Why we urgently need new approaches to monitoring biodiversity in the sea

Imagine boarding a bus that is about to drive off a cliff, knowing that the driver is partially sighted. Sadly, this scenario reflects the current state of affairs. Humanity has acknowledged that we are the primary drivers of the sixth mass extinction in the history of our planet. And, although we know that we are the driver of this phenomenon, the global response is insufficient.

The United Nations recognizes biodiversity loss as one of the most pressing challenges of the 21st century, on par with climate change and closely intertwined with it.

Well, most of what we call 'ecosystem services,' food production, for example, or protection from storms in coastal areas, rely on biodiversity. The crops we grow need pollinating by insects, the fish we eat need shelter and food to grow and coral reefs need different types of fish to protect them from predators. Each time a species disappears from an ecosystem, one or more species may suffer from it. This is why the loss of biodiversity can also accelerate the extinction of species by a chain of reaction.

Predictions show that depending on future global emissions, we will lose 6 to 10% of global biodiversity by 2050 and up to 27% by 2100. That means that about one million species are facing extinction. Many of them we rely on directly. The big question you may ask yourself now is: is the human race part of the one million that is facing extinction?

What measures are taken in the next two decades will be most decisive.

The Intergovernmental Science-Policy Platform on Biodiversity and Ecosystem Services (IPBES) reports that the five biggest culprits for this loss of biodiversity are, in order, of importance:

1. Changes in land and sea use (loss of habitat)

2. Direct exploitation of organisms

3. Climate change

4. Pollution

5. Invasive alien species

For the sea, climate change may play a more important role than for the land since many species are very sensitive to changes in temperature, but where we can have the fastest and most direct positive impact is on the loss of habitat and overexploitation. Reducing emissions is an important step, but if we can make sure that marine species have healthy habitats to strive in when the temperature rises, we are offering them better chances of survival. Unfortunately, the lack of understanding of the distribution of life in the ocean is a significant barrier to restoring its biodiversity and health.

Nowadays, a lot of projects are happening in the ocean. Some impact biodiversity more negatively, such as land reclamation or dredging, and some are trying to have a positive impact, such as ecosystem restoration and conservation through marine-protected areas and policies.

To be efficient in the way we protect and restore marine biodiversity, it is necessary to understand what is happening and this is a challenge when looking at marine biodiversity. Standardization of monitoring protocols and data analysis methods is essential to ensure consistent and comparable data over time and across different locations. It’s the only way to promote effective marine conservation and provide decision-makers with valuable information on the best practices and measures to implement. If we don’t have this information, we are like that passenger on the bus going down the cliff with the partially-sighted driver.

Today, the most traditionally used method to monitor marine biodiversity is visual surveys. This consists of a diver swimming in the water and counting fish or filming and then identifying them on the video. This method provides information on the abundance, distribution and diversity of fish populations. It is time-consuming, however, and requires skilled observers, which are not available everywhere and for every project. It’s expensive, difficult to scale up and misses a lot of valuable information in between surveys. Additionally, it can lead to the 'Pandora's box' effect, where the presence of the observer influences what it is trying to monitor.

Fishing surveys are often used to get data on the biomass and diversity of commercial fish species. They offer insights into the size, age and precise biomass of these fish populations. They are biased, however, towards larger and more easily caught fish, have a negative impact on non-target species, can be destructive to the ecosystem and are considered the most invasive way to monitor, as most fish caught in the net die. This type of data can be collected from specialized survey ships, but also from voluntary-based recreational fishing and commercial fisheries, as they have to report on their catch. The reliability of the results is directly linked to the integrity of the person filing the report.

Passive acoustic sensors register the ambient noise of a reef or other aquatic environment. The recorded audio is processed using artificial intelligence (AI) to identify and quantify some of the species present near the sensor. The species that can be identified include most marine mammals, some fish and also other types of animals, such as sea urchins. This method provides information on the abundance, distribution and behaviour of organisms at a relatively low cost. Its accuracy is limited in areas with high ambient noise, however, or complex seafloor topography and it can only identify species when they produce sounds.

Environmental DNA or eDNA refers to the genetic material shed by organisms into the environment, allowing for species detection and identification without direct observation or capture. This non-invasive method provides a comprehensive snapshot of species richness. It is non-quantitative, however, and requires specialized equipment and expertise and is expensive.

Additionally, there are some challenges when using this method for fish richness. A lot of water needs to be sampled and it is not certain to have all the species represented as some fish shed more DNA than others. Plus, the DNA collected in the water can be up to three months old, depending on environmental factors that can degrade DNA, such as UV, temperature, pH or the presence of other molecules that could bind to DNA. During these weeks or months, the DNA is transported in the currents, therefore, a fish present in an E-DNA sample may have lost this DNA hundreds of km away and inversely, a fish may have been nearby, but its DNA drifted far away.

A baited fish camera, as its name describes, is a system that attracts fish by placing bait in front of the camera. The camera usually records for a couple of hours. It is less labour-intensive than visual surveys, non-invasive and eliminates diver bias, making it suitable for remote areas inaccessible to divers. It has a short timeframe for data collection, however, and results can be biased by the addition of food and light. Only the fish attracted by the specific type of bait used will come to eat it in front of the camera. Regular on-site monitoring is still required over longer periods and the area sampled depends on the current direction and velocity.

An operational AI camera system is an underwater surveillance system that recognizes fish species. It’s a non-invasive method that utilizes AI and self-cleaning cameras to collect data all day, every day, providing long-term monitoring and engaging images. It works in remote locations as it can be solar-powered. This technology removes the bias induced by a human and quantifies any impact on fish biodiversity precisely over time and space.

Overall, each of these methods has strengths and weaknesses. Still, it is evident that there are some gaps in the spatial and temporal capacities of monitoring, as well as the quantification. Like the driver of the bus, we need to have a clear picture of the state and evolution of biodiversity in the marine environment to avoid the worst and for this, the traditional methods of monitoring are insufficient. Fortunately, technological advancement is coming rapidly and the use of imagery and AI mark a turning point in underwater monitoring.

Stream Ocean AG, a winner of the UpLink Coastal Tourism Challenge, has developed the first underwater camera system that allows cameras to stay autonomously powered, connected to the internet and utilizes AI to establish real-time marine life surveillance. The underwater video monitoring system provides data on fish populations and their changes, giving 24/7 access to data and video via an online platform. This system has the potential to revolutionize the way we monitor and manage fish populations and could help ensure the long-term sustainability of marine ecosystems. By providing long-term, accurate data on fish populations, we can make informed decisions about conservation and management strategies that will help protect and restore the biodiversity of aquatic ecosystems.

Only by giving a better vision of the situation to the bus driver, will we be able to take the right path and avoid the cliff.