Disentangling small‐island multilayer networks: Underlying ecological and evolutionary patterns

Ecology, Journal Year: 2025, Volume and Issue: 106(4)

Published: March 31, 2025

Abstract This study provides a pioneering analysis of the structural and topological characteristics one nature's simplest food webs, using Montaña Clara islet (Canary Islands) as case study. Applying multilayer network approach, which assesses multiple interaction types, we examined plant–animal plant‐fungi interactions during two seasons (humid dry), comparing this oceanic island web to from Na Redona, small continental in Balearic Islands. Data were collected through field observations, flower visitation records, fecal analysis, DNA metabarcoding root‐associated fungi. The identified 63 animal species 367 fungal amplicon sequence variants interacting with 13 plant species, five (38%) structurally significant, indicated by high versatility values (>0.5). structure was modular, 23 modules primarily representing single ecological functions, most involved only type. Notably, 73% shifted roles between layers. Results reveal that Clara's is simpler but more modular versatile than island, aligning biogeography theory. suggests unique biodiversity composition islands, particularly islets, influences their structures.

Language: Английский

Sampling biases across interaction types affect the robustness of ecological multilayer networks

Ecological Informatics, Journal Year: 2025, Volume and Issue: unknown, P. 103183 - 103183

Published: May 1, 2025

Language: Английский

The Attack and Defense Researches on the Dual‐Layer Network of Multivariable Anomaly Causes

International Journal of Intelligent Systems, Journal Year: 2025, Volume and Issue: 2025(1)

Published: Jan. 1, 2025

Multivariate anomaly causes interpretation provides insight into the root cause of information system anomalies, identifying direct factors that trigger anomalies and revealing potential systemic flaws. However, current research generally focuses on two directions: one hand, diagnosis for nodes with high degree; other single‐layer graph construction based explicit features capturing locations their neighborhood structures. These approaches pay insufficient attention to attack defense graph, thereby weakening credibility reliability causation interpretation. Therefore, we systematically explore strategy mechanism multivariate graph. Firstly, propose an adaptive learning method constructing a dual‐layer The reduces dependence artificial priori assumptions by introducing realizes dynamic decoupling spatiotemporal coupling relationships data, thus providing diversified perspective Second, considering vulnerability correlation after structural characteristics further protection complex networks improve robustness resistance interference Finally, verify effectiveness proposed model testing various scenarios such as noise attack, gradient structure attack. experimental results show in this paper can effectively defend against multiple methods ensure integrity

Language: Английский