical. Is there a reason for that coincidence? Similar observation holds for H and Hf. Can you give us a hint for this?

The ADI and AEI divide the spectrogram into frequency bins and take the proportion of amplitude values in each bin above a dB threshold [15]. The ADI applies the Shannon diversity index to these values, whereas the AEI applies the Gini index. A higher Shannon’s index indicates higher diversity, whereas a higher Gini index indicates lower evenness and diversity. ADI values should correlate positively and AEI values negatively with bird species richness.

We calculated the inverse of AEI, herein referred to as the inverse Acoustic Evenness Index (1-AEI) to facilitate comparability between the two indices [21]. Higher ADI and 1-AEI indicate higher frequency occupation of the spectrum and higher sound diversity. Thus, Spearman’s correlations for ADI and 1-AEI look identical.

The H is calculated as the product of Ht and Hf, where Ht represents the Shannon entropy of the amplitude envelope, and Hf represents the Shannon entropy of the frequency spectrum [17]. The Spearman’s correlations for H and Hf look identical, as they are both based on spectral complexity, whereas the Ht is based on amplitude parameters.

To estimate the effects of noise on the acoustic indices, we generated 5-min Brownian noise in increasing 0.01 dB increments in Audacity and mixed each clip with 5 randomly selected raw 5-min sound samples from our dataset from each site from 8-9 am. This resulted in 70 mixed tracks per site. We calculated the acoustic indices for each mixed track to see if noise affected their performance. We found that the nACI, H, Ht, and Hf were relatively robust to noise.

(3) There is quite recent literature about birds sounds in urban environments that the authors seem not to be aware of. I recommend to check the following references:

(3a) Benocci, et al. Mapping of the Acoustic Environment at an Urban Park...

Sensors 22, 3528 (2022).

Similar to the current study, Benocci et al. (2022) used low-cost recorders. They evaluated the frequency response of each recorder in terms of its sensitivity at different frequencies. They did this by exposing the recorders to white noise generated by a loudspeaker from a fixed distance.

We characterized our recorder’s frequency response and included this in Line 153-160 of the revised manuscript and included the details in S1 Fig. We found that the frequency response was not flat over the entire frequency range, and higher frequencies were increasingly attenuated, except for recorder A.

(3b) Benocci, R., et al. Auto-correlations and long time memory of environment sound..

Ecological Indicators 134, 108492 (2022).

Benocci et al. (2022) used GLMMs to evaluate the relationship between acoustic indices and rainfall frequency and intensity.

We used GLMMs to determine how acoustic indices predict bird species richness depending on the season and time of day. We included this in Line 265-277 of the materials and methods section of the revised manuscript. The GLMM results can be found in Table 4 and S2 Table. The detailed results can be found in Line 344-351 of the revised manuscript.

We found that 1-AEI, Ht, and AR predicted bird species richness in pre-processed 5-min sound samples, depending on the season and time of day. The nACI also predicted bird species richness but independently of season and time of day. Conversely, season and time of day were stronger predictors for ADI, BI, H, and Hf.

Reviewer #2

The study examines effectiveness of acoustic indices in bird species richness estimation in urban area. Similar studies have been conducted in various environments, so it is not novel approach. However inconsistent results reported in various studies point out that we still need more examples of such comparisons to understand where and when acoustic indices can be used as an approximation of biodiversity. From this perspective the study is valuable.

Did you check how well work acoustic indices when you compare separately dawn chorus, day and dusk chorus, or breeding vs non-breeding season? I suggest re-analysing data that way. I do not know how sound samples are distributed across day and season, maybe it will be necessary to analyse more sound samples to get optimal distribution, but authors can examine how each acoustic index estimate bird species richness dependently on time in a day and season (at least breeding vs non-breeding season). They can do this in simple way applying GLMMs in which they will specify the index for each sound sample as a dependent variable, number of detected bird species, time in a season, time in a day and one-way interactions as fixed effects, point id as a random effect. At the current stage of analysis, in my opinion, it will be difficult to publish results of the study. Also description of the methods should be more precise. I suggest considering reorganization of discussion and describing all indices together instead each of them separately.

I suggest adding raw data as a supplementary materials.

Yes, we agree. We included the results of the acoustic index analysis for 840 raw and 840 pre-processed 5-min sound samples as supplementary materials (S3 and S4 Tables) and in Line 318-325 of the revised text.

More detailed comments:

We selected 840 5-min sound samples, equivalent to 70 hours, through stratified sampling and pre-processed them before conducting acoustic index analysis on the raw and pre-processed data. We measured Spearman’s correlation between each acoustic index and bird species richness obtained from manual spectrogram scanning and listening to recordings.

L36: …840, 5-min sound samples.

L36-37: You used raw and pre-processed data.

L37: Spearman’s correlation

L38: …manual spectrogram scanning and listening to recordings.

We included the suggestions and reworded our manuscript to read as:

“We selected 840 5-min sound samples, equivalent to 70 hours, through stratified sampling and pre-processed them before conducting acoustic index analysis on the raw and pre-processed data. We measured Spearman’s correlation between each acoustic index and bird species richness obtained from manual spectrogram scanning and listening to recordings.”

