Assessment of acoustic backscatter intensity surveying on deep-sea ferromanganese crust: Constraints from Weijia Guyot, western Pacific Ocean

Hui-qing Yo, Yong-gng Liu, Yong Yng, Jin-feng M, Huo-Di Zhng, Jing-bo Ren,Xi-gung Deng, Go-wen He

a Key Laboratory of Marine Mineral Resources, Guangzhou Marine Geological Survey, China Geological Survey, Ministry of Natural Resources, Guangzhou 510075, China

b Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), Guangzhou 510301, China

Keywords:Ferromanganese crust Seamount Acoustic backscatter intensity Manned deep submersible Jiaolong Resource exploration and assessment Magellan Seamounts Marine geological survey engineering Western Pacific Ocean

ABSTRACT Near-bottom observation data from the manned deep submersible Jiaolong with high-precision underwater positioning data from Weijia Guyot, Magellan Seamounts in the Western Pacific Ocean are reported.Three substrate types were identified: Sediment, ferromanganese crust, and ferromanganese crust with a thin cover of sediment. The ferromanganese crusts show clear zoning and their continuity is usually disturbed by sediments on areas of the mountainside with relatively gentle slope gradients. The identified substrate spatial distributions correspond to acoustic backscatter intensity data, with regions of high intensity always including crust development and regions of low intensity always having sediment.Therefore, acoustic backscatter intensity surveying appears useful in the delineation and evaluation of crust resources, although further more work is needed to develop a practicable methodology.

1. Introduction

The cobalt grades of hydrogenic ferromanganese crust(FC) can be up to 2%, and the average values of 0.5%–0.8%are usually four times those of polymetallic nodules and tens of times those of terrestrial primary cobalt deposits (Hein JR et al., 2000). Such FC is therefore an important mineral resource for the future. China, Japan, Russia, Brazil, and South Korea have so far registered as contractors with the International Seabed Authority (ISA) for the exploration of FC in “AREA” and are promoting the exploration of FC resources. The distribution area and thickness are two important parameters for evaluating FC (He GW et al., 2011;Hein JR et al., 2009). Thickness is usually obtained by sampling techniques such as dredging, drilling, and manipulator sampling, which are inefficient. The distribution area is usually inferred from a combination of methods such as cable-towed video surveys, sub-bottom profiler surveys,submersible remotely operated vehicles, and geological sampling (He GW et al., 2005; Kim J et al., 2013). These methods are also relatively inefficient.In situacoustic measurements represent an improved method for quickly obtaining the thickness of FC (Thornton B et al., 2012, 2013;Usui A and Okamoto N, 2010). However, measuring the distribution area efficiently remains a major obstacle in resource assessment by current and potential contractors.

In recent years, marine acoustic backscatter intensity(ABI) surveying has made rapid progress and has been widely used in various fields such as the assessment of benthic organism distributions (Ierodiaconou D et al., 2018; Lacharite M et al., 2018; Sen A et al., 2016), identification of substrate types (Ahmed KI and Demšar U, 2013; Cuff A et al., 2015;Dunlop KM et al., 2018; Yang Y et al., 2015, 2016; Zhang TW et al., 2012), and abundance detection and delineation of submarine polymetallic nodules (Chakraborty B and Kodagali V, 2004; Chakraborty B et al., 1997; Lee SH and Kim KH,2004; Ma Y et al., 1986; de Moustier C, 1985; Tyagi A et al.,2009; Zhu F and Yu ZZ, 2015). Usui A and Okamoto N(2010) first attempted to establish a distribution pattern of FC on a Micronesian seamount using conventional narrow-beam and side-scanning sonar mapping, dredge sampling, and video monitoring with geological interpretation (Usui A and Okamoto N, 2010). Subsequently, an ABI threshold of −20 dB was suggested to delineate FC areas within seamounts in the Northwest Pacific Ocean (Cheng YS, 2014).Unsupervised or supervised classification of ABI data can identify soft, medium-hardness, and hard bottom in different seamounts (Yang Y et al., 2015, 2016). Joo J et al. (2016)found a strong correlation between acoustic-backscatter data acquired from a sea-surface multi-beam survey and the spatial distribution of sediments and FC; they suggested the analysis of acoustic backscatter could as a practical method for selecting optimal areas of the FC from seamounts for future mining (Joo J et al., 2016).

