Using historical data to predict possible changing fish growth rates

Aquatic ecosystems are facing many different stressors that all interact with each other, according to Karen Alofs of the School for Environment and Sustainability at the University of Michigan. These include climate change, habitat degradation, pollution, exploitation and invasive species. All together, these things can create a diverse set of biological responses in ecosystems. These stressors can act from a molecular scale, all the way up to perpetuating ecosystem-scale effects. 

In a recent presentation titled, “Learning from the past to manage inland fisheries for the future,” Alofs looked at how climate change could affect fish populations. She looked specifically at abundance, timing of mass mortality events, and effects on individual fish growth.

“Since the 1950s, average temperatures across the Great Lakes have increased by around 2.3 degrees Fahrenheit, and they are predicted, by mid century, to have increased by three to six degrees Fahrenheit,” she said. By the end of the century, that increase could be six to 11 degrees Fahrenheit. Temperature changes, however, are not evenly distributed in time or space. For instance, winter temperatures have risen much faster than temperatures in other seasons. 

Alofs said she was able to use historical data to look at impacts of climate change on Michigan inland lakes. She looked to see if temperature changes would create shifts in abundance of largemouth bass. She also wanted to determine if the seasonality of mass mortality events had shifted over time and whether increasing temperatures decreased fish growth. 

She used more than a century’s worth of lake surveys from the Michigan Institute for Fisheries Research (IFR) to complete her study. IFR, she said, is a cooperative between the Michigan Department of Natural Resources (DNR) and the University of Michigan. Not only did her research include several interdisciplinary researchers, but leveraging the Zooniverse, a community science crowd sourcing platform, she was able to involve over 2,800 volunteers online in helping to digitize the 75,000 data cards from the IFR.

Temperature and fish abundance over time

Predicted climate change for Michigan is much like that of Wisconsin in both a high-emission scenario and a lower-emissions scenario. Between 2010 and 2029, the Upper Peninsula should start to feel like the Lower Peninsula did historically.

Drawing on data from the 2002-2007 Michigan DNR’s status and trends data, Alofs looked at relative abundance of largemouth bass, showing the abundance higher in the Lower Peninsula. With projected water temperature increases, she said, largemouth bass should become more abundant in the U.P., as they are a warm-water adapted species. However, she said, that idea did not take into account dispersal limitations between lakes, new species interactions, changes in prey or predator availability or other management actions that had taken, or will take, along the way. In order to test the ability to forecast these changes, Alofs said, she used hindcasting to test the forecasting ability to predict changes in abundance.

She estimated a relative abundance across lakes with a model that took into account different catching techniques that could produce different results. Environmental data was also taken into account when building this model, Alofs said.

The models were used to estimate historical abundance of largemouth bass from 1936-1964. The map showed higher relative abundance in the L.P. than in the U.P. The same prediction was made for the contemporary period from 2003-2019 and subtracted the two to look at changes in relative abundance. She saw the strongest increases in relative abundance in the northern lower peninsula and declines in abundance in the southern L.P. She said, during this period, warming of lakes in the southern L.P. was not seen as it was in the northern L.P. and the eastern U.P. 

Alofs found that while temperature had some effect on changes in abundance, other lake characteristics also had a hand in controlling abundance. Very small lakes may not have as drastic of a shift in abundance even with a relatively high change in temperature, for instance. 

By incorporating data on gear and effort, Alofs said they were able to examine changes in abundance rather than just presence or absence of a species. Historically, that is all models allowed researchers to determine. Also, by analyzing historical data, she could then interpret changes across long time periods and on regional scales. She also found that fish abundance into the future could be predicted fairly well using a space-for-time substitution.

Seasonality of mass mortality shifts

Coverage of fish kills, or mass mortality fish events, has increased in recent years, Alofs said. Mass mortality events globally, involving tens of thousands to millions of organisms have been reported more frequently and increasing sizes across a range of taxa, not just fish.

Fish mortality events, overall, have been poorly documented and poorly understood. They are ephemeral, only lasting a few days at most. They are also unpredictable, and are driven by a variety of interacting stressors. There is a reliance on public reports and those observations are often by chance. This also means there is likely a bias toward finding these events on bigger or more frequently visited water bodies, where these events are more likely to be reported. Also, Alofs said, understanding the species and causes that were effective takes research and resources to do that type of research are rarely available.

