Digging deeper into fire ant nest architecture

Fire ants (Solenopsis invicta) are highly invasive social insects that live in colonies of up to hundreds of thousands of ants. Excavated nests, which can extend 2+ m deep, play a critical role in their success. For instance, fire ants rapidly emerge from their nests to respond to threats and escape during floods, as well as relocate larvae within the nest throughout the day. We hypothesize that structural elements and their spatial organization are important for promoting these behaviors by supporting robust and efficient intra-nest transport; therefore, some aspects should be similar across nests for a given substrate. We used CT (computed tomography) scanning to create 3D quantitative representations of five near-complete nests (20 to 110 cm deep) extracted from fields in the Atlanta area. These data reveal highly-redundant and complex subterranean networks that we decompose into stereotypical tunnel-like and chamber-like features using local, shape-based descriptors fed into a machine learning model. We observe that tunnels often have stereotyped circular cross-sections and connect larger, irregularly shaped, horizontal chambers. Chambers can connect to tens of tunnels, which include self-loops and redundant connections to other nest elements. Across nests, we find network density is greater in shallower regions and tunnels are primarily vertical in deeper regions. Ongoing work will investigate how redundancy in ant-excavated networks facilitates intra-nest transport.