Search

Vegetation in East Antarctica, such as moss and lichen, vulnerable to the effects of climate change and ozone depletion, requires robust non-invasive methods to monitor its health condition. Despite the increasing use of unmanned aerial vehicles (UAVs) to acquire high-resolution data for vegetation analysis in Antarctic regions through artificial intelligence (AI) techniques, the use of multispectral imagery and deep learning (DL) is quite limited. This study addresses this gap with two pivotal contributions: (1) it underscores the potential of deep learning (DL) in a field with notably limited implementations for these datasets; and (2) it introduces an innovative workflow that compares the performance between two supervised machine learning (ML) classifiers: Extreme Gradient Boosting (XGBoost) and U-Net. The proposed workflow is validated by detecting and mapping moss and lichen using data collected in the highly biodiverse Antarctic Specially Protected Area (ASPA) 135, situated near Casey Station, between January and February 2023. The implemented ML models were trained against five classes: Healthy Moss, Stressed Moss, Moribund Moss, Lichen, and Non-vegetated. In the development of the U-Net model, two methods were applied: Method (1) which utilised the original labelled data as those used for XGBoost; and Method (2) which incorporated XGBoost predictions as additional input to that version of U-Net. Results indicate that XGBoost demonstrated robust performance, exceeding 85% in key metrics such as precision, recall, and F1-score. The workflow suggested enhanced accuracy in the classification outputs for U-Net, as Method 2 demonstrated a substantial increase in precision, recall and F1-score compared to Method 1, with notable improvements such as precision for Healthy Moss (Method 2: 94% vs. Method 1: 74%) and recall for Stressed Moss (Method 2: 86% vs. Method 1: 69%). These findings contribute to advancing non-invasive monitoring techniques for the delicate Antarctic ecosystems, showcasing the potential of UAVs, high-resolution multispectral imagery, and ML models in remote sensing applications.

Vava'u, Kingdom of Tonga, is a well-established whale-watching destination in the South Pacific. Between July and October, the waters around the archipelago represent one of the most important breeding grounds for Oceania humpback whales (Megaptera novaeangliae). The Tongan government allows tourist swimming activities with whales and tour operators strongly promote the practice of swimming-with-whales, focusing primarily on mother-calf pairs. However, there is increasing evidence, derived from empirical research on swim-with-cetacean tourism, that this kind of interaction affects cetacean behaviour and can lead to negative effects on the cetaceans involved. This study represents the first assessment of humpback whales' behavioural responses to vessel and swimmer approaches in Vava'u. Fifty-six surveys took place during the 2016 and 2017 whale breeding seasons aboard dedicated research and tour vessels. Whale dive time, number of reorientation events, and respiration rates were documented in both the absence and presence of boats and swimmers. Vessel approach type, swimmer placement, and whale avoidance responses were also recorded. Results indicate that the average diving time and the proportion of time spent diving in the presence of swimming activities increased significantly for mother-calf pairs (F2,36 = 18.183, P < 0.001; F2,36 = 5.462, P = 0.009, respectively). Moreover, avoidance responses of whales towards tour vessels were observed for one third of vessel approaches (33.5%) and the avoidance rate was significantly affected by the boat approach type (95% CI: 20.7-69.2%, z = 3.50, P < 0.001). Finally, low levels of compliance to the existing Tongan swim-with-whales regulations were documented, in particular the stipulated whale resting time between interactions with tour operator vessels and swimmers was often not respected (38.4%). Vava'u is an important calving ground for the Oceania humpback whale population and these findings should be carefully considered by stakeholders in Tonga and at other locations where swim-with-whales opportunities are being undertaken. Effective strategies to reduce the risk of detrimental effects on the whales targeted by swimming activities, especially mother-calf pairs, are needed., Competing Interests: The authors have declared that no competing interests exist.

Unmanned aerial vehicles (UAVs) represent a novel and cost effective research tool to investigate cetacean behaviour, as conventional aircraft are expensive, limited in the altitude they can fly at and potentially disturb sensitive wildlife. In addition, the aerial observation from the UAVs allows assessment of cetacean behaviour from an advantageous perspective and can collect high spatial and temporal resolution data, providing the opportunity to gather accurate data about group size, age class and subsurface behaviour. However, concerns have been raised about the potential risks of disturbance to animals caused by the UAV's visual and acoustic stimuli. Boat-based surveys were conducted to assess the short-term behavioural responses of resting bottlenose dolphins (Tursiops truncatus) to a lightweight Vertical take-off and landing (VTOL) UAV flown at 10, 25, and 40 m altitude. Changes in group swim direction and frequencies of surface and aerial behavioural events were recorded from an anchored research vessel before (control) and during the aerial survey. The number of reorientation and tail slap events increased significantly between controls and flights when the UAV was flown at 10 m over the animals. In contrast, no significant differences were detected when the aircraft was flown at 25 and 40 m altitude. However, a precautionary approach is recommended for research applications requiring lower flight altitudes, with further research recommended to assess how different cetacean species and age class may respond to the UAV presence.