A lack of underwater positioning data for cable-towed video surveying and insufficient sampling stations (Joo J et al., 2016; Cheng YS, 2014; Yang Y et al., 2016) may affect the reliability of any findings related to FC exploration for future mining. This work tested the use of ABI in assessing distributions of FC and sediments on the seamount. The method incorporated the advantages of high-precision positioning of near-bottom observations using the manned deep submersibleJiaolong. The potential applicability of the method in the evaluation of FC resources is assessed and found to be of great potential for FC resource exploration.

2. Study area

The study area is within Weijia Guyot (also known as Ita Mai Tai Guyot), part of the Magellan Seamounts in the Western Pacific Ocean. The guyot has an irregular shape with a length of about 120 km and a width of about 65 km. It lies at depths of 1400–5850 m (Fig. 1).

The guyot comprises a relatively flat platform, steep sides,and four spurs. Several small round mounds in the southwestern part of the platform have height differences of up to 300 m. The base of the guyot has the oldest oceanic crust, dated to the Middle–Late Jurassic. The formation age of the main body is 118.1–119.9 Ma (Koppers AAP et al., 2003).The platform is covered by Eocene–Pleistocene foraminiferal sediments and limestones; the guyot base is generally covered by Aptian-Miocene carbonates, with thick FC on slopes without sediments or at the edges of the platform (Asavin AM et al., 2010; Mel’nikova ME et al., 2012).

3. Data acquisition and processing

3.1. Data acquisition

Bathymetry and ABI data were collected by the R/VHai Yang Liu Haoat Weijia Guyot in 2015 (during cruise DY36)using an EM122 multi-beam system. The parameters for roll,pitch, yaw, and time delay for field data collection were set at 0.06°, 0.75°, 0.11°, and 0.0 s, respectively, and the operation speed was 8–10 knots, and with working frequency is about 12 kHz. The beam angles ranged from 45° to 60° depending on the terrain. The beam spacing was equidistant. Data acquisition such as coverage graphic, swath water-depth profile, ABI, and beam quality was monitored in real-time by Seafloor Information System inboard monitoring software.The sound velocity profile is collected by the AML SVP, the files were collected every 1°×1° which conforms to the DZ/T 0292-2016 specification.

Video data were collected by the manned submersibleJiaolongduring near-bottom observation and sampling in Weijia Guyot (on cruise DY37). Videos were recorded through an optical detection system on the submersible and near-bottom working photos can be seen in Fig. 2. Highprecision position information was obtained through the POSIDONIA II ultra-short baseline system mounted in the submersible. The maximum working depth of the system was 7000 m, its maximum working distance was 8000 m, and the ranging accuracy was 0.2%–1% (Zhang TW et al., 2016).



3.2. Data processing

The ABI data were processed by the Geocoder module of Caris Hips & Sips software (version 8.0). The processing flow was as follows: (1) Bathymetry data were processed in Caris Hips and abnormal and distortion points were removed from the raw data; (2) Time-Varying Gain (TVG) correction, which can compensate for the influence of acoustic attenuation, and already done during the data acquisition, data correction with the actual measured SVP; (3) Angular Varying Gain (AVG)correction, which is mainly to the removal of the residual amount of field incident angle correction. The flat correction model was used to remove the angle response of the substrate for this study; (4) despeckle effectively reduced isolated pixel noise and the high pass filtering was used; (5) the feathering algorithm achieved smooth splicing of the ABI data between strips, forming a mosaic map (Ma JF et al., 2015; Yang Y et al., 2016). To facilitate the analysis of the ABI characteristics,gray-level conversion of the backscatter intensity was not performed.

