Summers for the typical American have gotten much hotter since 1970. There are two reasons: the climate warmed, and the country's population growth landed in hot places. The Lower 48 went from 203 million people to 340 million, and most of the added 137 million live in the Sun Belt. That shift comes from births, immigration and people relocating, not just moving vans. Here, each county's share of the population for every year from 1970 to 2025 is weighted against its climate record, and the change is split exactly into those two parts, for the nation, each region and each state.
Each state is measured against its own residents. A state's "population shift" counts only how its growth was spread across its own counties, such as Texas growing fastest around Houston and in the Rio Grande Valley. How fast the state grew compared with the rest of the country shows up in the Lower-48 figure instead.
Each year gets three population-weighted averages across the counties in the area you picked:
What people lived through = each county's value that year × its share of that year's population.
If the 1970 population map had held = the same year's climate, weighted by each county's 1970 share of the population.
If the climate had stood still = each county's 1996–2025 normal, weighted by that year's population shares.
The headline split is exact. Each county's 1970–2025 linear trend supplies a 1970 value and a 2025 value, so one unusual summer can't swing the answer. The change then divides into
Σ (share₂₀₂₅ − share₁₉₇₀) × average value (population shift) plus
Σ average share × (value₂₀₂₅ − value₁₉₇₀) (climate). The two parts add up to the total.
Tornadoes and hurricanes are counted only for the population shift. Tornado reporting has changed a lot since 1970, so each county gets one fixed 1971–2025 rate, weighted by each year's population. That shows whether the population grew toward the hazard; it says nothing about whether the hazard itself changed.
The "pull on the average" map shows each county's own share of the population-shift effect: how much its gain or loss of population share raised or lowered the area's average. A fast-growing hot county pulls the average up. A slow-growing cold county pulls it up too, even if it gained people, because its share of the country shrank.
This is not a count of people moving. A county's share of the population changes with births, deaths, immigration from abroad and moves within the country. Since 1970 the Sun Belt has gained on all of those, so the split here is where the growth went, whatever its source.