A multi-index-based deep-water pre-arc basin stage discrimination method and device

Abstract

The application discloses a multi-index-based deep-water pre-arc basin stage discrimination method and device. The method takes high-resolution digital water depth model and multi-channel seismic profile as data basis, faces the tectonic-geomorphic response under the control of sea mountain subduction, and constructs a quantifiable, traceable and rule-driven discrimination system. The method comprises the following steps: multi-source data processing and target identification; morphological screening; tectonic correlation screening; four-stage evolution mode construction and discrimination; stage mapping and visualization, and finally realizes rapid, interpretable discrimination and visualization of the pre-arc basin evolution stage. The constructed mode can complete quantitative discrimination and spatial mapping under the condition of only relying on water depth and seismic. The application can directly serve the application scenarios of seismic rupture segmentation research and judgment, slope instability and landslide source area identification, pre-arc exploration target optimization, and multi-subduction zone rapid mapping and engineering decision, and has important theoretical significance and broad application prospect.

Description

Technical Field

[0001] This invention relates to the technical fields of seafloor topography and geomorphology mapping, marine geology, marine geophysics, seafloor tectonic geomorphology and marine tectonic analysis, and in particular to a method and device for identifying deep-water forearc basin stages based on multiple indicators. Background Technology

[0002] The deep-water forearc basin and seamount subduction erosion processes are highly coupled, and their geometric and tectonic properties directly influence the rupture mode and stress release zone distribution of strong earthquakes in the subduction zone. In practice, the identification of forearc phases often relies on empirical interpretation using single or a few geometric parameters (such as local water depth anomalies or reflection characteristics on a single seismic profile). However, in erosion-prone margins with scarce sediments and intense tectonic alteration (such as the Mariana Trench), there is significant morphological confusion between "deep-water concave angle—deep-water depression—forearc basin": relying solely on depth or slope thresholds can easily lead to false alarms that are "similar in appearance but not in reality," making it difficult to objectively distinguish different developmental stages, and even more difficult to standardize, map, and verify the identification results. This makes it impossible to stably quantify forearc geometric parameters and weakens the quantitative connection with earthquake hazard assessment.

[0003] Existing observations, physical simulations, and numerical simulations collectively demonstrate that seamount subduction is a crucial driver of the formation and evolution of deep-water forearc basins. Subduction and collision with the wedge can lead to damage to the fore-edge wedge, steepening of the slope, and large-scale landslides. This results in groove-like depressions or deep-water craters along the rear edge of the seamount, generating a radial fault network (Dominguez et al., 1998), representing an early morphological response in forearc depressions. During the subduction advance phase, the overlying plate undergoes basement erosion and thinning, and the forearc enters a subsidence-deepening process. The basin floor depth approaches that of a deep-sea plain, and the sedimentary center often exhibits a tendency to migrate inland (Noda, 2016). In sediment-scarce erosion margins, seaward gaps are not sealed by accretionary wedges but are often effectively blocked by outer ridges, promoting the transformation of deep-water craters into closed deep-water depressions and maintaining containment space (Fryer, 2012). Global compilations of seamount chains reveal that moving seamounts, over long periods, regulate subduction erosion rates and forearc reshaping, etching marker units such as scars, deep-water craters, and outer rim uplifts onto the overlying plate. Once a seamount has completely subducted, the system typically enters a phase of rapid subsidence and sedimentary infilling (Scholl and von Huene, 2007). Simultaneously, forearc geometry and tectonic properties are correlated with seismic rupture segmentation; strong earthquakes and high slip often correspond to low-coupling zones beneath the forearc basin (Wang and Bilek, 2011), highlighting the necessity of accurately identifying different evolutionary stages of forearc depressions.

