An example of ridge-proximal hydrothermal mineralisation: evidence from radiolarian chert-hosted Fe-Mn-oxide mineralisation within the İzmir-Ankara-Erzincan Neotethyan Ocean, Central Turkey

ABSTRACT Fe-Mn-oxide (oxyhydroxide) deposits are common within the late Cretaceous mélange units of the İzmir-Ankara-Erzincan Suture Zone, which marks the closure of part of the northern branch of the Neotethys Ocean in Turkey. In the Cihanpaşa region (Yozgat, central Turkey), Fe-Mn mineralisation occurs as interlayers within a sequence of mudstone and siliceous pelagic units that overlies altered pillow lavas belonging to the Artova ophiolitic complex. T-Fe-Mn mineralisations are characterised by moderate amounts of Ba (560 ppm), and low contents of Sr (243 ppm), Co (51 ppm), Zn (183 ppm), and Ni (221 ppm). Mn/Fe (0.69) and low Co/Ni (<0.40), Co/Zn (<0.50) and Y/Ho (25.7) ratios and ΣREE (233) values suggest that mineralisation formed in a submarine environment hydrothermal system. The low Ceanom. values (−0.60) are consistent with a mostly oxic depositional environment. Further, low LaSN/NdSN (<1), negative Ce anomaly (0.29), and low Al/(Al+Fe) (0.14), whereas high LaSN/CeSN (4.01) and Fe/Ti ratio (208) with moderately positive Eu anomaly (1.22), and Ti/V (1.99), LuSN/LaSN (0.89), ΣREE/Fe ratio (22.36 × 10–4) indicate that the Cihanpaşa Fe-Mn mineralisation formed in an oxygenised marine sedimentary environment with influence of a ridge-proximal (near spreading ridge) submarine hydrothermal vent system in the İzmir-Ankara-Erzincan Neotethyan Ocean. These results indicate that detailed geological and geochemical evidences determined in silicilastic rocks and submarine volcanics within mélange associations well-exposed along the suture belt could be applied to exploration studies for the discovery of ancient submarine deposits, including significant and precious metals such as polymetallic Fe-Mn deposits (Fe-Mn crust and nodules) and seafloor sulphide mineralisation.


Introduction
Diverse marine mineral deposits, such as hydrothermal Feand Mn-oxide deposits, ferromanganese nodules and crusts, and seafloor massive sulphide deposits have a range of different precious metal enrichments. These deposits are reported to occur in sedimentary basins of ancient and modern oceans throughout geological time Nicholson et al., 1997;Roy, 1997;and references therein), and are considered to be likely future global resources of metal deposits in modern oceanic settings (Petersen et al., 2016). Deep-sea metalliferous sedimentary rocks are important guide sequences in discovering metal enrichments away from the ridge axes of mid-ocean ridges, and also the largest metalliferous deposits which have been accumulated on many slowspreading ridges (Andersen et al., 2017). Owing to their element enrichment (REEs, Co, Ti, Hf, Nb, Ta, Zr), much focus has been placed on hydrothermal Fe and Mn-oxide deposits, principal markers of oceanic basins in modern mid-oceanic ridge settings (e.g. Roy 1 997;Hässig et al., 2013;Hein et al., 1997;Nicholson et al., 1997). These deposits are produced by precipitation from comparatively pure Mn-oxides and/or Fe-oxy(hydroxides) from ascending hydrothermal fluids (Hein, 2004;Hein et al., 2008), and occur in a variety of geological settings such as island arc systems, intraplate seamounts, volcanic chains, and mid-oceanic spreading centres where rate of sedimentation is greater than 1000 mm/Ma (Fitzgerald & Gillis, 2006;Prakash et al., 2012;Roy, 1997;Xiao et al., 2017). There, Fe-and Mn-oxide deposits occur as laminated and discrete dense strata-bound layers within beds.
The tectonic history of Turkey is primarily related to the consumption of Tethyan oceans during the evolution of the Alpine-Himalayan orogenic system, which marks the boundary between Gondwana to the south and Laurasia to the north (Şengör & Yılmaz, 1981;Okay & Tüysüz, 1999). At the southern margin of the Eurasian continent is concerned with the closure of Palaeo-Tethys and the opening of Neotethys oceans occurred during the Mesozoic. The geological signatures of Turkey are mainly linked to the continuing subduction, obduction and continental colli-represented by the ophiolitic remnants of the northern branch of the Neotethyan Ocean, namely the İzmir-Ankara-Erzincan Ocean. This zone comprises dismembered ophiolitic rocks, accretionary ophiolitic mélange complexes, metamorphic sole rocks, and locally welldeveloped island-arc sequences (Göncüoğlu et al., 2006a;Çelik et al., 2011;Sarıfakıoğlu et al., 2017;and references therein). The ophiolites occur mainly as peridotite massifs, lacking a complete ophiolitic sequence. In NW Anatolia, sub-ophiolite metamorphic rocks indicate Albian-Cenomanian ages for the intra-oceanic subduction (Önen and Hall, 2000).
The radiolarian cherts and pure cherts have been considered to be tectonically dispersed parts of the deep-sea sedimentary cover of a paleo-oceanic plate deposited at different distances from the seafloor spreading centre (Kemkin & Filippov, 2001). After the closure of the oceanic basins, different types of marine mineral deposits on the ancient seafloor outcrop within ophiolitic mélange associations in the form of large or small lenses. It is significant to shed light on the questions of whether Fe-and Mn ore deposits were formed far or near from the oceanic spreading ridge. Therefore, their discovery on land can be determined via a detailed overview of ophiolitic melange bodies belonging to ancient oceanic basin. Hence, the present study is an example of ridge-proximal hydrothermal mineralisation formed near the ridge of the İzmir-Ankara-Erzincan Neotethyan ocean. It documents the geochemical characteristics of these deposits, discusses their origin, elaborates their relationship with the hydrothermal processes along a mid-oceanic ridge and compares them with other Fe-and Mn-oxide mineralisations within the İzmir-Ankara-Erzincan suture zone.
