Viviane R. Barroso

Tropical reefs are among the most biodiverse and threatened marine ecosystems and face both local and global stressors. Understanding community dynamics and ecosystem functioning under these pressures requires long-term monitoring. Image-based surveys are increasingly used in reef ecology because they enable rapid data acquisition and storage. However, manual identification of benthic organisms is time-consuming and potentially subjective, creating a bottleneck between image collection and ecological analysis. Automated annotation tools represent promising solutions, but few studies have directly compared their performance with that of manual approaches, especially under varying image quality conditions. Here, we assess the usefulness of the CoralNet automated tool for classifying benthic cover in tropical reef images and evaluate how image quality influences both manual and automated annotations. Using expert-validated images, we trained and tested both approaches, measuring performance with Cohen’s Kappa coefficient and modeling the effects of image imperfections. We found that manual and automated annotators achieved similar performance across most taxonomic and morphofunctional groups, and both were similarly affected by image quality. Importantly, although the automated annotator showed a slightly lower performance than the manual annotator, image-quality imperfections had a stronger influence on annotation performance than annotator type itself. Our findings indicate that automated annotation can provide comparable performance to manual methods once appropriately trained, with strong potential to enhance large-scale monitoring, biodiversity assessments, and conservation strategies in reef ecosystems.

Acoustic communication plays an important role in social interactions in reef fishes, but how signals vary with behavioural context remains poorly understood for many species. We characterised the acoustic repertoire of the Brazilian damselfish Stegastes fuscus using synchronised underwater audio–video recordings and tested how sound features vary across social behaviours. Sounds were recorded during agonistic and courtship‐related interactions, and temporal and spectral acoustic parameters were measured to compare acoustic structure among contexts. Differences between behaviours were primarily found in temporal features: Courtship sounds were longer and contained more pulses than agonistic sounds. These results show that S. fuscus conveys social context mainly through modulation of temporal organisation rather than spectral properties. Our findings indicate that a single call type can be flexibly adjusted to signal behavioural context during territorial defence and courtship. By documenting the acoustic repertoire of a dominant territorial herbivore, this study highlights how social context structures sound production in damselfishes.

The growing demand for offshore oil and gas production in deep waters has motivated the development of technologies to enable the continuous, reliable, and cost-effective monitoring of subsea equipment. Traditional inspection techniques rely on ROVs and AUVs, leading to delays between data acquisition and recovery and high operational costs. Underwater acoustic communication systems represent an attractive alternative for transmitting monitoring data to the surface in real time. This work evaluates the feasibility of implementing an underwater acoustic communication link for data transmission in deep-water environments, considering environmental conditions and acoustic channel characteristics. Using the BELLHOP ray-tracing model, simulations were performed to predict transmission loss, multipath effects, ambient noise, and the resulting signal-to-noise ratio (SNR) for different modem configurations and operating frequencies. The results demonstrate that the performance of the underwater link is strongly dependent on frequency, distance, and environmental variability. The study identifies optimal frequency–range relationships, quantifies the limitations imposed by transmission loss and ambient noise, and provides guidance for selecting acoustic modem parameters for real subsea monitoring applications. The SNR for three modem models operating at different frequencies illustrates the signal detection capability in the marine environment. The differences between modems A, B, and C are defined by their technical specifications and how they perform within the underwater acoustic channel of the Campos Basin. The data transmission capacity is supported by the data rates provided by the analyzed modems. The low frequencies of modem A (9.75 kHz) achieve the highest SNR, enabling long-range monitoring. At higher frequencies, modem C (78 kHz) allows short-distance communication. Modem B (35 kHz) offers a good balance between the data rate and power consumption, consuming only 1 W, making it highly viable for monitoring systems that rely on batteries and require long-term operation. The findings support the feasibility of integrating underwater acoustic communication into subsea monitoring architectures, enabling a more efficient oversight of deep-water production systems. The analysis concludes that project viability depends on selecting a system where the SNR and range meet the specific monitoring requirements.

Fish produce a wide variety of sounds that contribute to the soundscapes of aquatic environments. In reef systems, these sounds are important acoustic cues for various ecological processes. Artificial intelligence methods to detect, classify and identify fish sounds have become increasingly common. This study proposes the classification of unknown fish sounds recorded in a subtropical rocky reef using different feature sets, data augmentation and explainable artificial intelligence tools. We used different supervised algorithms (naive Bayes, random forest, decision trees and multilayer perceptron) to perform a multiclass classification of four classes of fish pulsed sounds. The proposed models showed excellent performances, achieving 98.1% of correct classification with multilayer perceptron using data augmentation. Explainable artificial intelligence allowed us to identify which features contributed to predict each sound class. Recognizing and characterizing these sounds is key to better understanding diel behaviours and functional roles associated with critical reef ecological processes. This article is part of the theme issue ‘Acoustic monitoring for tropical ecology and conservation’.

Sun coral (Tubastraea spp.) is an invasive species that poses a considerable threat to coastal ecosystems. Therefore, early detection is essential for effective monitoring and mitigation of its negative impacts on marine biodiversity. This study presents a novel computer vision approach for automated early detection of invasive Tubastraea species in underwater images. We used the YOLOv8 object detection model, which was trained and validated on a manually annotated dataset augmented with synthetic images. The data augmentation addressed the challenge of limited training data that is prevalent in underwater environments. The model achieved performance metrics (in terms of precision accuracy, recall, mAP50, and F1 score) of over 90% and detected both open and closed coral stage classes. Test phase results were compared with expert validation, demonstrating the model’s effectiveness in rapid detection (16 ms) and its limitations in areas highly covered by Tubastraea. This study demonstrates the potential of deep learning with data augmentation to facilitate the rapid assessment of large image datasets in monitoring sun coral bioinvasion. This approach has the potential to assist managers, taxonomists, and other professionals in the control of invasive alien species.

Aquatic ecosystems are constantly changing due to anthropic stressors, which can lead to biodiversity loss. Ocean sound is considered an essential ocean variable, with the potential to improve our understanding of its impact on marine life. Fish produce a variety of sounds and their choruses often dominate underwater soundscapes. These sounds have been used to assess communication, behaviour, spawning location, and biodiversity. Artificial intelligence can provide a robust solution to detect and classify fish sounds. However, the main challenge in applying artificial intelligence to recognize fish sounds is the lack of validated sound data for individual species. This review provides an overview of recent publications on the use of machine learning, including deep learning, for fish sound detection, classification, and identification. Key challenges and limitations are discussed, and some points to guide future studies are also provided.