TermiteZone Termite Swarm Dataset: 11,411 US Termite Belt Places with Per-Species Swarm Phenology

Open data behind TermiteZone, a daily termite swarm-risk forecast for the United States Termite Belt. Method and sources: termitezone.app/methodology. Source repository: github.com/Nathanzo/termitezone-data. Termite swarms are seasonally synchronized: a colony releases alates in a narrow annual window, and within that window the flight is triggered by weather. This dataset publishes the two components that make that predictable at town level: where each termite species is present, and when its swarm season falls at that latitude and longitude. Coverage is the Termite Belt as defined by the IRC Termite Infestation Probability (TIP) map, Regions I ("Very Heavy") and II ("Moderate to Heavy") - the same boundary that drives VA and FHA wood-destroying-insect inspection requirements. 11,411 places across 26 states. Files. places.csv (11,411 rows): location, population, land area, TIP zone, per-species presence weight, per-species swarm window and peak date, peak seasonal swarm score. swarm-seasons.csv (4 rows): per-species swarm phenology reference - season center, width, latitude slip, time of day, weather trigger, US range. states.csv (26 rows): state coverage. Species. Eastern subterranean (Reticulitermes flavipes/virginicus), Formosan subterranean (Coptotermes formosanus), drywood (Cryptotermes brevis and Incisitermes snyderi in the east; Incisitermes minor in the west), and desert subterranean (Heterotermes aureus). Note that the eastern and western drywood groups swarm about four months apart (peaks around May 15 and September 15); treating drywood as a single season misplaces roughly half the range. Method and validation. Swarm phenology is modeled as a day-of-year Gaussian per species, latitude-adjusted for the spring-flight species and longitude-split for drywood. Ranges and swarm timing are drawn from extension entomology (UF/IFAS, LSU AgCenter, UC IPM, University of Arizona Extension, UGA/CAES, Texas A and M AgriLife, University of Missouri Extension). Timing was checked against 1,860 research-grade geolocated iNaturalist observations and against documented dated swarm events with reanalysis weather; observed peaks matched modeled peaks within days. The seasonal score is a synthesis of published biology, not an outcome-validated infestation count.