Strong and weak environmental perturbations cause contrasting restructure of ant transportation networks

Dynamic transportation networks are embedded in all levels of biological organisation. Ever-growing anthropogenic disturbances and an increasingly variable climate highlight the importance of understanding how these networks restructure under environmental perturbations. Polydomous wood ants provide a convenient model system to study the resilience of self-organising multi-source multi-sink transportation networks. We used ten years of longitudinal empirical data on both unperturbed and experimentally manipulated colony networks to develop and validate a comprehensive dynamic simulation model to study network restructuring after resource removal. We performed simulation experiments to study the effects of excluding food sources with varying importance, either temporarily or permanently, imitating pulse and press perturbations of the networks. We found that removing heavily used resources, corresponding to a strong targeted perturbation, persistently decreased network efficiency, unlike random or weak perturbations. We also found that strong perturbations had excessively adverse effects on robustness and function, reducing the networks’ ability to withstand potential future perturbations. When transportation networks develop around the efficient use of a few key resources, they may be unable to quickly recover from the loss of these through self-organized restructuring. Our findings highlight the importance of considering the interaction of perturbation strength and network structure in studying transportation network dynamics.

Methods
Methods: How was this dataset collected?
This dataset was collected from a long-term study on nine polydomous wood ant (Formica lugubris) colony networks at Longshaw Estate, Peak District, UK (N53°18’33”, E1°36’96”), spanning the period 2012–2021. Colony mapping followed the protocol of Ellis & Robinson (2015) and was conducted annually in late summer (July–August), coinciding with peak ant activity.

Data collection involved:

Nest and Foraging Mapping: Actively used nests and the trees they foraged on were recorded.

Trail Measurement:  Internest and foraging trail lengths were measured, and trail strength was assessed by recording the distance required to count ten ants at the midpoint of each trail. This recorded distance was then converted into ants per millimeter (ants/mm) for further analysis.

Nest Population Estimation: The mound-volume technique (Chen & Robinson, 2013) was used to estimate nest sizes.

Previous Empirical Perturbation Experiment (2017–2021): Five colonies underwent a tree exclusion experiment in a previously published study, where a major foraging tree was temporarily blocked for one year to analyze network adaptation. Five other colonies served as controls. See Burns et al. (2021) for details.

How has it been processed?
The dataset was pre-processed to extract relevant model parameters and validate simulations:

2012–2016 data were used for parameter estimation, defining network structures such as nest density, trail lengths, foraging behaviors, and colony size.

2017–2021 data were categorized for model validation, comparing control and exclusion-treated colonies to assess network adaptation.

The original data are stored in CSV format. The csv files (empirical_edges.csv and empirical_nodes.csv) are available for direct download and are also contained in Ant_networks_Code_and_Data.zip (see the 'Input' folder inside the ZIP file). 

For a complete description of the methodology and dataset structure, refer to the Supplementary Methods (S1.2) and the included README file.

References
Burns DDR, Franks DW, Parr C, Robinson EJH. 2021. Ant colony nest networks adapt to resource disruption. Journal of Animal Ecology 90, 143–152. doi:10.1111/1365-2656.13198

Ellis S, Robinson EJH. 2015. The role of non-foraging nests in polydomous wood ant colonies. PLOS ONE 10, e0138321. doi:10.1371/journal.pone.0138321

Chen Y-H, Robinson EJH. 2013. A comparison of mark–release–recapture methods for estimating colony size in the wood ant Formica lugubris. Insectes Sociaux 60, 351–359. doi:10.1007/s00040-013-0300-z