Kinematics of synthetically observed high-<i>z</i> rotating discs: reliability and biases of 3D fitting tools

ABSTRACT Resolved high-redshift galaxy gas kinematics is a rapidly evolving field driven by increasingly powerful instrumentation. However, the resolution and sensitivity still impose constraints on interpretation. We investigate the uncertainties inherent to high-z galaxy kinematical analysis by modelling a suite of rotating disc galaxies, generating synthetic interferometric ALMA observations, and fitting them with the 3D-kinematical tools $^{\text{3D}}$barolo, galpak$^{\text{3D}}$, and qubefit. We present the recovered 3D-fitted kinematical parameters to assess their reliability, quantify the range of values possible for individual source studies, and establish the systematic biases present for observed samples. The $V/\sigma _{\rm V}$ ratio, which indicates how dynamically cold a system is, is of particular importance and depends on the choice of 3D-fitting tool. On average, $^{\text{3D}}$barolo and qubefit slightly overestimates $V/\sigma _{\rm V}$ ($&amp;lt; 1\sigma$) and galpak$^{\text{3D}}$ underestimates it ($&amp;lt; 2\sigma$). Therefore, all three tools are reliable for kinematical studies of averages of high-redshift galaxy samples. The value range possible for individual sources is significant, however, even more so for samples of not purely rotation dominated sources. To determine whether an observed galaxy is rotation dominated enough to be fitted with a 3D-kinematical tool, $V/\sigma _{\rm V}$ can be extracted directly from the observed data cube, with some caveats. We recommend that the median offsets, value ranges, and tool-dependent biases presented in this paper are taken into account when interpreting 3D-fitted kinematics of observed high-redshift galaxies.