Origin and Composition of Ferromanganese Deposits of New Caledonia Exclusive Economic Zone

: Located in the South-West Paciﬁc, at the northern extremity of the mostly submerged Zealandia continent, the New Caledonian Exclusive Economic Zone (EEZ) covers 1,470,000 km 2 and includes basins, ridges and seamounts where abundant ferromanganese crusts have been observed. Several investigations have been conducted since the 1970s on the nature and composition of ferromanganese crusts from New Caledonia’s seamounts and ridges, but none have covered the entire EEZ. We present data from 104 ferromanganese crusts collected in New Caledonia’s EEZ during twelve oceanographic cruises between 1974 and 2019. Samples were analysed for mineralogy, geochemical compositions, growth rates, and through a statistical approach using correlation coefﬁcients and factor analysis. Crust thicknesses range from 1 mm to 115 mm, with growth rates between 0.45 mm/Ma and 102 mm/Ma. Based on textures, structures, discrimination plots, and growth rates, we distinguish a group of hydrogenetic crusts containing the highest mean contents of Co (0.42 wt%), Ni (0.31 wt%), and high contents of Mo, V, W, Pb, Zn, Nb, from a group of hydrothermal and/or diagenetic deposits showing high mean contents of Mn (38.17 wt%), Ba (0.56 wt%) and low contents of other trace metals. Several samples from this later group have exceptionally high content of Ni (0.7 wt%). The data shows that crusts from the southern part of the EEZ, notably seamounts of the Loyalty Ridge and the Lord Howe Rise, present high mineral potential for prospectivity owing to high contents of valuable metals, and constitute a great target for further investigation.


Introduction
Hydrogenetic ferromanganese (Fe-Mn) oxide deposits are known to be distributed widely in all oceans of the planet, the largest known fields being located in the Pacific Ocean [1]. They occur as crusts on sediment free surfaces like seamount flanks and summits, ridges, or any topographic reliefs located between 400 and 7000 m water depths [2,3]. Ferromanganese crusts grow by precipitation of metals from ambient cold seawater and accumulation on the seafloor, forming layers of Mn oxides and Fe-oxyhydroxides. Their thickness ranges from less than a millimeter up to 25 cm [4]. The hydrogenetic accumulation of Mn oxides and Fe-oxyhydroxides requires stable conditions over long periods of time (million years) to form thick crusts [5]. Their distribution, textures and composition are impacted by several parameters, such as surface bioproductivity, depth of the  [33]); (B) Age of basement formation of the South-West Pacific (modified after [33]).

Sample Collection
A total of 104 samples of Fe-Mn deposits were selected from dredge material collected during multiple oceanographic cruises between 1974 and 2019 (Table 1). Samples were collected at different water depths ranging from 430 m to 4677 m ( Figure 2). Samples from the Lord Howe seamount chain, Lord Howe Rise, Fairway Ridge, Norfolk Ridge, D'Entrecasteaux Basin, North D'Entrecasteaux Ridge and Loyalty Ridge were recovered in diverse settings with regard to depositional environments, age of structures, and nature of substrate rocks ( Figure 1). For all crusts, a representative bulk sample of the whole stratigraphy has been selected. If any macroscopic boundaries were observed within the crust, representative subsamples of the macro-layers were taken, so that all the crust thickness was sampled. In such context, sub-samples are sorted in stratigraphic order. Table 1 shows the sampling type, thickness, location, depth, cruise and substrate rock (vacuolar to amygdaloidal basalt, andesite, hyaloclastite breccia, shoshonite, polygenic breccia, bioclastic limestone, or mudstone). A large proportion of crusts were lacking substrate rocks. Symbol "-" means that there is no thickness information for the sample; Crusts with no substrate are marked as "-".

Mineralogical Analyses
X-ray diffraction (XRD) analyses were conducted with a BRUKER AXS D8 Advance diffractometer. Samples were top loaded into 2.5 cm diameter circular cavity holders, and all analyses were run between 5 • and 70 • 2θ, with 0.01 • 2θ step at 1 s/step (monochromatic Cu Kα radiation, 40 kV, 30 mA). Minerals were identified using the Diffrac.Suite EVA software. This methodology allows the quick identification of most minerals (e.g., silicates, carbonates, well-crystallised manganates, well-crystallised iron oxyhydroxides). δ-MnO 2 is barely visible on diffractograms even when it constitutes the main crystalline phase of a sample. Estimation of the proportion of δ-MnO 2 from other crystalline phases is made on the basis of a qualitative analysis of the diffractograms, i.e., ratio between δ-MnO 2 visible peaks (±37 • and ±66 • 2θ; 2.45 Å and 1.42 Å) and peak signals of other well-crystallised minerals.
Scanning electron microscopy (SEM) imaging was done with an FEI Quanta 200 SEM on C-coated polished thin sections. Backscatter images were acquired for textural charac-terisation of the Fe-Mn oxyhydroxides. Energy Dispersive Spectroscopy (EDS) analysis was performed with and OXFORD X-MAX N Silicon Drift Detector (detector size: 80 mm 2 ).

