I've extracted a portion of one of the pages on their site dealing with ground temperature variations with season, location, and depth below the surface. Very helpful material for Ground Source Heat pumps or Earth Tubes.
The material below is from this page on the Virginia Tech website.
Soil temperature varies from month to month as a function of incident solar radiation, rainfall, seasonal swings in overlying air temperature, local vegetation cover, type of soil, and depth in the earth. Due to the much higher heat capacity of soil relative to air and the thermal insulation provided by vegetation and surface soil layers, seasonal changes in soil temperature deep in the ground are much less than and lag significantly behind seasonal changes in overlying air temperature. Thus in spring, the soil naturally warms more slowly and to a lesser extent than the air, and by summer, it has become cooler than the overlying air and is a natural sink for removing heat from a building. Likewise in autumn, the soil cools more slowly and to a lesser extent than the air, and by winter it is warmer than the overlying air and a natural source for adding heat to a building.
At soil depths greater than 30 feet below the surface, the soil temperature is relatively constant, and corresponds roughly to the water temperature measured in groundwater wells 30 to 50 feet deep. This is referred to as the “mean earth temperature.” Figure 2 shows the mean earth temperature contours across the United States. In Virginia, the mean earth temperature ranges from 52ºF in the northern Shenandoah Valley and Winchester area to 62ºF in coastal Tidewater.
The amplitude of seasonal changes in soil temperature on either side of the mean earth temperature depends on the type of soil and depth below the ground surface. In Virginia the amplitude of soil temperature change at the ground surface is typically in the range of 20-25ºF, depending on the extent and type of vegetation cover. At depths greater than about 30 feet below the surface, however, the soil temperature remains relatively constant throughout the year, as shown in Figure 3, below.
Vertical closed-loop earth heat exchangers are installed in boreholes 200 to 300 feet deep, where seasonal changes in soil temperature are completely damped out. Well-based open-loop systems also extend to this depth or deeper. These ground loop configurations are thus exposed to a constant year-round temperature.
On the other hand, horizontal-loop, spiral-loop, and horizontal direct-expansion (DX) loops are installed in trenches that usually are less than 10 feet deep. For these types of ground loops, it is important to accurately know the expected seasonal changes in the surrounding soil temperature. The extra cost of installing such systems in deeper trenches may be outweighed by the gain in thermal performance, since deeper soils have less pronounced seasonal temperature changes and are thus closer to room temperature, which reduces the work load of the heat pump units.
Deeper soils not only experience less extreme seasonal variations in temperature, but the changes that do occur lag farther behind those of shallower soils. This shifts the soil temperature profile later in the year, such that it more closely matches the demand for heating and cooling. Referring to Figure 4 for example, the maximum soil temperature occurs in late August (when cooling demand is high) at a depth of 5 feet below the ground surface, but occurs in late October (after the heating season has begun) at a depth of 12 feet below the surface.
Thus a deeper ground loop installation would lower the annual operating cost for electrical energy to run the heat pumps, and over the life of a GHP system, these accumulated savings may more than offset the higher capital cost of burying the ground loop more deeply. In order to determine the optimal depth of burial, it is important to accurately know how the seasonal change in soil temperature varies with depth, which is mainly determined by the soil's thermal properties.
Heat capacity (also known as specific heat) indicates the ability of a substance to store heat energy; the greater its heat capacity, the more heat it can gain (or lose) per unit rise (or fall) in temperature. The heat capacity of dry soil is about 0.20 BTU per pound per ºF of temperature change, which is only one-fifth the heat capacity of water. Therefore, moist or saturated soils have greater heat capacities, typically in the range of 0.23 to 0.25 BTU/lb/ºF. As shown in Figure 3 above, light dry soils experience greater seasonal temperature swings at a given depth than wet soils. This is because their lower heat capacity causes their temperature to rise or fall more than wet soils for a given amount of heat energy gained in the spring or lost in the fall.
