bioRxiv Science⌕ Search

Biology subjects

Wegmann, A.

Publications and source records attributed to Wegmann, A..

2 recordsLinked to original sources

Fish Aggregating Devices could enhance the effectiveness of blue water MPAs

In the past two decades, drifting fish aggregation devices (FADs) have revolutionised pelagic fisheries, and are now responsible for the majority of tuna purse seine catches. Here, we argue that by taking advantage of the same proven aggregative properties, FADs could be used to enhance the benefits provided by blue water Marine Protected Areas (MPAs). Using models of commercially-targeted fish populations, we explore the potential benefits that could be achieved if unfished conservation FADs were positioned within blue water MPAs. Our results suggest that conservation FADs could deliver benefits, both to target species and the broader ecosystem. By increasing the residence time of exploited species, conservation FADs will reduce average mortality rates inside MPAs. By increasing the local density of species whose populations are depressed by exploitation, FADs can also improve the function of ecosystems in blue water MPAs. Conservation FADs could therefore amplify the benefits of blue water MPAs. We find this amplification is largest in those contexts where blue water MPAs have attracted the most criticism - when their area is small compared to both the open ocean and the distribution of fish stocks that move through them.

ecology↗

A general deep learning model for bird detection in high resolution airborne imagery

Advances in artificial intelligence for computer vision hold great promise for increasing the scales at which ecological systems can be studied. The distribution and behavior of individuals is central to ecology, and computer vision using deep neural networks can learn to detect individual objects in imagery. However, developing supervised models for ecological monitoring is challenging because it needs large amounts of human-labeled training data, requires advanced technical expertise and computational infrastructure, and is prone to overfitting. This limits application across space and time. One solution is developing generalized models that can be applied across species and ecosystems. Using over 250,000 annotations from 13 projects from around the world, we develop a general bird detection model that achieves over 65% recall and 50% precision on novel aerial data without any local training despite differences in species, habitat, and imaging methodology. Fine-tuning this model with only 1000 local annotations increase these values to an average of 84% recall and 69% precision by building on the general features learned from other data sources. Retraining from the general model improves local predictions even when moderately large annotation sets are available and makes model training faster and more stable. Our results demonstrate that general models for detecting broad classes of organisms using airborne imagery are achievable. These models can reduce the effort, expertise, and computational resources necessary for automating the detection of individual organisms across large scales, helping to transform the scale of data collection in ecology and the questions that can be addressed.

ecology↗