The difficulty of isolating stressors for marine science research

Imagine this scenario: your baby is crying in the library. You realize this is a problem, based on irritated looks from other library-goers, and so you wish to solve this problem. Babies cry to let you know that something is wrong, but they can't tell you exactly what the problem is. Therefore, you have two main choices: you can just randomly go through the list of possible reasons the kid is crying, or you can use a bit of scientific deduction to guess the most likely cause. In this situation, which do you think is the best use of your important resource: time?

I think it's faster to analyze the problem, and then try to tackle it, rather than just relying on trial and error. For instance, changing a perfectly dry diaper is a waste of time, and won't calm the baby down.

The same logic holds for tackling environmental problems. Conservation funds and energy are not inexhaustible, and there's no point fixing something that isn't broken. Therefore, a significant amount of research (including much of my own) is focused on identifying and ranking the importance the stressors affecting a particular ecosystem in distress. Once the reason for ecosystem decline is identified, management efforts can be focused on fixing that problem and the ecosystem may recover.

However, nature is messy and individual stressors are often difficult to isolate. For instance, physical impacts from diving tourists might be correlated with fishing pressure, just as your baby might be hot and hungry; in that case, measuring and addressing just one issue may not solve your problem.

Another problem is that stressors that are often considered and managed singly may actually be composed of a number of different factors, or sub-stressors. Here's an example using a slightly older child: your toddler may be hungry, but he may have an insatiable desire for blueberries, not cereal. While offering cereal is perfectly acceptable at other times, in this case it won't result in a contented child.

As an example of this second complication in the marine realm, sewage pollution is often identified as a problem for coral reefs. However, as Drs. Stephanie Wear and Rebecca Vega-Thurber point out in their recent article, sewage - even treated sewage - can introduce a variety of sub-stressors to coral reefs. These individual stressors may interact with one another to intensify the negative effects on corals (just like a hot and hungry baby is way more unpleasant than one who is simply overdressed). Nutrients and metals are both components of sewage and can increase pathogenic microbe populations. Each is also stressful to corals on its own, which might result in the corals being even more susceptible to microbial diseases. In that case, treating sewage to remove metals but not nutrients won't solve the problem.

Wear and Vega-Thurber suggest that sewage not be treated as an individual stressor in coral reef studies, but as a suite of multiple stressors with the potential for unexpected effects. Different components of sewage might (a) interact in an additive fashion to drive coral health in the same direction; (b) they may offset one another; or (c) they could synergize - like the hot and hungry child, the presence of one sub-stressor could make another worse. Indeed, because sewage in different places undergoes varying levels of treatment and may contain different compounds owing to variation in human lifestyles around the world, it's likely that the resulting "sewage impact" in one location varies greatly from another.

As an example, Sheila Walsh and I used changes in the community of benthic foraminifera to test the impact of fishing and sewage on nutrient availability on Kiritimati Island in the central Pacific. At the time, Kiritimati did not have any sewage treatment facilities, but much of the population also had no sanitary facilities, traditionally using the beach or open ocean nearshore waters for personal needs. The resulting dilution spread over a large area may explain why we found a weak relationship between estimated sewage inputs and nutrification, in contast to another study that found small villages in Indonesia introduced significant nutrients to nearshore waters.

Isolating stressors in nature is difficult, because there is a lot going on. The seasons, weather, and wave activity change over time and might affect how sewage discharge dissipates along the coast. Because nature is complicated, many stressor-specific studies have been conducted in controlled laboratory settings. For instance, if we want to know precisely how nitrogen affects corals, all else being equal, it's easiest to test this in an aquarium where other factors like temperature can be kept constant. However, results from controlled settings are not always completely translatable to the real world, particularly when stressors interact with one another. One way to tackle this is by collecting long-term observational data of as many factors as possible, and then testing for relationships between various stressors and observed changes in the community.

One problem with this get-all-the-data approach is that it's difficult to measure everything that might matter in a system. In addition to logistical challenges, there might be other stressors or other ways of measuring ecosystems that we haven't even thought of yet. For example, coral cover used to be the main "health" metric used to assess reefs, but degraded reefs may still have high coral cover because "weedy" corals have replaced more sensitive species. More sophisticated data are therefore increasingly replacing coral cover as a health metric.

Some retrospective "wishes" for data that were not previously collected can be satisfied by proxy-based reconstructions, where something is measured that represents the original dataset. For instance, the width of tree rings can be a proxy for past rainfall, with wider rings occurring during wetter years that allowed the trees to grow better. Coral cores - similar to tree rings - are a good place to start for coral reef studies. Cores can be used to reconstruct water temperature, salinity, pH, nutrients, sedimentation, and coral growth rates (which theoretically will respond to these environmental drivers), and potentially other stressors as well. However, proxy records are not exactly perfectly understood yet.

References

Carilli, J., & Walsh, S. (2012) Benthic foraminiferal assemblages from Kiritimati (Christmas) Island indicate human-mediated nutrification has occurred over the scale of decades. Marine Ecology Progress Series, 456, 87-99.

Graham, N. A., Cinner, J. E., Norström, A. V., & Nyström, M. (2014) Coral reefs as novel ecosystems: embracing new futures. Current Opinion in Environmental Sustainability, 7, 9-14.

Wear, S. L., & Thurber, R. V. (2015) Sewage pollution: mitigation is key for coral reef stewardship. Annals of the New York Academy of Sciences.