2025MNRAS.540.3827A

We present a census of the Compton-thick (CT) active galactic nucleus (AGN) population and the column density ($N_{\rm {H}}$) distribution of AGN in our cosmic backyard using a mid-infrared selected AGN sample within 15 Mpc. The column densities are measured from broad-band X-ray spectral analysis, mainly using data from Chandra and NuSTAR. Our sample probes AGN with intrinsic 2-10 keV luminosities of $L_{\rm 2-10, int} = 10^{37}$-$10^{43}$ erg s$^{-1}$, reaching a parameter space inaccessible to more distant samples. We directly measure a 32$^{+30}_{-18}\, \rm per\, cent$ CT AGN fraction and obtain an $N_{\rm {H}}$ distribution that agrees with that inferred by the Swift-BAT survey. Restricting the sample to the largely unexplored domain of low-luminosity AGN with $L_{2-10,\mathrm{int}} \le 10^{42} \, \mathrm{erg} \, \mathrm{s}^{-1}$, we found a CT fraction of $19^{+30}_{-14}\, \rm per\, cent$, consistent with those observed at higher luminosities. Comparing the host-galaxy properties between the two samples, we find consistent star formation rates, though the majority of our galaxy have lower stellar masses (by $\approx 0.3$ dex). In contrast, the two samples have very different black hole mass ($M_{\rm BH}$) distributions, with our sample having $\approx$1.5 dex lower mean mass ($M_{\rm BH}$ $\sim$ 10$^{6} \, \rm M_\odot$). Additionally, our sample contains a significantly higher number of LINERs and H II-type nuclei. The Eddington ratio range probed by our sample, however, is the same as Swift-BAT, although the latter dominates at higher accretion rates, and our sample is more evenly distributed. The majority of our sample with $\lambda _{\rm Edd} \ge$ 10$^{-3}$ tend to be CT, while those with $\lambda _{\rm Edd} <$ 10$^{-3}$ are mostly unobscured or mildly obscured.