Geochemical Analyses
X-ray fluorescence analyses were conducted with a wavelength dispersive X-ray fluorescence spectrometer (WD-XRF; BRUKER AXS S8 TIGER) on fusion beads or compressed pellets (for major and trace elements, respectively). After data acquisition, measured net peak intensities corrected from inter-element interferences were converted into concentrations using calibration curves generated from the analysis of certified reference material powders (using BHVO-2 [60]), measured under identical analytical conditions. Additional trace elements (Sr, Y, Zr, Nb, Th, REE) were analysed for 17 hydrogenetic crusts from the Loyalty, Norfolk and D'Entrecasteaux ridges ( Figure 2) by inductively coupled plasma mass spectrometry using an ELEMENT II magnetic field ICP-MS at Institut Universitaire Européen de la Mer (IUEM) in Brest. The dissolution procedure was as follows; 0.1 g of sample powder was digested in a Teflon bottle with 4 mL of 6 mol/L hydrochloric acid for 24 h on a hot plate (120 • C) with a Tm spike [61,62]. If present, the residual phase composed of mostly silicates and refractory minerals was extracted by centrifuge and digested by a mixture of hydrofluoric and hydrochloric acid (3:1) for 48 h on hot plate (120 • C), evaporated and then remixed with the previously digested phase. Then, 0.5 µL of the solution was evaporated on a hot plate and the residue was made up to 10 mL with a 2% nitric and 0.05% hydrofluoric acid solution for trace element analysis by ICP-MS. Samples were corrected using internal calibrations, BHVO-2 reference material, and a Tm spike correction [62]. Every concentration later in the text expressed as % represents weight %.
The Co-chronometer method considers that the supply of Co in the ocean is constant over time and that Fe and Mn oxides are the main scavengers of this element [63,64]. Considering these hypotheses, a proportional relationship can be established to estimate growth rates. Using crusts' thicknesses, it is possible to derive minimum crusts ages from growth rates. However, this method cannot account for post-depositional events like phosphatisation, dissolution, or erosional events that are known to affect Co concentration, preservation of the stratigraphic record, and could therefore alter calculated ages [1,65]. The minimal age of crusts was determined using the empirically derived cobalt chronometer method defined as: GR = 0.68/(Co n ) 1.67 [66], where GR is the growth rate in mm/Ma, and Co n = Co × (50/Fe + Mn) with elements in wt.%. The equation of [64] was not considered to compute the growth rate as several samples exhibit Co content lower than 0.24%, which is the threshold needed to apply this method.
Several methods were used to examine statistically significant variations in major and minor elements concentrations for selected crusts samples. A Pearson correlation coefficient matrix was computed using chemical data to evaluate the strength of linear dependence between variables. To investigate possible chemical factor variations and biases in the data set, a matrix was produced using hydrogenetic macro-layers and bulk samples (n = 89). Bulk samples that have been subsampled were not considered to avoid duplicating data. All correlation coefficients in bold are significant at the 99% confidence level (CL). Factor analysis of the major and minor elements data was also run on the same data set (n = 89) to study element relationships and to determine groups of elements with the same behaviour. Using X-ray diffraction mineralogy and correlation coefficient matrices, each resultant factor of this analysis can be interpreted as a specific mineral or group of minerals in the Fe-Mn crusts and elements correlated with those factors to be part of the mineral group or mineral.

Sample Description
Two types of samples can be distinguished from the macroscopic study: (1) Brown to black Fe-Mn encrustations which are referred here as Fe-Mn crusts, and (2) Grey to dark and rarely brownish grey Mn-rich (±Ca-Fe) samples. Fe-Mn crusts show a large diversity of surface and layered textures ( Figure 3A-D). The surface can be smooth, granular and botryoidal, with botryoids ranging from millimeters to centimeters. Layers can be well separated from others with interstitial sediment and porous, columnar, dendritic, or very dense and well laminated. The thickness of Fe-Mn crusts vary from 2 mm to 115 mm, with a mean thickness of 27 mm (from 61 hydrogenetic bulk samples). The thickest crusts can contain up to four distinct macroscopic layers, but no uniform sequence of texture has been found between these crusts. Mn-rich (±Ca-Fe) samples (DW778B, DW778D, DW778D2, DW4998E, DW4998D, and DW2482, Figure 3E-H) present different morphologies and textures. These samples are denser and harder than Fe-Mn crusts. Some samples ( Figure 3F,H) display a strong imbrication of a metallic black zone and a pale white/reddish zone, showing colloform to dendritic-like growth textures in parts of the sample. Other samples ( Figure 3E,G) present comparable macro-layers with metallic black and blue/grey units, as well as colors ranging from pale grey/blue and white/reddish to metallic black. Alternation of macro-layers is visible in Figure 3E, with an innermost imbrication of pale grey, blue and white layers, presenting in some areas a more or less porous and colloform texture, followed by a black metallic layer present on both sides of the sample. Whilst most previous analysis reports samples of a hydrogenetic nature, these morphological and visual characteristics match criterion proposed by [11] of a hydrothermal nature or influence.

