Strategic directions for agent-based modeling: avoiding the YAAWN syndrome

Area of Science:

• Land Change Science
• Computational Social Science
• Environmental Modeling

Background:

• Agent-based modeling (ABM) is increasingly prevalent in land change science for developing localized case studies.
• There is a risk of fragmented learning and under-shared insights across the research community regarding ABM applications.
• The proliferation of 'yet another model' without broader theoretical grounding or rigorous evaluation hinders scientific advancement.

Purpose of the Study:

• To address the challenge of effectively leveraging agent-based modeling (ABM) in land change science.
• To foster a more cohesive and insightful use of ABMs by exploring key methodological questions.
• To guide the land change science community towards "doing better science with models" rather than simply generating more models.

Main Methods:

• Conceptual exploration of methodological trade-offs in ABM.
• Discussion of model evaluation strategies considering parameter and structural uncertainties.
• Examination of hybrid model structures for enhanced system dynamics understanding.

Main Results:

• Identifies critical questions regarding the balance between model realism and theoretical grounding.
• Highlights the need for robust model evaluation approaches adaptable to uncertainty.
• Suggests exploring hybrid models to capture complex system dynamics effectively.

Conclusions:

• Moving beyond "yet another model" requires careful consideration of model complexity, evaluation, and structure.
• Effective use of ABMs can yield significant new insights for stakeholders when these methodological challenges are addressed.
• Promoting shared learning and best practices is crucial for advancing land change science through modeling.