Exception Becomes Routine
Germany is warmer. But an average establishes direction while concealing the change institutions actually have to manage. A district that moves from six to twelve hot days has not merely warmed. It has crossed from an occasional disruption into a recurring operating condition.
This first act moves from place to recurrence, then tests whether the headline survives alternative definitions. It uses the same 1 January 2025 district geography and one fixed Zensus 2022 population surface throughout. The claim is not that every summer is hotter everywhere. It is that exceptional heat became a normal planning condition across a large part of Germany.
A large estimate is not an exact rank
Rankings answer where to look first, but neighbouring estimates often differ by tenths of a day. The interval view separates evidence of a material local shift from false precision about whether district number four truly belongs above district number seven. The bootstrap uses 2,000 common-year resamples with identical sampled-year indices across districts.
Exact values · Showing 20 of 400 rows
| District | Absolute Change | CI low | CI high | P(top decile) | Rank Ci Low | Rank Ci High |
|---|---|---|---|---|---|---|
| Südliche Weinstraße | +1.55 | +0.63 | +2.46 | 0.999 | 1 | 8 |
| Landau in der Pfalz | +0.97 | +0.05 | +1.89 | 0.940 | 2 | 78 |
| Mainz | +0.85 | -0.44 | +2.27 | 0.844 | 2 | 332 |
| Mainz-Bingen | +0.85 | -0.02 | +1.80 | 0.936 | 3 | 150 |
| Pirmasens | +0.81 | +0.30 | +1.36 | 0.895 | 2 | 73 |
| Merzig-Wadern | +0.80 | -0.04 | +1.68 | 0.934 | 2 | 157 |
| Chemnitz | +0.68 | +0.14 | +1.33 | 0.864 | 2 | 88 |
| Frankfurt am Main | +0.67 | -1.53 | +2.59 | 0.670 | 1 | 393 |
| Brandenburg an der Havel | +0.60 | +0.03 | +1.40 | 0.775 | 3 | 115 |
| Germersheim | +0.60 | -0.32 | +1.52 | 0.728 | 4 | 340 |
| Potsdam | +0.59 | -0.91 | +2.22 | 0.644 | 1 | 382 |
| Bad Kreuznach | +0.56 | +0.11 | +1.05 | 0.846 | 6 | 68 |
| Südwestpfalz | +0.55 | +0.21 | +0.97 | 0.809 | 4 | 71 |
| Wiesbaden | +0.54 | -0.46 | +1.59 | 0.707 | 8 | 334 |
| Main-Taunus-Kreis | +0.53 | -0.92 | +1.89 | 0.660 | 5 | 372 |
| Groß-Gerau | +0.52 | -0.87 | +1.89 | 0.650 | 5 | 371 |
| St. Wendel | +0.48 | -0.36 | +1.28 | 0.703 | 8 | 291 |
| Wetteraukreis | +0.46 | -0.14 | +1.11 | 0.764 | 12 | 217 |
| Saalekreis | +0.45 | +0.05 | +0.98 | 0.685 | 9 | 100 |
| Gera | +0.45 | -0.03 | +1.08 | 0.628 | 5 | 189 |
One extreme summer is not a regime
Germany's resident-weighted mean rose from 7.84 to 13.81 hot days per year between the two decades. The annual path matters because endpoint means can otherwise be mistaken for a smooth trend. The question is not whether 2003 or 2018 was exceptional; it is whether the background frequency shifted after those exceptional summers.
Exact values · Showing 2 of 2 rows
| Metric | Metric Label | 1991–2000 | 2016–2025 | Absolute Change | Unit | Comparison |
|---|---|---|---|---|---|---|
| hot days population | Hot days · daily Tmax ≥30 °C | 7.84 | 13.81 | +5.96 | days per year | 2016–2025 mean minus 1991–2000 mean |
| tropical nights population | Tropical-night proxy · daily Tmin ≥20 °C | 0.39 | 0.42 | +0.03 | days per year | 2016–2025 mean minus 1991–2000 mean |
237 districts crossed into routine heat
A qualifying year has at least ten resident-weighted days with Tmax of 30°C or more. Rare means the threshold appears in zero to two years of a decade, intermittent in three to five, routine in six to eight, and near-annual in nine or ten. Under that declared rule, 237 districts moved from rare or intermittent in 1991–2000 to routine or near-annual in 2016–2025. The alluvial uses actual district cohorts and conserves all 400 districts.
