Top 100 Countries by Average Annual Temperature, 1991–2020 vs 2025
Climate Normals vs. 2025: What the 1991–2020 Baseline Reveals
This ranking compares average annual near-surface air temperature (2-meter air temperature) at the country level across two reference frames: the standard 1991–2020 climate normal and the 2025 annual estimate. The gap between them is the anomaly — how much warmer (or cooler) the current year was relative to the recent baseline.
Thank you for reading this post, don't forget to subscribe!A climate normal is a 30-year average. The World Meteorological Organization updates it every decade, and the 1991–2020 period is the operational normal used in current climate reporting. It smooths year-to-year noise while still reflecting the recent climate state — which makes country-to-country comparisons meaningful.
Reading the values: The 1991–2020 normals come from the World Bank Climate Change Knowledge Portal aggregation of Climatic Research Unit (CRU) gridded climatology. The 2025 figures are derived from ERA5 monthly reanalysis output through year-end 2025 and remain subject to revision until the dataset is finalised. All values are rounded to one decimal place.
Top 10 warmest countries by the 1991–2020 normal
The leaders are concentrated in the West African Sahel and along the Sahara’s southern margin, where solar irradiance is intense, cloud cover is limited, and the dry season produces a long stretch of hot days and warm nights.
| Rank | Country | 1991–2020 normal (°C) |
|---|---|---|
| 1 | Burkina Faso | 30.4 |
| 2 | Mali | 29.2 |
| 3 | Senegal | 28.9 |
| 4 | Mauritania | 28.8 |
| 5 | Tuvalu | 28.6 |
| 6 | Djibouti | 28.5 |
| 7 | The Gambia | 28.4 |
| 8 | United Arab Emirates | 28.2 |
| 9 | Maldives | 28.1 |
| 10 | Niger | 28.0 |
Top 10 countries by the 1991–2020 climate normal (°C), rounded to one decimal.
A high annual average is not just about being near the equator. The leaderboard mixes two climate regimes: continental Sahel (Burkina Faso, Mali, Niger), where the dry-season heat dominates the annual cycle, and maritime tropical (Tuvalu, Maldives), where seasonality is minimal and ocean heat keeps temperatures consistently warm. The Gulf states add a third pattern — desert continentality moderated by warm coastal waters.
Absolute Heat vs. Anomaly Heat: Two Different Maps
Discussions of climate often blur two distinct questions: where is it hottest in absolute terms?, and where is the current year most unusual relative to its baseline? These rankings can look very different. Absolute temperature is fixed by latitude, sunshine, and the length of the warm season; anomalies depend on regional circulation, ocean state, and how much the baseline itself was already cool.
In 2025 the Top 20 by absolute temperature stayed concentrated in the Sahel, the Horn of Africa, and the Arabian Peninsula — the same regions as the 1991–2020 leaders. The Top 20 by anomaly, however, sat far away: in central Asia, Fennoscandia, and northwestern Europe, where Copernicus reported record or near-record annual values for 2025.
National-level 2-metre air temperature is aggregated from a gridded product to country borders with area weighting. The 1991–2020 normal uses CRU climatology via the World Bank portal; the 2025 estimate uses ERA5 monthly means. Both products are widely cross-referenced in climate monitoring and converge on the same broad rank order for major countries.
Chart 1 — Top 20 countries by 2025 annual mean temperature
Ranked by 2025 annual mean (°C). 2025 is provisional pending dataset finalisation.
Chart 2 — Top 20 countries by 2025 anomaly vs. 1991–2020
Anomalies (°C) for 2025 relative to the 1991–2020 baseline. Central Asia and far-northern Europe show the strongest positive departures — consistent with Copernicus regional summaries for 2025.
For travellers, the absolute ranking is the relevant one. For infrastructure planners, agronomists, and insurers, the anomaly map matters more: systems built around the 1991–2020 baseline face the most stress where the baseline is being exceeded most strongly, even if the country itself is not “hot” by global standards.
