Title: Hydrogen Column Density Variability in a Sample of Local Compton-Thin AGN

Abstract: We present the analysis of multiepoch observations of a set of 12 variable, Compton-thin, local (z<0.1) active galactic nuclei (AGN) selected from the 100-month BAT catalog. We analyze all available X-ray data from \chandra, \xmm, and \nustar, adding up to a total of 53 individual observations. This corresponds to between 3 and 7 observations per source, probing variability timescales between a few days and $\sim 20$~yr. All sources have at least one \nustar observation, ensuring high-energy coverage, which allows us to disentangle the line-of-sight and reflection components in the X-ray spectra. For each source, we model all available spectra simultaneously, using the physical torus models \myt, \bor, and \uxc. The simultaneous fitting, along with the high-energy coverage, allows us to place tight constraints on torus parameters such as the torus covering factor, inclination angle, and average column density. We also estimate the line-of-sight column density ($N_{\rm H}$) for each individual observation. Within the 12 sources, we detect clear line-of-sight $N_{\rm H}$ variability in 5, non-variability in 5, and for 2 of them it is not possible to fully disentangle intrinsic-luminosity and $N_{\rm H}$ variability. We observe large differences between the average values of line-of-sight $N_{\rm H}$ (or $N_{\rm H}$ of the obscurer) and the average $N_{\rm H}$ of the torus (or $N_{\rm H}$ of the reflector), for each source, by a factor between $\sim2$ to $>100$. This behavior, which suggests a physical disconnect between the absorber and the reflector, is more extreme in sources that present $N_{\rm H}$ variability. $N_{\rm H}$-variable AGN also tend to present larger obscuration and broader cloud distributions than their non-variable counterparts. We observe that large changes in obscuration only occur at long timescales, and use this to place tentative lower limits on torus cloud sizes.