How to estimate soil temperature from air temperature 1

[MartinLe]

I hope this is the right forum.
I need to calcualte the heat loss of tanks into the ground, tanks will be standing on the ground or be built up to 6m into the ground. Tank diameters are anything between 16 and 40m. We have all the k-values etc. So far, we basically guessed the soil temperature based on some experiences. Basically wee assume the ground temp in summer wo be a bit cooler than air, and in winter a bit warmer. Air temperature is most of the times easily attained from the internet. We also assume a uniform ground temperature. To our best knowledge, we were not far off so far.


I still want to get better estimates, mainly because we build plants all over the globe soon and experience from middle europe will not hold true elsewhere. Also, I'm a bit of a nerd and want to get this right.

 

[LittleInch]

I've studied this quite a bit and you can often find details down to 1m below ground from met survey stations.

Analysis by others has shown that the temperature at 1m depth more or less tracks average air temperature. This isn't always noted and often all you can get is mean minimum and mean maximum. Average is about mid point between the two for 1m depth. The soil temp lags by a moonth or two, but mostly you want min and max for design purposes. The key is getting mean or average temps, not peak max or peak min.

The deeper you get, the less annual variation there is unless the groundwater level changes between the seasons. At 6m my best estimate is that you're at average air temperature plus or minus a couple of degrees C.

There is no sibstitute for actual readings over a year period in the location but the above should give you a good start.

My motto: Learn something new every day

Also: There's usually a good reason why everyone does it that way

 

[Althalus]

In temperate zones away from bodies of water, the soil at about 6 feet down is fairly constant at 55 degrees. From there upward, there is a fairly arithmetic gradient to the surface at near ambient temperature.

Variables:
1) The surface temperature will lag behind the ambient temperature by an hour or so.
2) The surface temperature will vary with the soil type (emissivity) and moisture content.
3) The "surface" level may be considered the bottom of a very open graded layer of soil. i.e. If the top 1 ft is an open graded sand or gravel layer, it will allow a lot of circulation between the ambient air and the bottom of the gravel. Thus, the gradient will be much steeper in that layer.
4) Your latitude or underground water motion may cause the 55 degrees to vary a lot.

 

[BigInch]

I'd change that 55F to equal the average yearly temperature at the location of interest. Remember that elevation can make a considerable difference in average values. Do not compare temperatures of points with different elevations, even if the latitude and longitudes are nearly the same.

Independent events are seldomly independent.

 

[racookpe1978]

Other than that one mention of emissivity above, you have not addresses the very significant solar radiation effect on the top of the soil.

That is, top-of-soil temperature will always follow air temperature by several hours, slowing the rate of increase and slowing the decrease, moderating the rapidly changing air temeprature by the 1000x greater physical mass and thus thermal mass of the ground. Air temperatures are re-heated by the ground re-radiating energy "up" and so air temperature is about 2-3 hour "after solar time: hottest part of the day is 1:30 to 3:30 in the afternoon, but the sun is highest overhead at noon. Coolest at night is 3:00 AM to 5:00 AM.

As noted above, the 55 deg F average soil temperature 5 feet underground is valid for Euro-American mid-latitudes only. In lower Idaho for example, the 6 inch soil temperature does not get above freezing (32 F/0 C) until Mid-May. (The local farmers use that 6 inch soil temp for planning potato planting and potato harvesting times.) Consider the "average" surface (upper) soil temperatures will "average" daily "average" air temperatures over a month's long period as a good starting point. Local solar conditions (asphalt, grass, concrete, shaded, not-shaded, east-south-west exposure will change the actual hour-to-hour upper soil temperature greatly day-to-day. I don't know if that ever-changing hourly top-of-soil temperature is going to matter to your analysis though.

So it is very likely that your tank "floor" will be at one fairly constant temperature, the south wall will be much higher (exposed to the sun), the east and west wall at changing temperatures, and the north wall at a mid-point between air temperature and fluid temperature. The top (roof) will be exposed to vapor inside (low thermal mass) and air temperature outside (low thermal mass) , and so will almost immediately change as solar exposure changes: getting much higher at midday than the bottom of the tank. That op, however, will freely radiate energy at night, and so lcool much more on clear nights than the bottom of thet ank.

 

[oldestguy]

Saying the location might help. For instance in winter in northern (or southern) areas the depth of freezing depends on many things, such as water content, surface insulation, etc. Adding near-surface insulation will help to stabilize things also.

 

[MartinLe]

Thanks all for the input.

LittleInchs approach is basically what we did so far, a bit more fleshed out. I'll work with that.
Part of the reason I'm asking is that we have to deal with more and more plants outside of europe. So I'd rather go with assuming the ground a few m. down has the mean yearly temp. of air, than taking the 55°F/12°C.

As I mostly need an average to calulate heat losses , the day to day variation is no big concern - and we are talking tanks about 1000m³, filled with mostly water = lots of thermal mass, plus isolation. Still, thanks for the detailed input to racookpe1978

 

[BigInch]

Average temperature will also vary with elevation, roughly -4F / 1000 ft.

Independent events are seldomly independent.