seeding6 min readUK & US

How soil temperature is measured and modelled

Nobody has a probe in your lawn, so the figure is modelled. How that works, at what depth, how far to trust it, and how to check it yourself.

Published 4 May 2026 ·Last reviewed 3 September 2026

Every soil temperature figure you see online for your postcode is modelled. There is no network of probes sitting in suburban lawns. Knowing what that means is the difference between using the number sensibly and trusting it further than it deserves.

What gets measured

Real soil temperature observations come from a fairly small number of weather stations, where thermometers sit at standard depths, commonly 5cm, 10cm, 20cm, 50cm and 100cm, under a defined surface that is usually bare soil or short grass.

Those are good measurements at those particular points. There are simply not many of them, and yours is not one.

How the figure for your postcode is produced

Weather models solve for the land surface as well as the atmosphere. A land surface scheme tracks the energy balance at the ground, which means incoming radiation, outgoing longwave, sensible and latent heat, and conduction downwards, and steps soil temperature and moisture forward at several depths across a grid.

That grid is coarse. A cell is typically a few kilometres across even in the better models, and the number you get for your postcode is an interpolated value from whichever cell happens to contain it.

So it carries three separate sources of error. The model has its own, because the land surface scheme is a simplification and it inherits whatever error is in the forecast driving it. Spatial averaging adds more, since one value has to represent a few square kilometres of roads, buildings, woodland, water and your lawn. Then there are the surface assumptions: the model assumes a generic land surface for the cell, while your lawn has its own aspect, shading, drainage, thatch layer and soil texture.

Which depth to use

We report 6cm, because that is germination depth. It is roughly where sown seed sits and where new roots start working, and most published germination thresholds are quoted at 5–10cm, so the figure and the threshold refer to the same thing.

Depth matters more than people expect. The surface swings wildly, and can be 15°C above air temperature in full sun and below freezing the same night, which makes it close to useless for decisions. Between 5 and 10cm you have the seeding and rooting zone, damped, lagging air temperature by a couple of days, and the right depth for any germination question. At 20cm and below the figure lags by weeks and is essentially the season’s average, so it tells you about the trend rather than about this week.

What makes your lawn differ from the model

The differences are fairly predictable, and most of them push the same way.

A south-facing slope runs warmer than the cell average in spring; a north-facing one lags, sometimes by a week or more. Under a tree or on the north side of a fence, soil warms much later than the model thinks. Wet soil warms slowly, because water has a high heat capacity and evaporation carries energy away, so a poorly drained lawn can sit several degrees behind a well-drained one a street away. A thick thatch layer insulates and slows spring warming. And a lawn ringed by paving and walls runs warmer, as does anything in a town centre.

Practically: treat the modelled figure as a good estimate for your area, and expect your shaded, badly drained back corner to be behind it.

Checking it yourself

A soil thermometer is cheap, and it settles the argument for your particular lawn.

  1. Push the probe to about 6cm (most have a depth marking).
  2. Take the reading at the same time each day. Mid-morning is a reasonable proxy for the daily mean at that depth; early afternoon reads warm.
  3. Take it in the part of the lawn you care about, then take a second in the worst corner.
  4. Compare against the modelled figure over a week or two. Once you know your lawn runs, say, 1.5°C behind, you can apply that offset from then on.

The honest summary

The modelled figure is genuinely useful, and far better than guessing from air temperature or from the calendar. It is not a reading from your lawn, and anyone presenting it as one is overselling it.

See the modelled figure and trend for your area.

Sources

  1. Weather forecast API documentation: soil variables
  2. Soil temperature measurement guidance
  3. Soil temperature and its effect on turfgrass
  4. Guide to Meteorological Instruments and Methods of Observation

Check your own lawn

One email when the app is ready, and nothing else. Unsubscribe in one click. See the privacy notice.