The depth profile was extracted from multibeam bathymetry data, with a grid of 50 m × 50 m based on the locations of individual screenshots along the track of HOV.The slope gradient is calculated from the same multibeam bathymetry data by using the Spatial Analysis Tool in the Software ArcGIS V10.1, and the slope gradient profile was extracted from the calculated slope gradient map according to the locations of individual screenshots along the track of HOV. The length of each section was calculated based on the spatial location from the start point to the endpoint of the section through the Software ArcGIS V10.1.

Screenshots from the videos for dives 105 and 106 were taken every 30 s, and each was analyzed for the substrate.Positional information for each screenshot was obtained from the interpolation of ultra-short baselines according to the time;the jumping point of the ultra-short baseline needed to be eliminated before the interpolation. The two dives generated about 2333 and 2552 positioning data points with 157 and 239 jumping data points, which accounted for 6.73% and 9.36%,respectively.



Fig. 1. The schematic diagram showing the study area. a–the schematic diagram shows the location of Weijia Guyot, western Pacific Ocean;b–bathymetric map of Weijia Guyot and red solid lines are tracks of the manned deep submersible Jiaolong.

Fig. 2. Near-bottom working photos of Jiaolong submersible. a–collecting the cobalt nodules; b–collecting the gravel-like ferromanganese crust.

4. Results and discussion

4.1. Near-bottom observation results

4.1.1. Dive 105

The total length of the dive 105 track was about 4 km and could be divided into seven sections according to the distribution of substrates (Fig. 3; Table 1).

Section A–B: The deepest section (2287–2138 m) has steep terrain (31.5°–41.8°) with FC continuously occurring and usually covered by a thin layer of sediment (Fig. 4a).

Section B–C: At the depths of 2138–2091 m, the terrain is less steep than the previous section (17.0°–25.6°). FC continuously develops with no sediment cover. Clearly visible scour marks on the FC surface are attributable to near-bottom current (Fig. 4b).

Section C–D: The depth (2091–1965 m) and slope(11.7°–22.9°) continue to become shallower; wavy sediment continuously develops, and no FC is observed (Fig. 4c).

Section D–E: At the depths of 1965–1938 m, the gradient(20.7°–23.4°) is slightly steeper than in the previous section.FC continuously develops, usually with a thin cover of sediment, showing obvious scour marks on the surface (Fig. 4d).

Section E–F: At the depths of 1938–1565 m, the mountainside gradually becomes the top platform as the slope increases to 24.5°, and then gradually reduces to 2.7°. FC continuously develops with no sediment covering (Fig. 4e).

Section F–G: At the depths of 1565–1553 m and the gradients of 0.6°–2.9°, FC develops with a relatively poor continuity and is usually covered by a thin layer of sediment.Gravel-like FCs become more frequent (Fig. 4f).

Fig. 3. Substrate distribution characteristics along the track of dive 105. a–FC and sediment distributions: Red indicates FCs (C1); pink indicates FCs with thin sediment covering (C2); black indicates sediment (S); b–depth and slope gradient profiles.

Table 1. Deep-sea ferromanganese crust and sediments distribution characteristics along the track of dive 105.

Fig. 4. Typical screenshots for the dive 105. a–section A–B, FC with thin sediment covering; b–section B–C, FC with scour marks;c–section C–D, wavy sediment; d–section D–E, FC with thin sediment covering and scour marks; e–section E–F, FC; f–section F–G,FC with poor continuity and thin sediment layer, and development of gravel-like FC.

Section G–H: At the depths of 1553–1541 m and the gradients of 0.6°–1.7°, wavy sediment continuously develops,and gravel-like FC appears occasionally.

In general, continuously distributed FC is often interrupted by sediment (section C–D) in areas along the mountainside with a lower slope gradient. Although section A–B is relatively steep, the influence of near-bottom currents causes the FC to be covered by a thin layer of sediment. FC continuously develops along with section B–C with no sediment cover. This may be related to local near-bottom currents and the scour marks on the FC surface. Note that the continuity of FC becomes worse, and the probability of gravel-like FC increases within the transition zone from mountainside to platform (F–G), at which the gradient reduces and often does not exceed 3°. The slope gradient is further reduced to mainly below 2° on the platform itself, which is covered by wavy sediment with the occasional appearance of gravel-like FC.