Exact values · Showing 16 of 16 rows
| Class 1991 2000 | Class 2016 2025 | Districts | Share Of 400 |
|---|---|---|---|
| Rare · 0–2 | Rare · 0–2 | 38 | 0.095 |
| Rare · 0–2 | Intermittent · 3–5 | 86 | 0.215 |
| Rare · 0–2 | Routine · 6–8 | 63 | 0.158 |
| Rare · 0–2 | Near-annual · 9–10 | 14 | 0.035 |
| Intermittent · 3–5 | Rare · 0–2 | 0 | 0.000 |
| Intermittent · 3–5 | Intermittent · 3–5 | 4 | 0.010 |
| Intermittent · 3–5 | Routine · 6–8 | 92 | 0.230 |
| Intermittent · 3–5 | Near-annual · 9–10 | 68 | 0.170 |
| Routine · 6–8 | Rare · 0–2 | 0 | 0.000 |
| Routine · 6–8 | Intermittent · 3–5 | 0 | 0.000 |
| Routine · 6–8 | Routine · 6–8 | 1 | 0.003 |
| Routine · 6–8 | Near-annual · 9–10 | 32 | 0.080 |
| Near-annual · 9–10 | Rare · 0–2 | 0 | 0.000 |
| Near-annual · 9–10 | Intermittent · 3–5 | 0 | 0.000 |
| Near-annual · 9–10 | Routine · 6–8 | 0 | 0.000 |
| Near-annual · 9–10 | Near-annual · 9–10 | 2 | 0.005 |
The headline is sensitive, but not fragile
No threshold is ordained by nature. Across annual cutoffs of eight to twelve hot days and recurrence requirements of five to seven years, the crossing count ranges from 154 to 253 districts. Removing one year from each comparison decade produces 196 to 247 crossings, with a median of 219. These are sensitivity ranges, not confidence intervals; 237 remains the predeclared estimate.
Exact values · Showing 24 of 115 rows
| Spec Family | Hot Day Threshold | Routine Years Required | Years In Window | Early Year Omitted | Late Year Omitted | Crossing Districts | Share Of 400 |
|---|---|---|---|---|---|---|---|
| nearby threshold grid | 8.00 | 5 | 10 | — | — | 210 | 0.525 |
| nearby threshold grid | 8.00 | 6 | 10 | — | — | 253 | 0.632 |
| nearby threshold grid | 8.00 | 7 | 10 | — | — | 248 | 0.620 |
| nearby threshold grid | 9.00 | 5 | 10 | — | — | 230 | 0.575 |
| nearby threshold grid | 9.00 | 6 | 10 | — | — | 246 | 0.615 |
| nearby threshold grid | 9.00 | 7 | 10 | — | — | 232 | 0.580 |
| nearby threshold grid | 10.00 | 5 | 10 | — | — | 233 | 0.583 |
| nearby threshold grid | 10.00 | 6 | 10 | — | — | 237 | 0.593 |
| nearby threshold grid | 10.00 | 7 | 10 | — | — | 206 | 0.515 |
| nearby threshold grid | 11.00 | 5 | 10 | — | — | 230 | 0.575 |
| nearby threshold grid | 11.00 | 6 | 10 | — | — | 217 | 0.542 |
| nearby threshold grid | 11.00 | 7 | 10 | — | — | 177 | 0.443 |
| nearby threshold grid | 12.00 | 5 | 10 | — | — | 207 | 0.517 |
| nearby threshold grid | 12.00 | 6 | 10 | — | — | 188 | 0.470 |
| nearby threshold grid | 12.00 | 7 | 10 | — | — | 154 | 0.385 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,016 | 231 | 0.578 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,017 | 242 | 0.605 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,018 | 206 | 0.515 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,019 | 206 | 0.515 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,020 | 231 | 0.578 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,021 | 247 | 0.618 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,022 | 206 | 0.515 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,023 | 223 | 0.557 |
| paired leave one year out | 10.00 | 6 | 9 | 1,991 | 2,024 | 213 | 0.532 |
Fourteen thousand district-summers reveal persistence
The barcode preserves every district-year rather than compressing the record into two endpoints. Coral cells mark years reaching ten resident-weighted hot days. Keyboard, pointer and district lookup expose exact values while the national reach lines show how widely each summer spread. The cooler 2021 season leaves a visible interruption before the 2022–2025 run.
