*2.3. Resolution*

Several methods were used to determine whether the inversion results are reliable for interpretation or not. Firstly, the Checkerboard Resolution Test (CRT) was used to test the resolution of the tomography model by making a synthetic velocity model. The CRT results showed a pattern of high and low-velocity anomalies relative to the 1D initial model, which is ±10% relative to the 1D initial model. Secondly, the Derivative Weight Sum (DWS) calculated the density of ray paths that pass through the grid nodes by giving them a certain weight. A high DWS value represents a high-resolution zone (Figure S6). Thirdly, the Ray Hit Count (RHC) was used to locate the number of ray paths that pass through the grid nodes. An increased number of ray paths represents high resolution zones that can be well-interpreted (Figure S7). Fourthly, the Diagonal Resolution Element (DRE) was used, as shown in Figure S8, which relates to the resolution matrix R (damped least-squares problems); these have an R-value between 0 and 1 [39,43,44].

The focus of this research on the vertical cross-section for the resolution test and tomography inversion consisted of four areas in the north–south cross-section: profile A-A (longitude 123.80◦) and profile B-B (longitude 127.58◦), and the west–east cross-section, profile C-C (latitude 3.26◦) and profile D-D (latitude 1.10◦). The horizontal grid size used in this study was 60 km × 60 km; the vertical grid size ranged from 30 km to a depth of 330 km, as shown in Figure 4.

**Figure 4.** The horizontal and vertical seismic map of the research location in the Molucca Collision Zone. The bold black lines A-A (south–north), B-B (south–north), C-C (west–east), D-D (west–east) denote the areas of the vertical cross-section tomography. The black plus sign represents the distribution of the grid nodes for tomographic inversion; the colored dots indicate the distribution of earthquakes; the blue triangles are BMKG stations, and the black triangles signify volcanoes.
