How we compute climate averages.
Every number on PageWeather is derived from 11 complete historical years (2015–2025) of daily observations from the Open-Meteo Archive. No forecasts, no simulations, no proprietary blends. This page explains exactly how the pipeline works, what each metric means, and where the limits are.
Eleven years, one archive
Our climate baseline is 2015–2025 — a 11-year window of complete years, deliberately chosen to reflect recent conditions rather than a 20th-century baseline. Every destination page, every “best time to visit” verdict, and every monthly average shares this same source window.
The current year is excluded from the average because it is incomplete — including it would bias the numbers toward whichever season we're currently in.
From raw archive to monthly average
- 01Source
Open-Meteo Archive API — a global hourly/daily historical weather dataset built on ECMWF ERA5 reanalysis with station-level corrections.
- 02Coordinates
Each destination has a fixed WGS84 latitude/longitude sourced from GeoNames. We query Open-Meteo at that point (not the surrounding admin polygon).
- 03Fields fetched
temperature_2m_max, temperature_2m_min, precipitation_sum, sunshine_duration — daily resolution, one row per day per destination.
- 04Aggregation
Daily rows → monthly stats per year (min/max/mean/sum) → climate average across the 11-year window. Rainy-day counts are computed at the daily level before aggregation.
- 05Storage
~380,000 daily rows in Postgres (weather_daily) + one pre-computed monthly climate row per destination-month in weather_monthly. Site pages read from the monthly table only.
What each number means
| Metric | Field | Unit |
|---|---|---|
| Average high | avg_tmax_c | °C |
| Average low | avg_tmin_c | °C |
| Rainfall | avg_precip_mm | mm |
| Rainy days | avg_rainy_days | days |
| Sunshine hours | avg_sunshine_hours | hrs/day |
Mean of daily maximum temperatures across all days in that month, across all 11 years.
avg( daily_tmax ) grouped by month, over 2015–2025
Mean of daily minimum temperatures across all days in that month, across all 11 years.
avg( daily_tmin ) grouped by month, over 2015–2025
Sum of daily precipitation across a full month, then averaged across all 11 years.
avg( sum( daily_precip ) per month-year ) over 2015–2025
Count of days in a month with precipitation ≥ 1 mm, then averaged across all 11 years.
avg( count( daily_precip ≥ 1mm ) per month-year ) over 2015–2025
Daily sunshine duration (converted from seconds to hours), averaged across the month and all 11 years.
avg( daily_sunshine_seconds / 3600 ) grouped by month, over 2015–2025
Where averages fail you
Averages are useful for planning. They are not weather forecasts. Three specific caveats you should keep in mind:
- Microclimates. Our queries hit the destination's city-center coordinates. Coastal, mountain, or valley micro-climates can deviate 5–15% from what we report. If you're going somewhere with dramatic elevation (Cusco, Denver, Kathmandu) or a strong sea-breeze pattern (San Francisco, Lima), treat the numbers as directional.
- Extremes blend into the mean. A brutal heatwave in July or a freak storm week in April becomes a small nudge in an 11-year average. Averages tell you what to pack for; they don't tell you the top or the bottom of the range.
- Climate change. A 11-year window ending in 2025 is warmer than a 20- or 30-year baseline would be — typically by +0.3 to +0.8°C depending on latitude. We consider this a feature: it reflects the climate you're actually about to travel into, not the one your parents remember.
How often the numbers change
| Full recompute | Annually — every February, once the previous calendar year is complete in the archive. |
| Content refresh | Quarterly — narrative copy on high-traffic destination and month pages is re-read and lightly re-edited. |
| Live forecast | Never. PageWeather is a climate archive, not a 7-day forecast service. For the actual weather this Saturday, check a forecast provider. |
Clear boundaries
- · 7-day or 14-day forecasts — use a forecast provider.
- · Hurricane, storm, or severe-weather alerts — use your local meteorological service.
- · Real-time observations or radar — use the Windy, Meteoblue, or your national weather agency.
- · Climate-change projections beyond 2025 — see the IPCC or Copernicus C3S.
- · Sponsored or paid rankings — every verdict on this site is data-driven.
How to cite us
If you cite PageWeather in academic or editorial work, please reference us as:
PageWeather (2026). Climate averages 2015–2025, Open-Meteo Archive. Retrieved from https://pageweather.com.