Close to the studied area, the oldest radiolarian cherts associated with OIB-type basaltic rocks are Late Triassic (late Norian). These basaltic rocks are related to cherts of Late Jurassic (middle-late Oxfordian to late Kimmeridgian-early Tithonian) and Early Cretaceous (late Valanginian to late Hauterivian) ages (Bortolotti et al., 2013). E-MORB type rocks in other blocks are related to radiolarian cherts of Cretaceous age (middle late Barremian-early early Aptian). Although concentrated in Valanginian to middle Aptianearly Albian time, the age of N-MORB type rocks associated with radiolarian cherts ranges between Late Jurassic (early-early late Tithonian) and Early Cretaceous (late Valanginian-early Barremian). P-MORB type rocks yielded Middle Jurassic ages (early-middle Bajocian to late Bathonian-early Callovian age) (Bortolotti et al., 2013). Also, Bortolotti et al. (2018) described blocks comprising oceanic lavas (e.g. garnet-influenced MORB, enriched-MORB, plume-type MORB and alkaline basalts) associated with pelagic sediments just in the western part of the Artova Ophiolitic Mélange. The Tekelidağ Mélange, in the eastern extension of the Artova Ophiolitic Mélange, was studied in detail by Özkan et al. (2020) who obtained Middle Jurassic to Early Cretaceous ages from radiolarian assemblages in the cherts coexisting with the basaltic rocks. The first group of radiolarian cherts observed within the MORB-type basalts yielded Late Bajocian (Middle Jurassic) to Tithonian (Late Jurassic) ages. The second group associated with supra-subduction zone-type basalts yielded upper Aalenian (Middle Jurassic) and lower Aptian (Lower Cretaceous) ages (Özkan et al., 2020). Recently, Akçay and Beyazpirinç (2017) reported Late Tithonian-Berriasian and late Albian-early Cenomanian foraminifer ages from the Artova Ophiolitic Mélange, albeit from the epiophiolitic cover of the mélange. In brief, the depositional ages of the studied radiolarian cherts hosting Fe-Mn-oxides within the immediate supra-Artova Ophiolitic Mélange ranges from late-(Middle-late Oxfordian to late Kimmeridgian-early Tithonian)-middle Jurassic (Early-middle Bajocian to late Bathonian-early Callovian) to Early Cretaceous (Middle late Barremian-early early Aptian) (Bortolotti et al., 2018).
The Fe-Mn-oxide mineralisations were studied mainly in the Cihanpaşa Region (northern Yozgat, Central Anatolia) and its surroundings ( Figure 2) where the Artova Ophiolitic Mélange tectonically overthrust the Darmik Formation (Senomanian-Maastrichtian). The Upper Palaeocene-Middle Eocene Boğazköy Formation, consisting dominantly of volcano-sedimentary rocks (Akçay & Beyazpirinç, 2017), is overthrust by the Darmik Formation ( Figure 2). The host rocks (radiolarian chert) in the study area are commonly observed as blocks in a serpentinite matrix of mélange on dominant hills (e.g. Otluk, Eyerci) because of their resistance to alteration, and rarely in valleys ( Figure 2). Thin-bedded (2 mm thick, rarely up to 10 cm) and laminated radiolarian cherts have the original depositional structures and flat or less frequently undulating bedding surfaces (Figure 3a). Cherts are interbedded with mm-to cm-thick beds of the radiolarites. The deposit is located approximately 30 km to the N of the Yozgat district (Otluk area 39°57ʹ59.26"N, 34° 47ʹ38.68"E and Eyerci Hill 39°57ʹ47.00"N, 34°48ʹ19.87"E) ( Figure 2). Fe-Mn-oxides in the region occur as massive and interlayered within host rocks (Figure 3b, c), and are characterised by folding and fracturing. The contact between the Fe-Mn ores and host rock is rarely sharp (Figure 3d). In the Cihanpaşa region in the northern section of the study area, some massive Mn-ores with host rocks are inverted. Economically, these important Mn-ores have been previously mined by private companies as underground mines within the radiolarian chert layers around Cihanpaşa village (Aydoğan, 2022). The studied Fe-Mn-oxide mineralisation is hosted by violet-reddish radiolarian chert assemblages tectonically overlying highly serpentinised ultramafic rocks and mafic lavas of the Artova Ophiolitic Mélange. Altered mafic lavas exposed within the ophiolitic mélange host some sulphide-rich mineralisation (Figure 3e, f).
Mineralogically, the host rocks comprise dominantly silica-rich radiolarian, in which radiolarians are recrystallised to microcrystalline quartz, but some are filled with chalcedony ( Figure 4a, b, c). The matrix is composed mainly of cryptocrystalline quartz. Moderately preserved radiolarians are filled with quartz ( Figure 4e). Rocks are impregnated by Fe-Mn-oxides. Fe-Mn-ores have dominant mineral assemblages of braunite, haematite, goethite, quartz, and rarely pyrolusite based on XRD and petrographic studies (Figure 4f, g). Goethite is determined by XRD. Haematite is commonly observed as disseminated grains and veinlets in the host rock. Quartz is found as crystocrystalline quartz in the matrix, and vein/veinlets. XRD data of two samples are given in the Supplementary data.

Analytical methods
A total of eight samples were collected from bottom to top of the Fe-Mn ores from two different areas in the Cihanpaşa region (Otluk area and Eyerci Hill) (see Figure  2). The geographical co-ordinates of the sampling sites are given in Table 1. Mineralogical identification and textural relations of the Fe-Mn ores were studied by optical examination of polished-, thin sections and X-Ray powder diffraction (XRD). XRD analyses were performed with BRUKER D8 with a Cu anode and Ni filtration operating at a generator voltage of 40 kV and a current of 40 mA; 2θ scans varied from 2° to 72° 2-theta. Analyses were carried out at the İstanbul Technical University (Turkey), and samples for analysis were powdered to less than 100 µm. Polishedsections were prepared in 3 × 2 × 1 cm size of Fe-Mn ore samples. These sections were observed under the Meiji MT-9930 microscope at the University of Balikesir.