Exact values · Showing 35 of 35 rows
| Year | Event Districts | District Share | Represented Population Weight | Represented Population Share | National Majority | Comparison Period |
|---|---|---|---|---|---|---|
| 1991 | 53 | 0.133 | 12003592.45 | 0.145 | no | 1991-2000 |
| 1992 | 200 | 0.500 | 41323580.98 | 0.500 | no | 1991-2000 |
| 1993 | 10 | 0.025 | 1840519.06 | 0.022 | no | 1991-2000 |
| 1994 | 356 | 0.890 | 76266251.19 | 0.922 | yes | 1991-2000 |
| 1995 | 228 | 0.570 | 55956581.93 | 0.677 | yes | 1991-2000 |
| 1996 | 1 | 0.003 | 306315.02 | 0.004 | no | 1991-2000 |
| 1997 | 38 | 0.095 | 11001069.67 | 0.133 | no | 1991-2000 |
| 1998 | 140 | 0.350 | 26520547.62 | 0.321 | no | 1991-2000 |
| 1999 | 60 | 0.150 | 15045666.50 | 0.182 | no | 1991-2000 |
| 2000 | 30 | 0.075 | 5036491.17 | 0.061 | no | 1991-2000 |
| 2001 | 134 | 0.335 | 30693938.45 | 0.371 | no | middle |
| 2002 | 24 | 0.060 | 5276432.81 | 0.064 | no | middle |
| 2003 | 361 | 0.902 | 75741323.13 | 0.916 | yes | middle |
| 2004 | 64 | 0.160 | 13811028.34 | 0.167 | no | middle |
| 2005 | 122 | 0.305 | 21684101.17 | 0.262 | no | middle |
| 2006 | 339 | 0.848 | 74213013.92 | 0.897 | yes | middle |
| 2007 | 26 | 0.065 | 5637458.82 | 0.068 | no | middle |
| 2008 | 109 | 0.273 | 21836460.49 | 0.264 | no | middle |
| 2009 | 36 | 0.090 | 8034295.27 | 0.097 | no | middle |
| 2010 | 287 | 0.718 | 63443291.78 | 0.767 | yes | middle |
| 2011 | 26 | 0.065 | 6709152.91 | 0.081 | no | middle |
| 2012 | 154 | 0.385 | 33332006.50 | 0.403 | no | middle |
| 2013 | 292 | 0.730 | 56816758.39 | 0.687 | yes | middle |
| 2014 | 46 | 0.115 | 10872718.67 | 0.131 | no | middle |
| 2015 | 347 | 0.868 | 70905340.29 | 0.857 | yes | middle |
| 2016 | 229 | 0.573 | 52151304.00 | 0.631 | yes | 2016-2025 |
| 2017 | 148 | 0.370 | 30329254.67 | 0.367 | no | 2016-2025 |
| 2018 | 364 | 0.910 | 77781921.37 | 0.940 | yes | 2016-2025 |
| 2019 | 367 | 0.917 | 77868731.13 | 0.942 | yes | 2016-2025 |
| 2020 | 273 | 0.682 | 62240744.98 | 0.753 | yes | 2016-2025 |
| 2021 | 18 | 0.045 | 6622271.87 | 0.080 | no | 2016-2025 |
| 2022 | 358 | 0.895 | 76479391.04 | 0.925 | yes | 2016-2025 |
| 2023 | 262 | 0.655 | 56762618.88 | 0.686 | yes | 2016-2025 |
| 2024 | 272 | 0.680 | 55857179.28 | 0.675 | yes | 2016-2025 |
| 2025 | 270 | 0.675 | 52867652.25 | 0.639 | yes | 2016-2025 |
Routine-heat districts account for 70% of fixed exposure
Districts classified as routine or near-annual account for 8.6% of the fixed 2022 represented-population surface in the early period and 70.2% in the recent period. The surface assigns 50.9 million represented-person weight to districts that crossed from occasional to routine or near-annual recurrence. This is standardized exposure accounting, not a reconstruction of migration or individual indoor exposure.
Exact values · Showing 8 of 8 rows
| Period | Recurrence Class | Class Rank | Districts | Represented Population Weight | Represented Population Share |
|---|---|---|---|---|---|
| 1991-2000 | Rare | 0 | 201 | 38811268.02 | 0.469 |
| 1991-2000 | Intermittent | 1 | 164 | 36774517.32 | 0.445 |
| 1991-2000 | Routine | 2 | 33 | 6581362.12 | 0.080 |
| 1991-2000 | Near-annual | 3 | 2 | 539308.54 | 0.007 |
| 2016-2025 | Rare | 0 | 38 | 5602710.18 | 0.068 |
| 2016-2025 | Intermittent | 1 | 90 | 19045033.41 | 0.230 |
| 2016-2025 | Routine | 2 | 156 | 34272826.52 | 0.414 |
| 2016-2025 | Near-annual | 3 | 116 | 23785885.89 | 0.288 |
Half the added exposure is concentrated in 77 districts
The distribution of change is broad, but its contribution is unequal. Seventy-seven districts account for half of the positive hot-day resident-day equivalents. Across the fixed exposure surface, the net change is about 493 million hot-day resident-day equivalents per year. This is not a health-burden estimate and does not count distinct people.