[0004] However, current models for forearc basin evolution struggle to cover diverse convergent margins, particularly in erosive margins characterized by sediment scarcity and significant subduction erosion. Existing identification methods often rely on single geometric or local reflection facies features, resulting in inconsistent thresholds and sensitivity to resolution and noise, hindering cross-regional reproduction and mapping verification. Different stages often exhibit morphological similarities but different origins, lacking tectonic constraints such as azimuth-distance-direction with subducting seamounts, easily leading to misjudgments of origin mechanisms. There is an urgent need for a standardized process relying solely on high-resolution DBM and multi-channel seismic data, employing a dual-gate approach of multi-index hard thresholds and tectonic correlations to output visualized stage mapping and auditable criteria. This would reduce false alarms, improve cross-regional comparability and operational timeliness, and support rapid risk assessment of different stages in forearc basins and erosive margins. Summary of the Invention

[0005] This invention discloses a multi-index-based method and apparatus for identifying deep-water forearc basin stages, aiming to address the difficulty in objectively identifying and standardizing the mapping of forearc basin evolution processes and different stages in existing studies. By unifying coordinates and quality control, employing an auditable rule system, and integrating multi-source water depth and seismic information, target identification, feature selection, and tectonic constraints are completed, ultimately achieving rapid, interpretable identification and standardized mapping output of forearc basin evolution stages.

[0006] This invention is achieved through the following technical solution: A multi-index-based method for determining the stage of a deep-water forearc basin includes the following steps: 1) Multi-source data processing and target recognition Multi-source depth data are fused to generate a high-resolution digital depth model (DBM), and multi-channel seismic profiles are stacked, time-migrated, and corrected for two-way travel time (TWTT). Based on the DBM, candidate deep-water depressions and deep-water concave corners are extracted by overlay analysis of isobath closed loop identification and terrain rendering map; wherein, deep-water depressions are closed loop terrain and deep-water concave corners are open loop terrain. 2) Morphological screening: For the candidate regions, determine whether they simultaneously meet the following four conditions: (a) Ellipticity Φ ≥ 0.7, where Φ = 4πM / z², M is the area of ​​the region, and z is the perimeter; (b) Equivalent diameter Dr ∈ [15, 100] km, where Dr = 2√(M / π) (based on the area equivalent circle diameter); (c) Boundary slope S > 5°, where S = arctan(ΔV / ΔH), ΔV is the vertical elevation difference, and ΔH is the horizontal distance; (d) The difference between the bottom water depth B_d and the average water depth B_c of the adjacent deep-sea plain, |B_d - B_c| ≤ 500m; 3) Constructing association filtering: Determining whether candidate regions have a spatial association with subducting seamounts, including: (a) The Euclidean distance between the candidate region and the projection center of the nearest subducting seamount is ≤ 30 km; (b) The candidate region is located within the projection zone of the seamount subduction direction, which is calculated by the MORVEL global plate motion model; The region is marked as a candidate region for "forearc basin inception" only if both steps 2 and 3 are met simultaneously; 4) Construction and discrimination of four-stage evolution model: Based on the topographic, seismic and tectonic features of the candidate areas, the following discrimination rules are used to determine their evolution stages: Compression stage: if (i) slope S > 8°, (ii) terrain roughness RMS > 25 m, (iii) seismic profile shows that the leading edge wedge reflection structure is broken and the distance from the seamount is ≤20 km; Uplift stage: if (i) the fracture density E_fault = N_fault / M > 0.3 fractures / km², (ii) there is uplifted topography and the elevation change > 100 m, (iii) the seismic profile shows an extensional-strike-slip fault network; Settlement stage: If (i) 250 m ≤ |B_d - B_c| ≤ 500 m and B_d ≥ 5200 m, (ii) the seismic profile shows a tensile fault at the basement and no sediment cover; Depositional stage: if (i) the seaward gap is closed, (ii) there is a serpentinized mud volcano (which appears as a chaotic reflection of submarine multiple waves in the seismic profile), and (iii) the sedimentary layer thickness is ≥ 0.2 s TWTT; 5) Stage Mapping and Visualization: The discrimination result F_stage ∈ {Ⅰ, Ⅱ, Ⅲ, Ⅳ} for each candidate region is mapped to color encoding and output as a GIS vector layer. F Ⅰ Corresponding to the first stage (red). F Ⅱ Corresponding to the second stage (green), F Ⅲ Corresponding to the third stage (yellow), F Ⅳ This corresponds to the fourth stage (blue).

[0007] Step 1) describes the identification of closed-loop isobaths, which includes: locating closed-loop isobaths around a common center and with progressively increasing water depth on the isobath map generated by DBM; if the difference in water depth between the deepest isobath within the loop and the adjacent outer isobath is ≤ 50 m, and there is no interruption in the loop connecting to the outer isobaths, then it is determined to be a valid closed loop.