XRD and SEM Mineralogy
XRD mineralogical analyses are reported in Table 2. δ-MnO2 is the most dominant phase detected in Fe-Mn crusts. Given no Fe mineral was identified on XRD, we conclude that most of the Fe is in the form of X-ray amorphous Fe-oxyhydroxides [1,2] and/or Ferich δ-MnO2. In most crusts, quartz, feldspar, calcite, and Mg-calcite represent the main detrital components. Other detrital phases include mica, clay minerals (DR19K, DR19K-1, DR21Biii, GO310 and DR08C), gypsum and amphibole (114D and GO310). Samples where the only Mn phase is δ-MnO2 are characteristic of hydrogenetic ferromanganese crusts.

XRD and SEM Mineralogy
XRD mineralogical analyses are reported in Table 2. δ-MnO 2 is the most dominant phase detected in Fe-Mn crusts. Given no Fe mineral was identified on XRD, we conclude that most of the Fe is in the form of X-ray amorphous Fe-oxyhydroxides [1,2] and/or Fe-rich δ-MnO 2 . In most crusts, quartz, feldspar, calcite, and Mg-calcite represent the main detrital components. Other detrital phases include mica, clay minerals (DR19K, DR19K-1, DR21Biii, GO310 and DR08C), gypsum and amphibole (114D and GO310). Samples where the only Mn phase is δ-MnO 2 are characteristic of hydrogenetic ferromanganese crusts. They generally show complex internal structures including laminated layers of Fe-Mn oxides with varying porosities, cuspate texture (i.e., more porous and chaotic structure) or massive jointed columnar texture with only small amounts of detrital minerals ( Figure 4A-C). Only one Fe-Mn crust sample (DN5080B) contains detectable amounts of fluorapatite. Table 2. XRD and SEM mineralogy of ferromanganese deposits (n = 104) from New Caledonia.

Fe-Mn Samples Classification
A ternary plot of Fe, Mn and (Co + Cu + Ni) × 10 [70] ( Figure 5A) shows that the majority of our samples (including DW4998E sample) fall within the hydrogenetic field [71]. This is in good agreement with macroscopic and mineralogical results, which point out that most of the samples (n = 98/104) can be referred to as hydrogenetic Fe-Mn crusts. However, one sample (i.e., DW4998E) with hydrothermal macroscopic and mineralogical characteristics is plot in the hydrogenetic field due to high Co, Ni and Cu concentrations. The last five samples (DW2482, DW788B, DW778D, DW778D2 and DW4998D) fall within the overlap of the diagenetic and the hydrothermal part of the diagram, consistent with the observed morphologies and mineralogy. The second diagram [72] ( Figure 5B) plots (Fe + Mn)/4, 100 × (Zr + Y + Ce), 15 × (Cu + Ni) and shows a major clustering of samples in the lower part of the hydrogenetic area as well. Two samples (DN5085A and GO16D) plot slightly below the hydrogenetic field, due to high contents of Cu and Ni and relatively low Zr, Ce and Y concentrations. Four samples (DW778D, DW778D2, DW778B and DW4998D) are located within the hydrothermal fall-out crusts and impregnations, strengthening our interpretation of a hydrothermal origin. DW4998E and DW2482 are distributed between the diagenetic and the hydrothermal field. The classification of [73] requires full REE determination which was only produced on 17 of the 104 samples of this set. Nonetheless, the 17 analysed samples plot in the hydrogenetic field are in good agreement with other classification for this subset ( Figure 6).  The classification of [73] requires full REE determination which was only produced on 17 of the 104 samples of this set. Nonetheless, the 17 analysed samples plot in the hydrogenetic field are in good agreement with other classification for this subset ( Figure 6).

Hydrogenetic Fe-Mn Crusts
For the combined hydrogenetic data set (n = 98), chemical composition of bulk and macro-layers is presented as mean ± 2σ% (Table 3; Supplementary Material Table S1). The mean Fe/Mn ratio for the combined data set is 1.26 ± 0.66 and the mean Si/Al ratio is 3.76 ± 2.37. The highest values of Si/Al ratio are found in layers of the thickest crusts (e.g., GO327D-4, GO327D-3 and DR11Ai-3). The mean combined Cu + Ni + Co (%) concentration is 0.81 ± 0.56%. This wide variation in metals with the greatest economic interest is mainly controlled by cobalt concentrations ranging from 0.16% to 1.02%. However, few Fe-Mn crusts samples (e.g., DN5085A, GO16D) are dominated by nickel enrichment with concentrations up to 0.61%.

