The seafloor can be so rich and diverse in animals and habitats that it has often been compared to tropical rainforests but, conversely, the seafloor can be so lacking in larger more visible animals that it has also been likened to deserts (Snelgrove and Grassle, 1995). Not only is the biodiversity hugely variable between areas of the seabed, but the particular species and communities that inhabit these areas can be almost unpredictable, almost…
Consider our current predicament and the predicament of all deep-sea biologists when trying to locate a specific species or habitat in such an extremely varied and expansive ecosystem – we’re going to need a map!
But how do we create a map when we know so little about what is at the bottom of our oceans? You’ve probably heard that we know more about the surface of the moon than our planet’s seafloor, that’s because there is relatively little obstructing our view of the moon when compared to the miles of water between us and the seafloor. Although it isn’t possible to see the seafloor from the surface, we can hear it. By using multibeam sonar, we can detect the depth of the seafloor below a vessel and use this information to create detailed maps.
So, now we have a map of the sea floor. By combining this information with any information we have on the conditions below the surface (temperature, salinity, currents, etc.), we can attempt to predict what types of animals may be there.
Mathematicians have weaved sorcery that allows you to determine which environmental conditions certain species prefer, purely based on the conditions in areas these species have previously been found. These mathematical wizards call this a model, which can be used to understand many different processes in many different fields. We, however, use these models to locate deep-sea creatures in the interests of science and conservation.
Deep-sea coral reefs, much like their shallow water cousins, only thrive under particular conditions. Using everything we know about the seafloor, we have created maps of where our models predict these conditions – and therefore the coral reefs – will occur.
We are particularly interested in collecting corals from various sites and depths in order to estimate the level of connectivity among these different populations. This means we have to take an educated guess as to where these corals are using our models and then send the ROV in to investigate and hopefully collect some specimens.
Seeing first-hand if what a model predicts is the same as what is actually on the seafloor is called ‘groundtruthing’. This groundtruthing procedure is the ultimate test of a predictive model and so far we’ve had a lot of success. If we send the ROV down to an area where we’ve predicted coral, more often than not we will come across spectacular reefs that have formed entire landscapes over thousands of years of growth. Of course, no model is perfect and we have stumbled into sandy plains where we have predicted coral reefs. Even a mistake made by a predictive model can offer interesting insights into what conditions species prefer and what has been overlooked by the model. For example there may be some physical barrier, such as a seamount, that prevents certain species from colonising an area.
Information on the ability of certain species to populate areas of the deep sea is particularly limited and is dependent on many aspects of the species biology as well as the local conditions, in particular the currents along the seafloor. Fortunately, our present expedition aims to answer various questions about the distribution and connectivity of deep-sea populations.
Now, we find ourselves in a strange but not uncommon scenario where we use predictive models to guide our study in order to collect information to guide our models. Although improving our current models is not the primary focus of our research it may one day lead to increasing success in future expeditions.
Snelgrove PV and Grassle JF (1995). The Deep Sea: Desert and Rainforest. Oceanus-Woods Hole Oceanographic Institute, Massachusetts, vol 38, pages 25-29.
Text by Oti Brunner, Plymouth University