At first, all samples were selected using a binocular microscope. Then, selected samples were given to the ACME Analytical Laboratories (Vancouver, Canada) where they were examined. Samples were ground finer than 700 mesh for the chemical analyses. The major oxides (SiO 2 , TiO 2 , Al 2 O 3 , MnO, MgO, CaO, K 2 O, Na 2 O, P 2 O 5 ) and LOI of the samples were analysed by inductively coupled plasma-atomic emission spectrometry (ICP-AES) following lithium metaborate/tetraborate fusions following dilute nitric acid digestion. The trace elements were analysed by inductively coupled plasmamass spectrometry (ICP-MS), with rare earth and incompatible elements determined following LiBO 2 fusion, and precious and base metals determined from an aqua-regia digestion. After ignition at 1000°C, loss-on-ignition (LOI) was calculated by weight difference. The detection limits are usually 0.1-8 ppm for trace elements, 0.01-0.1 wt.% for major elements, and 0.1-0.01 ppm for REEs. For evaluation of the accuracy of analytical data, the geochemical standard STD SO-18, STD DS7 and STD CSC were used. The quality of analyses has been studied via a range of reference materials (http://acmelab.com/services/ method-descriptions/soiltill-and-sediment/). Shale-normalised REY patterns were obtained using the standard values given for Post-Archaean average Australian Shale (McLennan, 1989). Standard values of the PAAS (Post-Archaean average Austuralian Shale, McLennan, 1989) and NASC (North America Shale Composite, Gromet et al 1984) were used in the present study. For calculating value of Ce anomaly, PAAS-and NASC-normalised values were determined by the following equation: Ce/Ce* = Ce SN / (0.5La SN + 0.5Pr SN ) (Bau et al., 2014). Pr anomaly was calculated as [(Pr/Pr*) PAAS = Pr PAAS /(0.5Ce PAAS + 0.5Nd PAAS )] according to Bau et al. (1996a, b). Ce anomaly (Ce anom. ) = log [3 × Ce SN /(2× La SN + Nd SN )] (shalenormalised McLennan, 1989) was proposed by Wright et al. (1987). Eu anomaly was identified as Eu/Eu*, Eu PAAS /(0.67 × Sm PAAS + 0.33 × Tb PAAS ) (Bau & Dulski, 1996;Bolhar & Kranendonk, 2007).

Major elements
The bulk geochemical composition of eight representative samples of the Cihanpaşa Fe-Mn-oxide mineralisations are given in Table 1. In Fe-Mn ores, the content of Fe 2 O 3tot ranges from 11.29 wt% to 30.01 wt%; with an average of 16.05 wt%). The MnO concentrations of studied samples vary from 3.90 wt% to 17.14 wt%, with an average of 9.39 wt%. The concentration of SiO 2 varies between 42.51 and 75.39, with an average of 64.60 wt%. The studied samples are characterised by low Al 2 O 3 (1.79 wt% to 3.23 wt%; with mean of 2.42 wt%), and TiO 2 (0.05 wt% to 0.14 wt%; with mean of 0.08 wt%). The average contents of K 2 O (0.31 wt%), Na 2 O (0.05 wt%), CaO (0.62 wt%) and P 2 O 5 (0.30 wt%) are minimal. LOI (loss on ignition) ranges from 4 wt% to 8.80 wt% (avg: 5.54 wt%) ( Figure 5a).

Rare earth elements
The total concentrations of ΣREE in the Cihanpaşa Fe-Mn ores varies between 128 and 320 ppm, with an average value of 233 ppm (Table 1, Figure 5b). Compared with the hydrogenous deposits (ΣREE>1000 ppm), the hydrothermal Mn-oxides have low REE-compositions (ΣREE<100 ppm) (Prakash et al., 2012). These results are slightly higher than those of hydrothermal deposits, but much lower than hydrogenetic ones, and they have low LREE contents. A strongly positive correlation (r ΣREE-LREE = 0.99) between ΣREE and LREE concentrations in the studied rocks reveals the preferential uptake of REEs during the ore formation in numerous Fe-and Mn-oxide deposits extending along the Neotethyan suture (Zarasvandi et al., 2016;Öksüz, 2011a, b).

Element correlations
The modes of occurrence of elements in the Cihanpaşa Fe-Mn mineralisations were investigated using cluster analysis and correlation coefficients. Element correlations in the mineralisation are indicated by the clustering diagram ( Figure 5c). Cluster analysis showed a strong negative correlation between Si and Fe+Mn (r Si-Fe+Mn = −0.99), reflecting the addition of deep-sea silicious pelagic sediments of biogenic origin. The correlation coefficient between Ca and P (r Ca-P = 0.95) with yttrium suggests phosphatic occurrences in the sedimentary environment.
A correlation between Cs and Rb (r Cs-Rb = 0.96) together with K (r K-Cs = 0.89, r K-Rb = 0.85) indicates a fraction of clay-rich particles. A weak correlation coefficient of Mn-Pb (r Mn-Pb = 0.64), Mo-Sb (r Mo-Sb = 0.54) and Ba-Ga (r Ba-Ga = 0.62) may indicate similar emplacement behaviour. As seen in Figure 5c, it should be noted that Fe correlated with Cu+Co+Ni (r Fe-Cu+Co+Ni = 0.94) with Zn. This indicates that Cihanpaşa Fe and Mn mineralisations are related to the mid-oceanic fluids near a spreading . Photomicrographs of thin-and polished-sections. a radiolarian-rich host rock and Fe-Mn ore. All radiolarians are recrystallised to quartz, b Reddish-brownish slightly bioturbated radiolarite and quartz veinlets, c radiolarian test filled with quartz in the inner parts, d sharp contact between radiolarian chert and Fe-Mn ore, e Some radiolarians filled by quartz, f-g Fe-Mn ore consisting of haematite (hem), pyrolusite (pst), braunite (brt) and quartz (qtz) within the host rock.  ridge. There is also a strong correlation coefficient between Mg and V (r Mg-V = 0.86). Both elements are enriched in mafic precursors, and may be connected to the alteration of basaltic oceanic crust with a ridge-proximal hydrothermal input. Moreover, the presence of mafic volcanic rocks (pillow lavas) in the vicinity of the Cihanpaşa region provides further evidence for a nearby source for the hydrothermal exhalations (proximal hydrothermal source) (Aydoğan, 2022).

REE + Y signatures
A PAAS-normalised diagram shows that REE-Y concentrations of the Cihanpaşa Fe-Mn ores are lower than hydrogenetic ones, and also with the imprint of a negative Ce anomaly and positive Eu (1.17 to 1.25) and Gd (1.16 to 1.30) anomalies (Figure 6a)  They have also a positive Gd and negative Ce anomaly, similar to low T-hydrothermal deposits, but show a different pattern from high T-hydrothermal vent fluids ( Figure  6a). Normalising to shale composites, oxygenised modern seawater displays a remarkable negative Ce anomaly, whereas suboxic and anoxic waters lack negative Ce signatures (Byrne & Sholkovitz, 1996;German & Elderfield, 1990). Interpretation of the presence or lack of a Ce anomaly is often complex due to the anomalous behaviour of La (Bau & Dulski, 1996). Earlier studies Bolhar & Kranendonk, 2007;Zhang et al., 1994) indicate that modern seawater shows characteristics of positive La, Y, Gd and negative Ce anomalies, LREE depletion, HREE enrichment and also hydrothermal Fe-Mn sediments on oceanic crust. The Cihanpaşa Fe-Mn ore samples reflect positive La and pronounced negative Ce anomalies ( Figure 6a). Bau et al. (1996a, b) introduced the use of the Ce SN /Ce SN * vs. Pr SN /Pr SN * plot to discriminate the 'true' from 'false' negative Ce (SN) anomalies in Fe-Mn-oxides due to the possible anomalous behaviour of La. In this diagram, all of the Cihanpaşa Fe-Mn ore samples plot in field IIIb (true negative Ce anomaly, Figure 6b).