4.1.2. Dive 106

The total track length during dive 106 was about 5 km and could be divided into ten sections based on the distribution of substrate (Fig. 5; Table 2).

Section A–B: At the depths of 2634–2519 m, the relatively steep terrain has gradients of 22.7°–29.7°. The FC develops continuously, usually with a thin cover of sediment(Fig. 6a); areas of gravel-like FC are visible.

Section B–C: At the depths of 2519–2466 m, the gradient increases to 31.9°. Wavy sediment develops, and no FC is observed, possibly due to near-bottom currents.

Section C–D: At the depths of 2466–2069 m and the slopes of 25.0°–35.6°, FC develops continuously and is usually covered with a thin cover of sediment. Scour marks are visible on the FC surface (Fig. 6b).

Section D–E: At the depths of 2068–1992 m and the slopes of 13.0°–22.7°, the terrain is less steep than in the previous section, and wavy sediments occur with no FC development.

Section E–F: At the depths of 1992–1970 m and the gradients of 11.5°–15.0°, the terrain continues to flatten. FC is developed continuously and covered with a thin cover of sediment.

Section F–G: At the depths of 1970–1858 m and the slopes of 7.2°–15.8°, the flattening progresses and wavy sediments develop with no FC.

Section G–H: Ascending to 1858–1816 m, the gradients of 7.5°–17.3° are slightly steeper than in the preceding section. FC develops continuously with thin sediment covering and scouring marks on the FC surface.

Section H–I: At the depths of 1816–1585 m and the gradients of 4.7°–28.5°, this section has steeper slopes than the previous section. FC continuously develops with no thin sediment covering (Fig. 6c).

Section I–J: At the depths of 1585–1542 m and the slopes of 1.3°–4.7°, the terrain is relatively flat. FC develops with less continuity than in the previous section. FC is covered with a thin cover of sediments, and the occurrence of gravellike FC increases (Fig. 6d).

Section J–K: At the shallowest depths (1542–1524 m) and gentlest slopes (0.9°–1.5°), the terrain is relatively flat. Wavy sediments develop with no FC.

In general, the continuously distributed FC is often interrupted by sediment (sections D–G) in areas of the lower gradient along the mountainside. This is similar to the observations during dive 105. Although the terrain in sections A–D is steep, FC develops with a thin cover of sediment caused by near-bottom currents, with section B–C being completely covered by wavy sediments without the appearance of FC. Note that the FC continuity worsens and the occurrence of gravel-like FC increases within the transition zone from mountainside to platform (I–J), where the slope reduces to often not exceeding 5°. Further flattening to mostly less than 2° on the platform is accompanied by wavy sediment without the appearance of FC.

4.2. Relationship between acoustic-backscatter intensity and substrate

The above near-bottom observation results identified three types of substrate: FC, FC covered with thin sediment, and sediment, which is labeled as C1, C2, and S, respectively.Table 1, Table 2, and Fig. 3, Fig. 5 summarize the results.

4.2.1. Dive 105

Fig. 5. Substrate distribution characteristics along the track of dive 106. a–FC and sediment distributions: Red indicates FCs (C1); pink indicates FCs with thin sediment covering (C2); black indicates sediment (S); b–depth and slope gradient profiles.

Table 2. Deep-sea ferromanganese crust and sediments distribution characteristics along the track of dive 106.

Fig. 6. Typical screenshots for dive 106. a–section A–B, FC with thin sediment layer and visible scour marks; b–section C–D, continuous distribution of FC, visible scour marks; c–section H–I, continuous distribution of FC; d–section I–J, poor continuity of FC, thin layer of sediment, and gravel-like FC.

