Rao, Higgs, and Kingi (2602.21491) show that kiwifruit vine disease in New Zealand spread through human mobility networks — not wind, not water, not insect vectors. The pathogen hitchhiked on the people who tend the orchards.
They integrate high-resolution data on human movements across New Zealand with a metapopulation model to simulate the 2010 Psa-V outbreak that devastated the kiwifruit industry. The model reproduces the observed spatiotemporal spread pattern accurately, confirming that human travel between orchards was the primary transmission mechanism.
The finding that matters: most dispersal is local, which matches the standard plant-disease narrative. But sporadic long-range connections — workers traveling between distant regions — are necessary to reproduce the nationwide scope of the outbreak. Remove the long-range connections and the model produces contained local epidemics. Keep only the long-range connections and you get the wrong spatial pattern. The real outbreak required both scales simultaneously.
The practical implications cut against intuition. Enhanced surveillance accelerates detection, as expected. But outbreak severity is highly sensitive to where and when the pathogen first arrives — a timing and location dependence that can't be predicted from the pathogen's biology alone. You need the mobility network. There's also a seasonal labor signal: high-traffic seasons (harvest, pruning) correlate with epidemic risk because that's when the most workers move between the most orchards.
Plant epidemiology has traditionally modeled spread through abiotic mechanisms — wind dispersal kernels, water flow, insect flight ranges. The human mobility channel adds a fundamentally different topology: long-range, seasonal, and structured by economic activity rather than geography.