Exact values · Showing 20 of 400 rows
| ARS | District | Type | State Code | Population 2022 Grid | Hot Days Population Change | Resident Day Contribution | Included In Lorenz |
|---|---|---|---|---|---|---|---|
| 11000 | Berlin | Kreisfreie Stadt | 11 | 3570942.18 | +6.03 | 21532495.68 | yes |
| 09162 | München | Kreisfreie Stadt | 09 | 1468894.53 | +7.89 | 11595218.41 | yes |
| 05315 | Köln | Kreisfreie Stadt | 05 | 1021119.85 | +7.15 | 7300149.19 | yes |
| 06412 | Frankfurt am Main | Kreisfreie Stadt | 06 | 755335.27 | +9.08 | 6855294.47 | yes |
| 03241 | Region Hannover | Landkreis | 03 | 1129164.20 | +5.82 | 6570120.92 | yes |
| 08226 | Rhein-Neckar-Kreis | Landkreis | 08 | 552790.91 | +11.33 | 6261186.26 | yes |
| 08317 | Ortenaukreis | Landkreis | 08 | 436710.09 | +13.44 | 5869034.24 | yes |
| 08111 | Stuttgart | Stadtkreis | 08 | 604271.14 | +9.57 | 5780916.99 | yes |
| 08118 | Ludwigsburg | Landkreis | 08 | 529409.58 | +10.88 | 5762358.56 | yes |
| 08116 | Esslingen | Landkreis | 08 | 530926.39 | +10.45 | 5547968.41 | yes |
| 14612 | Dresden | Kreisfreie Stadt | 14 | 562358.46 | +9.73 | 5470454.42 | yes |
| 08215 | Karlsruhe | Landkreis | 08 | 449914.98 | +11.56 | 5200477.26 | yes |
| 14713 | Leipzig | Kreisfreie Stadt | 14 | 598971.99 | +7.92 | 4743319.12 | yes |
| 08119 | Rems-Murr-Kreis | Landkreis | 08 | 440525.81 | +10.49 | 4622349.26 | yes |
| 09564 | Nürnberg | Kreisfreie Stadt | 09 | 515636.03 | +8.49 | 4376883.61 | yes |
| 02000 | Hamburg | Kreisfreie Stadt | 02 | 1796777.80 | +2.43 | 4367427.80 | yes |
| 06435 | Main-Kinzig-Kreis | Landkreis | 06 | 411455.59 | +9.70 | 3991740.49 | yes |
| 10041 | Regionalverband Saarbrücken | Landkreis | 10 | 327700.20 | +10.74 | 3520024.46 | yes |
| 08125 | Heilbronn | Landkreis | 08 | 347039.82 | +10.09 | 3502360.54 | yes |
| 06438 | Offenbach | Landkreis | 06 | 345077.13 | +10.09 | 3481931.76 | yes |
Heat Is Not One Variable
The recurrence headline is useful only if it does not become a new composite score. Daytime frequency, nighttime thresholds, spell persistence and population exposure answer different questions. This act keeps those mechanisms separate and makes their disagreements visible.
Day and night do not share one geography
The district scatter separates change in hot days from change in tropical-night-proxy days. Both increased in 202 districts; 193 gained hot days while the night proxy did not. The weak correlation means that one axis cannot stand in for the other. Threshold tails and seasonal means also remain conceptually distinct.
Exact values · Showing 18 of 400 rows
| District | Type | Hot Days Population Change | Proxy-night change · residents | Quadrant |
|---|---|---|---|---|
| Freiburg im Breisgau | Stadtkreis | +4.89 | -3.06 | Hot days increased; proxy nights did not |
| Südliche Weinstraße | Landkreis | +10.92 | +1.55 | Both increased |
| Heidelberg | Stadtkreis | +12.07 | -1.28 | Hot days increased; proxy nights did not |
| Landau in der Pfalz | Kreisfreie Stadt | +10.14 | +0.97 | Both increased |
| Germersheim | Landkreis | +11.38 | +0.60 | Both increased |
| Regensburg | Kreisfreie Stadt | +13.64 | -0.10 | Hot days increased; proxy nights did not |
| Merzig-Wadern | Landkreis | +10.13 | +0.80 | Both increased |
| Ortenaukreis | Landkreis | +13.44 | -0.13 | Hot days increased; proxy nights did not |
| Rastatt | Landkreis | +11.29 | +0.32 | Both increased |
| Kaiserslautern | Kreisfreie Stadt | +12.53 | -0.07 | Hot days increased; proxy nights did not |
| Karlsruhe | Landkreis | +11.56 | +0.25 | Both increased |
| Baden-Baden | Stadtkreis | +11.44 | +0.23 | Both increased |
| Mainz-Bingen | Landkreis | +8.36 | +0.85 | Both increased |
| Saarlouis | Landkreis | +10.45 | +0.43 | Both increased |
| Offenbach am Main | Kreisfreie Stadt | +9.32 | -0.64 | Hot days increased; proxy nights did not |
| Frankfurt am Main | Kreisfreie Stadt | +9.08 | +0.67 | Both increased |
| Würzburg | Kreisfreie Stadt | +11.23 | +0.23 | Both increased |
| Konstanz | Landkreis | +11.32 | +0.20 | Both increased |
Heat has a calendar, not only an annual total
Three algorithmically selected district calendars expose timing and clustering: a night-dominant shift, a day-dominant shift and a balanced high shift. Similar annual counts can arise from an early burst, a late episode or a prolonged sequence. Operational stress depends on that chronology.