L47: Please remove solely.

Yes, we have removed “solely.” in this line.

L52-53: Mobile network also serves such possibility and seems to be more useful than wi-fi.

Thank you for the suggestion. We have added “mobile network.”

L60-61: Not only these packages and not only R, but I agree that calculation is easy.

We used the soundecology and seewave R packages, as they were the most used in studies to calculate acoustic indices although there are other ways to calculate these as the reviewer pointed out.

L64-65: Please remember that some indices correlate negatively with soundscape complexity.

Yes, we agree with the reviewer. We want to clarify that the reference studies for acoustic indices in our study reported positive correlations.

L66-67: Please look at this review https://doi.org/10.1111/brv.12890, it can be useful.

We thank the reviewer for this suggestion, and we added this reference as suggested and considered this in our rewording of our texts.

L71-80: Consider mentioning about complexity and variation of birds vocalization – two different species (or even individuals within a species) will generate different values of acoustic indices.

This is true. However, we are doing a community-level analysis. We take into account the call or song variation among individuals within a species by taking several samples over different times and different 5-min samples.

L78-80: Great point. But also specific species composition may be one of the main factors responsible for effectiveness of acoustic indices.

The birds in the NSC are mostly passerines, so their frequency ranges have a high chance of overlap due to their relatively similar body sizes and vocal capacities.

L91: Not many papers reported strong correlation, it is rather moderate or weak.

Yes, we agree. We reworded our hypothesis to this:

“We hypothesized that if acoustic index values are significantly correlated with the number of species heard, then the selected indices can be used as proxies for bird species richness in the field.”

L96: ...is home to 42 bird species, including endemic...

We have corrected this.

L96-98: How many of them are breeding species, how many are sedentary? Please write how long is a breeding season and when it has place.

In our results, we reported that out of 21 species heard in the NSC, seven are Philippine endemics, thirteen are resident species, and only one is a migrant, the Brown Shrike.

The breeding season for most birds in the Philippines aligns with the dry season, from November to April. The wet season is from May to October. We used the dry and wet seasons as a proxy for breeding and non-breeding season. We used this as a factor in our GLMMs.

L121-122: Please write about frequency response, signal to noise ratio and detection distance of recorders. Did you calibrate recorders before using?

We tested the response of each low-cost recorder across a range of frequencies. Using Audacity, we generated a 30-sec tone from 1 Hz to 10,000 Hz and played it through a loudspeaker held at breast level, gradually increasing the distance from the recorder. We repeated this several times and we plotted the spectrum data of the 30-sec generated tone against the average response of each recorder (S1 Fig). The frequency response was not flat over the entire frequency range, except for recorder A, and higher frequencies were increasingly attenuated (S1 Fig).

L124:125: Impressive. Do you use these data in silent cities project?

The data was not used in the Silent Cities Project. However, initially, we set-up recorders in the National Science Complex to study the effects of the anthropause on the bird community. But, we still have insufficient pre-pandemic and post-pandemic data. The three recorders were continuously recording until July 2022. The Site A system resumed recording in November 2022 and is still recording now.

L130: …daily activity period, including dawn and dusk chorus.

We randomly selected 840 .wav files from the database based on the YYYY-MM-DD-HH-mm timestamp. Through stratified sampling, we selected 20 sound samples from each hour between 5:00 AM and 6:00 PM to cover the daily activity period, including the dawn and dusk chorus. This resulted in a total of 280 samples for each recording point.

The distribution of the sound samples across months is variable but fully represented (see Figure in attached Response to Reviewers). Yes, the months of July to January were sampled twice.

We viewed each pre-processed 5-min sound sample in Audacity, using Spectrogram View set to Linear Scale, with the Minimum Frequency set to 0 kHz and the Max Frequency set to 8 kHz (Fig 2). S. Diaz was the primary observer who analyzed all the recordings to reduce variability between observers.

We recognized species based on their calls and songs since the seasonal variation is less pronounced in the Philippines than in temperate countries. The Philippines has two seasons: dry and wet. We used the dry and wet seasons as a proxy for breeding and non-breeding season.

L136-137: How commonly vocalizations of animals other than birds were recorded? The information about other biological sources of sound in your study area is missing.

L138: Please look at this review: https://doi.org/10.1080/09524622.2021.2010598

We considered this in rewriting the manuscript and included it in the references.

L140: Similar effect you can get modifying setting of acoustic indices calculation.

We wanted to use the default settings of the acoustic indices to streamline the processing of multiple .wav files.

L150-154: Please add a table (e.g., as a supply mat) in which you will show settings for each index.

We have added the summary of parameters used in the calculation of acoustic indices as supplementary materials (S1 Table).

We used the 840 pre-processed sound samples for manual spectrogram scanning and listening to recordings.

We used two data sets for the acoustic index analysis: 840 raw and 840 pre-processed .wav files. The only difference between the two sets of 840 5-min sound samples is that the pre-processed samples were subjected to processing steps (i.e., high-pass filter, noise reduction, amplification). Both datasets have the same YYYY-MM-DD-HH-mm timestamp and recording point.

These were written into the revised manuscript in Line 174-175 and Line 196-197.

L175-176: So these two indices should correlate negatively and positively with bird species richness.

