Apply

Introduction

Freshwater ecosystems are the most degraded and threatened of all ecosystems on the planet(1).

They also host a high diversity of vertebrate species, many of which are critically endangered (e.g. cetaceans, reptiles, amphibians, and fish)(2). The high biodiversity and threats from human activity create “biodiversity hotspots”(3).
Threats are posed from a range of anthropogenic activities that include overexploitation for food, habitat degradation and destruction, water pollution, and fragmentation of rivers through the construction of infrastructure (e.g. large dams) required to support increased urbanisation. Technologies and management strategies are developed to mitigate environmental impacts of these activities, such as the construction of fish passes designed to facilitate the negotiation of dams by migratory fish, or screening systems the prevent them entering dangerous areas, such as hydropower turbines.
Unfortunately, the efficiencies of fish passes and screens can be highly variable, ranging from those that work relatively well, to others that totally fail – and in the worst cases can have negative impacts on the species they were designed to protect(4). To advance the next generation of technologies. There is a need to understand why current system fail through effective monitoring of their operations and outcomes. The results can then be used to inform future design. A range of techniques are used to monitor the performance of environmental impact mitigation technology, including the use of video imagery and photography, acoustic cameras when visual techniques are inappropriate (e.g. when dark or under turbid conditions), telemetry used to describe trajectories of the animals themselves (e.g. radio or acoustic tagging), or other techniques such as aerial or satellite imagery to track animal movements. However, such data is often poorly analysed due to ineffective techniques needed to handle big data with considerable amounts of material archived never to be reviewed. AI is likely to disrupt this field and be a major game changer through enhancing the efficiency of current environmental monitoring techniques(5).

Working with the Environment Agency and the water industry this project aims to advance effective monitoring of environmental impact mitigation technologies designed to protect freshwater ecosystems through the use of AI. The key objectives are:

Review the range of data collected by the Environment Agency for the purpose of monitoring the effectiveness of environmental impact mitigation technology, such as fish passes or screening systems, or more generic water quality information available to identify pollution events.
Prioritise the greatest needs of the regulatory agency to close information gaps in key areas related to protecting freshwater ecosystems.
Identify available data that would help address the challenges identified.
Build appropriate AI solutions to resolve the data challenges selected to improve UK environmental monitoring programmes operating in freshwater ecosystems.