Top 100 countries by 2025 annual mean temperature
Format: 1991–2020 normal → 2025 (anomaly). Ranking is by the 2025 value.
| Rank | Country | 1991–2020 → 2025 (Δ, °C) |
|---|---|---|
| 1 | Burkina Faso | 30.4 → 31.0 (+0.6) |
| 2 | Mali | 29.2 → 29.8 (+0.6) |
| 3 | Senegal | 28.9 → 29.5 (+0.6) |
| 4 | Mauritania | 28.8 → 29.5 (+0.7) |
| 5 | Tuvalu | 28.6 → 29.0 (+0.4) |
| 6 | Djibouti | 28.5 → 29.1 (+0.6) |
| 7 | The Gambia | 28.4 → 29.0 (+0.6) |
| 8 | United Arab Emirates | 28.2 → 28.9 (+0.7) |
| 9 | Maldives | 28.1 → 28.5 (+0.4) |
| 10 | Niger | 28.0 → 28.7 (+0.7) |
| 11 | Qatar | 28.0 → 28.7 (+0.7) |
| 12 | Benin | 28.0 → 28.5 (+0.5) |
| 13 | Marshall Islands | 28.0 → 28.4 (+0.4) |
| 14 | Guinea-Bissau | 28.0 → 28.6 (+0.6) |
| 15 | South Sudan | 28.0 → 28.6 (+0.6) |
| 16 | Sudan | 28.0 → 28.7 (+0.7) |
| 17 | Palau | 27.9 → 28.3 (+0.4) |
| 18 | Nauru | 27.8 → 28.2 (+0.4) |
| 19 | Kiribati | 27.8 → 28.2 (+0.4) |
| 20 | Bahrain | 27.7 → 28.4 (+0.7) |
| 21 | Singapore | 27.7 → 28.1 (+0.4) |
| 22 | Ghana | 27.7 → 28.2 (+0.5) |
| 23 | Oman | 27.6 → 28.3 (+0.7) |
| 24 | Chad | 27.6 → 28.3 (+0.7) |
| 25 | Samoa | 27.6 → 28.0 (+0.4) |
| 26 | Cambodia | 27.4 → 27.9 (+0.5) |
| 27 | Togo | 27.3 → 27.8 (+0.5) |
| 28 | Nigeria | 27.3 → 27.9 (+0.6) |
| 29 | Micronesia | 27.3 → 27.7 (+0.4) |
| 30 | Sri Lanka | 27.3 → 27.7 (+0.4) |
| 31 | Antigua and Barbuda | 27.2 → 27.6 (+0.4) |
| 32 | Seychelles | 27.1 → 27.5 (+0.4) |
| 33 | Philippines | 27.1 → 27.6 (+0.5) |
| 34 | Saint Lucia | 27.0 → 27.4 (+0.4) |
| 35 | Brunei | 27.0 → 27.4 (+0.4) |
| 36 | Somalia | 27.0 → 27.5 (+0.5) |
| 37 | Thailand | 26.9 → 27.4 (+0.5) |
| 38 | Côte d’Ivoire | 26.8 → 27.3 (+0.5) |
| 39 | Eritrea | 26.6 → 27.2 (+0.6) |
| 40 | Barbados | 26.6 → 27.0 (+0.4) |
| 41 | Trinidad and Tobago | 26.6 → 27.0 (+0.4) |
| 42 | Sierra Leone | 26.5 → 27.0 (+0.5) |
| 43 | Grenada | 26.5 → 26.9 (+0.4) |
| 44 | Malaysia | 26.4 → 26.8 (+0.4) |
| 45 | Kuwait | 26.3 → 27.1 (+0.8) |
| 46 | Saint Vincent and the Grenadines | 26.2 → 26.6 (+0.4) |
| 47 | Indonesia | 26.0 → 26.4 (+0.4) |
| 48 | Saudi Arabia | 25.9 → 26.7 (+0.8) |
| 49 | Solomon Islands | 25.9 → 26.3 (+0.4) |
| 50 | Jamaica | 25.9 → 26.3 (+0.4) |
| 51 | Nicaragua | 25.9 → 26.3 (+0.4) |
| 52 | Guinea | 25.9 → 26.3 (+0.4) |
| 53 | Cuba | 25.8 → 26.2 (+0.4) |
| 54 | Bangladesh | 25.7 → 26.3 (+0.6) |
| 55 | Venezuela | 25.7 → 26.1 (+0.4) |
| 56 | Belize | 25.7 → 26.1 (+0.4) |
| 57 | Panama | 25.6 → 26.0 (+0.4) |
| 58 | Bahamas | 25.6 → 26.1 (+0.5) |
| 59 | Yemen | 25.5 → 26.2 (+0.7) |
| 60 | Central African Republic | 25.5 → 26.0 (+0.5) |
| 61 | Liberia | 25.5 → 25.9 (+0.4) |
| 62 | Brazil | 25.4 → 25.9 (+0.5) |
| 63 | El Salvador | 25.2 → 25.6 (+0.4) |
| 64 | Gabon | 25.2 → 25.6 (+0.4) |
| 65 | Kenya | 25.1 → 25.6 (+0.5) |
| 66 | Tonga | 25.0 → 25.4 (+0.4) |
| 67 | Colombia | 25.0 → 25.4 (+0.4) |
