Ana Širović
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Ana Širović has a total of 89 co-authors. Here are the top co-authors by number of shared publications:
- John A. Hildebrand
Scripps Institution of Oceanography, University of California San Diego, La Jolla, USA
13 shared publications - Simone Baumann-Pickering
Scripps Institution of Oceanography, University of California San Diego, San Diego, USA
7 shared publications - Sean M. Wiggins
Scripps Institution of Oceanography, University of California San Diego 1 , La Jolla, California 92037
6 shared publications - Marie A. Roch
Department of Computer Science, San Diego State University, San Diego, USA
5 shared publications Sean M. Wiggins
Scripps Institution of Oceanography University of California San Diego 9500 Gilman Drive MC 0205 La Jolla California 92093‐0205 U.S.A
5 shared publications- Erin M. Oleson
NOAA Fisheries—Pacific Island Fisheries Science Center 2 , Honolulu, Hawaii 96818
4 shared publications
Latest Publications
- Open Access
- 18 September 2026
- Published by Springer Nature in Scientific Reports
Application of deep learning to estimate blue and fin whale call density in the southern California Current Ecosystem
Blue ( Balaenoptera musculus ) and fin whales ( Balaenoptera physalus ) are dominant contributors to low-frequency ocean soundscapes, yet reliably extracting their calls from long-term passive acoustic recordings is methodologically challenging. Here, we train a multi-class deep-learning detector to identify five principal blue and fin whale call types (A, B, D, 20 Hz, and 40 Hz) from low-frequency spectrograms using a Faster R-CNN architecture combined with three rounds of iterative human review and hard-negative mining, progressively expanding and rebalancing the training set using California Cooperative Oceanic Fisheries (henceforth, CalCOFI) sonobuoy and moored hydrophone recordings from the southern California Current Ecosystem. The detector was evaluated on four independent test datasets spanning multiple years, seasons, and recording platforms and then deployed on CalCOFI sonobuoy recordings collected quarterly over two decades (2004–2024). The final model achieved consistently high mean precision, recall and F1 scores for most call types (e.g., A: 0.71/0.71/0.71; B: 0.83/0.59/0.63; D: 0.79/0.84/0.80; 20 Hz: 0.87/0.74/0.78), while 40 Hz calls remained challenging (0.42/0.69/0.51), primarily due to confusion with spectrally overlapping humpback whale downsweeps. Detections were post-processed using call-specific characteristics and received-level thresholds and normalized by recording effort and detection area to derive standardized indices of call density $$\left( \frac{\text {calls}}{\text {h} \cdot 1000 \text { km}^2}\right) $$ with uncertainty estimates. Densities were aggregated annually and show call-specific differences between inshore and offshore habitats and interannual variability associated with periods of anomalous oceanographic conditions. Inter-call interval analyses suggested seasonal stability in blue whale song, high variability in blue and fin whale social calls, and seasonal and interannual variability in fin whale song repetition rates. This study is among the first to use deep-learning to estimate baleen whale call density from decades of passive acoustic recordings in a complex soundscape.- 18 September 2026
- Published by Springer Nature in Scientific Reports
Open AccessApplication of deep learning to estimate blue and fin whale call density in the southern California Current Ecosystem
Blue ( Balaenoptera musculus ) and fin whales ( Balaenoptera physalus ) are dominant contributors to low-frequency ocean soundscapes, yet reliably extracting their calls from long-term passive acoustic recordings is methodologically challenging. Here, we train a multi-class deep-learning detector to identify five principal blue and fin whale call types (A, B, D, 20 Hz, and 40 Hz) from low-frequency spectrograms using a Faster R-CNN architecture combined with three rounds of iterative human review and hard-negative mining, progressively expanding and rebalancing the training set using California Cooperative Oceanic Fisheries (henceforth, CalCOFI) sonobuoy and moored hydrophone recordings from the southern California Current Ecosystem. The detector was evaluated on four independent test datasets spanning multiple years, seasons, and recording platforms and then deployed on CalCOFI sonobuoy recordings collected quarterly over two decades (2004–2024). The final model achieved consistently high mean precision, recall and F1 scores for most call types (e.g., A: 0.71/0.71/0.71; B: 0.83/0.59/0.63; D: 0.79/0.84/0.80; 20 Hz: 0.87/0.74/0.78), while 40 Hz calls remained challenging (0.42/0.69/0.51), primarily due to confusion with spectrally overlapping humpback whale downsweeps. Detections were post-processed using call-specific characteristics and received-level thresholds and normalized by recording effort and detection area to derive standardized indices of call density $$\left( \frac{\text {calls}}{\text {h} \cdot 1000 \text { km}^2}\right) $$ with uncertainty estimates. Densities were aggregated annually and show call-specific differences between inshore and offshore habitats and interannual variability associated with periods of anomalous oceanographic conditions. Inter-call interval analyses suggested seasonal stability in blue whale song, high variability in blue and fin whale social calls, and seasonal and interannual variability in fin whale song repetition rates. This study is among the first to use deep-learning to estimate baleen whale call density from decades of passive acoustic recordings in a complex soundscape. - Open Access
- 11 February 2026
- Published by The Company of Biologists in Journal of Experimental Biology
- Vol. 229 (Suppl_1)
Physiological effects of anthropogenic sound on aquatic animals: where are we and what is next?