Exact values · Showing 6 of 6 rows
| ARS | District | Archetype | Comparison Period | Majority Hot Days Total | Majority Proxy Nights Total | Both Thresholds Total | Years |
|---|---|---|---|---|---|---|---|
| 06434 | Hochtaunuskreis | balanced high shift | 1991–2000 | 87.00 | 5.00 | 4.00 | 10 |
| 06434 | Hochtaunuskreis | balanced high shift | 2016–2025 | 180.00 | 10.00 | 10.00 | 10 |
| 07337 | Südliche Weinstraße | night-dominant high shift | 1991–2000 | 104.00 | 3.00 | 3.00 | 10 |
| 07337 | Südliche Weinstraße | night-dominant high shift | 2016–2025 | 214.00 | 15.00 | 13.00 | 10 |
| 09562 | Erlangen | day-dominant high shift | 1991–2000 | 85.00 | 2.00 | 1.00 | 10 |
| 09562 | Erlangen | day-dominant high shift | 2016–2025 | 203.00 | 1.00 | 1.00 | 10 |
Residents and territory answer different questions
Land weighting asks how a district's territory changed. Resident weighting asks what the fixed 2022 represented-population surface would have experienced under each period's climate. The two agree closely in most places but diverge where settlements occupy a climatically distinct part of a heterogeneous district. Neither denominator should silently replace the other.
Exact values · Showing 18 of 400 rows
| District | Hot Days Area Change | Hot Days Population Change | Hot Days Resident Minus Area Gap | Tropical Nights Resident Minus Area Gap |
|---|---|---|---|---|
| Waldshut | +6.85 | +9.99 | +3.14 | +0.01 |
| Reutlingen | +5.32 | +8.39 | +3.06 | +0.04 |
| Berchtesgadener Land | +2.50 | +5.43 | +2.93 | -0.19 |
| Breisgau-Hochschwarzwald | +6.05 | +8.79 | +2.74 | -0.11 |
| Lörrach | +6.89 | +9.12 | +2.23 | +0.07 |
| Bad Kreuznach | +6.36 | +8.59 | +2.23 | +0.21 |
| Jena | +8.00 | +10.22 | +2.22 | +0.04 |
| Karlsruhe | +8.94 | +6.81 | -2.13 | -0.35 |
| Hochtaunuskreis | +6.19 | +8.32 | +2.13 | +0.14 |
| Göppingen | +6.63 | +8.74 | +2.11 | +0.10 |
| Ortenaukreis | +11.34 | +13.44 | +2.10 | -0.09 |
| Odenwaldkreis | +7.99 | +10.03 | +2.04 | -0.18 |
| Forchheim | +8.52 | +10.54 | +2.02 | -0.07 |
| Rastatt | +9.31 | +11.29 | +1.98 | +0.20 |
| Nürnberger Land | +6.91 | +8.85 | +1.94 | -0.08 |
| Garmisch-Partenkirchen | +1.74 | +3.65 | +1.91 | -0.06 |
| Rems-Murr-Kreis | +8.68 | +10.49 | +1.81 | -0.11 |
| Sächsische Schweiz-Osterzgebirge | +4.19 | +5.99 | +1.80 | +0.08 |
More hot days usually mean longer runs, with exceptions
Frequency and persistence move together strongly, but not perfectly. The descriptive district line has a correlation near 0.85. Residuals identify places where the longest hot run lengthened more or less than their increase in total hot days would suggest. The line summarizes association; it is not a causal or predictive model.
