Scylla VI: Parsec-Scale Dust Extinction Maps in the SMC and LMC

Abstract

We present a novel methodology for mapping dust extinction in nearby galaxies at parsec-scale resolution. We apply it to HST 68 fields within the Small and Large Magellanic Clouds (23 fields in the SMC and 45 fields in the LMC) using multi-band HST photometry from the Scylla and METAL surveys. Our technique leverages \textit{kriging}, a geostatistical interpolation method built on the principles of Gaussian Process regression, combined with Gaussian mixture modeling to statistically isolate background stellar sources and account for line-of-sight depth effects. 3D dust simulations demonstrate the method's capability to recover column densities to an accuracy of $A_V \approx 0.1$ mag in fields with at least 1000 sources. The resulting $4^{\prime\prime}$ resolution ($\sim1$-pc) dust maps reveal detailed structure and strong spatial correlation with ancillary ISM tracers, especially in star-forming regions like 30 Doradus. Global extinction of total column densities follows log-normal profiles in both galaxies, with the SMC exhibiting slightly higher mean extinction ($e^{\mu}=0.47$ mag) than the broader LMC ($e^{\mu}=0.43$ mag), likely due to significant line-of-sight depths. We find systematic offsets between dust mass surface densities ($\Sigma_{D}$) derived from extinction versus FIR emission in both galaxies, with $\Sigma_{D, FIR}/\Sigma_{D, A_V}$ ratios ranging from $0.6-1.8$. This work provides the highest-resolution dust extinction maps in SMC and LMC to date, which offer a vital independent benchmark for constraining dust emissivity, $\text{CO}$-dark gas fractions, and the multi-scale structure of the ISM in low-metallicity environments.