Many aquatic animals have a well-developed sense of hearing as sound is important for communication underwater. However, this trait leaves them susceptible to injury, and physiological and behavioral impacts from exposure to intense or persistent anthropogenic sounds. We provide an overview of the current state of knowledge on the physiological effects of five main sources of anthropogenic sound: marine traffic, seismic exploration, pile driving, other industrial activity and sonar. Our understanding of impacts varies greatly by sound type and taxon, although the studied species do not represent the full taxonomic diversity. Exposure to ship sounds has been best studied in fish and it generally leads to responses along the stress response cascade, while few studies have been conducted on its physiological effect on invertebrates or marine mammals. Effects of exposure to seismic sound show mixed impact across taxa. Pile driving sounds have been shown in captive studies to result in hearing impairment in marine mammals and can cause injury to fishes. Lethal impacts have been documented from naval sonar on marine mammal species. Currently, physiological impacts from other industrial sound sources are poorly documented across taxa. Overall, given the limited number of species examined in sound impact experiments, it is crucial to establish categorizing principles and guidelines and modeled response pathways to improve management strategies, especially as new sound threats continue to emerge in our changing world.
- 11 February 2026
- Published by The Company of Biologists in Journal of Experimental Biology
- Vol. 229 (Suppl_1)
Open AccessPhysiological effects of anthropogenic sound on aquatic animals: where are we and what is next?
Many aquatic animals have a well-developed sense of hearing as sound is important for communication underwater. However, this trait leaves them susceptible to injury, and physiological and behavioral impacts from exposure to intense or persistent anthropogenic sounds. We provide an overview of the current state of knowledge on the physiological effects of five main sources of anthropogenic sound: marine traffic, seismic exploration, pile driving, other industrial activity and sonar. Our understanding of impacts varies greatly by sound type and taxon, although the studied species do not represent the full taxonomic diversity. Exposure to ship sounds has been best studied in fish and it generally leads to responses along the stress response cascade, while few studies have been conducted on its physiological effect on invertebrates or marine mammals. Effects of exposure to seismic sound show mixed impact across taxa. Pile driving sounds have been shown in captive studies to result in hearing impairment in marine mammals and can cause injury to fishes. Lethal impacts have been documented from naval sonar on marine mammal species. Currently, physiological impacts from other industrial sound sources are poorly documented across taxa. Overall, given the limited number of species examined in sound impact experiments, it is crucial to establish categorizing principles and guidelines and modeled response pathways to improve management strategies, especially as new sound threats continue to emerge in our changing world.
- Open Access
- 21 August 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (8) , e0329398
Blue and fin whales in the Northern Mariana Islands: Their call characteristics and occurrence
Many baleen whale vocalizations are species-, or even population-specific and can be used to monitor their occurrence. Although baleen whale occurrence has been well studied in parts of the Pacific Ocean, little is known about the seasonal distribution of blue and fin whales in the Western Pacific. Since 2010, a more concerted visual and acoustic survey effort has occurred around a very remote region of the western Pacific Ocean: the Northern Mariana Islands. Passive acoustic data were collected at two locations in the Northern Mariana Islands, Saipan and Tinian, from 2015–2017 and they were analyzed for call characteristics and occurrence of blue whale and fin whale calls. Low levels of Central North Pacific blue whale tonal calls were detected year-round with peaks in the winter (December) and summer (June). There was a clear seasonal pattern in fin whale calls (20 and 40 Hz calls), with the majority of detections occurring during winter and spring. Moreover, two unknown low-frequency sounds were detected, one tonal and one pulsed. The former was more common at Tinian and the latter at Saipan. By providing their acoustic features and occurrence patterns, we aim to facilitate future identification of their source. Additionally, the observed seasonal patterns in blue and fin whale call occurrences may offer insights into their movement patterns in this remote region.