Non-Hydrogenetic Mn-Rich (±Ca-Fe) Samples Deposits
The six samples characterised by non-hydrogenetic morphologies and mineralogy (DW778B, DW778D, DW778D2, DW4998D, DW4998E and DW2482) fall in either the hydrothermal or diagenetic field of common classification schemes [72,73]. These samples have very low Fe/Mn ratios (<0.07), except for sample DW4998E (0.97) where the analysis incorporates both Mn-dominated mineralization and Fe-Ca-rich, hydrothermally-altered hyaloclastite. They have low content of elements characteristic of the aluminosilicate

Growth Rates and Ages
Estimated growth rates have been calculated using an empirical Co chronometer [66]. Growth rates vary from 0.45 to 102 mm/Ma ( Figure 8) and show no geographic correlations. Bulk hydrogenetic crusts (n = 74) have a mean growth rate of 2.2 ± 2.5 mm/Ma. Growth rates for macro-layers sub-samples (n = 24) are on average higher, at 3.1 ± 2.9 mm/Ma. Considering stratigraphic variations ( Figure 8B), most samples (DR14F, DR21F, DR38C, GO302D) present an increasing growth rate towards the most recent period, whereas only one crust exhibits a decreasing growth rate (DR19K) towards its top two layers. These trends highlight why bulk samples, which could not be divided in macro-layers, have a lower average growth rate as they might only represent the most recent growth period. Hydrothermal/diagenetic deposits are characterised by higher growth rates than typical hydrogenetic crusts [75]. Our data set reveals values of 18.3, 38, 64, 64.7, and 102 mm/Ma, for samples DW2482, DW778D, DW778D2, DW4998D, and DW778B, respectively consistent with a hydrothermally or diagenetically influenced growth. Considering an apparent mean thickness of crust samples, it is possible to extrapolate a period of oxide accumulation, which represents the time it would have taken the crusts to form assuming no hiatuses, and could be assimilated with great care as a minimal age of initiation of growth. Assuming the surface of each sample represents present-day, the minimal age of initiation of growth ranges from 0. 79

Growth Rates and Ages
Estimated growth rates have been calculated using an empirical Co chronometer [66]. Growth rates vary from 0.45 to 102 mm/Ma ( Figure 8) and show no geographic correlations. Bulk hydrogenetic crusts (n = 74) have a mean growth rate of 2.2 ± 2.5 mm/Ma. Growth rates for macro-layers sub-samples (n = 24) are on average higher, at 3.1 ± 2.9 mm/Ma. Considering stratigraphic variations ( Figure 8B), most samples (DR14F, DR21F, DR38C, GO302D) present an increasing growth rate towards the most recent period, whereas only one crust exhibits a decreasing growth rate (DR19K) towards its top two layers. These trends highlight why bulk samples, which could not be divided in macrolayers, have a lower average growth rate as they might only represent the most recent growth period. Hydrothermal/diagenetic deposits are characterised by higher growth rates than typical hydrogenetic crusts [75]. Our data set reveals values of 18.3, 38, 64, 64.7, and 102 mm/Ma, for samples DW2482, DW778D, DW778D2, DW4998D, and DW778B, respectively consistent with a hydrothermally or diagenetically influenced growth. Considering an apparent mean thickness of crust samples, it is possible to extrapolate a period of oxide accumulation, which represents the time it would have taken the crusts to form assuming no hiatuses, and could be assimilated with great care as a minimal age of initiation of growth. Assuming the surface of each sample represents present-day, the minimal age of initiation of growth ranges from 0.79 to 34.3 Ma. It was not possible to obtain ages for samples showing no signs of stratigraphic polarity without substrate.

Element Correlations
A Pearson correlation coefficient matrix was calculated for the hydrogenetic macrolayers and bulk crusts (n = 89). Bulk samples that have been subsampled were not considered in the data set. In addition to 27 elements, the analysis contains growth rates and the Fe/Mn ratios (Table 5). Based on statistically significant (CL > 99%) correlations and iden-

