Why do well water and spring water feel cold in summer and warm in winter?
― It's the air around you that's moving, not the water
It's shockingly cold when you drink it in summer, yet feels faintly warm when you dip your hand in during winter. It seems as if the water changes with the seasons, but it's actually the opposite. The water's temperature barely changes at all. What's really changing a lot is you.
You're walking a mountain trail in midsummer and come across a spring by the path. You dip your hand in and it's cold as ice water. "Mountain water really is cold," you think.
Now come back to the same spot in the depths of winter. Everything around is white with frost, yet a faint mist rises from the spring. Put your hand in, and it feels far warmer than the outside air.
There's no device underground switching the water's temperature by season. So what's actually going on?
There are really just two reasons
Metal carries heat almost instantly, but soil and rock are orders of magnitude slower. The heat and cold of the surface can only creep downward.
By the time summer's heat has crept a few metres down, winter's cold is already chasing after it. The two cancel each other out, so at depth the temperature barely moves at all.
Together, these two effects mean underground water stays at almost the same temperature all year round. That temperature ends up very close to the location's yearly average air temperature. Across much of Japan's lowlands, that's roughly 15°C.
What you're feeling as "cold" is a difference, not the water itself
Touch 15°C water on a 33°C midsummer day, and your body reads it as "cold." Touch that same 15°C water on a 3°C winter day, and this time you feel it as "warm." The water hasn't changed a thing. What changed is the air you're comparing it to.
In other words, spring water isn't moving in step with the seasons — it's refusing to move with them. While everything around it swings wildly up and down, that one thing stays still. So it always feels like it's going the opposite way.
How deep can heat and cold actually reach?
Surface temperature swings widely over the year. That wave travels down into the ground, but it grows weaker the further it goes. Figure 1 shows this.
As a rough rule, underground, for roughly every 2 metres of depth, the temperature swing shrinks to about a third of what it was. So by a depth of 10 metres, a yearly swing of over 20°C at the surface has shrunk to less than 1°C.
On top of that, the seasons arrive "late"
It's not just weaker — underground, the heat and cold themselves arrive later. Summer's warmth doesn't reach a depth of several metres until well after summer has ended on the surface.
That's why, near the surface, you get a season lag — "autumn on the surface is actually the warmest point underground." The chill you feel inside tunnels and caves is a relative of the same effect.
The fact that underground temperature stays almost constant all year is itself a useful resource. If you can exchange heat with something that's colder than outside air in summer and warmer than outside air in winter, both cooling and heating take less energy. Systems that bury pipes underground to exchange heat use exactly this property.
Deep underground, the Earth itself holds heat. As a rough guide, temperature rises by about 3°C for every 100 metres of depth. Down to some tens of metres it stays close to the yearly average air temperature, but beyond that it starts gradually warming again.
Summary
Spring water's temperature isn't changing with the seasons. Because soil conducts heat so slowly, neither the surface's heat nor its cold reaches deep down — the temperature there stays fixed near the location's yearly average. It feels cold in summer and warm in winter because of the air around it, not because of the water.
Spring water isn't chasing the seasons.
It's simply being left behind by them.
Follow the trail of underground heat further and you reach Why are hot springs hot?. For how the same temperature can feel different depending on context, see Why does metal feel colder than wood?, and for how a gap between air and ground temperature creates water, see Why is the grass wet in the morning?.
- Dig two holes in your garden soil, one 5cm deep and one 30cm deep, and stick a cooking thermometer into each. Wait 10 minutes. (If digging anywhere other than your own land, get the owner's permission first.)
- Take readings twice: once at midday on a clear day, and once after sunset. The shallow hole should swing widely; the deep one should barely move.
- If there's a spring or well nearby, measure the water temperature twice, in different seasons. You'll find the air temperature differs by tens of degrees, while the water's difference is tiny.
Even just 30cm down, the difference is clear. If you can't dig deep, simply comparing shaded ground with sunlit ground will show you how soil stores heat.
Want to know more? ― Terms, formulas, and textbook connectionsWe've labelled which level each part belongs to, from junior-high science to university specialist subjects
- JHSCovered in junior high school science
- HSCovered in high school "Physics Basics / Earth Science Basics"
- HS+High school advanced content, or textbook sidebar material
- Univ.Not taught in high school — university specialist content (heat transfer engineering, hydrology)
- ResearchNot yet settled fact even at university — what researchers are currently investigating
JHSTerms: this phenomenon has names
- Underground temperature (ground temperature): the temperature inside the ground. Weather stations keep long-term records at various depths.
- Thermal diffusivity: a number describing how fast a temperature change spreads through a material. Soil has a small value; metal has a large one.
- Yearly average air temperature: the average air temperature for a location over a full year. Deep underground temperature is known to settle close to this value.
- Geothermal gradient: the term for how temperature rises with depth due to heat from inside the Earth.
JHSHSCheck with a formula: how deep can heat and cold reach?