Discussion
Marine Fe-Mn oxides (hydroxides) deposited by chemical and sedimentary processes from ambient seawater are mainly characterised by various subtypes such as hydrothermal (diffuse-flow), hydrogenous and diagenetic based upon their mineralogy, composition and depositional sites (Bolton et al., 1988;Hein et al., 1997Hein et al., , 2008  al., 2012). In the formation of these deposits, the effect of volcanic influence, hydrothermal activity and terrigenous fluxes plays a key role in the submarine environment (Gao et al., 2018). Some elements (e.g. Nb, Sc, Th, Al, Ti, Y, and Zr) are relatively insoluble in seawater at surface temperatures, and also tend not to be integrated with geochemical processes in ocean water (Boström, 1970;Gurvich, 2006). Therefore, some ratios of these elements can be extensively monitored as a specific indicator to show the occurrence of some effects in the origin of the Fe-Mn deposits (Maynard, 2010;Mohapatra et al., 2009;Zarasvandi et al., 2016). Al 3+ and Ti 4+ are considered as proxies of terrigenous fluxes (Crear et al., 1982;Maynard, 2010;Murray & Leinen, 1996;Tribovillard et al., 2006). Remarkably, sedimentary contribution during the deposition of Fe-Mn ores can be explained by high concentrations of Al (Choi & Hariya, 1992). Also, high Al concentrations indicate an excessive amount of detrital input to the deposition site (Ganno et al., 2017). The compositions of both elements are used in the recognition of significant effects (e.g. submarine volcanism, hydrothermal activity, detrital phases) on ferrigenous and/or manganiferous units in submarine systems (Crear et al., 1982). The Al 2 O 3 /TiO 2 ratio helpful in identifying these effects as in sedimentary rocks it is less affected during weathering, diagenesis and sedimentary transportation (Hayashi et al., 1997;Sugitani, 1996). Previous investigations have revealed that the Al 2 O 3 /TiO 2 values of mafic volcanic units range from 3 to 14 (Hayashi et al., 1997;Sugitani, 1996) Bau and Dulski (1999), b Ce SN /Ce SN * versus Pr SN /Pr SN * discrimination diagram for Ce and Pr anomalies (Bau & Dulski, 1996). Note that all analysed samples of the Cihanpaşa Fe-Mn ores display true negative Ce anomalies, indicative of oxic conditions. Fields for modern deep and shallow oxic seawater from Alibo and Nozaki (1999)  In this study, some classical and new discrimination diagrams were used to reveal the genetic type of Fe-Mn ores from the Cihanpaşa region. Hydrothermal deposits have low concentrations of trace elements (Cu, Ni, Co (<300 ppm)) Ingram et al., 1990), whereas hydrogenetic ones contain the highest Cu, Co, Ni values. Bonatti (1975) argued that both SiO 2 and Al 2 O 3 are indicative for assessing the origin of Mn-bearing sediments. A SiO 2 vs. Al 2 O 3 Figure 7. Bivariate and ternary diagrams discriminating the hydrothermal components of the Cihanpaşa region. a SiO 2 vs. Al 2 O 3 discrimination diagram indicating the hydrothermal affinity of the studied Fe-Mn samples (Bonatti et al., 1972;Toth, 1980). b Co +Ni (wt%) vs. As+Cu+Mo+Pb+V+ Zn (wt%) plot showing supergene and hydrothermal Mn-oxides (Nicholson, 1992). c (Co + Ni + Cu) x 10-Fe-Mn ternary diagram (Bonatti et al., 1972). d 15*(Cu+Ni)-100*(Zr+Ce+Y)-(Mn+Fe)/4 ternary diagram (Josso et al., 2017). e (Ce SN /Ce SN *) vs Nd (ppm) and f Ce SN /Ce SN * vs Y SN /Ho SN diagrams illustrating discrimination between different genetic types of Fe-Mn deposits (Bau et al., 2014). diagram shows that all Cihanpaşa Fe-Mn ore samples are hydrothermal in origin (Figure 7a). Hydrothermal deposits have notably low concentrations of Co and Ni in compared with hydrogenous ones. The studied Fe-Mn ores have low contents of Co and Ni (Table 1). Therefore, a Co+Ni (wt.%) vs. As+Cu+Mo+Pb+V+ Zn (wt.%) diagram reveals distinctions between hydrogenetic, diagenetic, or supergene and hydrothermal precipitation depending on some element mobility in hydrothermal solutions (Nicholson, 1992). This diagram shows that the studied Fe-Mn ores plot within the field of hydrothermal Mn-oxides (Figure 7b). Likewise, the Fe-Mn-(Ni + Co + Cu)*10 triangular diagram of Bonatti et al. (1972) indicates that all the studied samples fall within the field of hydrothermal origin (Figure 7c). However, Josso et al. (2017) suggested that these classical diagrams do not precisely discriminate Fe-Mn deposits formed by especially mixed genetic processes, i.e. nodules with different hydrogenetic/diagenetic quantities. Alternative classification schemes have been discussed recently for the discrimination of Fe-Mn deposits by Josso et al. (2017). The ternary 15*(Cu+Ni)-100*(Zr+Ce+Y)-(Mn+Fe)/4 classification diagram reveals that the Cihanpaşa Fe-Mn ores plot towards a mixed-type origin between hydrothermal and hydrogenetic field (Figure 7d). Similarly, binary diagrams Y SN /Ho SN vs. Nd (Figure 7e) and Ce SN / Ce SN * vs. Y SN /Ho SN (Figure 7f) proposed by Bau et al. (2014) reflect submarine hydrogenetic, diagenetic, and hydrothermal Fe-Mn deposits. In both diagrams, the Cihanpaşa Fe-Mn samples plot within both the diagenetic and hydrothermal fields. Compared with classical diagrams, newly proposed diagrams show that the studied Fe-Mn ores are of mixed-type in genetic origin indicating that it was diagenetic and hydrothermal or hydrogeneous/hydrothermal. La/Ce and Y/Ho ratios provide important data which reveal the origin of hydrothermal activities in a submarine environment. The La/Ce values of hydrothermal ferromanganese crust are close to that of seawater (~2.8). La/Ce ratios of the ferromanganese crusts from the Juan de Fuca ridge, the East Pacific Rise and mid-Atlantic Ridge have low values (<1), suggesting the relative enrichment of Ce by adsorption (Yang et al., 1981). Nath et al. (1997) revealed that La/Ce ratio is low (~0.25) in hydrogenetic Fe-Mn crusts and pelagic sediments (~0.61, Toyoda et al., 1990). The La/Ce ratio of marine hydrothermal Mn deposits in the NW Pacific has an average of 1.67 (Usui & Someya, 1997). The La/ Ce ratios of the studied Fe-Mn ores ranging between 0.65 and 3.07 with average of 1.82 show values close to those of marine hydrothermal deposits in the NW  (Kılıç et al., 2018), Kasımağa Mn (Koç et al., 2000), Büyükmahal Fe-Mn (Öksüz & Okuyucu, 2014) deposits from Turkey and some materials of the Earth (Bau & Dulski, 1999;Bolhar & Kranendonk, 2007;Nozaki et al., 1999)  Pacific, and are quite different from hydrogenetic ferromanganese crusts. Yttrium is very similar to REE, especially, Ho (Bau & Dulski, 1999). The behaviour of Y +3 is similar to Ho +3 as both elements have very similar to identical effective ionic radii, charge density and trivalent oxidation state (Bau et al., 1996a, b;Bau & Dulski, 