Along the track of dive 105, ABI data were extracted from the ABI map (Fig. 7) at intervals of about 50 m as listed in Table 1. In general, the backscatter intensity weakened in the order C1 > C2 > S, except section B–C (C1 type). As section B–C was relatively short (about 50 m, similar to the measurement interval), its data may have included ABIs from adjacent sections. For the C2 type substrate, except for section D–E, the ABIs were approximately equal on sections with very different slopes (i.e., Section A–B at 31.45°–30.16° and Section F–G at 0.6°–2.9°). This indicates the small influence of the slope gradient on ABI. While the S-type substrates showed slightly decreasing ABIs from section C–D (from–45.89 dB to –28.46 dB, with an average of –38.37 dB) to section G–H (from –45.85 dB to –31.1 dB, with the average of –40.45 dB) as the slope gradient decreased from the range of 11.7°–22.9° to that of 0.6°–1.7°, it was inferred that it may be attributed to the increasing thickness of the sediment layer,rather than to the gradient directly.

Fig. 7. Substrate and backscatter intensity characteristics along the track of dive 105. a–substrate distribution: Red, C1; pink, C2; and black, S.b–acoustic backscatter intensity profile.

The box-whisker plots for dive 105 show gradually decreasing ABI in the order C1 > C2 > S (Fig. 8). The corresponding interquartile ranges were –21.58 dB to –25.32 dB, –28.27 dB to –30.41 dB, and –35.14 dB to –45.85 dB,with medians of –23.36 dB, –30.21 dB, and –41.52 dB,respectively. The ABI for substrate type C1 was greater than that of C2 by about 7 dB, which in turn was greater than that of S by about 11 dB.

4.2.2. Dive 106

Fig. 8. Box-whisker diagrams showing the relative backscatter intensity for each seabed type along the track for dive 105. Brackets represent the standard range of data = 1.5 × inter-quartile range; red crosses represent the means; solid blue circles represent the maximum and minimum values.

Fig. 9. Substrate and backscatter intensity characteristics along the track of dive 106. a–substrate distribution: Red, C1; pink, C2; black, S.b–acoustic backscatter intensity profile.

Similarly extracted ABIs for approximately every 50 m along the track of dive 106 are shown in Fig. 9 and Table 2.The ABI ranges for substrates C1 and C2 were similar. The maximum ABI for C2 was slightly higher than that of C1,while the C1 ABIs had a narrow range than that of the C2(Fig. 9). Comparison of ABIs for the different sections containing C2 type substrate indicates similar ranges and average values in sections with differing slopes (i.e., section C–D at 26.0°–35.6° and the section I–J at 1.3°–4.7°). This suggests that the slope has little influence on the ABI. The S type substrate, like that in the track of dive 105, showed decreasing ABI for sections D–E, F–G, and J–K as the slope decreased (except for the short, 30 m, section B–C).

The box-whisker plots for ABI during dive 106 show similar values for types C1 and C2, which were higher than those of substrate type S (Fig. 10). The corresponding interquartile ranges were –23.08 dB to –28.72 dB, –22.97 dB to –29.08 dB, and –29.35 dB to –40.35 dB, with median values of –26.13 dB, –25.13 dB, and –35.63 dB, respectively.While C1 and C2 had similar values, C1 had a slightly lower mean and median than C2. Similar to the results for dive 105,the C2 ABI was about 10 dB greater than that for the S type substrate.

4.3. Assessment of ABI used for FC exploration

Fig. 10. Box-whisker diagrams of relative backscatter intensity for each seabed type along the track of dive 106. Brackets represent the standard range of data = 1.5 × inter-quartile range; red crosses represent means; blue solid circles represent maximum and minimum values.

Fig. 11. ABI inter-quartile ranges for each seabed type along the track of Dive 105 and Dive 106. Substrate distribution: red, C1; pink,C2; and black, S