| 68 | Haiti | 25.0 → 25.4 (+0.4) |
| 69 | India | 24.9 → 25.5 (+0.6) |
| 70 | Costa Rica | 24.8 → 25.2 (+0.4) |
| 71 | Cameroon | 24.8 → 25.3 (+0.5) |
| 72 | Vietnam | 24.8 → 25.4 (+0.6) |
| 73 | Republic of the Congo | 24.7 → 25.2 (+0.5) |
| 74 | Papua New Guinea | 24.7 → 25.1 (+0.4) |
| 75 | Honduras | 24.7 → 25.1 (+0.4) |
| 76 | Fiji | 24.7 → 25.1 (+0.4) |
| 77 | Equatorial Guinea | 24.7 → 25.1 (+0.4) |
| 78 | Timor-Leste | 24.6 → 25.1 (+0.5) |
| 79 | Dominican Republic | 24.6 → 25.0 (+0.4) |
| 80 | São Tomé and Príncipe | 24.5 → 25.0 (+0.5) |
| 81 | Vanuatu | 24.4 → 24.8 (+0.4) |
| 82 | Mozambique | 24.4 → 24.9 (+0.5) |
| 83 | DR Congo | 24.4 → 24.9 (+0.5) |
| 84 | Laos | 24.2 → 24.8 (+0.6) |
| 85 | Paraguay | 23.9 → 24.4 (+0.5) |
| 86 | Myanmar | 23.8 → 24.4 (+0.6) |
| 87 | Comoros | 23.7 → 24.2 (+0.5) |
| 88 | Guatemala | 23.7 → 24.1 (+0.4) |
| 89 | Algeria | 23.6 → 24.4 (+0.8) |
| 90 | Ethiopia | 23.4 → 23.9 (+0.5) |
| 91 | Mauritius | 23.3 → 23.8 (+0.5) |
| 92 | Uganda | 23.3 → 23.8 (+0.5) |
| 93 | Egypt | 23.1 → 23.9 (+0.8) |
| 94 | Iraq | 23.0 → 23.9 (+0.9) |
| 95 | Tanzania | 22.9 → 23.4 (+0.5) |
| 96 | Libya | 22.8 → 23.6 (+0.8) |
| 97 | Malawi | 22.7 → 23.2 (+0.5) |
| 98 | Madagascar | 22.6 → 23.1 (+0.5) |
| 99 | Cape Verde | 22.5 → 23.0 (+0.5) |
| 100 | Zambia | 22.2 → 22.8 (+0.6) |
Two columns, two questions. The first column (1991–2020 normal) tells you whether a country is structurally warm. The anomaly tells you whether 2025 strayed from that structure. In 2025 the tropics moved less than they did in 2023–2024 — the El Niño signal had faded by mid-year — while continental and high-latitude regions kept warming above their baselines.
What the Ranking Says — and What It Doesn’t
An annual mean is a coarse statistic. It tells you nothing about a heatwave, a humid-tropics afternoon, or the difference between a coastal city and an inland plateau in the same country. What it does describe is the background energy state of a place: how much cooling capacity its buildings need, how long its growing season is, how its labour productivity is constrained in summer, how warm its nights tend to be.
The 1991–2020 normal anchors that background. The 2025 value tells you whether the year was business-as-usual or stress-test territory. Both numbers matter, and they answer different questions.
What 2025 looked like in regional terms
- Polar amplification was the headline. Copernicus put the Antarctic annual anomaly at a record +1.06 °C, and the Arctic at +1.37 °C — the second-highest on record. This is consistent with the long-running pattern of stronger warming at high latitudes.
- Central Asia and far-northern Europe set regional records. The northwestern and southwestern Pacific, the northeastern Atlantic, and the western flanks of the Eurasian continent all registered record annual temperatures in 2025.
