Vertical Temperature Structure in Io's Atmosphere from ALMA SO$_2$ Observations

TL;DR

This work probes Io's atmospheric vertical structure by retrieving temperature profiles from four ALMA observations of SO$_2$ lines, employing a forward radiative-transfer model that integrates LOS Doppler velocity maps and a sub-beam wind-dispersion term within a Bayesian, multi-line framework. The analysis compares one-, two-, and three-node temperature parameterizations (1T/2T/3T) using PSIS-LOO-CV for model selection, revealing robust vertical structure on Io's leading hemisphere: a quasi-isothermal lower layer around $T \\approx 124$–$137$ K up to $P \\sim 0.5$ nbar, followed by a thermospheric rise to several hundred kelvin by $P \\sim 10^{-2}$ nbar. In contrast, the trailing hemisphere is sensitivity-limited, with consistently lower SO$_2$ columns that confine emission to near the surface and yield isothermal-consistent fits across TP models. The results provide a quantitative benchmark for Io's thermal energy balance and offer vertically resolved constraints that can guide future multi-wavelength campaigns and magnetospheric heating models.

Abstract

The structure of Io's atmosphere is controlled by competing processes, from volcanic outgassing and sublimation to radiative cooling and plasma heating. Yet, the lack of an observationally-derived temperature profile has left this balance unconstrained. We used four epochs of Atacama Large Millimeter/submillimeter Array (ALMA) Band 7 (275-373 GHz) and Band 8 (385-500 GHz) SO$_2$ spectroscopy to retrieve Io's vertical atmospheric temperature profiles. To mitigate longstanding degeneracies common in atmospheric retrievals, we performed a simultaneous multi-line analysis combined with line-of-sight disk-resolved Doppler velocity maps and a forward model that included a sub-beam velocity-dispersion term. This modeling approach enabled the separation of thermal and dynamical line-shape contributions. On the leading hemisphere, we retrieved a cold, quasi-isothermal lower atmosphere ($\sim$124-137~K up to $\sim$0.5~nbar), followed by a thermospheric rise reaching hundreds of kelvins by $\sim10^{-2}$~nbar. On the trailing hemisphere, our fits yielded qualitatively similar profiles but consistently retrieved lower SO$_2$ column densities. The lower column densities confined line formation to the first few kilometers, making the trailing hemisphere spectra statistically consistent with an isothermal atmosphere. Across datasets, we retrieved fractional gas coverages of $\sim$35-50$\%$ and sub-beam velocity dispersions of $\sim$25-85$\mathrm{~m~s^{-1}}$, encoding line-of-sight velocity dispersion within a beam element in excess of the disk-resolved Doppler velocity map. Together, these retrievals deliver the first vertically resolved temperature profiles of Io's atmosphere, reveal robust vertical structure on the dayside leading hemisphere, and offer new constraints on Io's thermal energy balance.

Figures (9)

• Figure 1: Reproduction of Figure 3 from Wong2000, showing modeled vertical temperature profiles at the subsolar point for high-density ($T_\mathrm{ss} = 120$ K) and low-density ($T_\mathrm{ss} = 113$ K) atmospheres. Here, $T_\mathrm{ss}$ denotes the subsolar surface temperature, which sets the boundary condition for SO$_2$ sublimation. Panel (a) corresponds to western elongation (Io's sunlit trailing hemisphere), where plasma heating on the dayside leads to a more extended atmosphere. Panel (b) shows eastern elongation (Io's sunlit leading hemisphere), where reduced dayside plasma heating results in a cooler, more compact atmosphere.
• Figure 2: Model Comparison for the SO$_2$ transition at 430.194 GHz from M22-L. Observed spectrum (black points with error bars) overlaid with best-fit model spectra for the three tested temperature profiles: 3T (left), 2T (middle), and 1T (right). The red bar at the top right of the third column indicates the ALMA spectral resolution (FWHM). Residuals are shown in the lower row.
• Figure 3: Retrieved temperature profiles for leading hemisphere datasets: M22-L (a, top row) and J21-L (b, bottom row). Columns show the 3-node (3T), 2-node (2T), and 1-node (1T) parameterizations. The black curve is the best-fit TP; the gray band marks the 1$\sigma$ credible interval. The legend for each profile provides the best-fit SO$_2$ column density, fractional coverage, wind dispersion parameters, and, when applicable, the transition pressure (with $1\sigma$ errors). Left axes give pressure; right axes show the corresponding altitude. The dashed horizontal black line indicates the modeled surface pressure, quoted with uncertainties calculated from the upper and lower bounds of the SO$_2$ column density.
• Figure 4: Retrieved temperature profiles for trailing hemisphere datasets: M22-T (a, top row) and J21-T (b, bottom row). Columns show the 3-node (3T), 2-node (2T), and 1-node (1T) parameterizations. The black curve is the best-fit TP; the gray band marks the 1$\sigma$ credible interval. The legend for each profile provides the best-fit SO$_2$ column density, fractional coverage, wind dispersion parameters, and, when applicable, the transition pressure (with $1\sigma$ errors). Left axes give pressure; right axes show the corresponding altitude. The dashed horizontal black line indicates the modeled surface pressure, quoted with uncertainties calculated from the upper and lower bounds of the SO$_2$ column density.
• Figure 5: Retrieved temperature-altitude profiles for Io from our four ALMA datasets (M22-L, J21-L, M22-T, and J21-T) each fit with the same three-node temperature parameterization (solid lines). The left panel shows the retrieved temperature profiles as a function of altitude (km), while the right panel shows the same profiles as a function of pressure (nbar). The horizontal bars in the right panel mark the surface pressure associated with each model. For context, dashed curves show representative model profiles from Strobel1994 and Wong2000. Specifically for Strobel1994, we plot the high-density dayside solutions from their model that includes solar, plasma, and Joule heating. For Wong2000, we plot the high-density dayside solutions at western elongation (trailing hemisphere) and eastern elongation (leading hemisphere).