Yes, ADI values should correlate positively and AEI values negatively with bird species richness. Thus, to facilitate comparability between the two indices, we decided to get the inverse of AEI, herein referred to as the 1-AEI.

We thank the reviewer for this suggestion and have used generalized linear mixed models (GLMMs) to examine how acoustic indices predict bird species richness in pre-processed 5-min sound samples depending on season and time of day. For each sound sample, we specified the acoustic index as the dependent variable, bird species richness, season, and time of day as fixed effects, and the recording point as the random effect. We fitted a set of three models for each acoustic index, which included the full model and fixed effects models in R using the lme4 and glmmTMB packages (S2 Table).

L202-203: 840 5-min sound samples? I suggest adding histogram (or modifying Table 1) to show how many sound samples each bird species was detected.

L206-209: You do not need mention them here, please add marker in Table 1 which species are endemic.

L213: I suggest to show in how many sound samples each species was detected in each recording point instead markers.

We thank the reviewer for this suggestion and have modified Table 1 to reflect how many sound samples each bird species was detected in each recording point. In Table 1, we marked Philippine endemic species with asterisk (*).

L205-206: Why?

There is insufficient eBird data for both species.

Thank you for this suggestion. We reorganized the discussion and reported the acoustic index correlations more holistically. We calculated Fisher's z-transformation to compare the correlations between the raw and pre-processed datasets.

Yes, we agree. We removed this part of the discussion.

L343-344: There was a few studies examining effectiveness of acoustic indices in urban areas from other regions. Please look at them.

We wanted to mention that this is the first study testing acoustic indices as proxies for bird species richness in an urban setting in the Philippines. We are aware that other studies used acoustic indices in urban areas in temperate regions.

Reviewer #3:

The paper “Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila” describes the correlation results between bird species richness and eco-acoustic indices.

The paper is clear but requires supplementary analysis before being accepted. It focuses just on a specific aspect of the soundscape recorded at three sites linked to bird species richness.

Generally, eco acoustic indices represent summaries of the whole soundscape and highlight specific features of the frequency spectrum. This concept should be included in the introduction.

Thank you for the suggestion. We’ve added this to the introduction: “Several ecoacoustic indices were proposed to summarize the acoustic features of soundscapes using the sound spectrum [8]. Moreover, acoustic indices are mathematical functions that measure the distribution of acoustic energy across time and frequency in a recording [9]. These indices enable the quick processing of sound data and measure the acoustic complexity, diversity, evenness, entropy, and richness of recordings. They are based on species diversity indices and are thought to characterize the variation in sound production in animal communities across time, serving as proxies for diversity metrics such as species abundance, richness, evenness, and diversity [10].”

Again, we thank the reviewer for this suggestion. In the Materials and methods section, we noted that we used a sampling rate of 44.1 kHz, and we also described the frequency response of the recorders.

Sect. Manual aural identification: Here, report also all other sound sources identified and their frequencies. This is important especially in urban green areas

Sect. Acoustic index analysis: when you introduce the pre-processing, please report an example of a spectrum for a raw and pre-processed audio file.

Yes, we agree. We added sample spectrograms for raw and pre-processed .wav files opened in Audacity (Fig 2).

All indices are based upon a FFT computation of audio recordings. Please specify the number of FFT points use, the frequency and time resolution.

We specified in the Materials and methods section that we used a sampling frequency of 44,100 Hz and a fast Fourier transform (FFT) window of 512 points (S1 Table). This corresponds to a frequency resolution of FR = 86.133 Hz and a time resolution of TR = 0.0116 s.

In Sect ADI and IAEI , it is reported that each frequency bin represents a particular bird species. This is NOT true!!

At line 176, I would say that higher ADI and 1-AEI indicate higher frequency occupation of the spectrum.

We thank the reviewer for finding this error, and we have corrected it. We changed it to what we originally meant: “Higher ADI and 1-AEI should indicate higher frequency occupation of the spectrum and higher sound diversity.”

If possible include a temporal trend of the mean bird richness as well as for the eco-acoustic indices.

We thank the reviewer for this suggestion. We modified the figure (See Figure 3) to reflect the differences among the sites in terms of richness, and we used a clearer title. We also included a temporal trend of the mean bird richness (See Figure 4).

Lines 264-266: this part needs to be improved

Yes, we agree. We changed it to this: “Ground-truthing of the acoustic indices should be conducted before applying them to studies in different environments as their performance depends on various factors including, the local noise conditions, bird species composition of the site, season, and time of day.”

Line 272-273. This is why abundance of birds should be included in the analysis.

The limitation of our study is that we did not measure bird species abundance because of the difficulty of telling apart individuals based simply on their calls and the pandemic limited on-site observations simultaneous with the recordings. Hence, the study focused on the relationship of acoustic indices with bird species richness.

Given the proximity of the three sites, the bird community across them is most likely shared. We assume that the bird abundance and bird richness are similar across the three sites.

Line 303-304: this is why you need to include a more detailed aural survey.

We mentioned in the results that in addition to bird sounds, we also detected various other biotic sounds, including those produced by mammals, amphibians, reptiles, and insects. Although we detected other non-avian sounds in our recordings, we did not take their frequencies into account, as we wanted to focus on bird vocalizations.