Exact values · Showing 20 of 400 rows
| ARS | District | Type | State Code | Hot Days Population Change | Longest Hot Run Population Change | Fitted Longest Hot Run Change | Residual |
|---|---|---|---|---|---|---|---|
| 09362 | Regensburg | Kreisfreie Stadt | 09 | +13.64 | +3.03 | +3.11 | -0.08 |
| 08317 | Ortenaukreis | Landkreis | 08 | +13.44 | +2.88 | +3.07 | -0.19 |
| 07312 | Kaiserslautern | Kreisfreie Stadt | 07 | +12.53 | +2.52 | +2.87 | -0.34 |
| 08221 | Heidelberg | Stadtkreis | 08 | +12.07 | +2.99 | +2.76 | +0.22 |
| 08215 | Karlsruhe | Landkreis | 08 | +11.56 | +2.79 | +2.65 | +0.14 |
| 09562 | Erlangen | Kreisfreie Stadt | 09 | +11.54 | +2.46 | +2.64 | -0.19 |
| 08211 | Baden-Baden | Stadtkreis | 08 | +11.44 | +2.56 | +2.62 | -0.06 |
| 07334 | Germersheim | Landkreis | 07 | +11.38 | +2.38 | +2.61 | -0.23 |
| 08226 | Rhein-Neckar-Kreis | Landkreis | 08 | +11.33 | +2.88 | +2.60 | +0.28 |
| 08335 | Konstanz | Landkreis | 08 | +11.32 | +3.45 | +2.60 | +0.85 |
| 08216 | Rastatt | Landkreis | 08 | +11.29 | +2.47 | +2.59 | -0.12 |
| 08128 | Main-Tauber-Kreis | Landkreis | 08 | +11.24 | +2.88 | +2.58 | +0.30 |
| 09663 | Würzburg | Kreisfreie Stadt | 09 | +11.23 | +2.42 | +2.58 | -0.16 |
| 07318 | Speyer | Kreisfreie Stadt | 07 | +11.11 | +2.76 | +2.55 | +0.21 |
| 07337 | Südliche Weinstraße | Landkreis | 07 | +10.92 | +2.16 | +2.51 | -0.34 |
| 08118 | Ludwigsburg | Landkreis | 08 | +10.88 | +2.97 | +2.50 | +0.47 |
| 09572 | Erlangen-Höchstadt | Landkreis | 09 | +10.85 | +2.45 | +2.49 | -0.04 |
| 10041 | Regionalverband Saarbrücken | Landkreis | 10 | +10.74 | +2.20 | +2.47 | -0.27 |
| 08316 | Emmendingen | Landkreis | 08 | +10.70 | +2.46 | +2.46 | -0.00 |
| 09474 | Forchheim | Landkreis | 09 | +10.54 | +2.34 | +2.42 | -0.08 |
Federal states do not contain the district variation
State intervals provide a governance lens without flattening local heterogeneity. Roughly half of descriptive district variation remains within states, so a Länder average is useful for coordination but insufficient for local planning. Each district remains inspectable inside its state's P10–P90 range.
Exact values · Showing 16 of 16 rows
| State Code | State Name | District Count | Lower Change Days | Inner Lower Change Days | Median Change Days | Inner Upper Change Days | Upper Change Days |
|---|---|---|---|---|---|---|---|
| 01 | Schleswig-Holstein | 15 | -0.12 | +0.05 | +0.34 | +0.99 | +1.38 |
| 02 | Hamburg | 1 | +2.43 | +2.43 | +2.43 | +2.43 | +2.43 |
| 03 | Lower Saxony | 45 | +0.49 | +2.03 | +3.52 | +4.40 | +5.07 |
| 04 | Bremen | 2 | +1.12 | +1.60 | +2.40 | +3.19 | +3.67 |
| 05 | North Rhine-Westphalia | 53 | +3.11 | +3.43 | +3.95 | +4.69 | +5.68 |
| 06 | Hesse | 26 | +4.78 | +5.80 | +6.73 | +9.26 | +9.64 |
| 07 | Rhineland-Palatinate | 36 | +3.58 | +4.68 | +8.26 | +9.58 | +10.62 |
| 08 | Baden-Württemberg | 44 | +5.97 | +6.86 | +8.77 | +10.46 | +11.32 |
| 09 | Bavaria | 96 | +3.65 | +4.85 | +6.03 | +7.85 | +9.46 |
| 10 | Saarland | 6 | +7.57 | +8.91 | +9.78 | +10.37 | +10.60 |
| 11 | Berlin | 1 | +6.03 | +6.03 | +6.03 | +6.03 | +6.03 |
| 12 | Brandenburg | 18 | +5.24 | +5.68 | +6.41 | +7.40 | +9.13 |
| 13 | Mecklenburg-Vorpommern | 8 | +1.60 | +1.86 | +2.50 | +3.22 | +3.94 |
| 14 | Saxony | 13 | +5.16 | +5.75 | +6.53 | +8.78 | +9.56 |
| 15 | Saxony-Anhalt | 14 | +5.88 | +6.57 | +7.94 | +8.80 | +9.57 |
| 16 | Thuringia | 22 | +4.42 | +5.19 | +6.37 | +7.86 | +8.98 |
Compound Climate, Qualified Evidence
Heat is only one part of the operating environment. Dry-spell persistence and concentrated rainfall can rise together, while gridded and station evidence impose different inferential limits. The final act makes those limits part of the argument rather than an appendix apology.
