The heat a thermometer misses.

Three measurements, three responsibilities.

Weather monitoring, exposure analysis and health surveillance answer related but distinct questions. DWD tells us about the climate of an area. The Lancet estimates where vulnerable populations overlap with persistent heat. A health-outcome model would ask a further question about illness or mortality, with its own assumptions and uncertainty.

For a municipality, the useful implication is therefore specific: a national hot-day total is not sufficient evidence for the local burden. Planning needs information about population, housing, night-time cooling, access to protection and health-service capacity. This article does not estimate how much any particular intervention reduces that burden.

The latest year is not always the common year.

The matched exposure record ends in 2024. DWD's hot-day file extends to 2025, when Germany recorded an area average of 11.06 days. Its summer-temperature file already includes the completed summer of 2026, at 19.55 degrees Celsius. These are newer weather observations, not substitutes for missing 2025 or 2026 exposure estimates.

The distinction matters because refreshing one source does not update the whole analysis. We retain the newer observations as separate context and keep every linked comparison within the years the sources actually share.

A discontinuity deserves a visible boundary.

The Lancet workbook warns of possible inconsistencies before and after 2000 because two population sources were used to obtain spatial and temporal information. The earlier history remains visible, but the principal comparison deliberately stays after that change: 2001-2010 against 2015-2024.

Those periods are not an age-standardization exercise. Both heat conditions and the size and distribution of the population can affect total person-days. The country export contains no fixed-population counterfactual, so it cannot identify the separate demographic contribution. The index is a way to compare change within a group, not to rank which age group is personally at greater risk.

Where the evidence stops

What these data do not show

• Exposure estimates combine climate reanalysis and population grids. They are not direct observations of each person's indoor or outdoor conditions.
• Person-days count accumulated exposure, not unique people, diagnosed illness, hospital admissions or deaths.
• German age-group denominators, per-person exposure, fixed-population comparisons and country exposure intervals are absent from the downloaded country sheet.
• The inspected Lancet mortality workbook has global and regional estimates, but no German country sheet or German uncertainty intervals. No regional figure is relabelled as Germany.
• The descriptive comparisons do not identify the effect of a heat-action plan, establish individual clinical risk, or attribute the two-year reversal to one mechanism.

The audit trail

Sources, methods and rights

The reproducible pipeline freezes the official source bytes and SHA-256 hashes, checks complete unique calendar years, reads the actual workbook headers and joins only national Germany records. The DWD reference is 1991-2020; the Lancet heatwave definition retains 1986-2005; the derived within-group exposure index uses the stated 2001-2010 benchmark. These reference periods serve different purposes and are never silently equated. No observations are interpolated, no density or regression is fitted, and no missing denominator is replaced with zero.

Exposure data and their adapted charts are credited to the Lancet Countdown, Federico Tartarini and Ollie Jay under CC BY-NC-SA 4.0. Adaptations comprise the Germany subset, arithmetic comparisons, indices and visual encoding. Commercial reuse requires separate permission; no endorsement is implied. DWD weather data retain DWD attribution.

• DWD Climate Data Center: regional annual hot days and summer mean temperature. Read with the published geography and reference periods.
• Lancet Countdown 2025, indicator 1.1.1: German vulnerable-population heatwave person-days. 2025 report.
• IPCC AR6 Synthesis Report, SPM A.2 and C.2: the impacts and adaptation framing, not a source for these newer German annual values.

Photograph: ArildV (Arild Vagen), Berlin-Mitte, June 24, 2023. CC BY-SA 4.0. Centered crop and WebP conversion. The image establishes place; it does not identify a heatwave or show measured temperature.