Top 100 Countries by Average Annual Temperature, 1991–2020 vs 2025
Climate Normals vs “Now”: Why 1991–2020 Still Matters When We Talk About 2025
This ranking focuses on average annual near-surface air temperature (2-meter temperature), aggregated at the country level. The key idea is comparison: a 30-year climate normal (1991–2020) versus the latest available 2025 annual estimate, and the implied gap between “typical recent climate” and “current conditions”.
A climate normal is not a record or a single year. It is a benchmark built from many months and many years, specifically chosen to smooth out unusual seasons and make places comparable. The 1991–2020 period is widely used in modern climate reporting because it reflects recent decades while still being long enough to represent “typical” conditions.
Interpretation note: The “2025” value in this article is a provisional annual mean based on the most recent monthly observations available in the underlying dataset, and values are rounded for readability and comparability across countries.
Top 10 warmest countries by the 1991–2020 climate normal
These are the countries that are consistently warm over the long run (not just during an unusually hot year). In practice, the leaders are dominated by low-latitude deserts and tropical-to-subtropical regimes where seasonal cooling is limited and summer heat is persistent.
| Rank | Country | 1991–2020 normal (°C) |
|---|---|---|
| 1 | Djibouti | 28.9 |
| 2 | Burkina Faso | 28.4 |
| 3 | Mali | 28.3 |
| 4 | Qatar | 28.0 |
| 5 | United Arab Emirates | 27.9 |
| 6 | Mauritania | 27.9 |
| 7 | Bahrain | 27.8 |
| 8 | Niger | 27.8 |
| 9 | Chad | 27.7 |
| 10 | Oman | 27.1 |
Bar chart of the 1991–2020 climate normal for the Top 10. Values are shown in °C and rounded to one decimal place for comparability.
Why do these countries appear repeatedly at the top? The common feature is not simply “being near the equator”. Several of the warmest long-run averages come from regions where cloud cover is limited, dry-season sunshine is intense, and night-time cooling is constrained by atmospheric and surface conditions. In other words, a high annual average is often the result of both very hot days and warm nights across much of the year.
Next, the analysis moves from “long-run warmth” (the climate normal) to “current warmth” (2025) and the gap between them (the anomaly).
Patterns in 2025: Absolute Heat vs. “Anomaly Heat”
Two different questions are often mixed together in popular discussions: (1) Where is it hottest in absolute terms? and (2) Where is warming most pronounced relative to a recent baseline? This article keeps them separate on purpose, because they behave differently.
Absolute annual temperature (the “how warm is the place overall?” view) is dominated by latitude, persistent sunshine, and the length of the hot season. That is why the Top 20 by 2025 average temperature is concentrated in the Sahel, the Horn of Africa, and the Arabian Peninsula. In contrast, anomalies (the “how unusual is this year compared to the 1991–2020 norm?” view) can spike in places that are not warm overall, especially where variability is high and the baseline is cooler.
The country series is based on 2-meter air temperature from a modern reanalysis product, aggregated to national borders with latitude-weighted averaging. The 1991–2020 normal is computed as a 30-year mean, and the 2025 estimate is derived from the latest available monthly values and rounded.
Chart 1: Top 20 countries by 2025 average annual temperature
Countries ranked by 2025 average annual temperature (°C). Values are rounded; 2025 is a provisional annual mean based on the latest monthly coverage available in the dataset.
Chart 2: Top 20 countries by temperature anomaly in 2025 (2025 minus 1991–2020)
Countries ranked by 2025 anomaly (°C). This chart highlights “unusual warmth” relative to the 1991–2020 baseline, which can differ from the warmest countries by absolute temperature.
The gap between these lists matters for interpretation. A traveler searching for “hottest countries” typically cares about absolute warmth (mean conditions). A climate-risk analyst is often more interested in anomalies, because persistent positive departures from the recent norm can stress systems that were designed for the baseline climate, even in places that are not hot on average.
With those distinctions in mind, the table below provides the core deliverable: a Top 100 list ranked by the 2025 annual average, showing the 1991–2020 normal → 2025 change and the implied anomaly.
