Causal inference in coupled human and natural systems.

Area of Science:

• Environmental social science
• Sustainability science
• Complex systems analysis

Background:

• Coupled human and natural systems (CHANS) exhibit complex feedback loops across social and environmental dimensions.
• Causal inference in CHANS is hindered by challenges like violated excludability and interference assumptions, which are under-explored in existing literature.
• Most causal variables in CHANS impact both environmental and human components, complicating analysis.

Purpose of the Study:

• To investigate the challenges of excludability and interference in causal inference within CHANS.
• To evaluate the plausibility of these assumptions using structural knowledge and explicit system recognition.
• To explore the relevance of interference in CHANS through a stylized marine simulation.

Main Methods:

• Review of nearly 200 studies in marine protected areas literature to assess the handling of causal assumptions.
• Development of a stylized simulation of a marine CHANS incorporating policy interventions, ecological disturbances, and technological disasters.
• Analysis of human and capital mobility as sources of interference and moderators of causal effects.

Main Results:

• Few studies in the marine protected areas literature adequately address the excludability assumption.
• Human and capital mobility in CHANS act as sources of interference, biasing causal effect inferences.
• Interference also moderates the magnitude of causal effects within CHANS.

Conclusions:

• No perfect solutions exist for satisfying excludability and interference assumptions in CHANS research.
• Elucidating causal relationships in CHANS necessitates employing multiple analytical approaches to identify and triangulate credible inferences.
• Advancing causal understanding in CHANS and sustainability science requires interdisciplinary expertise and academic-practitioner partnerships.