1999). Considering that the hydrothermal solutions do not have superchondiritic Y/Ho ratios (near chondritic with 27; Douville et al., 1999), this ratio is superchondritic for seawater (Bau & Dulski, 1999). Bau and Dulski (1999) claimed that there is no significant fractionation of Y and Ho during high-temperature basalt and seawater interaction. Also, the Y/Ho ratio is a strong indicator in defining Mn source (Hein et al., 1999). The Y/Ho ratio values of the Cihanpaşa Fe-Mn ores are similar to those of some Fe-and Mn deposits in Turkey. The Y/Ho values of the Büyükmahal Fe-Mn deposits (Central Turkey) vary between 13.06 and 30, with an average of 25.3 (Öksüz & Okuyucu, 2014); the Y/Ho ratios of the Eymir Mn deposit (Central Turkey) range from 20 to 51.82, with an average of 34.7 (Öksüz, 2011a); the Kasımağa Fe-Mn deposit (Central Turkey) has Y/Ho ratios ranging from 20.9 to 90, with an average of 48.8 (Koç et al., 2000); the Y/Ho ratio of Kula Mn deposit (Kula, western Turkey) ranges between 23 and 41, with an average of 29. The studied Fe-Mn ore samples show Y/Ho ratios ranging 22.04 to 32.20 (avg. 25.73). These values fall among ratios of midoceanic ridge hydrothermal fluids (Y/Ho: 28-30; Bau & Dulski, 1999;Bolhar & Kranendonk, 2007;, PAAS (McLennan, 1989, Y/Ho = 27.25), compared to chondrite (Sun & McDonough, 1989, Y/ Ho = 27.74) (Figure 8a). Transition metals (Cu, Co, Ni) with Fe are characteristically enriched in mid-ocean ridge fluids (Frei and Polat, 2007). Supporting evidence for this is that there is a strong positive correlation between Fe and metals (Cu+Ni+Co) in ore samples, and their correlation coefficient was 0.97. This indicates that the Cihanpaşa Fe-Mn mineralisations are closely associated with hydrothermal fluids in the mid-ocean ridge of the İzmir-Ankara-Erzincan ocean (Figure 8b). Meanwhile, a Zr vs. Y/Ho diagram (Figure 8c) for the studied Fe-Mn ores shows that they were not affected by terrigenous clastic input (Men et al., 2020).
In understanding the ambient conditions of a submarine environment, redox elements are critical for a correct paleoceanographic interpretation. Fe and Mn deposits were recognised in many hydrothermal systems of modern oceans, being best expressed in the low-temperature discharge sites. Both Fe and Mn display redox-sensitive behaviour in a sedimentary environment (Naeher et al., 2013). Thus, the Fe/Mn ratio of various genetic types is generally used to reveal the redox state. Fe/Mn ratios ranging 0.67 to 2.89 in the Cihanpaşa Fe-Mn ore samples indicate that these mineralisations formed in an oxidising sedimentary environment. In the sedimentary environments, Eh and pH conditions play an important role in determining the fractionation between Fe and Mn in deposition. At lower Eh, iron precipitates as iron sulphides due to lower solubility. In this context, iron sulphides are formed by the reaction of iron oxides/hydroxides with sulphides. However, Mn solubility is high and there is no comparable insoluble manganese sulphide, so it remains in solution. Subsequent decrease in Eh and/or pH in the depositional environment may result in manganesedepleted and iron-enriched occurrences.
In sedimentary environments, the distribution of some REEs (e.g. Ce, Eu) is very informative in relation to the geochemical distribution of other elements and the sediment deposition site (redox state). Of these, Ce is sensitive to oxygen fugacity, and is generally an indicator of the depositional location in modern and ancient Fe-Mn-oxide Table 2. REE indices of Jurassic chert assemblages (Franciscan Terrane, California) formed in different depositional environments (from Murray et al., 1991) and some Late Cretaceous radiolarian-chert-hosted Fe-Mn deposits in an ophiolitic melange complex belonging to the İzmir-Ankara-Erzincan ocean. deposits (Murray et al.1990;Wright et al., 1987;Piepgras & Jacobsen, 1992;Owen et al., 1999;Chen et al., 2006). Importantly, it has been used as a sensitive tracer for paleoredox conditions because of its two possible oxidation states (Ce 3+ and Ce 4+ ) (German & Elderfield, 1990;Wright et al., 1987). Ce 4+ is scavenged by Fe-Mn oxides leading to depletion of Ce in oxic seawater in an oxygenised marine sedimentary environment (Elderfield & Greaves, 1982). The reduction of Ce 4+ to Ce 3+ and the oxidation of Ce 3+ to Ce 4+ correspond to changes in the redox conditions (De Baar et al., 1988;German & Elderfield, 1990;Byrne & Sholkovitz, 1996;German et al., 1991a, b;Kato et al., 2006). According to Gadd et al (2016), the Ce anomaly is ~ 1 in surface water; but it is depleted by the reactive Fe-Mn oxide/oxyhydroxide particles in deeper areas of the sea. A positive Ce anomaly (Ce/Ce* > 1) because of gain of Ce 4+ by these particles is observed; however, a negative Ce anomaly (<1) results from the autogenic marine processes precipitating from Ce-poor seawater (Gadd et al. 2016). Also, the lack of a Ce anomaly presumably signals the influence of seawater on the REE composition, and it is indicator of precipitation from oxygenised seawater (Piepgras & Jacobsen, 1992). Wright et al. (1987) proposed the Ce distinction (Ce anom. ) as an index of an oxic or anoxic nature in ancient sedimentary environments. When the Ce anom. is lower than −0.1, it reflects oxidation-reduction conditions, while if the Ce anom. is higher than −0.1, it marks an anoxic nature during deposition. As shown in Table 1, the Ce anom. values (avg. −0.60) and Nd concentrations (with average of 66.6 ppm) of ore samples demonstrates that they are characterised by Ce anom. < −0.1, indicative of an oxic environment with a moderate depositional rate based on Nd concentrations (Wright et al., 1987). Murray et al. (1990Murray et al. ( , 1991Murray et al. ( , 1992 highlighted that shale-normalised Ce anomalies on the bedded chert series and shale from the Franciscan Terrane of California have a Ce value of ~0.3 for Lower Jurassic samples ranging up to a Ce value of ~1.0 for Middle Cretaceous samples. They argued that multiplicities in REEs of cherts and shales can reveal their depositional site. In this context, Ce values of chert and shale, formed in different marine environments of deposition, suggest that three depositional regimes can be identified: i) proximal to a spreading ridge [(within 400 km, 0.18-0.6, avg. 0.29)], ii) an ocean-basin floor setting or pelagic domains (0.5-0.76, avg. 0.60), and iii) continental margin regimes [(>2800 km, ~0.67-1.52, avg. 1.11) ( Table 2)]. Thus, the low Ce-anomaly ranging from 0.16 to 0.64 (average, 0.31) of the Cihanpaşa Fe-Mn ores indicates that these mineralisations formed near a spreading centre of the İzmir-Ankara Neotethyan Ocean during the Late Cretaceous.
