Prescott (Yavapai County) vs Seneca (Oconee County)

Mean air temperature of the coldest and the warmest QUARTER, from CHELSA V2.1 bioclimatic variables bio11 and bio10 over the 1981-2010 climatology, at 30 arcsec (~1 km), averaged over the town's own stored footprint polygon. The low figure is the coldest-quarter mean and the high figure is the warmest-quarter mean: TWO named measures, not a range, so neither may ever be ranked. Two caveats travel with every figure: these are MODELED climatologies rather than a station record at the town, and the figure describes the FOOTPRINT, so a large municipio spanning a coast and a sierra reports a wider pair than the town center experiences.

Total precipitation in the driest and the wettest QUARTER, from CHELSA V2.1 bioclimatic variables bio17 and bio16 over the 1981-2010 climatology, at 30 arcsec (~1 km), averaged over the town's own stored footprint polygon. The low figure is the driest-quarter total and the high figure is the wettest-quarter total: TWO named measures, not a range, so neither may be ranked. Two caveats travel with every figure: these are MODELED climatologies rather than a gauge record at the town, and the figure describes the FOOTPRINT, so a large municipio spanning a coast and a sierra reports a wider pair than the town center receives.

Mean near-surface relative humidity in the least humid and the most humid MONTH, from the twelve CHELSA V2.1 hurs climatologies over 1981-2010, at 30 arcsec (~1 km), averaged over the town's own stored footprint polygon. The low figure is the least humid month and the high figure is the most humid month: TWO named measures, not a range. Note the window differs from the rainfall and temperature axes, which use a QUARTER - a month is more extreme at both ends, so these figures are not interchangeable with a quarterly one. Modeled climatologies, not a station record, and measured over the footprint rather than the town center.

Drive time by road to the nearest pharmacy mapped on OpenStreetMap, searched within 30 km — a weekly errand, so a shorter radius than the hospital measure, on purpose. The route starts at the town's own map point (for a US county, the US Census Bureau's internal point of the named town) and follows OpenStreetMap roads (OSRM, driving profile), measured 2026-09-23. The 12 nearest pharmacies in a straight line are routed and the shortest drive wins, ties broken by road distance, so a re-run gives the same answer. A pharmacy more than 2 km from any road is set aside rather than routed to the wrong place. A town with no pharmacy within 30 km carries no figure; the radius is never widened for one town. Source: OpenStreetMap contributors (ODbL), via Geofabrik regional extracts.

Drive time by road to the nearest hospital Medicare (CMS) classes as Acute Care or Critical Access (Hospital General Information, 2026-07-22), located from the register's address with the US Census Bureau geocoder or matched by name to the hospital on OpenStreetMap. The route starts at the town's own map point (for a US county, the US Census Bureau's internal point of the named town) and follows OpenStreetMap roads (OSRM, driving profile), measured 2026-09-25. The 12 closest listed destinations within 60 km in a straight line are routed and the shortest drive wins, ties broken by road distance, so a re-run gives the same answer. A destination more than 2 km from any road, a route implying a speed under 5 or over 130 km/h, or a route shorter than the straight line is set aside as a mapping error. The 60 km radius is never widened for one town. Each country's own register decides which hospitals count, so this figure is most reliable when it compares towns in the same country.

Average residential electricity price from the US EIA, by state, in USD per kWh (all residential customers; no consumption tier and no climate assignment — a different construction from the EU band figure). Rendered as a monthly bill at a REFERENCE consumption of 250 kWh/month, a national figure for the state; US states are comparable among themselves.

Ground elevation from the Copernicus DEM GLO-30 (~30 m, EPSG:4326, ESA/Copernicus, 2021 release), computed over the town's own stored footprint polygon. Low and high are the minimum and maximum cell values inside that polygon. Two caveats travel with every figure: the figure describes the FOOTPRINT, not the town center, so a municipio-level footprint reads higher than a published town elevation wherever the surrounding land rises; and GLO-30 is a surface model, so dense canopy or buildings can lift the maximum a few meters above bare ground.

Ground slope from the Copernicus DEM GLO-30 (~30 m, EPSG:4326, ESA/Copernicus, 2021 release), computed over the town's own stored footprint polygon. Low and high are the 10th and 90th percentiles of per-cell slope inside that polygon — not the minimum and maximum, because over any real footprint the minimum is 0° and the maximum is a single cliff cell. THE BASELINE IS PART OF THE MEASUREMENT: the elevation grid is averaged to a ~90 m post before differencing, and a figure taken over a different baseline is a different measurement. Three caveats travel with every figure: it describes the FOOTPRINT, so a municipio-level footprint takes in surrounding hills the town itself is not built on; GLO-30 is a SURFACE model, so canopy and buildings are in it; and the DEM's own vertical noise puts a floor of roughly 0.5° under the figure on near-level ground, which is artefact rather than terrain. The floor shifts every town's figure in the same direction and does not reorder them — a flat town measures about a quarter of a gently rolling one at every baseline tested.