Benocci, R.; Roman, H.E.; Bisceglie, A.; Angelini, F.; Brambilla, G.; Zambon, G. Eco-Acoustic Assessment of an Urban Park by Statistical Analysis. Sustainability 2021, 13, 7857.

Benocci, R.; Brambilla, G.; Bisceglie, A.; Zambon, G. Eco-Acoustic Indices to Evaluate Soundscape Degradation Due to Human Intrusion. Sustainability 2020, 12, 10455.

We added these references as suggested and note their importance. In the Conclusions, we added, “The use of other analysis, such as principal component analysis (PCA) [55] and cluster analysis [56], also help discriminate among different soundscapes in urban areas based on ecoacoustic indices.”

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(64.5KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0289001.r003

Decision Letter 1

Luca Nelli

26 Jun 2023

PONE-D-22-34645R1Acoustic indices as proxies for bird species richness in an urban green space in Metro ManilaPLOS ONE

Dear Dr. Diaz,

Please remove gradient from the caption and specify min and max values in dB, in Fig. 2. Also, please address the comment from reviewer 3, with respect to differences among the used sensors.

Please submit your revised manuscript by Aug 10 2023 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.
A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.
An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

We look forward to receiving your revised manuscript.

Kind regards,

Luca Nelli, PhD

Academic Editor

PLOS ONE

Journal Requirements:

1. Please review your reference list to ensure that it is complete and correct. If you have cited papers that have been retracted, please include the rationale for doing so in the manuscript text, or remove these references and replace them with relevant current references. Any changes to the reference list should be mentioned in the rebuttal letter that accompanies your revised manuscript. If you need to cite a retracted article, indicate the article’s retracted status in the References list and also include a citation and full reference for the retraction notice.

Additional Editor Comments:

Please remove gradient from the caption and specify min and max values in dB, in Fig. 2.

Also, please address the comment from reviewer 3, with respect to differences among the used sensors.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #3: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

Reviewer #1: Yes

Reviewer #3: Yes

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: Yes

Reviewer #3: Yes

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

Reviewer #1: Yes

Reviewer #3: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

Reviewer #1: Yes

Reviewer #3: Yes

**********

6. Review Comments to the Author

Reviewer #1: The authors have satisfactorily responded to all pints raised. As such, I can recommend the publication of this paper in its present form.

Reviewer #3: The authors mostly addressed the reviewer’s suggestions. Still there are some other minor points to be clarified and stressed.

Fig 2: remove gradient from the caption and specify min and max values in dB.

The frequency response that you obtained reveals that there could be significant differences among the used sensors. This may imply that also the indices calculation could be affected by it. This limitation needs to appear in the Conclusions.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #3: Yes: Giovanni Zambon

**********

PLoS One. 2023 Jul 28;18(7):e0289001. doi: 10.1371/journal.pone.0289001.r004

Author response to Decision Letter 1

6 Jul 2023

Reviewer #1

The authors have satisfactorily responded to all points raised. As such, I can recommend the publication of this paper in its present form.

We thank the reviewer for their comments and suggestions, which greatly improved this manuscript.

Reviewer #3

The authors mostly addressed the reviewer’s suggestions. Still there are some other minor points to be clarified and stressed.

Fig 2: remove gradient from the caption and specify min and max values in dB.

Thank you for the suggestion. We have removed the gradient from the caption and specified the minimum and maximum values in dB. The amplitude ranges from -60 dB to -44 dB (raw) and -60 dB to -29 dB (pre-processed).

We have included this limitation of low-cost recorders in the Conclusions and it reads as follows “As the frequency response of our recorders was not flat over the entire frequency range, which impacts the calculation of acoustic indices, low-cost recorders should be calibrated before use in ecoacoustic studies to offset this limitation. We also recognize the importance of pre-processing sound data, such as applying high-pass filters, noise reduction, and signal amplification, to account for differences among recorders, remove unwanted ambient sounds, and improve the detection of bird species.”

Attachment

Submitted filename: Responses to Reviewers V2.docx

Click here for additional data file.^{(7.5KB, docx)}

PLoS One. doi: 10.1371/journal.pone.0289001.r005

Decision Letter 2

Luca Nelli

10 Jul 2023

Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

PONE-D-22-34645R2

Dear Dr. Diaz,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Luca Nelli, PhD

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

PLoS One. doi: 10.1371/journal.pone.0289001.r006

Acceptance letter

Luca Nelli

20 Jul 2023

PONE-D-22-34645R2

Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

Dear Dr. Diaz:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Luca Nelli

Academic Editor

PLOS ONE

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

S1 Fig. The frequency response of recorders A-C with respect to the generated tone (black line) in Audacity® [40].

(PDF)

Click here for additional data file.^{(109.4KB, pdf)}

S1 Table. Summary of parameters used in the calculation of acoustic indices.

(PDF)

Click here for additional data file.^{(130.4KB, pdf)}

S2 Table. The results of GLMM examining how eight acoustic indices reflect bird species richness, season, and time of day, using the dataset of 840 pre-processed 5-min sound samples.