Deep-sea hydrothermal venting systems driven by magmatic heat sources on the seafloor spreading axes are prominently observed in mid-ocean ridges and basins near volcanic island arcs. The fluids resulting from these vents exhibit a wide variety of compositions (Sander & Koschinsky, 2011). The rare-earth-element compositions of vent fluids strongly can yield powerful information as a tracer of sub-seafloor hydrothermal vent fluids reflecting oceanic basic crustal alteration by high-T hydrothermal fluids (Craddock et al., 2010). In this context, the Eu anomaly could be used to identify the high-T or low-T hydrothermal alteration of the basaltic oceanic crust (Douville et al., 1999;A Michard et al., 1983;Mitra et al., 1994). The presence of either negative or positive Eu anomalies is significative in determining the input of hydrothermal activities, due to the fact that Eu abundance in seafloor hydrothermal vent fluids is associated with solution temperature. In general, high-T (>350°C) hydrothermal fluids typically developed in mid-ocean ridges and back-arc spreading centres display positive Eu anomalies; nevertheless, weakly positive or negative Eu anomalies are typical of low-T (<350°C) hydrothermal system (Bau & Dulski, 1999;Bolhar & Kranendonk, 2007). Additionally, the weakly negative Eu anomaly (0.63 to 0.72) is typical of modern oxygenised deep-sea waters (>2000 m) (Kamber, 2010;Kato et al., 2006). In contrast, high-T hydrothermal fluids vented along spreading ridges have strong Eu anomaly of up to ∼ 70 (Douville et al., 1999;Schmidt et al., 2011Schmidt et al., , 2007, and the most markedly positive Eu anomalies are related to the hydrothermal systems in slow spreading ridges, especially in ultramafic rocks (Bau & Dulski, 1999;Schmidt et al., 2011Schmidt et al., , 2007. High-T black smoker fluids have the largest Eu anomalies (7.2 to 15.1) compared with the deep-sea waters (Kato et al., 2006). Depending on the thermal regime and geotectonic setting in modern marine environments, the noticeable positive Eu anomaly is observed only at mid-ocean ridges and back-arc spreading centres, where alteration of oceanic crust (seafloor basalts or mafic rocks) contributes to REE. However, a slight or no Eu anomaly indicates submarine hydrothermal fluids produced by low-T hydrothermal alteration (Bau & Dulski, 1996A Michard et al., 1983). The Cihanpaşa Fe-Mn-oxide ores have some input of low-T hydrothermal fluids, as suggested by moderately positive Eu anomalies (1.17-1.25, with an average of 1.22) with respect to being distant from the hydrothermal source and depositional location. It is noteworthy that mixing of seawater and submarine hydrothermal fluid is characterised by the positive Eu, La and Y anomaly and light REE depletion, and heavy REE enrichment (Bolton et al., 1988;Men et al., 2020). Pioonering studies of Sugitani (1992) proposed that (La/Yb) SN ratios >1 indicate a hydrothermal fluid with greater LREE enrichment relative to HREE, whereas (La/Yb) SN ratios <1 reflect the mixing of seawater and hydrothermal fluids. In the Cihanpaşa Fe-Mn ores, (La/Yb) SN ratios ranging between 0.60 and 1.02, with an average of 0.84 indicate mixing of seawater and hydrothermal fluids.
As mentioned before, uncontaminated hydrothermal deposits have very little Al. Contamination of such deposits in a deep-sea environment by the input of pelagic and terrigenous sediments increases their content of Ti and Al. Therefore, their Fe/Ti ratios shows considerably decrease, while their Al/(Al+Fe+Mn) ratio increases (Ganno et al., 2017). Boström and Peterson (1969) emphasised that the Al/(Al+Fe+Mn) value is an indicator of detrital sediment sources and the relative contribution of submarine hydrothermal environments, and this ratio also shows a gradual decrease with decreasing distance from spreading centre. An Al/ (Al + Fe + Mn) >0.4 is thought to reflect a detrital source in marine sediments (Boström, 1973;Boström & Peterson, 1969), but it plots in the hydrothermal area with Al/(Al + Fe + Mn) values <0.1. The Fe/Ti vs. Al/(Al + Fe + Mn) diagram (Figure 9), introduced by Boström (1973), could thus be used to distinguish hydrothermal from hydrogenetic and detrital deposits or show volcanic input. Hydrothermal sediments deposited by high-T (300-400°C) hydrothermal fluids have low Al contents and high Fe/Ti values (>100) (Boström & Peterson, 1969). Strakhov (1976) considered that hydrothermal metal-bearing sediments have [(Fe + Mn)/Ti] > 25. According to Lisitzin et al. (1976), metalliferous sediments have Fe concentrations higher than 10 wt.% (on a carbonate-or biogenic silica-free basis).
