Agriculture climate layer

Last Spring Frost Map

Explore the average date of the last modeled spring frost during 2011–2025. This map helps compare the typical timing of the final modeled freezing day in spring.

Climate period
2011–2025 conditional frost-date average
Representation
Calendar date

What this map shows

Each value is the conditional mean date of the last valid local day below 0 °C in the first half of the agricultural year. Only years in which a spring frost occurred contribute a date to that mean.

It is not the guaranteed final frost date for a future year and not a probability threshold such as a 10% or 50% frost-risk date.

How it is calculated

MetriClima converts hourly ERA5-Land temperature into local solar days, identifies the last day with a minimum below 0 °C in the first agricultural half, and averages those dates over valid years in which a spring frost occurred.

Dates are stored on a normalized 365-day calendar. Northern agricultural years start January 1; southern agricultural years start July 1 so spring remains in the first half.

Data source

MetriClima processes ERA5-Land surface data on the 0.1° latitude–longitude download grid. The physical spacing of this geographic grid varies with latitude. ERA5-Land is a land-surface reanalysis dataset providing consistent hourly climate variables.

This layer uses local-day minimum 2 m air temperature with daily and agricultural-year completeness rules. Years without a spring frost are tracked but do not contribute a calendar date to the conditional mean.

Provider
Copernicus Climate Change Service (C3S), produced by ECMWF

Climate period and representation

2011–2025 conditional frost-date average

The map averages frost dates only across valid years in which a spring frost occurred. Frost-free years are not assigned an artificial date.

Calendar date

The map displays average dates in 15-day bands on a normalized 365-day agricultural calendar.

How to interpret the map

A later mapped date indicates a later average final spring frost among years in which spring frost occurred. Because the date is an average, frost occurred both earlier and later in individual years.

Risk-sensitive planting decisions need year-to-year distributions, forecasts, and local observations; this layer supplies climatological context rather than a safe planting guarantee.

Applications

  • Comparing spring frost timing between locations
  • Providing climatic context for planting, budbreak, flowering, and transplant schedules
  • Providing frost-timing context for orchards, vineyards, and gardens
  • Understanding one boundary of the frost-free period

Limitations

  • The current product is a conditional mean, not a percentile-based frost-risk date.
  • The conditional mean does not show how frequently spring frost occurred across the source years.
  • Plant tissue can freeze even when gridded 2 m air temperature remains above 0 °C, especially under radiative cooling.
  • Topography and cold-air pooling create strong local variation that a 0.1° grid may smooth.

Explore this map by country

Open Last Spring Frost for a country to start the same interactive map focused on that geography.

Africa (54)
Asia (50)
Australia (14)
Europe (47)
North America (24)
South America (12)