Reports

This summer 2022's advection setup that needed at least 4 soil heat flux plates, so I tested 10 different ones in the lab sandbox before we deployed them at BA. Joe has also been exploring using potted (waterproof) Peltier chips as inexpensive heat flux plates.

For reference I had two brand-new Hukseflux HFP01 borrowed for the calibration from the Elkhorn project. I tested three Hukseflux HFP01 that are lent (permanently) to us from Allen Goldstein's lab. Based on their serial numbers, they are probably ~15 years old but look to be in good shape. Finally, I tested 5 Peltier chips that I soldered to 10' cables. The sensors were wired as single-ended measurements to a CR6. Sensors were measured every 10sec and 1-min averages were saved.

The sensors were tested in a wooden sandbox in the lab. The inner wall of the sandbox has dimensions slightly larger than 11"x11". The outer wall of the sandbox is about 15"x15". In between the walls is foam insulation.To setup the experiment, I layered a heater at the bottom, aluminum foil, about half the sand, the sensors, and then the other half of the sand.

At the bottom of the sandbox I placed a aluminum plate glued to a silicone flexible heater, OmegaLux SRFG-1111/10-P. The heater has dimensions 11"x11" and produces a max of 10 W/m2. There was also a fine wire thermocouple attached to the aluminum plate to measure the temperature of the heater. The heater is powered with a Variac, an adjustable transformer that can control how much AC voltage is supplied. Running it at around 15-20% seems to increase the heater temperature to 60C in about a few hours.

I originally tried using turkey oven bags to hold the sand in the sandbox, but the bags were not large enough for the box. 2 bags side by side could work, but it was annoying to make the sand level equal between the 2 bags. I also burned the bags when I heated the sand at 450F (220C) for too long (6 hours).

Instead, heavy-duty aluminum foil worked much better as a heat-proof, flexible lining for the sandbox. I used 2 layers, like lining a baking pan. The second layer is turned 90deg relative to the first layer, so you have 2 overlapping slings lining the pan/sandbox.

On top of the aluminum foil, I poured in about half of the sand and leveled it. I placed the SHF plates on top of the sand. Then, holding the plates in place, I poured in the remaining sand and leveled the surface.

I did the SHF plate calibration in two rounds. In the first round, the plates were oriented as they would be in the field, with a positive flux when heat is moving from surface into the soil column (red side up for the Hukseflux plates). In the second round, the plates were oriented "upside-down," with a positive flux when heat is moving from bottom to top of soil column.

The sand was not super level for the first round. It was better for the second round when I used a small level to check.

For the data analysis, I used just the cool down data since the heating could be uneven as I adjusted the variac.

Figure 1. Time-series from first round of SHF plate calibration. Hukseflux plates are plotted on the left y-axis and Peltier chips are plotted on the right y-axis. Overall very consistent between all the sensors. It took about 12 hours for the sand to cool back down to room temperature.

Figure 2. Linear regression of Hukseflux mV data against the W/m2 of the reference Hukseflux plate. I restricted the regression to the first 908 rows to match the data count of the second round rows to the first

Figure 3.

Figure 4.

Figure 5.

Figure 6.