(PDF)

Click here for additional data file.^{(149.7KB, pdf)}

S3 Table. The results of the acoustic index analysis for 840 raw 5-min sound samples.

(PDF)

Click here for additional data file.^{(524.1KB, pdf)}

S4 Table. The results of the acoustic index analysis for 840 pre-processed 5-min sound samples.

(PDF)

Click here for additional data file.^{(509.5KB, pdf)}

Attachment

Submitted filename: Response to Reviewers.docx

Click here for additional data file.^{(64.5KB, docx)}

Attachment

Submitted filename: Responses to Reviewers V2.docx

Click here for additional data file.^{(7.5KB, docx)}

Data Availability Statement

All relevant data are within the manuscript and its Supporting Information files.

[pone.0289001.ref001] 1.Byers BE, Kroodsma DE. Avian vocal behavior. Cornell Lab of Ornithology. Handbook of Bird Biology. 2016:355–405. [Google Scholar]

[pone.0289001.ref002] 2.Bardeli R, Wolff D, Kurth F, Koch M, Tauchert KH, Frommolt KH. Detecting bird sounds in a complex acoustic environment and application to bioacoustic monitoring. Pattern Recognition Letters. 2010. Sep 1;31(12):1524–34. doi: 10.1016/j.patrec.2009.09.014 [DOI] [Google Scholar]

[pone.0289001.ref003] 3.Penar W, Magiera A, Klocek C. Applications of bioacoustics in animal ecology. Ecological complexity. 2020. Aug 1;43:100847. doi: 10.1016/j.ecocom.2020.100847 [DOI] [Google Scholar]

[pone.0289001.ref004] 4.Abrahams C. Bird bioacoustic surveys-developing a standard protocol. In Practice. 2018(102):20–3. [Google Scholar]

[pone.0289001.ref005] 5.Zwart MC, Baker A, McGowan PJ, Whittingham MJ. The use of automated bioacoustic recorders to replace human wildlife surveys: an example using nightjars. PloS one. 2014. Jul 16;9(7):e102770. doi: 10.1371/journal.pone.0102770 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref006] 6.Aide TM, Corrada-Bravo C, Campos-Cerqueira M, Milan C, Vega G, Alvarez R. Real-time bioacoustics monitoring and automated species identification. PeerJ. 2013. Jul 16;1:e103. doi: 10.7717/peerj.103 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref007] 7.Gibb R, Browning E, Glover‐Kapfer P, Jones KE. Emerging opportunities and challenges for passive acoustics in ecological assessment and monitoring. Methods in Ecology and Evolution. 2019. Feb;10(2):169–85. doi: 10.1111/2041-210X.13101 [DOI] [Google Scholar]

[pone.0289001.ref008] 8.Eldridge A, Casey M, Moscoso P, Peck M. A new method for ecoacoustics? Toward the extraction and evaluation of ecologically-meaningful soundscape components using sparse coding methods. PeerJ. 2016. Jun 30;4:e2108. doi: 10.7717/peerj.2108 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref009] 9.Sueur J. Sound analysis and synthesis with R. Cham: Springer; 2018. Jun 6. doi: 10.1007/978-3-319-77647-7 [DOI] [Google Scholar]

[pone.0289001.ref010] 10.Sueur J, Farina A, Gasc A, Pieretti N, Pavoine S. Acoustic indices for biodiversity assessment and landscape investigation. Acta Acustica united with Acustica. 2014. Jul 1;100(4):772–81. doi: 10.3813/AAA.918757 [DOI] [Google Scholar]

[pone.0289001.ref011] 11.Villanueva-Rivera LJ, Pijanowski BC. Soundecology: soundscape ecology. R package version 1.3.3. 2018. Available from: https://cran.r-project.org/package=soundecology [Google Scholar]

[pone.0289001.ref012] 12.Sueur J, Aubin T, Simonis C. Seewave: sound analysis and synthesis. R package version 2.2.0. 2022. Available from: https://CRAN.R-project.org/package=seewave [Google Scholar]

[pone.0289001.ref013] 13.Bateman J, Uzal A. The relationship between the Acoustic Complexity Index and avian species richness and diversity: a review. Bioacoustics. 2022. Sep 3;31(5):614–27. doi: 10.1080/09524622.2021.2010598 [DOI] [Google Scholar]

[pone.0289001.ref014] 14.Pieretti N, Farina A, Morri D. A new methodology to infer the singing activity of an avian community: The Acoustic Complexity Index (ACI). Ecological indicators. 2011. May 1;11(3):868–73. doi: 10.1016/j.ecolind.2010.11.005 [DOI] [Google Scholar]

[pone.0289001.ref015] 15.Villanueva-Rivera LJ, Pijanowski BC, Doucette J, Pekin B. A primer of acoustic analysis for landscape ecologists. Landscape ecology. 2011. Nov;26(9):1233–46. doi: 10.1007/s10980-011-9636-9 [DOI] [Google Scholar]

[pone.0289001.ref016] 16.Boelman NT, Asner GP, Hart PJ, Martin RE. Multi‐trophic invasion resistance in Hawaii: bioacoustics, field surveys, and airborne remote sensing. Ecological Applications. 2007. Dec;17(8):2137–44. doi: 10.1890/07-0004.1 [DOI] [PubMed] [Google Scholar]