Six indicators do not tell one identical story
Small-multiple maps retain tropical-night proxies, hot days, longest hot runs, warm-season Tmin, consecutive dry days and Rx5day in their natural units. The purpose is comparison without an opaque index: alignment is evidence, and disagreement is evidence too.
Exact values · Showing 18 of 400 rows
| ARS | District | Type | NUTS3 | State Code | Proxy-night change · residents | Hot Days Population Change | Longest Hot Run Population Change |
|---|---|---|---|---|---|---|---|
| 01001 | Flensburg | Kreisfreie Stadt | DEF01 | 01 | -0.37 | +0.29 | +0.18 |
| 01002 | Kiel | Kreisfreie Stadt | DEF02 | 01 | -0.41 | +0.08 | -0.13 |
| 01003 | Lübeck | Kreisfreie Stadt | DEF03 | 01 | +0.06 | +0.03 | -0.25 |
| 01004 | Neumünster | Kreisfreie Stadt | DEF04 | 01 | -0.45 | +0.92 | -0.15 |
| 01051 | Dithmarschen | Kreis | DEF05 | 01 | -0.26 | +0.90 | +0.52 |
| 01053 | Herzogtum Lauenburg | Kreis | DEF06 | 01 | +0.05 | +2.94 | +0.59 |
| 01054 | Nordfriesland | Kreis | DEF07 | 01 | -0.32 | +0.29 | +0.48 |
| 01055 | Ostholstein | Kreis | DEF08 | 01 | +0.01 | -0.37 | -0.34 |
| 01056 | Pinneberg | Kreis | DEF09 | 01 | -0.34 | +1.43 | +0.35 |
| 01057 | Plön | Kreis | DEF0A | 01 | -0.28 | -0.22 | -0.34 |
| 01058 | Rendsburg-Eckernförde | Kreis | DEF0B | 01 | -0.32 | +0.03 | -0.09 |
| 01059 | Schleswig-Flensburg | Kreis | DEF0C | 01 | -0.23 | +0.34 | +0.15 |
| 01060 | Segeberg | Kreis | DEF0D | 01 | -0.21 | +1.02 | +0.07 |
| 01061 | Steinburg | Kreis | DEF0E | 01 | -0.22 | +0.95 | +0.53 |
| 01062 | Stormarn | Kreis | DEF0F | 01 | -0.09 | +1.31 | +0.02 |
| 02000 | Hamburg | Kreisfreie Stadt | DE600 | 02 | -0.58 | +2.43 | +0.41 |
| 03101 | Braunschweig | Kreisfreie Stadt | DE911 | 03 | -0.12 | +4.40 | +1.18 |
| 03102 | Salzgitter | Kreisfreie Stadt | DE912 | 03 | -0.21 | +4.31 | +1.21 |
Longer dry spells and heavier rain can coexist
The longest May–September dry spell increased in 273 districts, while maximum warm-season five-day rainfall increased in 243. Both rose in 175 districts. Persistence and concentration are distinct mechanisms; calling a district simply wetter or drier would erase the compound signal.
