Estimating Soil Electrical Parameters in the Canadian High Arctic from Impedance Measurements of the MIST Antenna Above the Surface

TL;DR

This work tackles the challenge of accurately modeling the MIST antenna beam for global $21$-cm cosmology by deriving soil electrical properties from antenna impedance measurements at a Canadian High Arctic site. Using EM simulations with the Feko package and a $\,\chi^2$-minimization fit, the authors compare single-layer and two-layer soil models and establish a two-layer interpretation as the nominal solution, yielding a thawed top layer of approximately $t \sim 50$ cm overlying near-frozen permafrost. Key parameter estimates include $\sigma_1 \approx 0.01$--$0.02$ S m$^{-1}$, $\epsilon_{r1} \approx 16$--$18$, $\sigma_2 \lesssim 0.002$ S m$^{-1}$, and $\epsilon_{r2} \approx 5$--$7$, with a residual inductance $L$ of a few nanohenries; these values meet the precision targets for robust sky-background extraction and beam modeling. The study demonstrates that antenna impedance measurements offer a practical, autonomous route to soil characterization in polar regions and holds promise for long-term, high-cadence soil monitoring that can inform both cosmology and cryospheric geophysics.

Abstract

The MIST experiment aims to detect the cosmological 21-cm signal through sky observations at 25-125 MHz using a wide-beam antenna. The antenna is mounted above the soil and the beam characteristics are highly dependent on the soil's electrical properties. Accurate models for the beam obtained from electromagnetic simulations are crucial for detecting the 21-cm signal. Determining the soil properties to inform the beam simulations is therefore a very high priority for MIST. Here we report the first electrical characterization of the MIST observation site in the Canadian High Arctic, which was conducted in July, 2022. The electrical parameters were estimated using impedance measurements of the instrument's antenna, which is a very advantageous approach for MIST. Our best-fit soil model is consistent with a thawed active layer underlain by permafrost, and the parameters were estimated with a precision close to the requirements for the detection of the cosmological 21-cm signal.

Figures (9)

• Figure 1: (a) The MIST instrument at the study site in the Canadian High Arctic. The instrument consists of: (1) a horizontal blade dipole antenna formed by two aluminum panels of size $1.2$ m $\times$$60$ cm $\times$$3$ mm, separated horizontally by $2.15$ cm, and mounted $52$ cm above the soil; (2) a $5$-cm-long balun attached to the antenna excitation port at its input, and to the input of the receiver box at its output; (3) a $40.5$-cm-tall receiver box containing all the electronics (except for the balun) and four $12$-V batteries; (4) a support frame consisting of fiberglass and plastic parts. The antenna impedance measurements used in this analysis are calibrated at the antenna excitation port located between the two aluminum panels. (b) The simulated MIST instrument in Feko. The frame is not included in the simulations as it has a negligible contribution to the antenna impedance. The inset shows a zoomed-in view of the area around the excitation port. In Feko, the port is defined at the center of the two antenna panels and connected to them with two $1$-mm wires of length $1.075$ cm. These wires do not exist in the real instrument and their reactance is analytically removed from the simulated impedance before using the simulations to fit the measurements. The simulated soil extends to infinity in both horizontal directions and in depth, though this is not shown in the 3D view rendered by Feko.
• Figure 2: Measurements of the MIST antenna impedance at the study site in the Canadian High Arctic. In the top row, panels (a) and (b) show, respectively, the resistance and reactance of the first measurement. In the second row, panels (c) and (d) show, respectively, the variations of the resistance and reactance across our twelve-day measurement period relative to the first measurement. On the $y$-axis of panels (c) and (d), the ticks marking a given "day" correspond to local noon, and the dashed horizontal lines to local midnight. No measurements were conducted between day $4$ and the first three quarters of day $8$ due to bad weather, and the final measurement ended halfway through day $12$. Panel (e) shows the variations of the resistance and reactance relative to the first measurement at two frequencies with large variations. Panel (f) shows the temperature of the $50$-$\Omega$ load located inside the MIST receiver and used for impedance calibration, as well as air temperature and solar radiation measured by a weather station $\approx1$ km northwest of the study site. Panel (g) shows the relative humidity and wind speed measured by the weather station.
• Figure 3: Summary of the antenna impedance fits. Using the black line, panels (a) and (b) show, respectively, the resistance and reactance from the first impedance measurement. The dashed blue and dotted orange lines show, respectively, the best-fit single- and two-layer impedance models. Panels (c) and (d) show, for the resistance and reactance respectively, the fit residuals for all $85$ measurements. The blue (orange) lines show the residuals for the single-layer (two-layer) models. Panels (c) and (d) also show the total uncertainty, $\delta$, accounted for in the fits. The green dashed (dotted) lines correspond to the total uncertainty for the single-layer (two-layer) fits.
• Figure 4: Two-layer soil parameter estimates for the MIST site in the Canadian High Arctic. (a) Top-layer thickness, $t$. (b) Top-layer conductivity, $\sigma_1$, and relative permittivity, $\epsilon_{r1}$. (c) Bottom-layer conductivity, $\sigma_2$, and relative permittivity, $\epsilon_{r2}$. (d) Residual inductance, $L$. (e) Temperature of the internal $50$-$\Omega$ load, air temperature, and solar radiation. (f) Relative humidity and wind speed. In panel (a), the green dot and error bars represent the average and sample standard deviation, respectively, of the $28$ thaw depth measurements done at the study site on July 22, 2022, with a metal probe. The green dashed line and band are used to extend, for reference, the green dot and error bars across the measurement period. In panel (c), the blue downward-pointing arrows represent upper limits for $\sigma_2$ at the $68\%$ confidence level.
• Figure 5: Map of the MIST study site ($79.37980^{\circ}$ N, $90.99885^{\circ}$ W) in Expedition Fjord, Umingmat Nunaat (Axel Heiberg Island), Nunavut, Canada. This site is accessible from the McGill Arctic Research Station (MARS). (a) Satellite image of the Canadian High Arctic and Greenland, with Umingmat Nunaat (Axel Heiberg Island) identified with a pink box. (b) Umingmat Nunaat (Axel Heiberg Island), with red box identifying the MARS region. (c) The MARS region, with the MIST study site and MARS identified as yellow and blue dots, respectively. Maps created using Google Satellite data in the QGIS software.