[pone.0289001.ref017] 17.Sueur J, Pavoine S, Hamerlynck O, Duvail S. Rapid acoustic survey for biodiversity appraisal. PloS one. 2008. Dec 30;3(12):e4065. doi: 10.1371/journal.pone.0004065 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref018] 18.Depraetere M, Pavoine S, Jiguet F, Gasc A, Duvail S, Sueur J. Monitoring animal diversity using acoustic indices: implementation in a temperate woodland. Ecological Indicators. 2012. Feb 1;13(1):46–54. doi: 10.1016/j.ecolind.2011.05.006 [DOI] [Google Scholar]

[pone.0289001.ref019] 19.Machado RB, Aguiar L, Jones G. Do acoustic indices reflect the characteristics of bird communities in the savannas of Central Brazil?. Landscape and Urban Planning. 2017. Jun 1;162:36–43. doi: 10.1016/j.landurbplan.2017.01.014 [DOI] [Google Scholar]

[pone.0289001.ref020] 20.Mammides C, Goodale E, Dayananda SK, Kang L, Chen J. Do acoustic indices correlate with bird diversity? Insights from two biodiverse regions in Yunnan Province, south China. Ecological Indicators. 2017. Nov 1;82:470–7. doi: 10.1016/j.ecolind.2017.07.017 [DOI] [Google Scholar]

[pone.0289001.ref021] 21.Dröge S, Martin DA, Andriafanomezantsoa R, Burivalova Z, Fulgence TR, Osen K, et al. Listening to a changing landscape: Acoustic indices reflect bird species richness and plot-scale vegetation structure across different land-use types in north-eastern Madagascar. Ecological Indicators. 2021. Jan 1;120:106929. doi: 10.1016/j.ecolind.2020.106929 [DOI] [Google Scholar]

[pone.0289001.ref022] 22.Retamosa Izaguirre M, Barrantes-Madrigal J, Segura Sequeira D, Spínola-Parallada M, Ramírez-Alán O. It is not just about birds: what do acoustic indices reveal about a Costa Rican tropical rainforest?. Neotropical Biodiversity. 2021. Jan 1;7(1):431–42. doi: 10.1080/23766808.2021.1971042 [DOI] [Google Scholar]

[pone.0289001.ref023] 23.Moreno-Gómez FN, Bartheld J, Silva-Escobar AA, Briones R, Márquez R, Penna M. Evaluating acoustic indices in the Valdivian rainforest, a biodiversity hotspot in South America. Ecological Indicators. 2019. Aug 1;103:1–8. doi: 10.1016/j.ecolind.2019.03.024 [DOI] [Google Scholar]

[pone.0289001.ref024] 24.Alcocer I, Lima H, Sugai LS, Llusia D. Acoustic indices as proxies for biodiversity: a meta‐analysis. Biological Reviews. 2022. Dec;97(6):2209–36. doi: 10.1111/brv.12890 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref025] 25.Gasc A, Pavoine S, Lellouch L, Grandcolas P, Sueur J. Acoustic indices for biodiversity assessments: Analyses of bias based on simulated bird assemblages and recommendations for field surveys. Biological Conservation. 2015. Nov 1;191:306–12. doi: 10.1016/j.biocon.2015.06.018 [DOI] [Google Scholar]

[pone.0289001.ref026] 26.Eldridge A, Guyot P, Moscoso P, Johnston A, Eyre-Walker Y, Peck M. Sounding out ecoacoustic metrics: Avian species richness is predicted by acoustic indices in temperate but not tropical habitats. Ecological Indicators. 2018. Dec 1;95:939–52. doi: 10.1016/j.ecolind.2018.06.012 [DOI] [Google Scholar]

[pone.0289001.ref027] 27.Fairbrass AJ, Rennert P, Williams C, Titheridge H, Jones KE. Biases of acoustic indices measuring biodiversity in urban areas. Ecological Indicators. 2017. Dec 1;83:169–77. doi: 10.1016/j.ecolind.2017.07.064 [DOI] [Google Scholar]

[pone.0289001.ref028] 28.Pijanowski BC, Farina A, Gage SH, Dumyahn SL, Krause BL. What is soundscape ecology? An introduction and overview of an emerging new science. Landscape ecology. 2011. Nov;26(9):1213–32. doi: 10.1007/s10980-011-9600-8 [DOI] [Google Scholar]

[pone.0289001.ref029] 29.Rutz C, Loretto MC, Bates AE, Davidson SC, Duarte CM, Jetz W, et al. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife. Nature Ecology & Evolution. 2020. Sep;4(9):1156–9. doi: 10.1038/s41559-020-1237-z [DOI] [PubMed] [Google Scholar]

[pone.0289001.ref030] 30.Vallejo BM Jr, Aloy AB, Ong PS. The distribution, abundance and diversity of birds in Manila’s last greenspaces. Landscape and Urban Planning. 2009. Feb 15;89(3–4):75–85. doi: 10.1016/j.landurbplan.2008.10.013 [DOI] [Google Scholar]