As the yearly temperature wave soaks into the ground, its swing shrinks with depth. Let's find "the depth at which the swing weakens to about 30% of its original size." The thermal diffusivity of moist soil is taken to be about 0.5 square millimetres per second.
| In symbols | d = √( 2 × κ ÷ ω ), ω = 2π ÷ P, swing = A₀ × e^(−z ÷ d) |
| In words | penetration depth = the square root of ("twice the thermal diffusivity" divided by "the wave's rate of progress (one full turn ÷ one year)") |
| Where the formula comes from | Solving the heat conduction equation (rate of temperature change = thermal diffusivity × curvature of temperature) under the condition that the surface temperature rises and falls with a one-year period. At depth z, the swing shrinks to e^(−z ÷ d) times its surface value |
| Symbol | Meaning and unit |
| κ (symbol A) | Soil's thermal diffusivity. Unit: square millimetres per second |
| P (symbol B) | Period of the temperature wave, here one year. Unit: seconds |
| d | Penetration depth — the depth at which the swing shrinks to about 37% (roughly 30%) of its original size. Unit: millimetres |
| A₀, z | Swing at the surface (°C), and depth below the surface |
| Symbol A: soil's thermal diffusivity | 0.5 (unit: square millimetres per second) |
| Twice symbol A | 1.0 (unit: square millimetres per second) |
| Symbol B: length of one year | 31,500,000 (unit: seconds) |
| Twice the value representing one full turn of a circle | 6.28 (no unit) |
| Convert one year into a measure of the wave's rate of progress | 31,500,000 ÷ 6.28 ≈ 5,020,000 |
| Find the square of the penetration depth | 1.0 × 5,020,000 = 5,020,000 |
| Find the number whose square gives this value | 2240 × 2240 = 5,017,600 |
| Convert millimetres to metres | 2240 ÷ 1000 = 2.24 |
In symbols, the penetration depth is the square root of "twice symbol A divided by the wave's rate of progress." The unit is length, i.e. metres. The result, about 2.24 metres, is the depth at which the swing weakens to roughly 30% of the surface value.
| How many times that depth is 10 metres | 10 ÷ 2.24 ≈ 4.5 |
| How much the swing weakens at 4.5 times that depth | 0.011 (the estimated ratio) |
| What's left of a 12°C surface swing | 12 × 0.011 ≈ 0.13 |
In other words, at a depth of 10 metres, only about 0.13°C of the yearly swing remains. On a thermometer scale, that looks like almost no movement at all.
HSHS+Why does a "lag" also occur?
HSHeat flows from higher temperature to lower temperature, at a speed set by the temperature difference. When the surface warms, heat flows into the layer just below it; once that layer warms, heat flows further down still. Because it's passed on like a bucket relay, the deeper the layer, the later it arrives.
HS+This lag grows in proportion to depth. For every roughly 2.24 metres calculated above, the season is thought to arrive about 2 months late. At a depth of about 7 metres, the lag reaches half a year — so when the surface is in the height of summer, that depth is experiencing the depths of winter, a full reversal.
Univ.The heat conduction equation pins down a single answer
Treating the ground as a uniform semi-infinite solid, with the surface temperature oscillating on a one-year period, the internal temperature can be written as a solution of the heat conduction equation. That solution takes the form of a wave whose swing shrinks exponentially with depth while its phase lags in proportion to depth. The defining feature of this solution is that both the rate of shrinkage and the rate of lag are set by the same length scale — the roughly 2.24 metres calculated above is that scale. Real ground has layers with differing properties, plus rainfall and snow cover, so observed values deviate somewhat from this solution.
📖 For the derivation and further reading: Heat Conduction Equation (Japanese Wikipedia, "Heat Conduction")
ResearchWhat's still not fully understood
- Can underground temperature serve as a climate record? The temperature profile in deep wells is thought to preserve a "smeared-out" trace of surface temperature changes over the past few centuries. Attempts continue to reconstruct past air temperatures from this, but how much groundwater flow disturbs that record is still debated.
- How far has the ground under cities warmed? Heat from pavement, buildings, and underground equipment has been reported to raise urban underground temperatures above the surrounding area. How far this spreads, and its effect on creatures living in underground water, isn't yet well understood.
- How much underground heat can we actually use? Packing many heat-exchange systems close together underground gradually shifts the ground's own temperature year by year. How to balance this over the long term is still an active research question.
In other words, this article too describes things "as currently understood." There's no guarantee the ground near your own home behaves exactly according to this calculation.
Textbook connections (by level)
| Level | Subject / unit | Where in this article |
|---|---|---|
| JHS | Science — everyday phenomena (heat transfer) | The point about soil conducting heat slowly |
| HS | Physics Basics (heat and temperature) / Earth Science Basics | The explanation of heat flow and seasonal lag |
| HS+ | Physics (advanced treatment of heat conduction) | Calculating the penetration depth and the half-year lag |
| Univ. | Heat transfer engineering / hydrology (heat conduction equation) | How the shrinking swing and the phase lag share the same scale |
| Research | Paleoclimatology / urban subsurface environments | Attempts to read past air temperatures from underground temperature |
| ― | Everyday connections | Spring water temperature, heating and cooling using underground heat |
- Japan Meteorological Agency, "Observation of Underground Temperature (Ground Temperature)" (気象庁「地中温度（地温）の観測」), and regional underground temperature statistics
- National Astronomical Observatory of Japan (ed.), Rika Nenpyo [Chronological Scientific Tables] (『理科年表』), Maruzen Publishing — sections on air/ground temperature and thermal properties of materials
- National Institute of Advanced Industrial Science and Technology, Geological Survey of Japan, "Data on Japan's Geothermal Gradient and Crustal Heat Flow" (産業技術総合研究所 地質調査総合センター「日本の地温勾配・地殻熱流量に関する資料」)
- H. S. Carslaw and J. C. Jaeger, Conduction of Heat in Solids, Oxford University Press (solution for a semi-infinite solid with periodic surface temperature)
※This article is a general-audience science explainer. The figures given are estimates meant to aid understanding of the underlying mechanism. Well water and spring water may be unsafe to drink even when they look clear. Before drinking, check your local authority's guidance and water quality test results, and follow their instructions.