- 21 August 2025
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 20 (8) , e0329398
Open AccessBlue and fin whales in the Northern Mariana Islands: Their call characteristics and occurrence
Many baleen whale vocalizations are species-, or even population-specific and can be used to monitor their occurrence. Although baleen whale occurrence has been well studied in parts of the Pacific Ocean, little is known about the seasonal distribution of blue and fin whales in the Western Pacific. Since 2010, a more concerted visual and acoustic survey effort has occurred around a very remote region of the western Pacific Ocean: the Northern Mariana Islands. Passive acoustic data were collected at two locations in the Northern Mariana Islands, Saipan and Tinian, from 2015–2017 and they were analyzed for call characteristics and occurrence of blue whale and fin whale calls. Low levels of Central North Pacific blue whale tonal calls were detected year-round with peaks in the winter (December) and summer (June). There was a clear seasonal pattern in fin whale calls (20 and 40 Hz calls), with the majority of detections occurring during winter and spring. Moreover, two unknown low-frequency sounds were detected, one tonal and one pulsed. The former was more common at Tinian and the latter at Saipan. By providing their acoustic features and occurrence patterns, we aim to facilitate future identification of their source. Additionally, the observed seasonal patterns in blue and fin whale call occurrences may offer insights into their movement patterns in this remote region.
- 1 July 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (1) , 548-556
The role of accelerometer hardware limitations in focal caller identification from acoustic recording tags attached to mysticetes
Multi-sensor acoustic tags have revolutionized our understanding of the behavior of large whales. One limitation, however, is the inability to reliably distinguish calls produced by the tagged whale from those produced by other nearby whales. One proposed solution has been to detect calls using both hydrophone and accelerometer data to identify signals produced by the tagged animal. Some high-amplitude low-frequency calls can be detected with accelerometers, but the success in using this approach with all calls within and across species is variable. Here, we provide guidance on the role of the physics of sound propagation and the tag hardware's accelerometer capabilities for successful application of this method with examples from tag data collected from fin whales (Balaenoptera physalus), blue whales (B. musculus), and southern right whales (Eubalaena australis). Of 1190 high amplitude calls believed to likely be from the tagged animal, only 517 were also detected on the accelerometer. Reasons for lack of detection were primarily the frequency of the signal lying outside the usable frequency detection range of the accelerometer on the tag, indicating selection of appropriate hardware capabilities are critical for this approach.
- 1 July 2025
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 158 (1) , 548-556
The role of accelerometer hardware limitations in focal caller identification from acoustic recording tags attached to mysticetes
Multi-sensor acoustic tags have revolutionized our understanding of the behavior of large whales. One limitation, however, is the inability to reliably distinguish calls produced by the tagged whale from those produced by other nearby whales. One proposed solution has been to detect calls using both hydrophone and accelerometer data to identify signals produced by the tagged animal. Some high-amplitude low-frequency calls can be detected with accelerometers, but the success in using this approach with all calls within and across species is variable. Here, we provide guidance on the role of the physics of sound propagation and the tag hardware's accelerometer capabilities for successful application of this method with examples from tag data collected from fin whales (Balaenoptera physalus), blue whales (B. musculus), and southern right whales (Eubalaena australis). Of 1190 high amplitude calls believed to likely be from the tagged animal, only 517 were also detected on the accelerometer. Reasons for lack of detection were primarily the frequency of the signal lying outside the usable frequency detection range of the accelerometer on the tag, indicating selection of appropriate hardware capabilities are critical for this approach.