[pone.0289001.ref031] 31.Energy Development Corporation. EDC establishes threatened species arboretum in UP. 2014. December 11. Available from: https://www.energy.com.ph/2014/12/11/edc-establishes-threatened-species-arboretum-in-up/ [Google Scholar]

[pone.0289001.ref032] 32.Cabel IC. Walk and learn to save PH native tree species. Philippine Daily Inquirer. 2019. April 8. Available from: https://newsinfo.inquirer.net/1104179/walk-and-learn-to-save-ph-native-tree-species [Google Scholar]

[pone.0289001.ref033] 33.Villasper J. Destinations: The University of the Philippines Diliman Campus–Choose Your Own Adventure. Birdwatch. 2020. March 2. Available from: http://birdwatch.ph/2020/03/02/destinations-the-university-of-the-philippines-diliman-campus-choose-your-own-adventure-2/ [Google Scholar]

[pone.0289001.ref034] 34.Philippine Atmospheric, Geophysical and Astronomical Services Administration (PAGASA). Climate of the Philippines. Available from: https://www.pagasa.dost.gov.ph/information/climate-philippines [Google Scholar]

[pone.0289001.ref035] 35.OpenStreetMap Contributors, CC BY-SA 2.0. 2015. Available from: https://openstreetmap.org [Google Scholar]

[pone.0289001.ref036] 36.HueMan1—Own work, CC BY-SA 4.0. Blank map of the Philippines. 2021. Available from: https://commons.wikimedia.org/w/index.php?curid=108050493 [Google Scholar]

[pone.0289001.ref037] 37.Fletcher AC, Mura C. Ten quick tips for using a Raspberry Pi. PLoS computational biology. 2019. May 2;15(5):e1006959. doi: 10.1371/journal.pcbi.1006959 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref038] 38.Kysela J. arecord (1): command-line sound recorder and player for ALSA. [Cited 2021 August 31]. Available from: http://manpages.org/arecord [Google Scholar]

[pone.0289001.ref039] 39.Benocci R, Potenza A, Bisceglie A, Roman HE, Zambon G. Mapping of the Acoustic Environment at an Urban Park in the City Area of Milan, Italy, Using Very Low-Cost Sensors. Sensors. 2022. May 6;22(9):3528. doi: 10.3390/s22093528 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref040] 40.Team Audacity. Audacity®: free audio editor and recorder [Computer application]. Version 3.0.0. 2021. Audacity® software is copyright © 1999–2021 Audacity Team. Available from: https://www.audacityteam.org/. It is free software distributed under the terms of the GNU General Public License. The name Audacity® is a registered trademark. [Google Scholar]

[pone.0289001.ref041] 41.Planqué B, Vellinga WP, Pieterse S, Jongsma J, de By RA. Xeno-canto. 2005. Available from: https://www.xeno-canto.org [Google Scholar]

[pone.0289001.ref042] 42.Sullivan BL, Wood CL, Iliff MJ, Bonney RE, Fink D, Kelling S. eBird: A citizen-based bird observation network in the biological sciences. Biological conservation. 2009. Oct 1;142(10):2282–92. Available from: https://www.ebird.org/ [Google Scholar]

[pone.0289001.ref043] 43.R Development Core Team. The R Project for Statistical Computing. R Version 4.0.0. 2020. Available from: https://www.r-project.org/ [Google Scholar]

[pone.0289001.ref044] 44.Bates D, Maechler M, Bolker B, Walker S. lme4: Linear Mixed-Effects Models using ’Eigen’ and S4. R package version 1.1–32. 2023. Available from: https://cran.r-project.org/package=lme4 [Google Scholar]

[pone.0289001.ref045] 45.Magnusson A, Skaug H, Nielsen A, Berg C, Kristensen K, Maechler M, et al. glmmTMB: Generalized Linear Mixed Models using Template Model Builder. R package version 1.1.6. 2023. Available from: https://CRAN.R-project.org/package=glmmTMB [Google Scholar]

[pone.0289001.ref046] 46.Budka M, Sokołowska E, Muszyńska A, Staniewicz A. Acoustic indices estimate breeding bird species richness with daily and seasonally variable effectiveness in lowland temperate Białowieża forest. Ecological Indicators. 2023. Apr 1;148:110027. doi: 10.1016/j.ecolind.2023.110027 [DOI] [Google Scholar]

[pone.0289001.ref047] 47.Sánchez‐Giraldo C, Bedoya CL, Morán‐Vásquez RA, Isaza CV, Daza JM. Ecoacoustics in the rain: understanding acoustic indices under the most common geophonic source in tropical rainforests. Remote Sensing in Ecology and Conservation. 2020. Sep;6(3):248–61. doi: doi.org/10.1002/rse2.162 [Google Scholar]

[pone.0289001.ref048] 48.Mazerolle MJ. AICcmodavg: Model Selection and Multimodel Inference Based on (Q)AIC(c). R package version 2.3–2. 2023. Available from: https://CRAN.R-project.org/package=AICcmodavg [Google Scholar]

[pone.0289001.ref049] 49.Burnham KP, Anderson DR. Model selection and multimodel inference: A practical information-theoretic approach. New York, NY: Springer New York; 2002. doi: 10.1007/b97636 [DOI] [Google Scholar]