Element Correlations
A Pearson correlation coefficient matrix was calculated for the hydrogenetic macrolayers and bulk crusts (n = 89). Bulk samples that have been subsampled were not considered in the data set. In addition to 27 elements, the analysis contains growth rates and the Fe/Mn ratios (Table 5). Based on statistically significant (CL > 99%) correlations and identified mineralogical phases by XRD, the statistical analysis reflects the major distribution of elements between four phases. A Mn oxide phase (δ-MnO 2 ) contains Mn, Mo, Sr, Tl, Pb, Co, La, Ni, V, As, Y, Nd, Nb, Zn, and an Fe-oxyhydroxide phase with Fe and Zr, whilst a biogenic phase accounts for Ba, Zn and Ce. Presence of calcite and fluorapatite identified in SEM images and XRD is consistent with the correlation of Ca, P and Pb. Aluminosilicate elements Si, Al, Na, K, Cu and Zr are negatively correlated with the δ-MnO 2 and Fe-oxyhydroxide phases, which is a common observation in other areas of the Pacific Ocean [1,67]. The distinction between a fluorapatite phase and a residual biogenic phase is not obvious. Only samples CP5069 and DN5080B contain fluorapatite, but no correlation has been found between Ca and principal biogenic elements (Ba, Ce and Zn). The weak correlation between Ni, P and Ca could be explained by the presence of 10 Å manganates associated with fluorapatite in CP5069.

Factor Analysis
A factor analysis was performed for the 89 hydrogenetic macro-layers and bulk crusts. Four significant factors explain 75% of the variance in the data set ( Figure 9). Factor 1 is interpreted as δ-MnO 2 and accounts for 45.2% of the variance, factor 2 as Fe-oxyhydroxides and accounts for 12.8%, factor 3 as a Fe (+As) dominated phase accounting for 10.3%, and factor 4 as a Ti phase accounting for 5.9%. For each factor, elements with the highest scores are: δ-MnO 2 : Sr, Mn, Pb, Mo, Co, Tl, V, As, La, Ni, Zn, Y, P, Nb; Fe-oxyhydroxides: Zr, Ce, Fe, Nd; Fe (+As) phase: negatively correlated to Cu and Ti; Ti phase: negatively correlated to Ba.

Factor Analysis
A factor analysis was performed for the 89 hydrogenetic macro-layers and crusts. Four significant factors explain 75% of the variance in the data set ( Figure 9). F 1 is interpreted as δ-MnO2 and accounts for 45.2% of the variance, factor 2 as Fe-ox droxides and accounts for 12.8%, factor 3 as a Fe (+As) dominated phase accountin 10.3%, and factor 4 as a Ti phase accounting for 5.9%. For each factor, elements wit highest scores are: δ-MnO2: Sr, Mn, Pb, Mo, Co, Tl, V, As, La, Ni, Zn, Y, P, Nb; Fe-ox droxides: Zr, Ce, Fe, Nd; Fe (+As) phase: negatively correlated to Cu and Ti; Ti p negatively correlated to Ba.  Mineral associations and phases determined using factor analysis present important discrepancies compared to phases obtained with correlation matrices (Table 5). A dominant δ-MnO 2 phase is found, presenting associations with common Mn-associated elements (Mn, Ba, Co, Mo, Ni, Zn) and elements from Fe-oxyhydroxides (Pb, As, Nb, Y), and includes elements partitioned between both groups: Tl, La, Nd, Sr and V. This phase is also characterised by a strong opposition to aluminosilicate elements such as Si, Al, Mg, Na, K, Cu and Zr. The Fe-oxyhydroxides phase is different; Zr and Ce are the main elements and Nd is correlated to this factor. Negative correlations are found with likely biogenic related elements, Ca, Mg and Ni. Using factor analysis, no fluorapatite or residual biogenic phases were detected. Fe (+As) phase shows stark anti-correlations with Ti and Cu, and weaker ones with Mn-associated elements. The last factor is difficult to identify because its only significant correlation is an anti-correlation with Ba. Weaker anti-correlations with Ce and Zn are also found, possibly indicating an opposition to a biogenic phase composed of Ba, Ce and Zn [1,4]. As Y is already significantly correlated to δ-MnO 2 , the main element positively correlated to this factor is Ti.

Discussion
According to their macroscopic features, as well as mineralogical and geochemical compositions, we can distinguish a group of hydrogenetic Fe-Mn crusts (98 of 104 samples) from a group of deposits presenting hydrothermal characteristics (six samples).