The Fe/Ti ratios for the Cihanpaşa Fe-Mn ores range from 161 to 253 with average of 209. In Figure 9, the studied Fe-Mn ore samples plot up to 80% metalliferous sediments with no clastic input. On the mixing curve between the East Pacific Rise metalliferous and the terrigenous sediment end-members it is quite similar to the Kasımağa (Keskin, Kırıkkale) Mn-ores and slightly like the Büyükmahal (Yozgat) ores from Turkey. This indicates a contribution of sea floorderived, metalliferous hydrothermal fluids to the Cihanpaşa Fe-Mn ores. The Al/(Al + Fe + Mn) ratios ranging 0.06 to 0.14 (on average, 0.09) of the studied samples are like the SE Pacific Basin metalliferous sediments (Dekov et al., 2010). In addition, it is noteworthy that Fe-Mn ores of the Cihanpaşa region have consistently high Fe concentrations and low Al and Ti concentrations, and low Al/(Al+Fe) (on average, 0.14) ratios, indicating a strong hydrothermal influence (Adachi et al., 1986).
Of particular significance are the studies of the Fe and Mn-ore deposits related to the radiolarian chert assemblages in the ophiolitic mélange complex of the İzmir-Ankara-Erzincan Ocean (e.g. Kasımağa, Koç et al., 2000;Eymir, Öksüz, 2011a;Derbent, 2011b;Büyükmahal;Öksüz & Okuyucu, 2014;and Kula;Kılıç et al., 2018). In the present study, some geochemical data from these studies are used here for comparative purposes, and thus it has  Boström, 1970;Boström et al., 1976). The grey zone reflects ideal mixing curves between metalliferous and terrigenous sediments, and between metalliferous sediments and Pacific Ocean pelagic sediments. Percentages indicate magnitudes of endmember hydrothermal component on basis of data for East Pacific Rise metalliferous sediment.
been compared with results of this study in order to define the depositional environment (proximal ridge, distal or pelagic and continental margin) and location of deep marine sediments of the İzmir-Ankara-Erzincan ocean. This implication agrees well with the (La SN /Ce SN )-Al 2 O 3 /(Al 2 O 3 + Fe 2 O 3 ) diagram (Figure 10a) used to discriminate the depositional environment of cherts (Murray et al.1990). In this diagram, Al 2 O 3 /(Al 2 O 3 + Fe 2 O 3 ) ratios are given to be 0.05-0.40 at ridge-proximal, 0.40-0.70 at pelagic and 0.55-0.90 at continental margin environments for submarine cherts. In addition, ridge-proximal, ocean basin floor or pelagic areas and continental margin domains have La SN /Ce SN 3-4, 1-2.5 and 0.5-1.4, respectively (Murray et al.1990;Murray, 1994, (Figure 10b; Table 2). The radiolarian chert samples (unpublished data) from the mélange of the İzmir-Ankara-Erzincan ocean in the Kula  Murray (1994). e, f Diagrams of ratios of Ti/V and Lu SN /La SN (Murray, 1994).
(Manisa) region plot within the continental margin with terrigenous input field (Figure a-d). Additionally, 100*Fe/ Si vs. 100*Al/Si and Fe/100-Si vs. Al/100-Si diagrams reveal that the Fe-and Mn-ores of Kasımağa (Kırıkkale, central Turkey), Büyükmahal and Cihanpaşa (Yozgat, central Turkey) plot as ridge proximal (Figure 10c, d). All chemical data from the Fe-Mn ores from Kasımağa (Kırıkkale, central Turkey), Büyükmahal and Cihanpaşa (Yozgat, central Turkey) reveal that their mineralisations indicate a proximal hydrothermal source close to the mid-oceanic ridge system of the İzmir-Ankara-Erzincan Ocean.However, geochemical data from Mn-ores and their host rocks from Derbent (Yozgat), Eymir (Sorgun, Yozgat) and Kula (Manisa) regions from these studies suggest that all of these mineralisations are, by contrast, of pelagic and continental margin origin and indicate a distal hydrothermal source (Figure 10a-d). Murray (1990) also specified that cherts from spreading ridges and open-ocean basins have significantly high V contents; so Ti/V ratios are commonly used to categorise depositional environments. Although the ratio of Ti/V ≤ 7 defines the near-spreading ridge zone, a Ti/V ≥ 40 characterises the continental margin (Murray, ;Li, 2000;Kang et al., 2011;Kemkin & Kemkina, 2015, 2020. The Ti/V ratio of the pelagic zone ranges from 7 to 40. Data from the Cihanpaşa Fe-Mn ores suggest that their Ti/V ratios are characteristic of the spreading ridge zone (Figure 10). Kemkin and Kemkina (2020) highlighted that the LREE/HREE ratio in sea-floor sediments decreases progressively from continental margin to pelagic environment. Sediments are enriched in heavy REE relative to light REE from a distal to a proximal site of ocean floor. Jr JL (1978) reported from the East Pacific Rise that there is an enrichment of HREE in sediments near the ridge crest and also a LREE/HREE increase with increasing distance from the ridge crest. Thus, a Lu SN /La SN ratio can used to reveal the enrichment of HREE relative to LREE (Murray, ;Kemkin & Kemkina, 2015, 2020. That ratio in the studied Fe-Mn ores ranges from 0.69 to 1.20, with an average of 0.89, suggesting that these ores correspond to a ridge-proximal (nearly spreading ridge) depositional environment (Figure 10f).
Seafloor hydrothermal vent fluids are significantly enriched in REE and Fe relative to ambient seawater based on chemical analyses (Michard et al., 1983;Michard et al., 1984;Michard and Albarede 1986;Von Damm et al., 1985;Olivarez & Owen, 1989), and are diluted through mixing with ambient seawater. Fe is highly enriched in high-T hydrothermal fluids during hydrothermal circulation of seawater (Hawkes et al., 2013). Further, hydrothermal Fe precipitates as particulate sulphides or oxides near the hydrothermal vent site after mixing with seawater (Fitzsimmons et al., 2014;Rudnicki & Elderfield, 1993). As such, hydrothermal precipitation decreasing away from the source has a low ΣREE and Fe composition (Prakash et al., 2012). Thus, the ratio of ΣREE/Fe is an effective indicator of the distance to the hydrothermal source using both ΣREE and Fe. Systematically, the ΣREE/ Fe ratios for hydrothermal sediments increase with increasing distance away from the paleo-ridge crest (Olivarez & Owen, 1989;Ruhlin & Owen, 1986). This ratio can provide an independent test of the source region, and thus would increase in the hydrothermal sediment with increasing distance from the seafloor hydrothermal vent solutions (Murray, ;Ruhlin & Owen, 1986;Qiu & Wang, 2011). It has been recognised that the REE can continue to be removed from seawater, long after the particles have settled from the plume to be deposited on the seafloor due to the increased uptake of dissolved REE from Figure 11. Comparison of ΣREE/Fe values of Late Cretaceous radiolarian chert-hosted Fe-and Mn-oxide deposits from Turkey and modern hydrothermal sediments from the East Pacific Rise (Ruhlin & Owen, 1986). All data of the radiolarian chert-hosted Fe-Mn oxide deposits from the mélange complexes belonging to the İzmir-Ankara-Erzincan Neotethyan ocean are for comparison. (Kasımağa, Kırıkkale, Koç et al., 2000;Büyükmahal, Yozgat, Öksüz & Okuyucu, 2014). Chemical analyses of radiolarian chert are from the Vezirler mélange (Kula, Manisa, western Turkey) (unpublished data, Aytaçegerler, 2020).