[pone.0289001.ref050] 50.Ebird. eBird: An online database of bird distribution and abundance. eBird, Cornell Lab of Ornithology. UP Diliman—MSI Hotspot. 2022. Available from: https://ebird.org/hotspot/L10877450?yr=all&m=&rank=mrec [Google Scholar]

[pone.0289001.ref051] 51.Shamon H, Paraskevopoulou Z, Kitzes J, Card E, Deichmann JL, Boyce AJ, et al. Using ecoacoustics metrices to track grassland bird richness across landscape gradients. Ecological Indicators. 2021. Jan 1;120:106928. doi: 10.1016/j.ecolind.2020.106928 [DOI] [Google Scholar]

[pone.0289001.ref052] 52.Retamosa Izaguirre MI, Ramírez-Alán O, De la O Castro J. Acoustic indices applied to biodiversity monitoring in a Costa Rica dry tropical forest. Journal of Ecoacoustics. 2018. Feb 26;2. doi: 10.22261/JEA.TNW2NP [DOI] [Google Scholar]

[pone.0289001.ref053] 53.Kahl S, Wood CM, Eibl M, Klinck H. BirdNET: A deep learning solution for avian diversity monitoring. Ecological Informatics. 2021. Mar 1;61:101236. doi: 10.1016/j.ecoinf.2021.101236 [DOI] [Google Scholar]

[pone.0289001.ref054] 54.Marchal J, Fabianek F, Aubry Y. Software performance for the automated identification of bird vocalisations: the case of two closely related species. Bioacoustics. 2022. Jul 4;31(4):397–413. doi: 10.1080/09524622.2021.1945952 [DOI] [Google Scholar]

[pone.0289001.ref055] 55.Liu J, Zhang Y, Lv D, Lu J, Xie S, Zi J, et al. Birdsong classification based on ensemble multi-scale convolutional neural network. Scientific Reports. 2022. May 23;12(1):1–4. doi: 10.1038/s41598-022-12121-8 [DOI] [PMC free article] [PubMed] [Google Scholar]

[pone.0289001.ref056] 56.Benocci R, Roman HE, Bisceglie A, Angelini F, Brambilla G, Zambon G. Eco-acoustic assessment of an urban park by statistical analysis. Sustainability. 2021. Jul 14;13(14):7857. doi: 10.3390/su13147857 [DOI] [Google Scholar]

[pone.0289001.ref057] 57.Benocci R, Brambilla G, Bisceglie A, Zambon G. Eco-acoustic indices to evaluate soundscape degradation due to human intrusion. Sustainability. 2020. Dec 14;12(24):10455. doi: 10.3390/su122410455 [DOI] [Google Scholar]

PERMALINK

Acoustic indices as proxies for bird species richness in an urban green space in Metro Manila

Skyla Dennise U Diaz

Jelaine L Gan

Giovanni A Tapang

Roles

Abstract

Introduction

Materials and methods

Ethics statement

Study area

Fig 1. Map of three sound recording sites within the National Science Complex, University of the Philippines Diliman, and its relative location in the Philippines.

Acoustic data collection

Data pre-processing

Manual spectrogram scanning and listening to recordings

Fig 2. Sample spectrograms for raw and pre-processed.wav files opened in Audacity® Spectrogram View.

Acoustic index analysis

Normalized Acoustic Complexity Index (nACI)

Acoustic Diversity Index (ADI) and Inverse Acoustic Evenness Index (1-AEI)

Bioacoustic Index (BI)

Acoustic Entropy Index (H): Temporal Entropy Index (Ht) and Spectral Entropy Index (Hf)

Acoustic Richness Index (AR)

Statistical analysis

Spearman’s correlation

Generalized linear mixed models (GLMMs)

Results

Summary of bird species heard

Table 1. Summary of bird species heard and number of sound samples each species was detected out of 280 samples per recording point.

Fig 3. Bird species richness and number of recordings per site in 840 pre-processed 5-min sound samples.

Fig 4. Mean bird species richness across time from 5:00 AM to 6:00 PM in 840 pre-processed 5-min sound samples.

Spearman’s correlation between acoustic indices and bird species richness

Table 2. The Spearman’s rank correlation coefficients between acoustic indices and bird species richness.

Table 3. The results of Fisher’s z-transformation comparing Spearman’s rank correlation coefficients for the raw and pre-processed data sets.

Generalized linear mixed models (GLMMs)

Table 4. The results of GLMM examining how eight acoustic indices reflect bird species richness, season, and time of day, using the 840 pre-processed 5-min sound samples dataset.

Discussion

Acoustic indices and bird species richness

Conclusions

Supporting information

Acknowledgments

Data Availability

Funding Statement

References

Decision Letter 0

Cihun-Siyong Alex Gong

Roles

Author response to Decision Letter 0

Decision Letter 1

Luca Nelli

Roles

Author response to Decision Letter 1

Decision Letter 2

Luca Nelli

Roles

Acceptance letter

Luca Nelli

Roles

Associated Data

Supplementary Materials

Data Availability Statement

ACTIONS

PERMALINK

RESOURCES

Similar articles

Cited by other articles

Links to NCBI Databases