Using GIS and stakeholder involvement to innovate marine mammal bycatch risk assessment in data-limited fisheries.

Fisheries bycatch has been identified as the greatest threat to marine mammals worldwide. Characterizing the impacts of bycatch on marine mammals is challenging because it is difficult to both observe and quantify, particularly in small-scale fisheries where data on fishing effort and marine mammal abundance and distribution are often limited. The lack of risk frameworks that can integrate and visualize existing data have hindered the ability to describe and quantify bycatch risk. Here, we describe the design of a new geographic information systems tool built specifically for the analysis of bycatch in small-scale fisheries, called Bycatch Risk Assessment (ByRA). Using marine mammals in Malaysia and Vietnam as a test case, we applied ByRA to assess the risks posed to Irrawaddy dolphins (Orcaella brevirostris) and dugongs (Dugong dugon) by five small-scale fishing gear types (hook and line, nets, longlines, pots and traps, and trawls). ByRA leverages existing data on animal distributions, fisheries effort, and estimates of interaction rates by combining expert knowledge and spatial analyses of existing data to visualize and characterize bycatch risk. By identifying areas of bycatch concern while accounting for uncertainty using graphics, maps and summary tables, we demonstrate the importance of integrating available geospatial data in an accessible format that taps into local knowledge and can be corroborated by and communicated to stakeholders of data-limited fisheries. Our methodological approach aims to meet a critical need of fisheries managers: to identify emergent interaction patterns between fishing gears and marine mammals and support the development of management actions that can lead to sustainable fisheries and mitigate bycatch risk for species of conservation concern.

State-space modelling reveals proximate causes of harbour seal population declines.

Declines in large vertebrate populations are widespread but difficult to detect from monitoring data and hard to understand due to a multiplicity of plausible biological explanations. In parts of Scotland, harbour seals (Phoca vitulina) have been in decline for 10 years. To evaluate the contributions of different proximate causes (survival, fecundity, observation artefacts) to this decline, we collated behavioural, demographic and population data from one intensively studied population in part of the Moray Firth (north-east Scotland). To these, we fit a state-space model comprising age-structured dynamics and a detailed account of observation errors. After accounting for culling (estimated by our model as 14% of total mortality), the main driver of the historical population decline was a decreasing trend in survival of young individuals combined with (previously unrecognised) low levels of pupping success. In more recent years, the model provides evidence for considerable increases in breeding success and consistently high levels of adult survival. However, breeding success remains the most volatile demographic component of the population. Forecasts from the model indicate a slow population recovery, providing cautious support for recent management measures. Such investigations of the proximate causes of population change (survival, fecundity and observation errors) provide valuable short-term support for the management of population declines, helping to focus future data collection on those ultimate causal mechanisms that are not excluded by the demographic evidence. The contribution of specific ultimate drivers (e.g. shooting mortality or competitors) can also be quantified by including them as covariates to survival or fecundity.

Automatic detection and classification of odontocete whistles.

Methods for the fully automatic detection and species classification of odontocete whistles are described. The detector applies a number of noise cancellation techniques to a spectrogram of sound data and then searches for connected regions of data which rise above a pre-determined threshold. When tested on a dataset of recordings which had been carefully annotated by a human operator, the detector was able to detect (recall) 79.6% of human identified sounds that had a signal-to-noise ratio above 10 dB, with 88% of the detections being valid. A significant problem with automatic detectors is that they tend to partially detect whistles or break whistles into several parts. A classifier has been developed specifically to work with fragmented whistle detections. By accumulating statistics over many whistle fragments, correct classification rates of over 94% have been achieved for four species. The success rate is, however, heavily dependent on the number of species included in the classifier mix, with the mean correct classification rate dropping to 58.5% when 12 species were included.

The effects of acoustic misclassification on cetacean species abundance estimation.

To estimate the density or abundance of a cetacean species using acoustic detection data, it is necessary to correctly identify the species that are detected. Developing an automated species classifier with 100% correct classification rate for any species is likely to stay out of reach. It is therefore necessary to consider the effect of misidentified detections on the number of observed data and consequently on abundance or density estimation, and develop methods to cope with these misidentifications. If misclassification rates are known, it is possible to estimate the true numbers of detected calls without bias. However, misclassification and uncertainties in the level of misclassification increase the variance of the estimates. If the true numbers of calls from different species are similar, then a small amount of misclassification between species and a small amount of uncertainty around the classification probabilities does not have an overly detrimental effect on the overall variance. However, if there is a difference in the encounter rate between species calls and/or a large amount of uncertainty in misclassification rates, then the variance of the estimates becomes very large and this dramatically increases the variance of the final abundance estimate.