[0008] The terrain rendering analysis described in step 1) includes: overlaying the shadow / highlight rendering map with the contour map. If the light-dark transition zone is continuous and unbroken at the ring boundary, it is determined to be a deep water depression; if there is a unilateral bright band or shadow interruption, it is determined to be a deep water depression.

[0009] The subduction direction mentioned in step 3) is determined as follows: based on the MORVEL global plate tectonics model, the relative motion vector of the seamount with respect to the forearc basin is calculated, and a projection zone is formed by extending 30 km along this direction.

[0010] The seismic profile features used in each stage of the discrimination in step 4) include: leading edge wedge fracture, basement tension fault, mud volcano multiple wave chaotic reflectors, all of which are automatically identified through seismic amplitude anomalies and waveform characteristics.

[0011] The computer system execution process includes: using Python scripts to call the GDAL, NumPy, and Scikit-image libraries to automatically complete DBM preprocessing, contour line extraction, ellipticity calculation, slope analysis and stage discrimination, and output a stage spatial distribution map in GeoJSON format.

[0012] An apparatus for carrying out the method, comprising: The data receiving unit is used to acquire and store multi-source water depth data and multi-channel seismic profiles; The feature extraction unit is used to perform isobath loop identification, terrain rendering analysis, and ellipticity / slope calculation. Construct correlation cells to calculate the Euclidean distance and subduction direction projection between the candidate region and the seamount; A stage discrimination unit is used to automatically classify according to the four-stage rules; Visualization output unit, used to generate GIS-compatible stage color layers.

[0013] A computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, enable a computer to implement the deep-water forearc basin stage discrimination method.

[0014] The program instructions are Python scripts that call the GDAL, NumPy, and Scikit-image libraries to implement all the steps.

[0015] The beneficial effects of this invention are: This invention proposes and implements a multi-index-based method for identifying the developmental stages of deep-water forearc basins, and provides a complete workflow from model construction and stage identification to result visualization. Compared to traditional methods that rely on single geometric or local reflection facies features, this method can quickly, objectively, and reproducibly identify different developmental stages of forearc basins using only high-resolution DBM and multichannel seismic data. It outputs standardized stage distribution layers and profile annotations, significantly reducing false alarms and missed identifications, and supporting cross-regional applications and expert review. Its greatest advantage lies in the use of a multi-index hard threshold + a construction of an auditable rule system with a "double gate" for correlation, organically combining morphological features with the azimuth and distance constraints of "seamount-target," achieving near real-time identification and mapping with a low data threshold. Therefore, this invention provides a unified, quantitative, and transferable stage identification technology path for erosion-type convergence edges, enhances the traceability and comparability from data to conclusions, and can directly serve application scenarios such as seismic rupture segmentation analysis, slope instability and landslide source area identification, fore-arc exploration target optimization, and rapid mapping and engineering decision-making of multiple subduction zones. It has important theoretical significance and broad application prospects. Attached Figure Description

[0016] Figure 1 This is a flowchart illustrating the principle of the multi-index deep-water forearc basin stage discrimination method of the present invention.

[0017] Figure 2(a) is a schematic diagram of the compression stage in the forearc basin evolution model.

[0018] Figure 2(b) is a schematic diagram of the uplift stage in the forearc basin evolution model.

[0019] Figure 2(c) is a schematic diagram of the subsidence stage in the forearc basin evolution model.

[0020] Figure 2(d) is a schematic diagram of the sedimentary stages in the forearc basin evolution model.

[0021] Figure 3 This is a diagram showing the closure of the equal-depth coil and the delineation of the candidate region in an embodiment of the present invention.

[0022] Figure 4 This is a terrain rendering map in an embodiment of the present invention.

[0023] Figure 5 This is a typical terrain profile diagram in an embodiment of the present invention.

[0024] Figure 6 This is a construction association diagram in an embodiment of the present invention.

[0025] Figure 7 These are multichannel seismic profiles in embodiments of the present invention.

[0026] Figure 8 This is the final stage mapping result diagram in the embodiment of the present invention. Detailed Implementation

[0027] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0028] A multi-index-based method for determining the stage of a deep-water forearc basin includes the following steps: Please see the flowchart of the discrimination method of this invention. Figure 1 .

