We demonstrate that surface-wave magnitudes (M-s), measured at local, regional, and teleseismic distances, can be used as a rapid and robust estimator of seismic moment magnitude (M-w). We used the Russell (2006) variable-period surface-wave magnitude formula, henceforth called M-s(VMAX), to estimate the M-s for 165 North American events with 3.2 < M-w < 6.5 at distances ranging from 48 to 5268 km. Of the 7370 event-station pairs, more than half (4051) of the measurements were at distances less than 1000 km. M-w estimated from broadband waveform modeling (Herrmann et al., 2008) were regressed against M-s(VMAX). M-w can be estimated from M-s(VMAX) using the relationship: M-w = 1.91 + 0.66* M-s x (VMAX) for 2 < M-s < 6. The observed scatter of the M-w [M-s(VMAX)] with respect to M-w[Waveform Modeling] was approximately +/- 0.2 magnitude units (m.u). The residuals between M-w[M-s(VMAX)] and M-w[Waveform Modeling] show a significant focal mechanism effect, especially when strike-slip events are compared with other mechanisms. Validation testing of this method suggests that M-s(VMAX)-predicted M-w's can be estimated within minutes after the origin of an event and are typically within +/- 0.2 m.u: of the final M-w[Waveform Modeling]. While M-w estimated from M-s(VMAX) has a slightly higher variance than waveform modeling results, it can be measured on the first short-period surface-wave observed at a local or near-regional distance seismic station after a preliminary epicentral location has been formed. Therefore, it may be used to make rapid measurements of M-w, which are needed by government agencies for early warning systems.