Lee‐Yaw et al 2022 Species distribution models rarely predict the biology of real populations-min

Article Information

Title: Species distribution models rarely predict the biology of real populationsAuthors: Julie A. Lee-Yaw, Jenny L. McCune, Samuel Pironon, Seema N. ShethPublished by: John Wiley & Sons Ltd on behalf of Nordic Society OikosDate Accepted: 25 November 2021DOI: 10.1111/ecog.05877Keywords: abundance, ecological niche, genetic diversity, habitat suitability, independent data, occurrence, performance, population growth

Introduction to Species Distribution Models (SDMs)

  • Purpose of SDMs: The primary aim of Species Distribution Models (SDMs) is to capture the ecological niches of species and infer habitat suitability, which subsequently impacts population demography. These models play a crucial role in conservation biology by predicting how species distributions may shift in response to environmental changes and human activities.

  • Usage: SDMs are widely utilized in ecological research to estimate vital parameters such as occurrence rates, population abundance, and genetic diversity across different species and habitats.

  • Literature Review Objective: The review aimed to assess the predictive power of SDMs by examining their performance in relation to independent data across various aspects of population biology, focusing on their applicability and limitations.

Findings of the Human Study

General Findings

  • Conducted a systematic review of 201 studies that specifically evaluated the predictions made by SDMs.

  • Support for SDM Predictive Power: The review revealed that SDMs have a success rate of approximately 53% for predicting occurrence data successfully.

  • Performance Decline: SDM performance declines in a clear hierarchy: occurrence (~53%) ➔ abundance ➔ population fitness ➔ genetic diversity. This decline highlights the challenges in accurately predicting demographic parameters as one moves beyond basic occurrence probabilities.

  • Noteworthy differences were observed between studies evaluating single species versus multiple species, suggesting potential publication bias affecting the literature.

  • Hypothesis Framework: It's emphasized that SDM predictions should be treated as hypotheses rather than definitive population parameters, indicating a need for further empirical validation.

Theoretical Framework

  • Key Concepts:

    • Population Growth: Stable population growth is often correlated with suitable environmental conditions, highlighting the intimate relationship between habitat quality and species viability.

    • Habitat Suitability: This relates directly to population size, genetic diversity, and persistence, suggesting that suitable habitats are essential for long-term survival.

Model Construction and Limitations

Model Construction

  • SDM outputs can yield either the probability of occurrence or indicate relative habitat suitability based on various environmental features. Increasing access to diverse environmental and locality data has spurred interest in developing more robust SDMs.

Limitations of SDMs

  • Scale Issues: Utilization of coarse datasets may lead to a misrepresentation of finer-scale ecological interactions, which are often critical for understanding species dynamics.

  • Data Input Issues: Dispersal dynamics and other ecological factors can distort occurrence datasets, complicating niche modeling efforts.

  • SDMs have shown limitations in adequately predicting key population parameters beyond occurrence, indicating that other significant factors influence population biology and characteristics.

Literature Survey Methodology

  • The study reviewed literature from 1980 onwards, focusing exclusively on native species within their native ranges. A total of 1827 articles were initially considered, which were rigorously narrowed to 201 based on strict criteria for evaluating independent data regarding SDM performance.

Results of SDM Performance Assessments

Occurrence Predictions

  • A total of 101 papers assessed SDM predictions against independent occurrence data.

  • Approximately 52% of studies reported an Area Under the Curve (AUC) greater than 0.70, suggesting a reasonable ability to distinguish between occupied and unoccupied sites.

  • Additionally, about 64% of these studies reported significant results supporting the correlation between SDM predictions and empirical findings.

Abundance Predictions

  • In terms of abundance predictions, only 50% of studies found a positive correlation between SDM predictions and independent abundance data.

  • Notably, studies focusing on fewer species displayed stronger correlations with SDM predictions, hinting at the complexities introduced by diversity within datasets.

Population Mean Fitness and Performance

  • The analysis of 42 studies examining the relationship between SDM predictions and independent population mean fitness metrics found very limited evidence supporting a correlation, indicating that SDMs may not effectively predict demographic performance outcomes.

Genetic Diversity Predictions

  • Among the 18 studies investigating the relationship between SDMs and genetic diversity, only 2 out of 12 studies supported a positive relationship, which raises important questions about the integration of habitat suitability with genetic variation in populations. The apparent discrepancy complicates efforts to forecast genetic diversity based on habitat models.

Conclusions and Recommendations

  • Validity and Performance: The findings indicate that SDMs are more reliable for predicting occurrence rather than for broader population-level parameters, thus necessitating caution when utilizing these models in biological assessments.

  • Recommendations for Use: It is recommended that SDMs be validated with independent datasets; they should inform management decisions but should not replace more rigorous biological evaluations in conservation contexts.

  • Caution in Application: Emphasizing the necessity for careful parameterization and consideration of ecological dynamics in predicting demographic parameters is critical for improved model applicability.

Implications for Conservation

  • The study suggests that conservation applications must integrate more comprehensive demographic and genetic studies beyond mere habitat correlations to more accurately inform persistence potential.

  • While SDMs may assist in identifying potential habitats for populations requiring conservation attention, it is paramount that these findings are validated to ascertain their relevance and applicability.

Acknowledgements and Funding

  • The authors extend gratitude for the support received from various grants, as well as for the constructive peer review feedback that contributed to the manuscript's development.

Here are some interesting topics that could be explored in a larger discussion group based on the themes of the article, but which were not explicitly covered:

  1. Integration of SDMs with Other Ecological Models: Discuss how SDMs might be combined with other ecological modeling approaches (e.g., individual-based models, agent-based models) to provide a more comprehensive understanding of species dynamics.

  2. Impact of Climate Change on Species Distribution: Explore how changing climatic variables could further complicate SDM projections and what additional factors should be considered in models to account for rapid environmental changes.

  3. Technological Advancements in Data Collection: Consider the role of new technologies, such as remote sensing and machine learning, in improving data quality and quantity for SDMs. How could these advancements influence model accuracy and predictions?

  4. Ethics of Conservation Decisions: Debate the ethical implications of using SDMs for conservation planning. What are the responsibilities of researchers and policymakers when making decisions based on model predictions?

  5. Public Engagement and Education: Discuss strategies for communicating the limitations and utility of SDMs to the public and stakeholders involved in conservation efforts. How can we educate non-scientists about the importance of empirical validation?

  6. Role of Genetic Studies in Conservation: Delve into how genetic diversity assessments could be better integrated into conservation strategies alongside habitat models. What innovations could support this integration?

  7. Case Studies of SDM Applications: Invite discussion on successful or unsuccessful case studies where SDMs have been applied in conservation efforts, providing lessons learned from those experiences.

  8. Future Directions for Research: Identify potential areas for future research that can enhance the predictive capability of SDMs, including the need for interdisciplinary approaches that incorporate social, economic, and ecological factors.