Comparison of New Caledonia's Fe-Mn Crusts with Other Oceans Deposits
Fe-Mn deposits are found in all oceans, covering different types of geomorphological settings and environments, reflecting a large panel of chemical compositions and morphologies. The physiography within the New Caledonian EEZ is complex and contains several ridges and seamounts where crusts are expected to be found. In order to compare New Caledonia's Fe-Mn crusts composition with crusts from elsewhere in the global ocean, Figure 10 compiles compositions of Fe-M crusts and nodules from several oceans (after [1,76,77]) compared to New Caledonia's Fe-Mn crusts (this study). New Caledonian crusts' concentrations were normalised to other oceans' crusts' concentrations, and presented as ratios.
The New Caledonian crusts mostly resemble crusts from the Indian and Atlantic oceans, showing a mean Fe/Mn ratio greater than 1.2 that is generally suggesting a mixed hydrogenetic and hydrothermal, or continental margin hydrogenetic origin [1]. Studies on Atlantic crusts pointed a significant enrichment in terrigenous component (Fe, Pb, Al, and Si) compared to the Pacific Ocean crusts due to fluvial and eolian input [78]. Similar enrichments are observed in New Caledonian crusts and could be also associated with strong terrigenous components. Low K concentrations could reflect the nature of terrigenous elements coming from New Caledonia since the Eocene obduction of mafic and ultramafic nappes [28,43,44].
Considering the 17 hydrogenetic Fe-Mn crusts samples analysed with ICP-MS (Table 4), the mean ΣREE in New Caledonia's crusts is 1307 ppm, whilst it is ranging between 2352 ppm and 2541 ppm for the Atlantic, Indian and North Pacific oceans (South Pacific value is closer with 1634 ppm) [12,70]. The mean percentage of HREE is slightly higher than the global oceans, with a value of 22% compared to 16% to 21%.
Compared to polymetallic nodules from the Clarion-Clipperton Zone (CCZ), the Peru Basin and the Indian Ocean, Fe, Ca, Ti, P, As, Co, Cr, Pb, Sr, V, Y, Zr, La, Ce and a majority of REE concentrations are higher in New Caledonia's crusts. This observation confirmed that the Fe-Mn crusts are dominantly hydrogenetic, contrasting with nodules where diagenetic processes can lead to higher concentrations in Cu, Ni, Zn, Al, K and Cd [1].   [1,76,77]). Values greater than 1 are enriched compared to other oceans crusts, whereas values lower than 1 are depleted. Fe/Mn* and Si/Al* ratios are calculated using mean ocean values.

Crusts Chemical Changes with Water Depth
Chemical changes in Fe-Mn crusts with water depth are a common phenomenon that has been identified in several studies [2,64,[79][80][81][82][83]. Fe-Mn crusts selected for this study range from 430 m to 4677 m and allow us to observe changes in chemical composition with water depth (Figure 11). Non-hydrogenetic deposits are illustrated on the graphs but are not considered for this analysis. Manganese shows a large range of values at shallow depths and decreases with depth. On the contrary, Fe is more stable along the water column, with values close to 20%. As a result, the Fe/Mn ratio increases from 1 to 1.5 between 1000 m and 3000 m, which is phenomenon observed in several other locations [81,83]. Silicon, Al, K and Na exhibit a net increase with depth, emphasising an increase of the aluminosilicate fraction in Fe-Mn crusts with depth. Increasing Fe, Si, Al, K and Na contents in crusts with depth are also observed in other oceans and are generally explained by an increased supply of detrital phases, and/or a weaker input of Mn due to the distance with the Mn-rich OMZ, whilst Fe increases with the dissolution of biogenic calcite [1,81]. Phosphorus and Ca in crusts decrease with water depth, possibly also representing the effect of the lysocline on carbonates and their continuous dissolution with increasing pressure at depth. Elements of economic interest Ni and Co show decreasing concentrations with depth, ranging from 7000 ppm and 5000 ppm at 1000 m, respectively, to 4000 ppm and 3000 ppm at 3000 m. This trend is correlated to the changing concentrations of the dissolved metals along the water column, with higher values around 1000 m and a marked decrease that tends to reduce with increasing depth below 2000 m [2]. This reflects the relationships of some trace metals with Mn, which is explained by an enhanced supply of dissolved Mn 2+ near the OMZ [1]. Other elements presenting decreasing trends with increasing water depth are Mo, Pb, Zn, As, Sr, Tl and V. Contrarily to other metals, Cu shows a slight increase with depth, with values ranging from 500 ppm at 1000 m, to 1200 ppm at 3000 m water depth. This increase in Cu content can be explained by its role in biogeochemical cycles, depleting its dissolve form in shallower waters whilst sinking organic particles progressively release it at depth [82]. Other elements such as Ti, Nd, Ba, Nb, Ce, Cr, Y and La show no particular trends, suggesting they are neither especially enriched nor depleted with depths in New Caledonian crusts.
Minerals 2022, 2, x 23 of 30 3000 m water depth. This increase in Cu content can be explained by its role in biogeochemical cycles, depleting its dissolve form in shallower waters whilst sinking organic particles progressively release it at depth [82]. Other elements such as Ti, Nd, Ba, Nb, Ce, Cr, Y and La show no particular trends, suggesting they are neither especially enriched nor depleted with depths in New Caledonian crusts. Crust thicknesses have been measured for every sample where the right-way up was evident and for crusts with or without substrate. Final values are a mean of six measurements performed along crusts widths. The distribution of measured thicknesses highlights the presence of thicker crusts below 2000 m despite higher sampling density at shallower depth. The low number of valid thickness measurements for samples below 3000 m in our data set prevents definitive interpretation of the trend as continuously increasing, constant or decreasing. It is important to note that all samples were dredged during cruises that were not dedicated to study crusts and that sample recovery is strongly impacted by seabed and outcrop morphology. Fe-Mn crusts recovered by ROV along depth transects usually show no variation in crust thicknesses [84,85]. The two thickest crusts (GO327: 115 mm, DR11Ai: 97 mm) are situated respectively at 1820 m and 2375 m. Figure 11. Repartition of selected elements with water depth. Black dots are the bulk crusts (n = 74), red dots are the macro-layers crusts (n = 24), and blue dots are the non-hydrogenetic deposits. Graph of crust thickness only considers bulk crusts with a measurable/known thicknes.