Exact values · Showing 18 of 400 rows
| District | CDD change | Rx5day change | Quadrant |
|---|---|---|---|
| Flensburg | -0.66 | -3.63 | Both declined or unchanged |
| Kiel | -3.69 | -0.46 | Both declined or unchanged |
| Lübeck | -1.72 | +2.30 | Heavier five-day rain only |
| Neumünster | -0.80 | -1.78 | Both declined or unchanged |
| Dithmarschen | -0.73 | -1.58 | Both declined or unchanged |
| Herzogtum Lauenburg | -2.07 | -1.26 | Both declined or unchanged |
| Nordfriesland | -2.16 | +2.91 | Heavier five-day rain only |
| Ostholstein | -1.56 | -5.79 | Both declined or unchanged |
| Pinneberg | -2.89 | -5.78 | Both declined or unchanged |
| Plön | -2.29 | -4.66 | Both declined or unchanged |
| Rendsburg-Eckernförde | -2.32 | +2.23 | Heavier five-day rain only |
| Schleswig-Flensburg | -1.70 | +1.77 | Heavier five-day rain only |
| Segeberg | -2.10 | -1.74 | Both declined or unchanged |
| Steinburg | -1.10 | -0.73 | Both declined or unchanged |
| Stormarn | -1.20 | -4.19 | Both declined or unchanged |
| Hamburg | -3.08 | -3.38 | Both declined or unchanged |
| Braunschweig | -2.46 | +4.39 | Heavier five-day rain only |
| Salzgitter | -0.78 | +11.83 | Heavier five-day rain only |
The strict panel does not support a broad nighttime-faster hypothesis
Among 24 stable stations, seasonal mean Tmax rose much faster than Tmin. The broader low-relocation sample associates modeled urban exposure with a smaller daytime-over-nighttime warming gap, but the strict panel's interval is wide and crosses zero. The careful conclusion is descriptive: urban exposure may narrow the gap; it does not establish a nationwide nighttime lead or causal urbanization effect.
Exact values · Showing 5 of 5 rows
| Panel | Sample | Outcome | N Stations | N Years | Estimate | Standard Error | CI low |
|---|---|---|---|---|---|---|---|
| National trend | strict no move 33of35 | warm season tmin | 24 | 35 | +0.16 | 0.06 | 0.04 |
| National trend | strict no move 33of35 | warm season tmax | 24 | 35 | +0.70 | 0.11 | 0.49 |
| National trend | strict no move 33of35 | tmin minus tmax | 24 | 35 | -0.54 | 0.08 | -0.69 |
| UHI exposure association | strict no move 33of35 | night minus day slope c per decade | 24 | 35 | +0.07 | 0.21 | -0.34 |
| UHI exposure association | near fixed 100m 10m 33of35 | night minus day slope c per decade | 105 | 35 | +0.11 | 0.05 | 0.01 |
What This Atlas Measures
HYRAS-DE v6-1 daily temperature and precipitation grids are evaluated for 1991–2025. Thresholds and consecutive runs are calculated in each 1 km grid cell before aggregation. Comparisons use complete 1991–2000 and 2016–2025 windows; annual paths retain all 35 years.
Both periods use the 1 January 2025 district geography. Resident weighting uses one fixed 2022 represented-population surface; land weighting uses cell area. The represented surface totals 82,706,456 after suppressed census cells and boundary allocation, 7.05% below the official census population. It standardizes spatial exposure and does not reconstruct historical migration.
HYRAS is interpolated station evidence, not 1 km observational truth and not automatically homogenized. Station relocation, instruments, changing network support, the formal tropical-night time window and neighbouring-country precipitation inputs constrain interpretation. The article reports observed climate and standardized exposure, not household vulnerability or causal attribution.
Threshold first
Daily thresholds and consecutive runs are calculated inside each 1 km grid cell before district aggregation. Averaging first would suppress local exposure.
Two complete decades
The communicative comparison is 1991–2000 against 2016–2025. The complete 1991–2025 annual path prevents one extreme year from masquerading as a regime.
One fixed surface
Resident and area weights answer different questions. The fixed 2022 represented-population surface standardizes exposure; it does not reconstruct migration.
Bounded claims
The district atlas is observational. Station and UHI checks probe mechanism, while relocation, interpolation and daily-window differences bound what the evidence can establish.
Source and definition ledger
Climate grids. DWD HYRAS-DE v6-1 daily tasmin, tasmax and precipitation. Thresholds are evaluated at grid-cell/day grain before aggregation; grid spacing is not observational resolution.
Boundaries and weights. BKG VG250 districts fixed to 1 January 2025; one fixed Zensus 2022 1 km represented-population surface and cell-area weights.
Station check. DWD daily station observations, station histories and Project UHI-MAP are used as a separate descriptive robustness layer, not as the main district estimator.
Uncertainty. Common-year resampling retains the national annual covariance structure by applying the same resampled year indices to every district. Sensitivity ranges vary declared rules; they are not confidence intervals.
Publication boundary. This is an independent analysis of DWD products, not an official DWD publication. It describes outdoor climate regimes and standardized exposure, not household-level vulnerability or causal attribution.
Recurrence Changes the Planning Question
An average tells us that the climate moved. Recurrence tells a district whether it has entered a different operating regime. Local decisions should therefore measure recurrence, keep daytime and nighttime mechanisms separate, and treat dry-spell persistence and rainfall concentration as distinct demands on infrastructure.
The next useful layer is not another composite climate score. It is a transparent join to age, housing, canopy, hospital capacity and outdoor work—each visible enough to show which mechanism changes the decision.