seawater (Rudnicki & Elderfield, 1993). For example, ΣREE/ Fe compositions in the suspended particulate material from the TAG hydrothermal vent field on the Mid-Atlantic Ridge increases with increasing distances from their hydrothermal source, indicating that rare-earth elements must be continuously extracted from seawater as hydrothermal precipitates are dispersed through the water column. The ratio of ΣREE/Fe in hydrothermal sediments close to East Pacific Rise is about 6.2 × 10 −4 , while that in hydrothermal sediment approximately 802 km away from East Pacific Ridge reaches about 28.8 × 10 −4 (Olivarez & Owen, 1989, p. 759). Qiu and Wang (2011) highlighted that the average ratio of ΣREE/Fe of the Middle-Upper Permian Laibin chert in Guangxi (China) is 42.6 × 10 −4 , which might indicate that it was deposited at a point away from the hydrothermal centre, and was also a primary source of silica derived from submarine hydrothermal fluids. In an example from the Yangshuo basin (South China), an increase of ΣREE/Fe in different chert types indicates a major source of REE added to the siliceous deposits via adsorption from seawater. Thus, low abundances of REE were absorbed by the siliceous sediments; therefore, the high precipitation rate of silica-rich materials and the scavenging of REE by Fe-Mn oxyhydroxides in hydrothermal vents resulted in low abundances of REE and La (Ruhlin & Owen, 1986;Olivarez & Owen, 1989;Murray, ;Chen et al., 2006). According to Prakash et al. (2012), hydrothermal and hydrogenetic oxides containing dissolved rare earth elements from seawater include a high ΣREE/Fe ratio (1.5 × 10 −3 to 23.2 × 10 -3 ) close to the volcanic cratered seamounts. Ruhlin and Owen (1986), Olivarez and Owen (1989) presented paleodistance data [(9 km (most proximal) to 1139 km (most distal)] from paleo-rise crest and geochemical data (REE, Fe) from Quaternary hydrothermal sediments from the East Pacific Rise (EPR). In the present study, ΣREE/Fe data published by Ruhlin and Owen (1986) were compared to test whether some Fe-and Mn-oxide deposits (Kasımağa, Cihanpaşa and Büyükmahal) were close to or distant from the midocean ridge of the Izmir-Ankara-Erzincan Ocean. Figure  11illustrates that ΣREE/Fe ratio increases with increasing distance from paleorise crest in the EPR (Ruhlin & Owen, 1986). The average ΣREE/Fe ratios of especially the Kasımağa (3.99 × 10 -4 ), Cihanpaşa (22.36 × 10 -4 ) and partially the Büyükmahal (30.26 × 10 -4 ) samples suggest that these Fe-Mn deposits are close to the hydrothermal vent centre from mid-oceanic ridge of İzmir-Ankara-Erzincan Ocean. Other silicified Mn deposits (e.g. Kula, Derbent, Eymir) show high ΣREE/Fe ratios, suggesting a distal hydrothermal formation (Table 2). Consequently, the ΣREE/Fe ratio of the Fe-Mn ores from the Cihanpaşa region (central Turkey) ranges from 13 × 10 -4 to 35 × 10 -4 , with a mean of 22.36× 10 −4 , revealing that the studied ores were deposited at a point close to the centre of a hydrothermal vent related to the spreading ridge of the İzmir-Ankara-Erzincan Ocean. Consequently, all the geochemical data reveal that the Fe-Mn ores hosted by radiolarian cherts in the Cihanpaşa region (Yozgat, central Turkey) were formed under the influence of a hydrothermal vent system close to the mid-oceanic ridge of the İzmir-Ankara-Erzincan ocean.

Conclusions
This study reveals an example of ridge-proximal hydrothermal Fe-Mn-oxide mineralisation within the Artova ophiolitic melange, Central Turkey. The mineralisation is hosted by deep-sea silicified rocks within an ophiolitic fragment related to the closure of the İzmir-Ankara-Erzincan Neotethyan ocean. Geochemical evidence such as high iron content, La SN /Ce SN (1.39-6.58, averaging 3.83), Fe/Ti ratio (averaging 184) and, low Al/(Al+Mn +Fe) (averaging 0.09), moderate ΣREE/Fe ratio (22.36x10 −4 ), weak positive Eu (CN) anomaly (1.22) suggest a proximal source for the hydrothermal discharge. It is possible that the strong positive correlation between Fe and Cu+Ni+Co in the Cihanpaşa Fe-Mn mineralisations is associated with hydrothermal fluids in the mid-ocean ridge of the İzmir-Ankara-Erzincan ocean.
Marine ore deposits (hydrothermal Fe-Mn ore deposits, manganese nodules and cobalt-rich ferromanganese crusts) formed in ancient oceans occur in ophiolitic mélange, including blocks of numerous deep-sea siliciclastic rocks and subvolcanic assemblages along the ophiolitic suture belt. In the studied area, dispersed polymetallic sulphide mineralisations and manganese nodules (unpublished data) are well-exposed in the Artova ophiolitic mélange along the İzmir-Ankara-Erzincan belt. All the geochemical evidence from the Cihanpaşa Fe-Mn, Kasımağa Mn, Büyükmahal Fe-Mn-ore deposits emphasised that they were formed close to the ridge of the Izmir-Ankara Neotethys ocean basin. The different rock assemblages of the ophiolitic mélange complex (e.g. Artova mélange, Ankara mélange, Dağküplü mélange) that outcrops along the İzmir-Ankara-Erzincan belt should be evaluated in terms of discovering ancient marine deposits containing precious and base metal deposits (polymetallic Fe-Mn deposits and seafloor sulphide mineralisations) via detailed geological and geochemical exploration.