Nature of Non-Hydrogenetic Deposits
Based on macroscopic, mineralogical and geochemical characteristics, six samples from the data set are considered as non-hydrogenetic. DW2482, DW4998D and DW4998E Figure 11. Repartition of selected elements with water depth. Black dots are the bulk crusts (n = 74), red dots are the macro-layers crusts (n = 24), and blue dots are the non-hydrogenetic deposits. Graph of crust thickness only considers bulk crusts with a measurable/known thicknes. Crust thicknesses have been measured for every sample where the right-way up was evident and for crusts with or without substrate. Final values are a mean of six measurements performed along crusts widths. The distribution of measured thicknesses highlights the presence of thicker crusts below 2000 m despite higher sampling density at shallower depth. The low number of valid thickness measurements for samples below 3000 m in our data set prevents definitive interpretation of the trend as continuously increasing, constant or decreasing. It is important to note that all samples were dredged during cruises that were not dedicated to study crusts and that sample recovery is strongly impacted by seabed and outcrop morphology. Fe-Mn crusts recovered by ROV along depth transects usually show no variation in crust thicknesses [84,85]. The two thickest crusts (GO327: 115 mm, DR11Ai: 97 mm) are situated respectively at 1820 m and 2375 m.

Nature of Non-Hydrogenetic Deposits
Based on macroscopic, mineralogical and geochemical characteristics, six samples from the data set are considered as non-hydrogenetic. DW2482, DW4998D and DW4998E were dredged on the top and flanks of an Oligo-Miocene intraplate volcanic edifice on the Lord Howe Rise, part of a North/North-West oriented cluster of several seamounts ( Figure 2) [19,53]. DW778D, DW778D2 and DW778B were dredged on the summit of Mount K, a volcanic edifice of the Loyalty Ridge that is likely subduction related ( Figure 2B). The presence of both amorphous and crystalline 10 Å manganate (±pyrolusite), and the pseudolayered structure observed at the macro and microscopic scales in the DW778 samples, are consistent with other oceans hydrothermal deposits [11][12][13]86]. Hydrothermal deposits originating from ascending fluids are known to form in distal parts of the venting site in several geomorphological environments: back-arc basins [12,13], arc systems [11,69,87,88], or hot spot volcanoes [67,89,90]. They usually present a different mineralogy compared to hydrogenetic crusts, a strong partitioning between Fe and Mn, depleted trace metal contents (however, some deposits can exhibit notable enrichments in specific trace metals), and high growth rates [64].
Such specific characteristics are found in samples DW2482 and DW4998E that present high growth rates and unusual enrichments in Ni (up to 0.7%). Ni enrichment in hydrothermal deposits have been found in several places, such as the Yap volcanic arc [64], the submarine rift zones near Hawaii [91], or in the Wallis and Futuna back-arc system [9]. Same observations have also been pointed out by [19] in samples from the same group of volcanic seamounts on the Lord Howe Rise (samples DR01 were dredged from the same seamount as DW2482, DW4998D and DW4998E). In this study, Fe-Mn encrusted hyaloclastites and foraminifer-rich chalks are hydrothermally influenced, show strong Mn concentrations (up to 42.7%), low Fe concentrations (down to 0.53%), Ni enrichment (up to 0.65%), and globally depleted Co contents. These concentrations are similar to the ones we reported for DW2482, DW4998D and DW4998E, suggesting that they could possibly share the same origin.
Trace metal enrichments in these type of deposits are controlled by several parameters, such as the volume and type of leached rocks and sediments, the precipitation of sulfides at depth, the degree of partitioning between Mn and Fe, the amount of mixing with seawater and the distance from the vent sites [11,92]. It is likely that these parameters have influenced the formation of these samples, as notable differences in Ni, Ba, Zn and Cu concentrations between close dredging sites are observed. Based only on macroscopic description and XRD/XRF/SEM analyses of these six samples, it is not possible to decipher the origin of the Ni enrichment.