Thermal conductivity is another soil property that must be known in order to design a closed-loop or direct expansion GHP system. This indicates the rate at which heat will be transferred between the ground loop and the surrounding soil for a given temperature gradient. The thermal conductivity of the soil and rock is the critical value that determines the length of pipe required, which in turn affects the installation cost as well as the energy requirements for pumping working fluid through the ground loop.
Figure 5 indicates the thermal conductivity of different soil types. Heat transfer capability tends to increase as soil texture becomes increasingly fine, with loam mixtures having an intermediate value between sand and clay. As also shown in this figure, the thermal conductivity of any soil greatly improves if the soil is saturated with water. This effect is much greater for sandy soils than for clay or silt, since coarse soils are more porous and therefore hold more water when wet. Therefore, groundwater level is another important site factor in evaluating a potential GHP project and optimizing the depth at which horizontal and spiral ground loops should be installed.
As shown in Figure 6, soil thermal conductivity has a significant impact on the size of the earth-coupled heat exchanger. Thus in sandy soils for example (compare dry and saturated thermal conductivities of Figure 5 with tabulated values in Figure 6), the required length of the ground loop could be as low as 200 feet per system ton if the soil is saturated with water, or as high as 300 feet per ton if the soil is dry.
Soil thermal conductivity is of even greater importance to DX systems and designers might consider the deployment of a “soaker hose” for horizontal DX ground loops in dry areas or if the project site is higher than the sounding terrain.
The maps presented in the next section below enable rough estimates of soil properties for regional screening purposes, but any sort of detailed feasbility assessment or design study should engage a contractor for in-situ soil thermal conductivity testing. As shown in Figure 6, the range in ground loop lengths over the typcial range of soil thermal conductivities is 200 to 300 feet per system ton, which translates into a 30-50% difference in required land area, and a 10-20% difference in total system capital cost. In-situ conductivity testing minimizes the uncertainty in estimating this key thermal property and avoids undersizing or oversizing the ground loop.
As noted earlier, the thermal conductivity of dry soils tends to increase as their texture becomes increasingly fine. This simply is a consequence of the fact that the thermal conductivity of air is about one hundred times less than that of solid soil particles. Finer soils have more particle-to-particle contact and smaller insulating air gaps between particles than coarse soils, hence increased conductivity. The opposite is true for soils saturated with water, when the pore spaces between particles is filled with water rather than air, since the thermal conductivity of water is about two to three times greater than that of solid soil particles.
As a preliminary indication of likely soil texture at a potential GHP project site, Figure 7 (being developed at USDA) provides a soils map of Virginia that identifies general regions where various texture classes are to be found. Within a given region, however, the detailed distribution of soil textures can vary significantly from the regional norm, particularly in heavily built areas with a long history of construction activity. This map should not be used for project feasibility assessment or design, but is intended to provide rough guidance for preliminary screening.
As already noted above, the extent to which the soil is routinely saturated with water greatly influence a soil's thermal properties and the selection and design of an appropriate ground loop. Figure 8 shows the extent to which the elevation of the groundwater table can vary from month to month, and from a dry year to a wet year, at selected locations around Virginia. With this temporal variability in mind, Figure 9 (being developed at USDA) provides a map of the climatic normal pattern of state groundwater levels. As with the soil texture map, this should be used for regional guidance only, and not feasibility assessment or design.
Finally, depth to bedrock (i.e., the thickness of the soil layer) is an important factor that affects the feasibility of certain ground loop configurations. As explained on the Ground Loops page, standing column wells are only possible where bedrock is close to the surface, whereas vertical closed-loop systems require a depth to bedrock of at least 200 to 400 feet, depending on the texture and moisture content of the overlying soil. Figure 10 (being developed at USDA) provides a contour map showing the general depth of bedrock below the ground surface at the regional scale throughout Virginia. Like the other maps on this page, this is provided to aid in the use of our preliminary screening tool and should not be substituted for appropriate site surveys.