Keller-Lab-Studies
Managing the arrival and spread of aquatic invasive species
Invasive species are the most significant driver of biodiversity and economic losses in the Great Lakes ecosystem. The Keller lab, led by Professor Reuben Keller, PhD, published two papers that provide novel insights for managing the arrival and spread of invasive species in the Great Lakes.
The red swamp crayfish is an invasive species that can harm aquatic ecosystems in the Great Lake region.
Predicting the spread of invasive crayfish
SES graduate program alum Carter Cranberg collaborated with Keller and Loyola biology professor Joseph Milanovich, PhD, to publish a paper based on Cranberg's graduate research. In this study, published in Aquatic Invasions, the researchers used a machine learning approach to predict the spread of two invasive crayfish species under different climate change scenarios.
The study examined the spread of the rusty crayfish and the red swamp crayfish, both of which pose threats to aquatic ecosystems in North America. The researchers used a modeling tool called MaxEnt to predict the current and future distribution of the species under two different climate scenarios.
Their analysis of the current distribution identified areas where there are no known populations of the invasive crayfish, but conditions are suitable for their spread.
Projections under various climate scenarios indicated that climate change will shift the potential ranges of both the rusty and swamp red crayfish northward. The model showed a reduction in total high-quality habitat for rusty crayfish, but indicated substantial expansion of suitable habitat for the red swamp crayfish.
The researchers also noted that native crayfish are more sensitive to rising temperatures than the invaders, and the combined pressures of climate change and introduced species could exacerbate declines in the already dwindling native crayfish populations.
Comparing risk assessment tools
Climate change can increase the spread of invasive species by expanding the suitable habitat for introduced plants and animals. To combat the invaders, natural resource managers use a variety of risk assessment tools to predict which introduced species pose the greatest threat. In recent years, the assortment of available prediction tools has grown, raising questions about which methods are most accurate.
Former Loyola postdoctoral scholar Victoria Prescott, PhD, collaborated with Keller on research comparing methods of predicting which species are likely to survive the climate and conditions in the Great Lakes. The study, published in Fisheries, compared two machine learning methods with the Risk Assessment Mapping Program (RAMP), a computer program that generates predictive maps of suitable climates for various species.
The researchers tested how each tool performed in estimating the distribution of 30 invasive aquatic species in the Great Lakes region. The results showed wide variability in the tools' accuracy. RAMP was faster and simpler to use than the machine learning tools, but it overestimated the potential range of the introduced species. In contrast, generating predictions with machine learning tools required substantial expertise and time, but these models described the distribution of invasive species much more accurately.
The study concluded that while RAMP was less accurate, its relative speed and simplicity offer advantages. They also noted that, when it comes to combating invasive species, it may be safer to overestimate potential threats than to underestimate the risks.
Together, these papers offer insights to help natural resource managers predict and combat the spread of invasive species.
By Stephanie Folk
November 2025
Predicting the spread of invasive crayfish
SES graduate program alum Carter Cranberg collaborated with Keller and Loyola biology professor Joseph Milanovich, PhD, to publish a paper based on Cranberg's graduate research. In this study, published in Aquatic Invasions, the researchers used a machine learning approach to predict the spread of two invasive crayfish species under different climate change scenarios.
The study examined the spread of the rusty crayfish and the red swamp crayfish, both of which pose threats to aquatic ecosystems in North America. The researchers used a modeling tool called MaxEnt to predict the current and future distribution of the species under two different climate scenarios.
Their analysis of the current distribution identified areas where there are no known populations of the invasive crayfish, but conditions are suitable for their spread.
Projections under various climate scenarios indicated that climate change will shift the potential ranges of both the rusty and swamp red crayfish northward. The model showed a reduction in total high-quality habitat for rusty crayfish, but indicated substantial expansion of suitable habitat for the red swamp crayfish.
The researchers also noted that native crayfish are more sensitive to rising temperatures than the invaders, and the combined pressures of climate change and introduced species could exacerbate declines in the already dwindling native crayfish populations.
Comparing risk assessment tools
Climate change can increase the spread of invasive species by expanding the suitable habitat for introduced plants and animals. To combat the invaders, natural resource managers use a variety of risk assessment tools to predict which introduced species pose the greatest threat. In recent years, the assortment of available prediction tools has grown, raising questions about which methods are most accurate.
Former Loyola postdoctoral scholar Victoria Prescott, PhD, collaborated with Keller on research comparing methods of predicting which species are likely to survive the climate and conditions in the Great Lakes. The study, published in Fisheries, compared two machine learning methods with the Risk Assessment Mapping Program (RAMP), a computer program that generates predictive maps of suitable climates for various species.
The researchers tested how each tool performed in estimating the distribution of 30 invasive aquatic species in the Great Lakes region. The results showed wide variability in the tools' accuracy. RAMP was faster and simpler to use than the machine learning tools, but it overestimated the potential range of the introduced species. In contrast, generating predictions with machine learning tools required substantial expertise and time, but these models described the distribution of invasive species much more accurately.
The study concluded that while RAMP was less accurate, its relative speed and simplicity offer advantages. They also noted that, when it comes to combating invasive species, it may be safer to overestimate potential threats than to underestimate the risks.
Together, these papers offer insights to help natural resource managers predict and combat the spread of invasive species.
By Stephanie Folk
November 2025