Ground-based observations at terahertz (THz) frequencies are a newly explorable area of astronomy in the coming decades. We discuss science cases for a first-generation 10-m class THz telescope, focusing on the Greenland Telescope as an example of such a facility. We propose science cases and provide quantitative estimates for each case. The largest advantage of ground-based THz telescopes is their higher angular resolution (~ 4ยดยด for a 10-m dish), as compared to space or airborne THz telescopes. Thus, high-resolution mapping is an important scientific argument. In particular, we can isolate zones of interest for Galactic and extragalactic star-forming regions. The THz windows are suitable for observations of high-excitation CO lines and [N II] 205-?m lines, which are scientifically relevant tracers of star formation and stellar feedback. Those lines are the brightest lines in the THz windows, so they are suitable for the initiation of ground-based THz observations. THz polarization of star-forming regions can also be explored since it traces the dust population contributing to the THz spectral peak. For survey-type observations, we focus on "sub-THz" extragalactic surveys, the uniqueness of which is detecting galaxies at redshifts z ~ 1-2, where the dust emission per comoving volume is the largest in the history of the Universe. Finally we explore possibilities of flexible time scheduling, which enables us to monitor active galactic nuclei, and to target gamma-ray burst afterglows. For these objects, THz and submillimeter wavelength ranges have not yet been explored.