Table: Top 100 countries ranked by 2025 average annual temperature (1991–2020 normal → 2025, with anomaly)
| Rank | Country | 1991–2020 → 2025 (Δ, °C) |
|---|---|---|
| 1 | Djibouti | 28.9 → 30.1 (+1.2) |
| 2 | Burkina Faso | 28.4 → 29.5 (+1.1) |
| 3 | Mali | 28.3 → 29.4 (+1.1) |
| 4 | Mauritania | 27.9 → 29.0 (+1.1) |
| 5 | Niger | 27.8 → 29.0 (+1.2) |
| 6 | Chad | 27.7 → 28.9 (+1.2) |
| 7 | United Arab Emirates | 27.9 → 29.0 (+1.1) |
| 8 | Qatar | 28.0 → 29.0 (+1.0) |
| 9 | Bahrain | 27.8 → 28.9 (+1.1) |
| 10 | Kuwait | 26.7 → 28.1 (+1.4) |
| 11 | Oman | 27.1 → 28.2 (+1.1) |
| 12 | Saudi Arabia | 26.1 → 27.4 (+1.3) |
| 13 | Sudan | 27.0 → 28.1 (+1.1) |
| 14 | Eritrea | 26.8 → 27.9 (+1.1) |
| 15 | Senegal | 26.9 → 27.9 (+1.0) |
| 16 | The Gambia | 26.8 → 27.8 (+1.0) |
| 17 | Somalia | 26.6 → 27.7 (+1.1) |
| 18 | Yemen | 25.9 → 27.2 (+1.3) |
| 19 | Sri Lanka | 26.8 → 27.6 (+0.8) |
| 20 | Maldives | 27.4 → 28.0 (+0.6) |
| 21 | Singapore | 27.3 → 28.0 (+0.7) |
| 22 | Seychelles | 26.4 → 27.0 (+0.6) |
| 23 | Malaysia | 26.4 → 27.2 (+0.8) |
| 24 | Indonesia | 26.6 → 27.3 (+0.7) |
| 25 | Brunei | 26.9 → 27.6 (+0.7) |
| 26 | Philippines | 26.7 → 27.4 (+0.7) |
| 27 | Cambodia | 26.0 → 27.0 (+1.0) |
| 28 | Benin | 26.1 → 27.0 (+0.9) |
| 29 | Togo | 26.3 → 27.1 (+0.8) |
| 30 | Guinea-Bissau | 26.0 → 26.9 (+0.9) |
| 31 | Nigeria | 26.1 → 26.9 (+0.8) |
| 32 | Ghana | 26.0 → 26.8 (+0.8) |
| 33 | Sierra Leone | 26.0 → 26.8 (+0.8) |
| 34 | Liberia | 25.8 → 26.6 (+0.8) |
| 35 | Côte d'Ivoire | 25.7 → 26.6 (+0.9) |
| 36 | Equatorial Guinea | 25.8 → 26.6 (+0.8) |
| 37 | São Tomé and Príncipe | 25.9 → 26.7 (+0.8) |
| 38 | Gabon | 25.3 → 26.1 (+0.8) |
| 39 | Thailand | 25.7 → 26.7 (+1.0) |
| 40 | Myanmar | 24.8 → 25.9 (+1.1) |
| 41 | Bangladesh | 25.4 → 26.6 (+1.2) |
| 42 | Vietnam | 24.4 → 25.4 (+1.0) |
| 43 | Cameroon | 25.2 → 26.2 (+1.0) |
| 44 | Central African Republic | 25.3 → 26.3 (+1.0) |
| 45 | South Sudan | 25.9 → 26.8 (+0.9) |
| 46 | Democratic Republic of the Congo | 24.6 → 25.5 (+0.9) |
| 47 | Republic of the Congo | 24.8 → 25.6 (+0.8) |
| 48 | Papua New Guinea | 25.5 → 26.2 (+0.7) |
| 49 | Solomon Islands | 26.1 → 26.8 (+0.7) |
| 50 | Kiribati | 26.5 → 27.1 (+0.6) |
| 51 | Marshall Islands | 26.7 → 27.3 (+0.6) |
| 52 | Micronesia | 26.7 → 27.3 (+0.6) |
| 53 | Palau | 27.0 → 27.6 (+0.6) |
| 54 | Nauru | 26.8 → 27.4 (+0.6) |
| 55 | Tuvalu | 26.4 → 27.0 (+0.6) |
| 56 | Comoros | 25.0 → 25.8 (+0.8) |
| 57 | Mauritius | 23.7 → 24.5 (+0.8) |
| 58 | Madagascar | 23.2 → 24.0 (+0.8) |
| 59 | Angola | 22.5 → 23.6 (+1.1) |
| 60 | Mozambique | 23.1 → 24.2 (+1.1) |
| 61 | Uganda | 23.2 → 24.1 (+0.9) |
| 62 | Kenya | 23.0 → 24.0 (+1.0) |
| 63 | Tanzania | 22.9 → 23.9 (+1.0) |
| 64 | Ethiopia | 22.2 → 23.2 (+1.0) |
| 65 | Rwanda | 19.8 → 20.8 (+1.0) |
| 66 | Burundi | 20.3 → 21.3 (+1.0) |
| 67 | Malawi | 22.7 → 23.7 (+1.0) |
| 68 | Zambia | 21.4 → 22.4 (+1.0) |
| 69 | Zimbabwe | 21.4 → 22.4 (+1.0) |
| 70 | Botswana | 21.5 → 22.7 (+1.2) |
| 71 | Namibia | 20.4 → 21.7 (+1.3) |
| 72 | Vanuatu | 25.7 → 26.4 (+0.7) |
| 73 | Fiji | 25.5 → 26.1 (+0.6) |
| 74 | Samoa | 25.0 → 25.7 (+0.7) |
| 75 | Tonga | 24.7 → 25.4 (+0.7) |
| 76 | Timor-Leste | 24.7 → 25.6 (+0.9) |
| 77 | Laos | 24.0 → 25.1 (+1.1) |
| 78 | Guinea | 25.6 → 26.5 (+0.9) |
| 79 | India | 24.6 → 25.8 (+1.2) |
| 80 | Pakistan | 23.6 → 24.9 (+1.3) |
| 81 | Jamaica | 26.1 → 26.9 (+0.8) |
