Replication across space and time must be weak in the social and environmental sciences

Area of Science:

• Geospatial science
• Artificial Intelligence
• Scientific methodology

Background:

• Replicability is crucial in scientific research, especially for phenomena tied to specific locations.
• Spatial dependence and heterogeneity are key characteristics of Earth surface phenomena.
• Existing methods for spatial heterogeneity include place-based models.

Purpose of the Study:

• To review artificial intelligence (AI) applications for spatiotemporal phenomena.
• To explore how AI can address spatial heterogeneity in research.
• To introduce and discuss the measurement of "weak replicability".

Main Methods:

• Literature review of AI applications in spatiotemporal research.
• Conceptual analysis of spatial heterogeneity principles.
• Introduction of the concept of weak replicability.

Main Results:

• Artificial intelligence offers emerging solutions for analyzing spatiotemporal data.
• Spatial heterogeneity presents unique challenges for traditional replicability.
• A new concept, weak replicability, is proposed for phenomena with spatial variations.

Conclusions:

• AI holds promise for advancing research on spatially distributed phenomena.
• Addressing spatial heterogeneity is key to improving replicability in geospatial studies.
• Weak replicability offers a framework for understanding and measuring replicability in heterogeneous environments.