Higher-order interactions in ecology can be hidden in plain sight
Higher-order interactions in ecology can be hidden in plain sight
Authors: Violeta Calleja-Solanas, Santiago Lamata-Otín, Carlos Gómez-Ambrosi, Jesús Gómez-Gardeñes, Sandro Meloni Date: 2026-05-07 Paper ID: openalex:2605.06301
Summary
This paper investigates the limits of inferring higher-order ecological interactions from abundance time series, demonstrating that higher-order Lotka-Volterra dynamics can often be effectively represented by simpler pairwise models. The authors reveal a fundamental mechanism identifiability issue where different interaction structures produce near-indistinguishable dynamics. Consequently, they conclude that time-series data alone is insufficient for identifying complex ecological structures and advocate for the inclusion of additional ecological constraints.
Key Contributions
- Demonstrates that higher-order Lotka-Volterra dynamics can often be reproduced by effective pairwise models, rendering interaction inference from time-series alone ambiguous.
- Identifies a fundamental mechanism identifiability problem where distinct ecological interaction structures generate nearly identical time-series outputs.
- Argues that reliably inferring ecological interaction structures necessitates integrating external ecological information beyond abundance time series.
Open Questions & Future Work
Archivist Review
The paper identifies a fundamental identifiability constraint in time-series modeling where complex (higher-order) interactions are effectively masked by simpler (pairwise) models. I have approved the open question regarding the theoretical and practical conditions for overcoming this non-identifiability, as it represents a significant challenge for mechanistic inference in complex dynamical systems. No new concepts were approved as the primary contribution is a negative result regarding model identifiability rather than a novel, reusable algorithmic framework.
Approved Open Questions
- Identifiability of Ecological HOIs: This issue is critical because it addresses the core problem of model identification in complex systems; relying solely on predictive power from time-series fits can lead to fundamentally incorrect conclusions about species interactions, which has profound implications for conservation, management, and understanding ecological stability.
Links
Metadata & Links
- url
- https://arxiv.org/abs/2605.06301
- paper_id
- 2605.06301
- paper_source
- openalex
- domain
- time-series
- tags
-
- architectures
-
- datasets
-
- concept_slugs
-
- dataset_slugs
-
- skill
- TimeSeriesSkill
- created_at
- 2026-05-10T06:04:41Z