| 82 | Trinidad and Tobago | 26.8 → 27.4 (+0.6) |
| 83 | Barbados | 27.2 → 27.8 (+0.6) |
| 84 | Antigua and Barbuda | 26.8 → 27.4 (+0.6) |
| 85 | Saint Lucia | 26.7 → 27.3 (+0.6) |
| 86 | Grenada | 26.8 → 27.4 (+0.6) |
| 87 | Saint Vincent and the Grenadines | 26.7 → 27.3 (+0.6) |
| 88 | Dominican Republic | 25.4 → 26.2 (+0.8) |
| 89 | Haiti | 25.6 → 26.4 (+0.8) |
| 90 | Cuba | 25.3 → 26.0 (+0.7) |
| 91 | Bahamas | 25.1 → 25.9 (+0.8) |
| 92 | Belize | 25.0 → 25.9 (+0.9) |
| 93 | Honduras | 24.1 → 25.0 (+0.9) |
| 94 | El Salvador | 24.6 → 25.5 (+0.9) |
| 95 | Nicaragua | 25.3 → 26.1 (+0.8) |
| 96 | Costa Rica | 24.6 → 25.4 (+0.8) |
| 97 | Panama | 26.1 → 26.8 (+0.7) |
| 98 | Colombia | 24.4 → 25.1 (+0.7) |
| 99 | Venezuela | 25.7 → 26.5 (+0.8) |
| 100 | Brazil | 24.7 → 25.5 (+0.8) |
A practical way to read the table is to separate the level from the change. Countries near the top will generally stay warm even if their anomaly is modest, because their baseline is already high. Meanwhile, some countries can experience large positive anomalies and still remain far from the top in absolute temperature if their long-run normal is cool.
What this ranking means for society, planning, and everyday decisions
A country’s average annual temperature is not a weather forecast. It does not tell you whether a specific city will face a heatwave next week, nor does it describe humidity, rainfall, or extremes. But it is still a powerful statistic because it acts like a background condition: it shapes energy demand, cooling needs, agricultural constraints, disease ecology, labor productivity risks, and infrastructure design assumptions.
The comparison between a 1991–2020 climate normal and 2025 adds another layer: it distinguishes between places that are simply warm as a matter of geography and places where current conditions are unusually warm relative to what was “normal” in the recent past. This is why anomalies often get attention in climate journalism and risk communication: they speak directly to what systems were implicitly calibrated to handle.
Where warming “shows up” most clearly in a normals-vs-now view
Several broad patterns tend to appear when comparing a recent 30-year baseline to a current year:
- High-latitude amplification in anomalies: countries with cooler normals can still show large positive departures in a warm year. This does not mean they become the warmest countries; it means the year is unusually warm relative to their baseline.
- Desert and semi-arid regions dominate absolute rankings: the warmest long-run averages remain concentrated where the annual energy balance favors persistent heat and where seasonal cooling is limited.
- Tropical islands are warm but require caution at small scales: very small territories can be more sensitive to gridded-data resolution and boundary handling, so uncertainty is typically higher than for large landmass countries.
Regional breakdown: how to read the Top 100 without overgeneralizing
A useful “mental map” of the Top 100 by absolute annual temperature is to group them into three broad clusters:
- Hot arid belt (Sahel, Arabian Peninsula, Horn of Africa): these countries typically sit near the very top because the normal itself is high. The 2025 anomaly may be moderate or large, but the baseline already places them among the warmest environments on Earth.
