Simple models of subduction zone thrust earthquakes based on a single dip-slip dislocation embedded in an elastic half space produce a large surface uplift in near field, and a zone of small amplitude subsidence that slowly tapers to zero in the far field (Figure 1a, primary slip patch, e.g., Savage, 1983). Vertical displacements measured after most subduction earthquakes follow a similar pattern. However, some far field geodetic measurements of megathrusts earthquakes (M w > 8) detect a secondary zone of coseismic uplift (referred to as SZU in the text) a few hundred kilometers landward of the trench (for a summary, see van Dinther et al., 2019). In the years following the 1960 M w 9.5 Valdivia and 1964 M w 9.2 Alaska earthquakes (e.g., Kanamori, 1970;Plafker & Savage, 1970), uplifts of more than 1 m and 30 cm in amplitude, respectively, were measured in this secondary zone. After the 2010 M w 8.8 Maule and 2011 M w 9.0 Tohoku earthquakes, a few centimeters of secondary uplift were recorded in some data sets in the days to weeks following the mainshock (e.g., Ozawa et al., 2011;Vigny et al., 2011; Figure 1c). In Japan, unlike displacements measured in Chile, the region where the SZU occurred started to subside in the weeks to months following the mainshock (e.g., Fukuda & Johnson, 2021;Y. Hu et al., 2014). Whether the SZU is coseismic or very rapid postseismic is unknown at this time.The origin and consistency of the SZU remains ambiguous. None of the published coseismic slip models of the 2010 Maule event reproduce simultaneously the horizontal deformation, the near-field vertical displacements and the SZU (Figure 1c shows a selection of published slip models). Similarly, none of the published coseismic slip models for the 2011 Tohoku earthquake explain the observed SZU (e.g., Lay, 2017), whose amplitude is less than a twentieth of the near-field vertical displacement. Note that, for these two events, >1-year-postseismic SZU can be modeled with afterslip or viscoelastic processes (e.g.,