Resident population inside the town's own stored footprint polygon, from GHS-POP E2020 R2023A (~100 m, EPSG:4326, European Commission JRC). MODELED, NOT COUNTED: GHS-POP distributes national census counts across the built-up surface detected from satellite imagery, so it is an estimate of how many people live inside a shape, never a head count. What it buys in exchange is that it is computed identically in every country, which no census is — two national censuses share a definition only by coincidence. Cells on the boundary are weighted by the fraction of each one inside the polygon, so a small footprint is not rounded up or down by its own edge. THE FIGURE DESCRIBES THE FOOTPRINT, and for a town whose footprint is a settlement rather than a municipality that is a much smaller area than the one an official population refers to — which is the point of publishing it. Checked against the stored census figure for all 63 towns that carry one: the median difference is 5.5% for a municipality footprint and 8.7% for a municipio, and 57.6% for a settlement, where the two figures are correctly describing different places.

Shortest distance from any point of the town's own stored footprint polygon to the ocean shoreline, measured geodesically. The shoreline is OpenStreetMap's natural=coastline, which is mean high water; LAKES AND RIVERS ARE EXCLUDED, so a town on a lakeshore carries its distance to the sea rather than zero. THE FIGURE DESCRIBES THE FOOTPRINT, not the town center: a municipio whose boundary reaches the coast reads a smaller number than its built-up area would, and a municipio whose center is coastal but whose boundary is not reads zero. Zero means the footprint touches the shoreline. The nearest candidate is found in a local equirectangular frame and the winning pair of points is then re-measured with the haversine formula, so the published figure does not carry the projection's error. Nearer the sea is neither better nor worse — it is beach, humidity, salt and storm exposure together — so no figure on this axis is ever ranked.

What kind of place the town is — City, Town or suburb, or Rural — on the Degree of Urbanisation (DEGURBA) standard, from GHS-SMOD E2020 R2023A (~1 km, European Commission JRC). THE PLACE IS CLASSED BY WHERE ITS PEOPLE LIVE, NOT WHERE ITS LAND IS: every ~100 m population cell inside the town's own footprint is charged to the settlement class it falls inside, and the class holding the majority of the footprint's population is the answer. That is why a large municipio containing a city reads City rather than Rural. WHERE NO CLASS HOLDS A MAJORITY THE TOWN CARRIES NO FIGURE AT ALL, because a single word that is untrue of most of the people it describes is worse than a blank; three of the 74 towns are genuinely mixed in this way. THE FIGURE DESCRIBES THE FOOTPRINT, so a municipio whose historic center is a town but whose surrounding population is rural can read Rural — the area being measured is stated beside the figure for exactly this reason. Both the settlement grid and the population weighting are modeled from satellite imagery and census disaggregation, not surveyed. No class is better than another, so nothing on this axis is ever ranked.

Annual mean fine particulate matter (PM2.5) for 2024, in micrograms per cubic meter, from SatPM2.5 V6.GL.03 (~1 km, Atmospheric Composition Analysis Group, Washington University in St. Louis). MODELED, NOT MEASURED AT THE TOWN: satellite aerosol optical depth from MODIS, VIIRS, MISR and SeaWiFS is combined with a GEOS-Chem chemical-transport simulation and calibrated against ground monitors. There is no monitor in most of these towns, which is why a global model is the only way to compare them at all — and why this figure is comparable across a border where two national monitoring networks would not be. THE FIGURE IS POPULATION-WEIGHTED OVER THE TOWN'S OWN FOOTPRINT, not a plain average of the area: people are concentrated where the air is worst, so an areal mean would flatter any municipio that contains a city. Measured on this data, the two differ by about a fifth for a municipio containing a city and not at all for a settlement footprint. FOR SCALE, AND IT IS A PUBLISHED RULE RATHER THAN OUR JUDGEMENT: the World Health Organization's 2021 air quality guideline for annual mean PM2.5 is 5 micrograms per cubic meter, and its interim targets are 35, 25, 15 and 10. Cleaner air is better without qualification, so unlike the other figures measured over the footprint this one is ranked.