Data sources
Data required for creating To create L3 products are , the key data required is L2 MBES (reduced to EGM2008) in containing backscatter, typically stored in the proprietary format of the processing software used e.g. , such as Caris HDCS_Data or in gsf format.
Processing steps
Steps for determining Grid Resolution (using Teledyne Caris)
Grid bathymetry according to the AusSeabed Depth Ranges Guide 2019 (Table 12) and apply associated AusSeabed CUBE parameters (Table 13). In configuration method, select browse button and select CUBEParams_AusSeabed_2019.xml and select appropriate configuration according to grid resolution e.g., Configuration = AusSeabed_0.5m (Figure 2). Utilise the AusSeabed CUBE parameters for optimised results.
Choose one grid for one survey area (do not subdivide) except when:
areas have large depth variations (e.g. from a continental shelf to an abyssal plane while transiting), and/or the resultant dataset will be split into multiple geotiffs and calculated for the best resolution for the area based on the geometry.
When gridding, break up at the end of the line if exceeding the boundary.
Every grid must not exceed depth variations more than 5 depth range intervals of the AusSeabed Depth Ranges Guide 2019 e.g., -20m to - 160m (Table 12).
For each grid, run first an auto-resolution grid in HIPS which will determine the best resolution for the chosen grid. Then round it to the nearest band in Table 12 and re-grid accordingly to this grid size.
Table 11. Simplified AusSeabed depth bands and associated grid resolutions.
...
Depth (m)
...
Resolution (m)
...
-20
...
0.5
...
-40
...
1
...
-80
...
2
...
-120
...
3
...
-160
...
4
...
-200
...
5
...
-240
...
6
...
-280
...
7
...
-320
...
8
...
-360
...
9
...
-400
...
10
...
-440
...
11
...
-480
...
12
...
-520
...
13
...
-560
...
14
...
-600
...
15
...
-640
...
16
...
-1280
...
32
...
-2560
...
64
...
-5120
...
128
...
-12000
...
210
Table 12. AusSeabed CUBE parameters[1].
...
Grid Resolution (m)
...
Minimum Capture Distance[2]
...
Capture Distance Scale (<=1%)[3]
...
Horizontal Error Scalar[4]
...
0.5
...
0.3535
...
0.5
...
1.96
...
1
...
0.707
...
0.5
...
1.96
...
2
...
1.414
...
0.5
...
1.96
...
3
...
2.121
...
0.5
...
1.96
...
4
...
2.828
...
0.5
...
1.96
...
5
...
3.535
...
0.5
...
1.96
...
6
...
4.242
...
0.5
...
1.96
...
7
...
4.949
...
0.5
...
1.96
...
8
...
5.656
...
0.5
...
1.96
...
9
...
6.363
...
0.5
...
1.96
...
10
...
7.07
...
0.5
...
1.96
...
11
...
7.777
...
0.5
...
1.96
...
12
...
8.484
...
0.5
...
1.96
...
13
...
9.191
...
0.5
...
1.96
...
14
...
9.898
...
0.5
...
1.96
...
15
...
10.605
...
0.5
...
1.96
...
16
...
11.312
...
0.5
...
1.96
...
32
...
22.624
...
0.5
...
1.96
...
64
...
45.248
...
0.5
...
1.96
...
128
...
90.496
...
0.5
...
1.96
...
210
...
148.475
...
0.5
...
1.96
...
Decision tree for Processing
For some legacy data, CUBE cannot be performed, and Swath Angle is the preferred alternative. Decision tree is provided to help decide gridding algorithm.
Figure 3. Decision tree for gridding. *See dot points 3 and 4 in Section 6.2. ^Reduced to EGM2008, expecting L2 in EGM2008. This L2 data is essential for producing accurate backscatter mosaics and other L3 products, as it enables precise corrections and calibrations within the software. Alternatively, L2 data can be stored in open-source formats like Generic Sensor Format (GSF), provided the backscatter is included, improving interoperability across platforms. Regardless of format, L2 MBES data serves as the foundation for L3 products, which are crucial for seafloor mapping, habitat classification, and geological studies.
Processing steps
While some processing software retains L2 MBES backscatter data in proprietary formats during the mosaicking process, others only preserve the final L3 MBES backscatter. This variation in data retention can significantly influence how the data can be accessed, analysed, and reused in future projects.
Retaining L2 backscatter offers several advantages, such as the ability to revisit and refine the data with different processing techniques or to apply additional corrections that may be necessary as new methodologies emerge. This flexibility is particularly valuable in scientific research, where data integrity and the ability to reprocess information are critical for reproducibility and long-term studies.
On the other hand, software that retains only L3 backscatter offers a streamlined process aimed at delivering a final product—a backscatter mosaic ready for interpretation and analysis. Although this method is efficient, it may limit the ability to revisit the data for new requirements or additional validation without reverting to the original L0 MBES data.
Regardless of the specific software or data retention strategy, AusSeabed recommends following the standardised processing flowchart illustrated in Figure xx. This flowchart ensures that all steps, from data acquisition to final mosaic creation, are conducted systematically, promoting consistency and accuracy across different surveys and projects. By adhering to this recommended workflow, users can ensure that their backscatter data meets the highest standards of quality, whether it’s used for immediate analysis or archived for future research and development.
...
Figure xx. Recommended backscatter processing flowchart
Starting with L2 MBES data that includes backscatter measurements, the following steps outline the process for backscatter data processing (Figure xx):
Gain Correction: Adjust for any gain discrepancies in the backscatter data to ensure consistency across the dataset.
Propagation Effects Correction: Apply corrections for propagation effects, such as Time-Varying Gain (TVG) and insonified area adjustments, to account for variations in signal strength due to distance and angle.
Physical Properties Correction: Correct for physical factors like beam pattern distortions and apply dB offset calibration values to standardize the backscatter measurements.
Angular Dependence Removal: Address angular dependence by applying methods such as averaging, setting window sizes, and defining reference angles to minimize distortions related to the sonar's viewing angle.
Pre-Mosaicking Adjustments: Implement pre-mosaicking steps like anti-aliasing and despeckling to reduce noise and prepare the data for seamless integration.
Mosaicking: Create the backscatter mosaic by selecting grid size, choosing the appropriate operation and blending methods, and applying post-mosaicking enhancements to refine the final product.
L3 MBES Backscatter Product: The final L3 MBES backscatter grid, which integrates all corrections and adjustments to produce a high-quality backscatter mosaic ready for analysis.
Depending on the software, the first 4 or 5 steps may have already been completed for systems that retain L2 MBES backscatter, eliminating the need to repeat them. However, for software that does not retain L2 MBES backscatter, all 6 processing steps are required.
...
[1] Based on NOAA’s, AHO’s HIPP and correspondences with Dr Calder.
...