Soil management practices can have profound effects on the spatial and temporal variation of soil physical properties. Characterizing the soil thermal regime often involves measurement of heat transfer through the surface soil layers. For the last ~30 yrs, flux plates have been the method of choice for measurement of soil heat flux (G). Soil heat flux plates are small metal and/or plastic disks with an imbedded thermopile to measure the temperature difference across the sensor body. Using a laboratory calibration, this temperature difference is used to calculate a heat flux density. Soil heat flux plates have fixed thermal properties and are impermeable to liquid water and vapor flow. However, the surrounding soil has thermal properties that may change rapidly in response to changing water content. Recent investigations have demonstrated that the standard soil heat flux plate technique likely leads to systematic errors in G measurement. The objective of this study was to evaluate a new, perforated soil heat flux plate that was designed to reduce disturbance of liquid water and water vapor flow in the adjacent soil.

The new CAPTEC flux plate is thin (0.3 mm-thick) with a large face area (103 x 105 mm L x W) and 100 5 mm-square openings representing 23.1% of the plate face area. Laboratory and field tests compared the performance of three CAPTEC plates with a pair of standard REBS soil heat flux plates (HFT-3.1), which were round (38.6 mm-diameter), 3.9 mm-thick, and had a thermal conductivity of 1.22 W m^-1 K^-1. Laboratory measurements were completed with the plates embedded in dry and saturated sand inside an insulated cavity under steady-state, one-dimensional heat flux densities of 21, 43, 85, and 172 W m^-2. Each flux density was maintained for 2-4 days with sensor signals logged every 1 min. Comparisons were made using 24 hrs of hourly-average data for each flux plate. Field measurements were made over 10 wks in the summer/fall of 2005 in a Clarion loam soil (Typic Hapludoll) near Ames, Iowa USA. The gradient method was used to obtain an independent measurement of G at the flux plate depth (6 cm) during the field experiment. Three-needle heat dissipation probes were used to measure the soil thermal conductivity and temperature gradient necessary to calculate G using Fourier's Law.

In the dry sand, data from the CAPTEC and REBS plates were both below the known G but only by an average of 3 and 5%, respectively, over all flux densities. In the wet sand, G values from both plates were again lower than the known G but now by 30 and 9%. It is uncertain why the CAPTEC plates produced such low G values in the saturated sand. This experiment did, however, produce very small temperature gradients (0.15-0.6 °C cm^-1) in a media with a high thermal conductivity (2.2 W m^-1 K^-1) compared to the dry sand (0.65-4.7 °C cm^-1 and 0.35 W m^-11 K^-1). The CAPTEC plates performed very well under field conditions, providing G values within ~ 10 W m^-2 of both the REBS plates and the gradient method over a range of G from –100 to +150. There was less variation among the CAPTEC plates than between the REBS plates, likely due to the much greater sensing area of these plates. At the end of the experiment, soil samples collected just above and below each plate indicated slightly greater soil water content beneath the REBS plates. Further data analysis is in progress to investigate plate performance during and immediately after precipitation events as the wetting front passes the plate and after prolonged drying conditions.

The new CAPTEC soil heat plate has features including a perforated design and large sensing area that may prove very beneficial to the accurate measurement of G. Results from preliminary laboratory and especially field experiments were very promising as they indicated that the plate is accurate and durable. However, further research is needed to explain the poor performance in saturated sand and to characterize the liquid water flow patterns through and around the plate.