Biodiversity is an important aspect in environmental management and routinely monitoring benthic biodiversity allows us to better understand and manage human impacts on the aquatic ecosystems. Doing so also provides information about the presence of “indicator species", which, together with biodiversity provides a good indication of the environmental status and integrity of a particular site. In the European Union (EU), monitoring of marine and coastal ecosystems, relies heavily on benthic macroinvertebrates, and is regulated in the EU Water Framework Directive and Marine Strategy Framework Directive (MSFD).
In 2010, ICES BEWG stated that environmental monitoring had to adapt towards cost-effective, integrative strategies as the higher cost of traditional morphological monitoring prevents it being used on larger scales. ICES Working Group on Phytoplankton and Microbial Ecology (WGPME), through their work on sampling methods and diversity issues of planktonic microbes, have explored the use of indicators and provide recommendations for methods development. This work will potentially harmonize methodological approaches.
Environmental DNA (eDNA) metabarcoding does overcome the cost challenges presented by morphological monitoring and while some microorganisms show particular promise as bioindicators, a lack of knowledge on how these microorganisms react to different types of pressure has led members of WGPME to investigate de novo data-driven approaches for identification of bioindicators.
For a recently published study, members of WGPME collected sediment samples from the coast and estuaries of 44 routinely monitored stations in estuaries along the Basque coast in the Bay of Biscay. By targeting the prokaryotic diversity in these samples using eDNA, they then evaluated two distinct approaches: supervised machine learning using Random Forests, and Threshold Indicator taxa combined with quantile regression splines. Both were evaluated for multi-impact assessment, i.e. the ability to predict not only the total environmental impact, but also what type of impact is affecting a specific site (e.g. eutrophication, heavy metals, or hydrocarbons).
Anders Lanzén, AZTI, member of WGPME, and lead author in this study, explains the process behind choosing the two approaches. “Compared to, for example macroinvertebrates or fishes, microorganisms are incredibly diverse, with at least hundreds of thousands of species found in a single area. Almost none of them are named species and not much is known about the ecological role of most. So, it was clear that we needed to come up with a way to relate the diversity data itself to environmental impact rather than trying to predict it from what was known from each group of organisms."
Although the machine learning method and the more traditional one based on quantile regression splines are very different, both performed comparably, and both could predict the impact of a given site based on its prokaryotic diversity and composition. This was better than the authors expected because the studied ecosystem was very complex, with many different types of environmental degradation.
Lanzén comments that their model system is a good example of the complexity and vulnerability of coastal and estuarine habitats, with multiple stressors (different industrial and urban pollutants, recalcitrant contamination and hydrological alteration) interacting with complex natural variability in terms of salinity, tidal currents, air exposure in intertidal sites and sediment types, among others. “This highlights the benefit of our approach to estimate multiple rather than a single impact based on community composition."
“Our results can help bridge the science–policy divide", notes Lanzén, “as policy relies on well-studied organisms, in this case benthic macroinvertebrates". While many studies have focused on how to reproduce the composition and diversity of such organisms based on eDNA, the authors feel this approach has several challenges. “First, larger organisms have patchy distribution and although it is
possible to use community DNA extracted from bulk samples of organisms, you
cannot typically extract a representative sample of eDNA from several kilograms of sediment in the way that you study the diversity of large organisms from a complete sediment grab. DNA is instead typically extracted from 0.2 to 10 g of sediment, meaning that the eDNA captured is subjected to extreme sampling effects (variation and bias). Further, the amount of eDNA does not correspond to the number of individuals".
Targeting microorganisms instead, say the authors, can overcome many of these challenges, but as they are not traditional indicators and very little is known about their ecology, they need to use de novo methods to figure this out. “This brings us far outside the comfort zone of most regulators, notes Lanzén, “Therefore, I think it is important to focus instead on the predictive power of such de novo methods, i.e. can we predict the impact by using them. This, in the end, is the most important aim of environmental monitoring". However, he notes that more work is needed to improve and evaluate the
accuracy of this approach, especially across geographical regions.
Read the paper A microbial mandala for environmental monitoring: Predicting multiple impacts on estuarine prokaryote communities of the Bay of Biscay in Molecular Ecology.
ICES Working Group on Phytoplankton and Microbial Ecology (WGPME) provides reviews and guidance on the sampling methods and diversity issues of phytoplankton and other planktonic microbes. WGPME's work on exploring the use of indicators and providing recommendations for methods development addresses Ecosystem science and Emerging techniques and technologies, two of ICES scientific priorities.
Discover all seven interrelated scientific priorities and how our network will address them in our Science Plan: “Marine ecosystem and sustainability science for the 2020s and beyond".
The Urdaibai estuary - a natural region and a Biosphere Reserve of Biscay, Basque Country, Spain.
A microbial mandala for environmental monitoring: Predicting multiple impacts on estuarine prokaryote communities of the Bay of Biscay
Authors:Anders Lanzén, Iñaki Mendibil, Ángel Borja, Laura Alonso‐Sáez