- The tropics cooled relative to 2023–2024. With the 2023–2024 El Niño dissipated, tropical anomalies fell back to roughly +0.4–0.6 °C, well below the unprecedented values of the preceding two years. This is why several countries in the Top 20 absolute list have modest anomalies in the table above.
- Mixed conditions in parts of Asia. Eastern Russia, parts of Mongolia, and northern China recorded negative monthly anomalies during the year, pulling some continental Asian averages closer to (or below) baseline.
Reading the Top 100 by climate cluster
- Hot arid belt — Sahel, Sahara, Arabian Peninsula, Horn of Africa. Structurally warm thanks to high solar load, sparse cloud cover, and limited seasonal contrast. The baseline is already at the top of the global distribution; anomalies in 2025 were moderate.
- Humid tropics — equatorial Africa, Southeast Asia, tropical Americas. Annual means are high but day-to-day human heat stress is driven more by humidity and night-time temperatures than by the mean.
- Warm-island regimes — small island states across the Pacific, Caribbean, and Indian Ocean. Low seasonality, ocean-moderated. National averages are reliable as concepts, but small landmasses are more sensitive to grid resolution and coastline handling than continental countries.
Practical takeaways
- Use the baseline for design, the anomaly for stress. Infrastructure tolerances are usually built around climatology. Anomalies tell you how often that climatology will be exceeded.
- Don’t conflate annual means with heat risk. Heat mortality is driven by extremes, humidity, and night-time minima, not by the annual average. The ranking is context, not a risk assessment.
- Use one baseline consistently. Switching between 1961–1990, 1981–2010, and 1991–2020 produces different anomalies for the same country. The 1991–2020 normal is now the WMO operational baseline; sticking to it keeps comparisons honest.
- State the data vintage. A 2025 annual value built from data through November differs from one built after the December reanalysis. Provisional values get revised; published rankings should declare which version they used.
FAQ
What is a 1991–2020 climate normal?
A 30-year average of a climate variable, used as the operational reference point in current climate monitoring. The WMO recommends updating this baseline every decade. The current standard (1991–2020) replaced the previous 1981–2010 normal in 2021.
Why don’t all sources give the same country average?
Three reasons. Different gridded products (CRU, ERA5, Berkeley Earth, NOAA) produce slightly different values. Different weightings (area-weighted, population-weighted, land-only vs. land-plus-coast) shift the result. And different border definitions matter for small or coastline-heavy countries. For most large continental countries the agreement is within a few tenths of a degree.
How should a traveller read this ranking?
As a guide to background warmth, not to specific months. A country with an annual mean above 27 °C will feel warm year-round, but humidity, elevation, and the rainy season matter more for day-to-day comfort. Pair this list with monthly climatology before booking a trip.
Is a country average representative of where people live?
No. National averages mix mountains with coasts and deserts with savanna. For impact analysis, population-weighted or city-level metrics are more relevant. Russia and Canada illustrate the gap clearly: both are cold on average, but their populations live almost entirely in the warmer portions of those countries.
What are the main limits of reanalysis-based national averages?
Reanalysis blends observations with a physical model on a regular grid. For large continental countries the result is robust. For small island states, coastal microstates, or countries with very few in-situ observations, uncertainty is higher and the choice of grid resolution can shift the national mean by a few tenths of a degree.
Sources
World Bank — Climate Change Knowledge Portal (CRU climatology).
Country-level 1991–2020 normals based on Climatic Research Unit gridded climatology.
Copernicus Climate Change Service / ECMWF — ERA5 monthly means.
Reanalysis dataset providing gridded near-surface temperature used for the 2025 annual estimates and regional anomaly patterns.
https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels-monthly-means
Copernicus C3S — Global Climate Highlights 2025.
Official annual climate summary with global and regional anomalies relative to the 1991–2020 baseline.
https://climate.copernicus.eu/copernicus-2025-was-third-hottest-year-record
Our World in Data — Average annual surface temperature.
Country-level ERA5 aggregation with documentation and downloadable CSV.
https://ourworldindata.org/grapher/average-annual-surface-temperature
NOAA NCEI — Climate at a Glance.
Cross-reference for baseline conventions and global time-series context.
https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/global/time-series
The 1991–2020 normal is treated here as the operational baseline. The 2025 annual mean is provisional and rounded for comparative communication; it should not be read as a final audited national statistic.