[0029] 1) Multi-source data processing and target recognition Multi-source depth data are fused to generate a high-resolution digital depth model (DBM), and multi-channel seismic profiles are stacked, time-migrated, and corrected for two-way travel time (TWTT). Based on the DBM, candidate deep-water depressions and deep-water concave corners are extracted by overlay analysis of isobath closed loop identification and terrain rendering map; wherein, deep-water depressions are closed loop terrain and deep-water concave corners are open loop terrain. 2) Morphological screening For the candidate region, determine whether it simultaneously meets the following four conditions: (a) Ellipticity Φ ≥ 0.7, where Φ = 4πM / z², M is the area of ​​the region, and z is the perimeter; (b) Equivalent diameter Dr ∈ [15, 100] km, where Dr = 2√(M / π) (based on the area equivalent circle diameter); (c) Boundary slope S > 5°, where S = arctan(ΔV / ΔH), ΔV is the vertical elevation difference, and ΔH is the horizontal distance; (d) The difference between the bottom water depth B_d and the average water depth B_c of the adjacent deep-sea plain, |B_d - B_c| ≤ 500m; 3) Constructing association filtering Determining whether a candidate region has a spatial relationship with a subducting seamount includes: (a) The Euclidean distance between the candidate region and the projection center of the nearest subducting seamount is ≤ 30 km; (b) The candidate region is located within the projection zone of the seamount subduction direction, which is calculated by the MORVEL global plate motion model; The region is marked as a candidate region for "forearc basin inception" only if both steps 2 and 3 are met simultaneously; 4) Construction and identification of four-stage evolution model Based on the topographic, seismic, and tectonic features of the candidate areas, the following discrimination rules are used to determine their evolution stages: Compression stage (see Figure 2(a)): if (i) slope S > 8°, (ii) topographic roughness RMS > 25 m, (iii) seismic profile shows that the leading edge wedge reflection structure is broken and the distance from the seamount is ≤20 km; Uplift stage (see Figure 2(b)): if (i) the fracture density E_fault = N_fault / M > 0.3 fractures / km², (ii) there is uplifted topography and the elevation change > 100 m, (iii) the seismic profile shows an extensional-strike-slip fault network; Settlement stage (see Figure 2(c)): If (i) 250 m ≤ |B_d - B_c| ≤ 500 m and B_d ≥ 5200 m, (ii) the seismic profile shows a tensile fault at the basement and no sediment cover; Depositional stage (see Figure 2(d)): if (i) the seaward gap is closed, (ii) there is a serpentinized mud volcano (which appears as a chaotic reflection of submarine multiples in the seismic profile), and (iii) the sedimentary layer thickness is ≥ 0.2 s TWTT; 5) Stage Mapping and Visualization: The discrimination result F_stage ∈ {Ⅰ, Ⅱ, Ⅲ, Ⅳ} for each candidate region is mapped to color encoding and output as a GIS vector layer. F Ⅰ Corresponding to the first stage (red). F Ⅱ Corresponding to the second stage (green), F Ⅲ Corresponding to the third stage (yellow), F Ⅳ This corresponds to the fourth stage (blue).

[0030] Step 1) includes the identification of closed loops along contour lines: On the contour map generated by DBM, locate a closed loop of contour lines that surrounds a common center and gradually increases in depth. If the difference in depth between the deepest contour line in the loop and the adjacent outer contour line is ≤ 50 m, and there is no interruption in the loop connecting to the outer contour lines, then it is determined to be a valid closed loop.

[0031] The terrain rendering analysis described in step 1) includes: Overlay the shadow / highlight rendering map with the contour map. If the light-dark transition zone is continuous and unbroken at the ring boundary, it is determined to be a deep water depression; if there is a unilateral bright band or shadow interruption, it is determined to be a deep water concave corner.

[0032] The dive direction mentioned in step 3) is determined in the following way: Based on the MORVEL global plate tectonics model, the relative motion vector of the seamount with respect to the forearc basin is calculated, and a projection zone is formed by extending 30 km along this direction.

[0033] The seismic profile features used in each stage of the discrimination in step 4) include: Leading edge wedge fracture, basement tension fault, and mud volcano multiple wave chaotic reflectors are all automatically identified through seismic amplitude anomalies and waveform characteristics.

[0034] The computer system execution process includes: using Python scripts to call the GDAL, NumPy, and Scikit-image libraries to automatically complete DBM preprocessing, contour line extraction, ellipticity calculation, slope analysis and stage discrimination, and output a stage spatial distribution map in GeoJSON format.

