By Julia Cheresh, Aquatic and Fishery Sciences undergraduate

Dr. Éva Plagányi gave an insightful talk during the 2018 Bevan Series about her work as a fisheries modeler. At the end of her talk, she had a to-do list for students. One of the bullet points on this list was, “support data gathering and use”. She expanded on this, saying, “I don’t think we have nearly enough data. We can’t just rely on data from the past because some of those relationships are going to change.”

As a young, aspiring scientist, I thought about this quite a bit, and pondered several questions that this line of thinking led me to. I wondered where there are data deficiencies, and what kinds of data we need to fill in those gaps. Surrounded by some of the world’s leading fisheries modelers and data scientists, I looked no further than the SAFS faculty for some thoughts on this issue.

Surveying local experts

Dr. Andre Punt, director of SAFS and a leader in fisheries modeling, had an interesting perspective on data deficiencies. I learned that we are not necessarily data deficient in general, but that there are areas of the world that have little to no data whatsoever, including areas like northern Australia, where Dr. Plagányi works. Some of the most densely populated areas of the world are in fact very understudied.

For people like myself and Ray Hilborn, who look at world fisheries, that’s one of our biggest concerns—that there are places like Southeast Asia, where there are millions of fishing boats, and almost no data.

-Dr. Andre Punt

Photo: Worldwide Life

For example, Indonesia has a population of 261 million people, many of whom rely on fisheries for subsistence. This area is relatively poorly studied according to Dr. Punt.

There are a lot of areas where there is just no data. And where data have been collected, they haven’t necessarily been collected on the spatial and temporal scale you would want.

-Dr. Andre Punt

Robust time series data are needed

This partly answered my next question, which was, what kind of data do we need? According to Dr. Punt, modelers and managers generally rely on time-series data that are spatially and/or temporally robust. Because of the long-term nature of time series data, most modelers rely on data from the past.

Generally speaking, we can’t collect any new data. Unless you’ve started the survey 20 years ago, you’re kind of doomed.

-Dr. Punt

So to take stock so far, we have areas of the world that are data deficient, and are therefore poorly studied. Dr. Plagányi has called on young scientists like myself to collect more data, especially for areas of the world where these data deficiencies are pronounced. Dr. Punt has elucidated that the data most needed is time-series data that is spatially and temporally robust. At the surface, this seems simple enough. My generation of researchers should aim to set up long term studies in these parts of the world where fisheries data is lacking. However, digging into this idea reveals some complications to a seemingly simple solution.

What about rapidly changing ecosystems?

As we learned from Dr. William Cheung, another 2018 Bevan Series speaker, some of the areas expected to be hardest-hit by climate change are the most vulnerable. These areas include regions of the tropics, where wild fish is an important source of protein and nutrients. These are also some of the same areas that Dr. Punt and Dr. Plagányi have identified as data-deficient.

Dr. Ray Hilborn sampling a salmon. Credit: Sandra Hines

So, how do we go about setting up long-term monitoring studies, when these areas are changing so rapidly as a result of climate change? Students in SAFS today benefit from the long-term Alaska salmon studies maintained for decades by faculty members such as Dr. Ray Hilborn and Dr. Thomas Quinn. However, in the face of climate change, we don’t necessarily know if salmon will still be around in 50 years. In the face of such rapid change, establishing a long-term study to address species or systems that will look entirely different within a decade or two seems like a daunting task.

Dr. Punt had some thoughts on this conundrum.

There’s some real challenges to setting up long-term monitoring projects. You always know that what you collect in year 1 is not what you’re going to need in year 25. You have to try to second guess what the questions are going to be not just this year, but in 20 or 30 years’ time. In our school, a lot of our projects use data set up by people who had no idea we’d be answering these sorts of questions.

-Dr. Punt

Long term studies are always useful

Another perspective was offered by Dr. Tim Essington, a professor in SAFS and a fisheries modeler as well. He reminded me that long term ecological studies are rarely (if ever) useless. Although the data collected may not pertain to the original question that the researchers were interested in, something fascinating may appear from the data years later.

 Long term ecological projects are never useless. Trying to set one up that is perfect is a fool’s errand. But they never end up being useless.

-Dr. Tim Essington

My search for information on fisheries data deficits certainly left me with more questions than answers. However, I did find some takeaways, and important considerations if we are to fulfill Dr. Plagányi’s to-do item that calls for more data collection.

To do list to address data deficiencies:

1. Identify areas of the world that are particularly data deficient.
2. Scope the problem while keeping questions that scientists may have in 20 years in the future in mind.
3. Prepare the study to produce a dataset as temporally and spatially robust as possible, while maintaining a balance with time, money and effort.
4. Prepare for the dataset to fail to answer all of the questions you have now. But also monitor species and processes peripheral to the subject you’re interested in, so that your data can be useful in other capacities in the future.

References:

Gattuso, J. P., Magnan, A., Billé, R., Cheung, W. W., Howes, E. L., Joos, F., … & Hoegh-Guldberg, O. (2015). Contrasting futures for ocean and society from different anthropogenic CO2 emissions scenarios. Science, 349(6243), aac4722.