County-level disease trends

Lyme disease has an observed and modeled county series back to 1991. Every other year for the other six reportable tick-borne diseases comes from a state-level multi-year series (coverage varies by state, some back to 2010) and is shown on the map as a uniform fill across each state's counties — that's still true for most of the six-state domain. Two states are exceptions. Indiana publishes real county-level, year-by-year data for three of the six (anaplasmosis, babesiosis, tularemia), which paints its own counties with real texture on the map instead of a flat state color, and its clicked-county chart shows that real series instead of a state total. Iowa publishes real county-level data too, for two of the six (anaplasmosis, spotted fever rickettsiosis), but only as ten overlapping 5-year rolling windows rather than single years — not usable for the map's year picker, but its clicked-county chart shows that real windowed series, clearly marked as 5-year spans rather than annual figures. Pick the year above the map once one of the six is selected, or click a county for its full multi-year chart below. See the note below once a disease is selected.

The dashed and dotted series are model output, not observations: a fitted Lyme incidence hindcast and a one-year-ahead projection. The solid line, and the map above, are observed public-health case data.

Click a county on the map for its full series, including the Lyme model fit where available.

Hexagon-resolution explorer

The true underlying resolution behind the county map at left: roughly 165,000 individual 5 km hexagons across the six-state domain, versus 535 counties. Switch between climate, or either species' occurrence model prediction, uncertainty, and extrapolation flag — and click any hexagon (or county, once zoomed out) for its full summary. The individual-hexagon tiles are heavily thinned below roughly state-sized zoom to keep tile size down, so below that zoom the map shows a county-level aggregate instead of real gaps in the hex fill — true hexagon resolution resumes once zoomed in further. Three of the six-state domain's tick species have a validated hex-resolution occurrence model — lone star tick (AUC 0.928), American dog tick (AUC 0.717), and blacklegged tick (AUC 0.737) — all weaker discrimination than lone star tick's. American dog tick's first fit (2026-09-03) failed its own validation gate; a parameter sweep found min_samples_leaf=15 + balanced class weight cleared it with real margin, and improved blacklegged tick's score too, so both now use those settings. Every other species has too few records to model at any resolution. See the note below the map once a layer is selected for what each one actually measures.