[0035] An apparatus for carrying out the method, comprising: The data receiving unit is used to acquire and store multi-source water depth data and multi-channel seismic profiles; The feature extraction unit is used to perform isobath loop identification, terrain rendering analysis, and ellipticity / slope calculation. Construct correlation cells to calculate the Euclidean distance and subduction direction projection between the candidate region and the seamount; A stage discrimination unit is used to automatically classify according to the four-stage rules; Visualization output unit, used to generate GIS-compatible stage color layers.

[0036] A computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, enable a computer to implement the deep-water forearc basin stage discrimination method.

[0037] The computer-readable storage medium is described above, and the program instructions are Python scripts that call the GDAL, NumPy, and Scikit-image libraries to implement all steps. Example

[0038] To verify the effectiveness and correctness of "a multi-index-based method for determining the stage of deep-water forearc basins," an experiment was conducted in the forearc region of the central Mariana Trench as a demonstration area. The specific procedure is as follows: 1) Multi-source data processing and target identification: Based on high-resolution DBM and isobaths, closed loop identification and extraction of potential deep-water depressions and deep-water concave corner areas (isobaths spaced 20 m, see...) Figure 3 And the light-dark transition zone in the terrain rendering map confirmed that the deepwater depression's shape is close to a closed ring, while the deepwater depression's corners exhibit a curved, open shape (see...). Figure 4 ); 2) Morphological selection: Morphologically, the deepwater depression has a larger overall volume, while the deepwater concave angle is smaller. To study its seafloor topography, four typical topographic profiles that traverse both the deepwater depression and the deepwater concave angle were selected (locations are shown in...). Figure 4 ).

[0039] The deep-water depression is approximately crescent-shaped (ellipticity Φ approximately 0.87), concave towards the ocean side, with an equivalent diameter Dr of approximately 90 km. Its bottom contains multiple topographic highs and lows enclosed by equal-depth zones. The water depth is approximately 5500 m, differing from the average depth of the adjacent deep-sea plain (5700 m) by about 200 m (see...). Figure 5 The area is bounded by the Peacock Seamount to the northeast and extends to the continental slope of the Mariana Trench to the southeast. The western side is gentler, while the eastern side is bounded by a westward-curving chain of seamounts with an average slope of about 6-10°. This chain of seamounts includes several interconnected large seamounts (all rising about 2 km above the surrounding seabed), one of which is named Inazoma Seamount.

[0040] The deep-water concave angle (ellipticity Φ = 0.75) has a relatively small overall shape, approximately elliptical, with an equivalent diameter Dr of about 22 km. The bottom depth is approximately 5400 m, differing from the average depth of the adjacent deep-sea plain (5700 m) by about 300 m (see...). Figure 5 It is bordered by slopes or cliffs of approximately 6° to 10° on all sides, extending eastward to the inner slope of the Mariana Trench, bounded by thrust fault structures (see...). Figure 6 The western side is surrounded by an arc-shaped seamount that rises about 2,000 meters above the surrounding seabed. This chain of seamounts also forms the eastern boundary of the deep-water depression.

[0041] 3) Structural Correlation Screening: Based on topographic data measurements, a subduction seamount chain (Krasada Seamount Chain, see [link to relevant data]) exists approximately 25 km from the seafloor of the subducting Pacific Plate. This chain is directly related to the deep-water depression and deep-water angle, and is about 70 km long, 25 km wide, and rises about 2 km above the surrounding seafloor. Figure 4 and 6 Since the size of the deep-water depression and the deep-water angle are much larger than those of the Quesada Seamount, a portion of this seamount chain may have been subducted beneath the overlying plate. According to calculations using the MORVEL global plate tectonics model, the Quesada Seamount chain is moving with the Pacific Plate from 48° north-northwest towards the Philippine Sea Plate at a rate of 24 mm / yr. Due to the subduction effect of the seamount, the trench floor exhibits different topography, and the water depth along the trench axis is relatively shallow (~200-500 m) compared to the surrounding areas not affected by seamount subduction (see...). Figure 5 ).

