**3. Data**

As reference seismic catalog, we use the global centroid moment tensor (CMT) catalog [22,23] of shallow seismicity (depth ≤ 50 km) from 1980 to 2019. We do not decluster the catalog, to better exploit the potentiality of our method that does not assume any temporal distribution for earthquakes. As already commented above, this allows using more data and avoiding the introduction of the biases induced by declustering on the *β* estimation [14,24]. We selected the events in the Atlantic ridge approximately in the latitude range −60◦:60◦ (see Figure 5a).

Regarding the magnitude of completeness, we use *Mw* 5.5 from 1980 and *Mw* 5.0 from 2004, the ones suggested by the authors of the catalog ([23], see Figure 6). We then carefully test this choice of completeness: as suggested by Marzocchi et al. [25], if the catalog is complete the magnitudes must follow an exponential distribution, and the exponentiality of the magnitudes can be tested through the Lilliefors [26] test. To apply this test with multiple completeness levels, we can build a vector of variables by subtracting to each magnitude the corresponding magnitude of completeness (*M* − *MC*), and test the exponentiality of this dataset [27]. The hypothesis of exponential distribution cannot be rejected at any confidence levels, as we obtain a very large *p*-value (0.50). This demonstrates the robustness of the chosen magnitudes of completeness. A further check is also performed in Figure 5b by plotting the *M* − *MC* vs. the sequential number of events: as suggested by Zhuang et al. [28], a homogeneous pattern near the Y-axis (as it is possible to see in Figure 1b) suggests the correct selection of completeness values for the catalog. The final catalog contains 1168 events.

**Figure 5.** Panel (**a**): events in the Atlantic ridge selected from the CMT catalog (blue dots) inside the red polygon; panel (**b**): difference between the magnitude of events and the relative completeness (*M* − *MC*) vs. the sequential number of events.

**Figure 6.** Time vs. magnitude plot; blue dots represent seismic events, red lines the magnitude of completeness thresholds; the solid red line represents the reference level (*Mw* 5.5 from 1980 and *Mw* 5.0 from 2004), dashed red lines represent the conservative levels (+0.1 and +0.2 on the reference level).