Resource Considerations
Hydrogenetic Fe-Mn crusts can be strongly enriched in rare and critical metals, such as Co, Te, Mo, Bi, Pt, W, Zr, Nb, Y and REE [12]. These concentrations make Fe-Mn crusts potential resources for metals used in high and green technology [93]. As mentioned in Section 5.2, New Caledonian Fe-Mn crusts show typical metal compositions of deposits formed in the vicinity of a continental mass with moderate enrichment in metals of economic interest compared to values of PCZ and higher detrital content. Cobalt and Ni, considered of greatest economic interest [12,71], are in the range of crusts from the Atlantic or Indian oceans, but less concentrated than in North and South-Pacific crusts. Data from this study indicates that the highest Co and Ni concentrations are located at water depth ranging from 1000 m to 2000 m, whilst crust thickness tends to be the highest around 2000 m (Figure 11). Correlation analysis showed that Co and Ni are mostly bound to the δ-MnO 2 phase, which also displays significant correlations with elements of economic interest like Mo and Nb (see part 4.6). The Co + Ni + Cu (%) content can reach high values in New Caledonia in sites close to the ridges and seamounts of the southern part of the EEZ. There, clusters of crust samples exhibit values reaching up to 1.62% (Figure 12). The physiography of this area is particularly favourable to Fe-Mn crust exploration; (i) several seamounts of the Loyalty and Norfolk ridges combine acceptable exploration and mine-site parameters, such as a seamount area larger than 400 km 2 ; (ii) water depth above 2500 m; and (iii) large areas with slope values between 0 • and 20 • [2,94]. A summary of Fe-Mn crusts resource assessment is proposed in Figure 12 using cri teria cited above and samples analysed in this study. Surface areas and slopes were calcu lated using ArcMap's 3D analyst, ArcGIS ® , from a 100 m scale bathymetric map [95]. Pol ygons of seamounts and ridges above 2500 m and presenting one or more samples wer hand-drawn following bathymetric and slopes variations. This map underlines severa zones of the Loyalty and Lord Howe ridges with large areas (up to 2045 km 2 ), a relativel flat top (slopes < 20°) and samples with noticeable enrichment in Co + Ni + Cu (%). How ever, crusts used in this study are the result of opportunistic sampling during scientifi cruises of various origins (mostly biology or geology). This makes it difficult to go beyond a first-order resource assessment. To further characterise this potential, exploratio should be conducted, notably through a detailed sampling strategy, seamount-scale bath ymetric and backscatter mappings, and the study of physicochemical properties and mo tion of water masses.  A summary of Fe-Mn crusts resource assessment is proposed in Figure 12 using criteria cited above and samples analysed in this study. Surface areas and slopes were calculated using ArcMap's 3D analyst, ArcGIS ® , from a 100 m scale bathymetric map [95]. Polygons of seamounts and ridges above 2500 m and presenting one or more samples were hand-drawn following bathymetric and slopes variations. This map underlines several zones of the Loyalty and Lord Howe ridges with large areas (up to 2045 km 2 ), a relatively flat top (slopes < 20 • ) and samples with noticeable enrichment in Co + Ni + Cu (%). However, crusts used in this study are the result of opportunistic sampling during scientific cruises of various origins (mostly biology or geology). This makes it difficult to go beyond a first-order resource assessment. To further characterise this potential, exploration should be conducted, notably through a detailed sampling strategy, seamount-scale bathymetric and backscatter mappings, and the study of physicochemical properties and motion of water masses.

Summary and Conclusions
(1) Several deposit styles were identified within the EEZ: a group of hydrogenetic crusts with chemical, textural and mineralogical characteristics similar to other hydrogenetic deposits found elsewhere in the ocean, and two groups of hydrothermal and diagenetic deposits located on the Loyalty and the Lord Howe ridges. Ridge exhibit wider chemical and mineralogical compositions (10 Å manganates ± pyrolusite), as well as a significant enrichment in Ni for two samples. (4) New Caledonia's hydrogenetic crust compositions are in the range of typical hydrogenetic Fe-Mn crusts. The mean combined concentration of metals with high economic potential Co + Ni + Cu is 0.81%, which is higher than Indian and Atlantic oceans, but lower than the Pacific Prime Crust Zone and the South Pacific Ocean. Several seamounts in the Southern part of the EEZ present clusters of Co + Ni + Cu values above 1%. (5) Further investigations will be needed to constrain more precisely the depositional settings of the hydrothermal/diagenetic samples, and the economic potential of hydrogenetic Fe-Mn crusts inside New Caledonia's EEZ. Author Contributions: Conceptualization, P.S. and J.C.; methodology, P.S., P.J. and E.P.; validation, J.C., P.J. and E.P.; formal analysis, P.S., P.J. and E.P.; resources, S.C., A.B. and Y.G.; writing-original draft preparation, P.S.; writing-review and editing, J.C., P.J., E.P., S.E. and M.P.; supervision, J.C., S.E. and E.P.; project administration, J.C., S.E. and E.P. All authors have read and agreed to the published version of the manuscript.
Funding: This research received no external funding.

Data Availability Statement:
The data presented in this study are available in the main body of the paper and in Supplementary Materials.