[0042] Therefore, the identified deep-water depressions and deep-water angles simultaneously satisfy both steps 2 and 3, and are thus marked as candidate areas for “forearc basin prototype”. 4) Under the subduction of the seamount chain at different stages, the interaction between the seamount and the forearc region leads to fore-edge erosion and basement erosion processes dominated by compressional tectonic deformation. This process evolved in four stages, eventually developing into a forearc basin (see Figures 2(a)-2(d)): The identified deep-water depression is bordered to the east by serpentinized mud volcanoes such as Peacock Seamount and Inazoma Seamount, which close its seaward gap (see...). Figure 3 , 4 (and 7), forming a concave-angled seafloor shape, which is highly conducive to the accumulation of surrounding sediments. Multiple seismic data traversing the deep-water depression indicate that its bottom is typically a flat basin fill. Due to the obstruction of the arc-shaped seamount chain, the sediments in the central region are relatively thicker than those on the sides, but the overall sedimentary thickness is approximately 0.2–0.1 s TWTT (see TWTT for round-trip travel). Figure 7 Furthermore, folding and deformation occurred on the side near the seamount, and some sedimentary layers contained disordered fragments (e.g., Figure 7 (c) The fragments may have originated from scraped seamounts or forearc collapses. Therefore, the deepwater depression is in the sedimentary stage of the forearc basin evolution process.

[0043] The seabed of the deep-water concave is mostly relatively flat, with isolated seamounts and groups of seamounts of varying sizes present in some areas. These seamount rows are generally arranged in a near-northeast-southwest direction and are approximately parallel to the trench, while being roughly perpendicular to the subduction direction of the Quesada Seamount (see...). Figure 3 and 4 This suggests that these linear geomorphic units may be related to the subduction of the Quesada Seamount. The water depth at the bottom of the deep-water concave angle is approximately 5400 m, with a depth difference of about 300 m from the adjacent deep-sea plain. Seismic profile data also show that there is a tension fault at its basement and no sediment cover (see...). Figure 7 (part (b) and part (c) of 7). Therefore, the deep-water concave angle is in the subsidence stage of the forearc basin evolution process.

[0044] 5) After completing the stage discrimination of deep-water depressions and deep-water concave corners, stage mapping is performed, and different colors are assigned to each stage for visualization output (see...). Figure 8 ).

[0045] All the above steps are implemented through a computer system. By using Python scripts to call the GDAL, NumPy and Scikit-image libraries, DBM preprocessing, contour line extraction, ellipticity calculation, slope analysis and stage discrimination are automatically completed, and finally a staged spatial distribution map in GeoJSON format is output.

[0046] Therefore, this method, relying solely on high-resolution DBM and multichannel seismic data, introduces a dual gate of multi-index hard thresholds and structural correlation, preserving the evidence chain of topographic and seismic details, avoiding misjudgments and edge effects caused by single indicators, and forming auditable and reproducible standardized stage distribution layers and profile annotations, achieving near real-time mapping and cross-regional comparison, which has important theoretical value and prospects for promotion and application.

[0047] The embodiments described above can be further combined or replaced, and these embodiments are merely descriptions of preferred embodiments of the present invention, not limitations on the concept and scope of the present invention. Various changes and improvements made to the technical solutions of the present invention by those skilled in the art without departing from the inventive concept are all within the protection scope of the present invention. The protection scope of the present invention is given by the appended claims and any equivalents.