Cheng YS (2014) reported using ABI data, cable-towed camera videos, and sampling data to study the spatial distribution of two guyots and one apex seamount in the Pacific Ocean and found that ABI higher than –20 dB indicates a favorable area for FC development. Joo J et al.(2016) reported cable-towed camera video results and ABI map on the OSM11 seamount in the Western Pacific Ocean,finding that ABIs less than –25 dB indicated sediment distribution areas and values greater than –20 dB indicated FC distribution areas. While the study from Qianyu Guyot and Caiwei seamounts found that ABIs greater than –28 dB,–27dB matched FC development, respectively, and ABIs less than –33 dB, –34 dB matched sediment development,respectively (Yang Y et al., 2015, 2016). According to the ABI inter-quartile ranges from each seabed type from these two dives (Fig. 11), the authors suggested that ABIs greater than –29 dB indicated the FC development. Because the C2 seabed type has resource implications during the FC exploration for the FC occurring within the C2 seabed type is not covered fully by sediment. And the authors also suggested that ABIs less than –35 dB indicated the sediment development. ABIs from –35 dB to –29 dB indicates a more complicated situation, it needs more work to be carried out to obtain the substrate type.

It is interesting to note that the current work is consistent with the results from the Qianyu seamount in the mid-Pacific and the Caiwei seamount in the Western Pacific, but they differ from the findings of Joo J et al. (2016) and Cheng YS(2014). Differences in ABI of about 9 dB and 10 dB were observed here for the FC substrate and sediment substrate. A systematic difference between the two types of about 10 dB is therefore inferred. The difference may be due to background noise from the research vessel or equipment aging.

Yang Y et al. (2016) reported ABI data for the Qianyu Guyot in the mid-Pacific Ocean obtained by R/VHai Yang Liu Haoin 2011, and Yang Y et al. (2015) also reported ABI data for the Caiwei seamounts in the Western Pacific Ocean measured by the same vessel in 2015. The data used in this paper were collected in 2015 by the same vessel, indicating that equipment aging accounts for little; therefore, the difference was attributed to background noise from the research vessel. While the authors cannot explain the consistency of the results of Cheng YS (2014) and Joo J et al.(2016), which were collected by R/VDa Yang Yi Haoand R/VOnnuri, respectively (Cheng YS, 2014; Joo J et al.,2016), the only reasonable explanation is that these two ships had similar background noise levels. Therefore, further investigation is needed to explain the systematic difference.Nevertheless, it will not affect the delineation of FC resources, because using current technology, a seamount ABI survey could be completed in a few days without several ships having to work together. However, if a seamount is surveyed by multiple ships, the influences of the different ships should be considered.

Seamounts favorable for FC development tend to have platforms consisting of old formations and include three main terrain units: Platform, sides, and spurs. This paper reported data from the sides and platform of Weijia Guyot and found good correspondence between substrate type and backscatter intensity. However, limitations of this study include the relatively shallow depths considered (maxium reaching 2634 m on dive 106) and the consideration of only the platform and sides but not the spurs, which are known to be favorable for FC development (Usui A et al., 2017). Therefore, the development of an effective technical method for FC exploration requires further near-bottom floor observations with underwater positioning data in different specific terrain units.

5. Conclusions

(i) The FC at Weijia Guyot shows clear zonation and its continuity is usually disturbed by wavy sediments in areas of the sides of the mountains with relatively gentle slope gradients.

(ii) The backscatter-intensity data correspond well to the substrate type. Areas with high intensity always show FC development, whereas areas with low intensity have a sediment cover. Therefore, ABI surveying appears useful in FC resource delineation and evaluation, although further work is needed to develop a practicable methodology.

(iii) There may be systematic differences in intensity data collected by different research vessels, possibly due to the distinctive background noise from each vessel.

CRediT authorship contribution statement

Hui-qiang Yao and Gao-wen He conceived of the presented idea. Yong-gang Liu processed the video data. Xiguang Deng and Yong Yang collected the multibeam data,and Jin-feng Ma processed the multibeam data. All authors discussed the results and contributed to the final manuscript.

Declaration of competing interest

The authors declare no conflict of interest.

Acknowledgment

The authors are grateful to all team members of the scientific expeditions on cruises DY36 and DY37. This study was funded by the Resource and Environment COMRA Projects (DY135-C1-1-02, DY135-C1-1-01) and the China Geological Survey (DD20191009). The editor, associate editor and anonymous reviewers are thanked for their constructive comments that helped greatly improve the paper.