- 3 July 2022
- Published by Springer Nature
- p. 195-221
The Bioacoustics of Blue Whales—Global Diversity and Behavioral Variability in a Foraging Specialist
- 3 July 2022
- Published by Springer Nature
- p. 195-221
The Bioacoustics of Blue Whales—Global Diversity and Behavioral Variability in a Foraging Specialist
- Open Access
- 1 May 2022
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 151 (5) , 3052-3065
Grouper source levels and aggregation dynamics inferred from passive acoustic localization at a multispecies spawning site
Four species of grouper (family Epinephlidae), Red Hind (Epinephelus guttatus), Nassau (Epinephelus striatus), Black (Mycteroperca bonaci), and Yellowfin Grouper (Mycteroperca venenosa) share an aggregation site in Little Cayman, Cayman Islands and produce sounds while aggregating. Continuous observation of these aggregations is challenging because traditional diver or ship-based methods are limited in time and space. Passive acoustic localization can overcome this challenge for sound-producing species, allowing observations over long durations and at fine spatial scales. A hydrophone array was deployed in February 2017 over a 9-day period that included Nassau Grouper spawning. Passive acoustic localization was used to find positions of the grouper-produced calls recorded during this time, which enabled the measurement of call source levels and evaluation of spatiotemporal aspects of calling. Yellowfin Grouper had the lowest mean peak-to-peak (PP) call source level, and Nassau Grouper had the highest mean PP call source level (143.7 and 155.2 dB re: 1 μPa at 1 m for 70–170 Hz, respectively). During the days that Nassau Grouper spawned, calling peaked after sunset. Similarly, when Red Hind calls were abundant, calls were highest in the afternoon and evening. The measured source levels can be used to estimate communication and detection ranges and implement passive acoustic density estimation for these fishes.
- 1 May 2022
- Published by Acoustical Society of America (ASA) in The Journal of the Acoustical Society of America
- Vol. 151 (5) , 3052-3065
Open AccessGrouper source levels and aggregation dynamics inferred from passive acoustic localization at a multispecies spawning site
Four species of grouper (family Epinephlidae), Red Hind (Epinephelus guttatus), Nassau (Epinephelus striatus), Black (Mycteroperca bonaci), and Yellowfin Grouper (Mycteroperca venenosa) share an aggregation site in Little Cayman, Cayman Islands and produce sounds while aggregating. Continuous observation of these aggregations is challenging because traditional diver or ship-based methods are limited in time and space. Passive acoustic localization can overcome this challenge for sound-producing species, allowing observations over long durations and at fine spatial scales. A hydrophone array was deployed in February 2017 over a 9-day period that included Nassau Grouper spawning. Passive acoustic localization was used to find positions of the grouper-produced calls recorded during this time, which enabled the measurement of call source levels and evaluation of spatiotemporal aspects of calling. Yellowfin Grouper had the lowest mean peak-to-peak (PP) call source level, and Nassau Grouper had the highest mean PP call source level (143.7 and 155.2 dB re: 1 μPa at 1 m for 70–170 Hz, respectively). During the days that Nassau Grouper spawned, calling peaked after sunset. Similarly, when Red Hind calls were abundant, calls were highest in the afternoon and evening. The measured source levels can be used to estimate communication and detection ranges and implement passive acoustic density estimation for these fishes.
- Open Access
- 23 March 2022
- Published by Frontiers Media SA in Frontiers in Marine Science
- Vol. 9
Editorial: Innovation and Discoveries in Marine Soundscape Research
Editorial: Innovation and Discoveries in Marine Soundscape Research
- 23 March 2022
- Published by Frontiers Media SA in Frontiers in Marine Science
- Vol. 9
Open AccessEditorial: Innovation and Discoveries in Marine Soundscape Research
Editorial: Innovation and Discoveries in Marine Soundscape Research
- Open Access
- 10 March 2022
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 17 (3) , e0264214
Characterization of fin whale song off the Western Antarctic Peninsula
Song is produced by a variety of terrestrial and marine animals and is particularly common among baleen whales. Fin whale (Balaenoptera physalus) song is comprised of relatively simple 20 Hz pulses produced at regular intervals. The timing of these intervals, in addition to the presence and frequency of overtones, appears to be unique to each population. The purpose of this study was to characterize Western Antarctic Peninsula fin whale song and describe temporal pattern variations in song type and occurrence. Recordings were collected in the area from 2001–2004 and again 2014–2016. One song type was identified with a primary inter-pulse interval (IPI) of approximately 14 s and secondary IPI of 12.5 s. This song occurred in three pattern variants: singlet, doublet, and long triplet. The interval between pulses increased by 1.5 s between recording periods while the frequency of the overtones decreased from 89 Hz to 86 Hz. Song was never recorded in August and while it was recorded at other times in some years, it was consistently present in recordings from April through June across all years. While multiple pattern variants were present each year, singlets were generally the most prevalent variant. Doublets and triplets occurred from February through June, with highest levels of variants in February. In later years the triplet variant presence increased and in 2016 it comprised 53% of recorded song bouts. Further research is needed to understand the reasons why song changes over time and to examine the feasibility of using song to delineate and identify populations.