Claims

1. A multi-index-based method for determining the stage of a deep-water forearc basin, characterized in that, Performed by the computer system, it includes the following steps: 1) Multi-source data processing and target recognition Multi-source depth data are fused to generate a high-resolution digital depth model (DBM), and multi-channel seismic profiles are stacked, time-migrated, and corrected for two-way travel time (TWTT). Based on the digital water depth model, candidate deep-water depressions and deep-water concave corners are extracted by overlay analysis of isobath closed loop identification and terrain rendering map; wherein, deep-water depressions are closed loop terrain and deep-water concave corners are open loop terrain. 2) Morphological screening: For the candidate regions, determine whether they simultaneously meet the following four conditions: (a) Ellipticity Φ ≥ 0.7, where Φ = 4πM / z², M is the area of ​​the region, and z is the perimeter; (b) Equivalent diameter Dr ∈ [15, 100] km, where Dr = Based on the area equivalent circle diameter; (c) Boundary slope S > 5°, where S = arctan(ΔV / ΔH), ΔV is the vertical elevation difference, and ΔH is the horizontal distance; (d) The difference between the bottom water depth B_d and the average water depth B_c of the adjacent deep-sea plain, |B_d - B_c| ≤ 500 m; 3) Constructing association filtering: Determining whether candidate regions have a spatial association with subducting seamounts, including: (a) The Euclidean distance between the candidate region and the projection center of the nearest subducting seamount is ≤ 30 km; (b) The candidate region is located within the projection zone of the seamount subduction direction, which is calculated by the MORVEL global plate motion model; The region is marked as a candidate region for "forearc basin inception" only if both steps 2 and 3 are met simultaneously; 4) Construction and discrimination of four-stage evolution model: Based on the topographic, seismic and tectonic features of the candidate areas, the following discrimination rules are used to determine their evolution stages: Compression stage: if (i) slope S > 8°, (ii) terrain roughness RMS > 25 m, (iii) seismic profile shows that the leading edge wedge reflection structure is broken and the distance from the seamount is ≤15 km; Uplift stage: if (i) the fracture density E_fault = N_fault / M > 0.3 fractures / km², (ii) there is uplifted topography and the elevation change > 100 m, (iii) the seismic profile shows an extensional-strike-slip fault network; Settlement stage: If (i) 250 m ≤ |B_d - B_c| ≤ 500 m and B_d ≥ 5200 m, (ii) the seismic profile shows a tensile fault at the basement and no sediment cover; Depositional stage: If (i) the seaward gap is closed, (ii) there is a serpentinized mud volcano, the seismic profile shows a chaotic reflection of multiple waves on the seabed, and (iii) the thickness of the sedimentary layer is ≥ 0.2 s TWTT; 5) Stage Mapping and Visualization: The discrimination result F_stage ∈ {Ⅰ, Ⅱ, Ⅲ, Ⅳ} for each candidate region is mapped to color encoding and output as a GIS vector layer. F Ⅰ Corresponding to the first stage, F Ⅱ Corresponding to the second stage, F Ⅲ Corresponding to the third stage, F Ⅳ This corresponds to the fourth stage.

2. The method according to claim 1, characterized in that, Step 1) describes the identification of closed-loop isobaths, which includes: locating closed-loop isobaths around a common center and with progressively increasing water depth on the isobath map generated by DBM; if the difference in water depth between the deepest isobath within the loop and the adjacent outer isobath is ≤ 50 m, and there is no interruption in the loop connecting to the outer isobaths, then it is determined to be a valid closed loop.

3. The method according to claim 1, characterized in that, The terrain rendering analysis described in step 1) includes: overlaying the shadow / highlight rendering map with the contour map. If the light-dark transition zone is continuous and unbroken at the ring boundary, it is determined to be a deep water depression; if there is a unilateral bright band or shadow interruption, it is determined to be a deep water depression.

4. The method according to claim 1, characterized in that, The subduction direction mentioned in step 3) is determined as follows: based on the MORVEL global plate tectonics model, the relative motion vector of the seamount with respect to the forearc basin is calculated, and a projection zone is formed by extending 30 km along this direction.

5. The method according to claim 1, characterized in that, The seismic profile features used in each stage of the discrimination in step 4) include: leading edge wedge fracture, basement tension fault, mud volcano multiple wave chaotic reflectors, all of which are automatically identified through seismic amplitude anomalies and waveform characteristics.

6. The method according to claim 1, characterized in that, The computer system execution process includes: using Python scripts to call the GDAL, NumPy, and Scikit-image libraries to automatically complete DBM preprocessing, contour line extraction, ellipticity calculation, slope analysis and stage discrimination, and output a stage spatial distribution map in GeoJSON format.

7. An apparatus for carrying out the method according to claims 1–6, characterized in that, include: The data receiving unit is used to acquire and store multi-source water depth data and multi-channel seismic profiles; The feature extraction unit is used to perform isobath loop identification, terrain rendering analysis, and ellipticity / slope calculation. Construct correlation cells to calculate the Euclidean distance and subduction direction projection between the candidate region and the seamount; A stage discrimination unit is used to automatically classify according to the four-stage rule described in claim 1; Visualization output unit, used to generate GIS-compatible stage color layers.

8. A computer-readable storage medium having program instructions stored thereon, which, when executed by a processor, cause a computer to implement the deep-water forearc basin stage discrimination method as described in claims 1–6.

9. The computer-readable storage medium according to claim 8, characterized in that, The program instructions are Python scripts that call the GDAL, NumPy, and Scikit-image libraries to implement all the steps of claims 1–6.

Images