- Humid tropics (equatorial Africa, Southeast Asia, tropical Americas): annual averages are high, but humidity and rainfall patterns can be a more important determinant of human comfort and heat stress than temperature alone.
- Warm-island regimes (many small states): annual averages are high with relatively low seasonality. For planning and travel, these are often “consistently warm” climates, but local microclimates can vary sharply with elevation and coastal exposure.
Policy takeaways
- Design for the baseline and the deviation: 1991–2020 normals remain a practical reference for “typical” conditions, but planning that ignores anomaly behavior can underestimate strain in unusually warm years.
- Separate average warmth from heat risk: annual averages are informative, but heat risk depends on extremes, humidity, nighttime temperatures, and exposure. Use this ranking as context, not as a standalone heat-risk assessment.
- Use consistent baselines for communication: when comparing countries or publishing media-friendly leaderboards, a shared baseline (like 1991–2020) makes differences interpretable, especially when discussing anomalies.
- Be explicit about coverage and revisions: “current-year” values can be provisional until the year completes and the upstream dataset finalizes. Publishing the data vintage and coverage window prevents misinterpretation.
FAQ
What is a 1991–2020 climate normal?
It is the average of a climate variable over a 30-year period (here: annual mean near-surface air temperature) used as a benchmark. Thirty years is long enough to smooth short-term noise, while still representing “recent” climate conditions.
In this article, the normal is used as the reference point for computing the 2025 anomaly (2025 minus the 1991–2020 mean).
Why can 2025 be higher or lower than the normal even without “changing climate” arguments?
Even in a stable climate, year-to-year differences happen due to natural variability in atmospheric circulation, ocean patterns, and regional weather regimes. That is exactly why normals exist: they provide a stable reference so a single year can be interpreted as a departure from “typical” conditions.
How should travelers interpret this ranking?
Use it as a background warmth guide. If a country has a very high annual average, it is likely to feel warm most of the year, but comfort depends heavily on humidity, coastal winds, altitude, and the season of travel. For travel decisions, pair annual averages with monthly climatology, heat index/humidity data, and local elevation.
Is a country average the same as the temperature people experience?
Not necessarily. A national average mixes hot and cool regions (including mountains and coastlines) and does not reflect population distribution. Large countries can have climates that vary dramatically across regions. City-level or population-weighted metrics can be more relevant for impacts.
What are the main limitations of reanalysis-based country averages?
Reanalysis products combine observations with physical models on a grid. Country aggregation requires matching administrative borders and applying area/latitude weighting. This works well for large countries, but very small landmasses can be more sensitive to grid resolution and boundary handling.
Our World in Data — “Average monthly surface temperature” (country series).
Country-level 2-meter temperature derived from a Copernicus/ECMWF reanalysis and aggregated to national borders (includes processing notes and citation guidance).
https://ourworldindata.org/grapher/average-monthly-surface-temperature
Copernicus Climate Change Service / ECMWF — ERA5 monthly means (original data).
Reanalysis dataset providing gridded near-surface temperature used as the upstream source for the country aggregation.
https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels-monthly-means?tab=overview
Our World in Data — “Monthly temperature anomalies” (baseline explanation).
Documentation of anomalies computed relative to the 1991–2020 mean for each country, consistent with modern climate normals usage.
https://ourworldindata.org/grapher/monthly-temperature-anomalies
NOAA NCEI — Climate at a Glance (global time series; baseline context).
A widely used official portal for climate monitoring and time-series context (useful as a cross-reference for baseline conventions and anomaly framing).
https://www.ncei.noaa.gov/access/monitoring/climate-at-a-glance/global/time-series
Berkeley Earth — Data portal (independent temperature datasets for cross-checking).
Independent temperature products and documentation that can be used to compare broad-scale warming patterns with alternative methodologies.
Technical note: This article treats the 1991–2020 mean as the “normal” reference and reports 2025 as a provisional annual mean (rounded), which is appropriate for comparative communication but should not be read as a final audited national statistic.
DownloadDataset & charts — “Average Annual Temperature (1991–2020 vs 2025)”
One ZIP archive containing the ranking tables (Top 100 + supporting tables) and the PNG chart images used in the article.
- Tables: CSV files + one Excel workbook (XLSX) with multiple sheets
- Charts: PNG images (ready to reuse in press or presentations)
- Notes: README + data dictionary (column definitions)
Tip: If your browser opens the file instead of downloading, use right click → Save link as…