She also looked at various causes of fish kills. Oxygen depletion can be cause by ice and snow cover, but also algal blooms and decaying plant matter. Extreme temperatures, or rapid changes in temperatures, can play a role as well. Spawning depletion, infection and other disturbances such as contaminants or water withdrawals can also play a part in fish kills. Often it could be several variables acting together that cause these events. Winter kills have been more common in Michigan lakes, but summer kills are predicted to increase from 6 to 34-fold and winter kills, by the end of the century, could all but vanish.

Alofs said lake conditions are already changing. For instance, ice duration has already declined globally six times faster in the last 25 years than in the previous quarter century. Also, approximately 1,500 Northern Hemisphere lakes have been newly documented as experiencing ice-free years recently. There has also been a six-fold increase in frequency of lake heatwaves from 1995-2004 to 2009-2019, which have been associated with deoxygenation in both shallow and deep lakes.

That said, Alofs said she was interested in learning if the occurrence of these mass mortality events had already begun to shift. The dataset she used provided her with good spacial and temporal coverage of the state, as she showed in a graphic during her presentation.

Altogether, she collected 525 observations of fish kills. In the 1930s-50s, the peak observations occurred in March and April. In the 1960s-80s, that date shifted later, with peak observations happening in June. In more recent times, the 2010s to 2020s, fish kill observations were extending later into the fall. The overall trend, she said, showed a shift for early, median and late observations shifted by approximately five days per decade.

The long-term data set allowed Alofs allowed her and her team to show that previously unrecognized shift in mortality events. However, she said, clarifying the causes will be needed to forecast future change and also to advise any climate-resilient management practices in the future. 

Her specific next steps will be to evaluate the cause of the mortalities more closely. She said she would like to be able to predict, through her continued research, which lakes and species will be most vulnerable to mass mortality events. Another area of interest is downstream impacts for fish populations and communities that were adapted to winter kills, but are now experiencing more summer kills. She also looks to answer the question of what effects these later issues have if reproduction is shifting to become earlier.

Growth rates

The last question Alofs looked to answer was whether increasing water temperatures played a part in decreased fish growth. There is a prediction, she said, that species are expected to shrink as temperatures warm. This prediction includes not only fish, but trees and birds as well. The evidence from fish, she said, is mostly from fisheries catch data, and has been highly variable. She showed results from a study of 335 Australian coral reef fishes that showed those that lived in the warmest waters would be the most likely to have a small shift to smaller sizes. When arranged by body length, larger fish are more likely to increase in size with warmer water temperatures, and smaller fish are more likely to decrease in size. Another study from 2003 showed some salmonids increased in size and some decreased in size, even in the same genus.

She wanted to look into the historical data of Michigan lake fishes. She looked at 88,000 mean length-at-age observations from 1945-2020 for 12 game fish species. This work is still in progress, but she shared information about bluegill as a starting point. Over those 75 years, she said, the older age classes of bluegill are stable to even slightly increasing in size. Younger year classes, however, are showing a decrease that is becoming larger. 

A post doctoral researcher in Alofs’ lab, Peter Flood, has put together the same types of models for all 12 different species. 

“We might expect that warm water fishes would be less likely to decline, perhaps more likely to increase, while cold water species we might expect to have the strongest designs in size,” she said. “But this isn’t really what we found.” Overall, she said, when looking across species, the model predicted small, but negative changes in size-at-age for a variety of cool water fish as well.

She then added in different environmental predictors associated with lake attributes as well as biotic factors and presence/absence of other species and several other predictors. In age classes one through eight, lake area and depth are influential on bluegill growth. Other important factors include degree days in lakes and catch per unit effort. Degree day increases, longer growing seasons or warmer temperatures, are associated with less growth. So, too, is increasing surface temperature, she said. 

Bioenergetic models used to estimate what changes in temperature would do to growth, however, it was expected that warmer temperatures on average should increase consumption and therefore growth of bluegills. Alofs said this indicates that it is perhaps consumption that is limited, meaning that there is less forage available.

In her next steps, Alofs said, she would like to look at other fish species, and non-game species on which heavy emphasis has not been put. She is hoping to find growth-over-time data for 30 species that are native to Michigan and then to be able to discern whether or not those species are likely to grow or shrink over time with climate change. To do this, she will continue to combine historical environmental data with available museum specimens. Those interested in viewing Alofs’ full presentation, or to find her contact information, may do so by searching for her presentation on the Midwest Glacial Lakes Partnership YouTube page.

Beckie Gaskill may be reached via email at [email protected].