The RKI’s heat-related mortality estimates are statistical model outputs, not direct counts of people whose death certificates list “heat” as the cause.

The linked RKI page and its weekly reports make this explicit.

RKI researchers use a generalized additive model (GAM) on weekly all-cause mortality (from Destatis) and weekly mean temperatures (from DWD stations, averaged day+night). They identify a temperature-mortality relationship in which mortality rises once the weekly mean exceeds roughly 20 °C.

They then estimate a counterfactual “background” mortality assuming temperatures stayed below that threshold (also incorporating lag effects of up to three prior weeks, long-term trends, seasonality, and post-COVID adjustments in later versions).

The difference between the observed/modeled mortality and this counterfactual is labeled heat-related excess mortality.

Key points from RKI materials:

• Heat is almost never coded as the underlying cause on death certificates (most deaths involve cardiovascular, respiratory, or other pre-existing conditions exacerbated by heat).
• The approach is standard in environmental epidemiology for estimating temperature-attributable mortality.
• Figures are estimates with uncertainty intervals (e.g., ~9,800 [8,700–10,900] through calendar week 29 of 2026). The largest share is among people aged 75+, especially 85+.
• RKI notes limitations of the weekly resolution (daily data can capture sharper extremes better) and that the model can under- or over-estimate in particular situations.

This is not “every death in a week whose mean temperature is >20 °C is a heat death. Only the excess above the modeled non-heat baseline is attributed.

The RKI does not merely declare all excess in warm weeks to be heat deaths. It fits an exposure-response curve from multi-year data and computes a counterfactual. The correlation with summer means is expected if the underlying temperature-mortality relationship holds; that does not by itself prove the attribution is invalid.

Weekly mean temperature is a coarse metric. It can smooth intra-week extremes and does not incorporate humidity, so wet-bulb or heat-index measures would be physiologically more relevant for assessing thermoregulatory stress.

Attribution remains correlational/associational. The model assumes the observed temperature-mortality link is causal after adjusting for the included covariates; residual confounding (air pollution, behavioral factors, healthcare access, etc.) is possible.

Harvesting/mortality displacement is partially addressed via lags, but longer-term displacement or frailty effects remain debated in the literature.

Direct cause-of-death data cannot validate the totals because heat is rarely coded as underlying cause.

Observational data across many countries and decades show clear, consistent rises in mortality (especially cardiovascular) during hot periods, concentrated in the elderly and those with comorbidities.

Germany’s own historical analyses (1992–2021 and later) and independent European studies find similar patterns. The 20 °C weekly-mean threshold is data-driven for Germany’s climate and population, not arbitrary.

The RKI numbers are model-based estimates of excess deaths statistically associated with elevated temperatures, not verified individual heatstroke cases.

Similar debates occur with other temperature-attributable mortality studies worldwide.

_____________________________________________________________________________________

Wet-bulb temperature mortality studies

Wet-bulb temperature (Tw or WBT) is a combined measure of air temperature and humidity. It represents the lowest temperature to which air can be cooled by evaporating water into it.

It is physiologically relevant because human cooling relies heavily on sweat evaporation; when humidity is high, evaporation is limited and heat stress rises even if the dry-bulb (ordinary air) temperature is not extreme.

A widely cited 2010 paper by Sherwood & Huber proposed that a sustained wet-bulb temperature of 35 °C marks an upper limit for human survivability. At that point, even a healthy person resting in the shade with unlimited water and a fan cannot shed metabolic heat effectively, leading to rising core temperature and potentially fatal hyperthermia within hours.

This threshold has been used extensively in climate-impact projections. Real-world observations of Tw ≥ 35 °C remain rare and short-lived, though they have occurred briefly in places such as the Persian Gulf, South Asia, and parts of the Americas.

Laboratory and field studies show that uncompensable heat stress (the point at which core temperature begins an uncontrollable rise) occurs at substantially lower wet-bulb temperatures for most people under realistic conditions:

• Experiments on young, healthy adults performing light activity found critical Tw values typically in the 25–31 °C range, often 5–10 °C below the theoretical 35 °C limit. The exact threshold varies with humidity, air temperature, clothing, activity level, and wind. Hot-dry conditions can produce larger deviations because dry heat gain outpaces evaporative cooling.
• For older adults or those with comorbidities the limits are lower still (reported values around 22 °C in some modeling).
• Heat-index models that incorporate more realistic physiology (metabolic heat, radiation, limited wind, etc.) predict critical Tw values ranging from roughly 20–32 °C and fatal conditions from about 24–37 °C depending on humidity and exertion.

Historical deadly heat waves (Europe 2003, Russia 2010) occurred at peak Tw values ≤ 28 °C. Mortality and morbidity therefore begin well below the classic 35 °C threshold.

Population-level studies increasingly use wet-bulb temperature, wet-bulb globe temperature (WBGT, which also includes radiation and wind), heat index, or similar metrics:

• A Mexico study using high-resolution wet-bulb data found heat-related deaths concentrated among younger people (< 35 years accounted for ~75 % of recent heat deaths and most lost life-years), contrary to the usual emphasis on the elderly in dry-temperature studies. Minimum-mortality wet-bulb temperatures differed by age group.
• Japanese studies comparing WBGT and mean temperature found both metrics produce similar inverse-J mortality curves; mean temperature is often a reasonable proxy when WBGT data are unavailable, though WBGT can better capture humid-heat risk in some settings. researchgate.net
• Multi-city analyses (including a large European and global city networks) show that humidity-inclusive indices (Tw, WBGT, UTCI, heat index) are associated with elevated mortality, but the added predictive power over simple air temperature varies by climate and region. In some temperate or drier locations, the difference is modest; in humid subtropical or tropical settings humidity metrics can improve risk estimates. academic.
• Case-crossover and time-series studies link higher WBGT to increased heat-related hospitalizations and in-hospital mortality, with risk rising sharply above certain thresholds (e.g., WBGT > 27–31 °C).

Most national surveillance systems, including Germany’s RKI weekly reports, rely primarily on dry-bulb air temperature (often weekly means) because:

• Long historical temperature records are widely available.
• Temperature and humidity are correlated, so temperature alone captures much of the mortality signal in many mid-latitude climates.
• Weekly averaging further smooths extremes.

However, critics correctly note that pure temperature metrics omit humidity’s role in limiting evaporative cooling. In humid heat waves, risk can be understated; in very dry heat, the physiological stress profile differs. Studies that directly compare metrics find that humidity-inclusive indices sometimes (but not always) outperform simple temperature, depending on location, season, and population vulnerability. Daily or sub-daily resolution and inclusion of lags also improve estimates relative to weekly means.

Summary

• Wet-bulb temperature (and related indices such as WBGT or heat index) is physiologically more grounded than dry-bulb temperature alone for assessing heat stress.
• The classic 35 °C Tw survivability limit is an upper theoretical bound for ideal conditions; real-world uncompensable heat stress and elevated mortality risk begin at substantially lower values, especially for vulnerable groups.
• Epidemiological evidence confirms associations between high wet-bulb / humid-heat metrics and excess mortality or heat illness, but the incremental benefit over temperature-based models varies by climate.
• For surveillance and attribution in places like Germany (temperate climate with relatively moderate humidity extremes), weekly mean air temperature remains a pragmatic and reasonably effective proxy, yet it is an incomplete representation of the full physiological heat load.

Ongoing research continues to refine critical thresholds with human-subject experiments and high-resolution epidemiological analyses that incorporate humidity, radiation, activity, age, and comorbidities.