- 10 March 2022
- Published by Public Library of Science (PLoS) in PLOS ONE
- Vol. 17 (3) , e0264214
Open AccessCharacterization of fin whale song off the Western Antarctic Peninsula
Song is produced by a variety of terrestrial and marine animals and is particularly common among baleen whales. Fin whale (Balaenoptera physalus) song is comprised of relatively simple 20 Hz pulses produced at regular intervals. The timing of these intervals, in addition to the presence and frequency of overtones, appears to be unique to each population. The purpose of this study was to characterize Western Antarctic Peninsula fin whale song and describe temporal pattern variations in song type and occurrence. Recordings were collected in the area from 2001–2004 and again 2014–2016. One song type was identified with a primary inter-pulse interval (IPI) of approximately 14 s and secondary IPI of 12.5 s. This song occurred in three pattern variants: singlet, doublet, and long triplet. The interval between pulses increased by 1.5 s between recording periods while the frequency of the overtones decreased from 89 Hz to 86 Hz. Song was never recorded in August and while it was recorded at other times in some years, it was consistently present in recordings from April through June across all years. While multiple pattern variants were present each year, singlets were generally the most prevalent variant. Doublets and triplets occurred from February through June, with highest levels of variants in February. In later years the triplet variant presence increased and in 2016 it comprised 53% of recorded song bouts. Further research is needed to understand the reasons why song changes over time and to examine the feasibility of using song to delineate and identify populations.
- Open Access
- 15 October 2021
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 9 (10) , 1128
Applying Artificial Intelligence Methods to Detect and Classify Fish Calls from the Northern Gulf of Mexico
Passive acoustic monitoring is a method that is commonly used to collect long-term data on soniferous animal presence and abundance. However, these large datasets require substantial effort for manual analysis; therefore, automatic methods are a more effective way to conduct these analyses and extract points of interest. In this study, an energy detector and subsequent pre-trained neural network were used to detect and classify six fish call types from a long-term dataset collected in the northern Gulf of Mexico. The development of this two-step methodology and its performance are the focus of this paper. The energy detector by itself had a high recall rate (>84%), but very low precision; however, a subsequent neural network was used to classify detected signals and remove noise from the detections. Image augmentation and iterative training were used to optimize classification and compensate for the low number of training images for two call types. The classifier had a relatively high average overall accuracy (>87%), but classifier average recall and precision varied greatly for each fish call type (recall: 39–91%; precision: 26–94%). This coupled methodology expedites call extraction and classification and can be applied to other datasets that have multiple, highly variable calls.
- 15 October 2021
- Published by MDPI AG in Journal of Marine Science and Engineering
- Vol. 9 (10) , 1128
Open AccessApplying Artificial Intelligence Methods to Detect and Classify Fish Calls from the Northern Gulf of Mexico
Passive acoustic monitoring is a method that is commonly used to collect long-term data on soniferous animal presence and abundance. However, these large datasets require substantial effort for manual analysis; therefore, automatic methods are a more effective way to conduct these analyses and extract points of interest. In this study, an energy detector and subsequent pre-trained neural network were used to detect and classify six fish call types from a long-term dataset collected in the northern Gulf of Mexico. The development of this two-step methodology and its performance are the focus of this paper. The energy detector by itself had a high recall rate (>84%), but very low precision; however, a subsequent neural network was used to classify detected signals and remove noise from the detections. Image augmentation and iterative training were used to optimize classification and compensate for the low number of training images for two call types. The classifier had a relatively high average overall accuracy (>87%), but classifier average recall and precision varied greatly for each fish call type (recall: 39–91%; precision: 26–94%). This coupled methodology expedites call extraction and classification and can be applied to other datasets that have multiple, highly variable calls.