Journal of Geochemical Exploration 164 (2016) 122–135

Contents lists available at ScienceDirect

Journal of Geochemical Exploration

journal homepage: www.elsevier.com/locate/jgeoexp

Three-dimensional geochemical patterns of regolith over a concealed gold deposit revealed by overburden drilling in desert terrains of northwestern China

Bimin Zhang a,b,c,⁎, Xueqiu Wang b,c, Qinghua Chi b,c, Wensheng Yao b,c,HanliangLiub,c, Xin Lin b,c a School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, China b Institute of Geophysical and Geochemical Exploration, CAGS, Langfang 065000, China c Key Laboratory for Geochemical Exploration Technology, MLR, Langfang 065000, China article info abstract

Article history: Desert terrains are widespread in northwestern and northern China, and these areas present particular Received 22 December 2014 challenges for exploration. In recent years, partial extraction techniques have been proven to be effective in Revised 9 June 2015 the search for concealed deposits in arid desert terrains in some cases. However, we still lack an understanding Accepted 13 June 2015 of the dispersion patterns of ore-forming elements in regolith. In this study, air reverse circulation drillings Available online 19 June 2015 were used to create three-dimensional (3D) distribution patterns of elements in regolith over the Jinwozi gold deposit in China, which is covered by tens of metres of regolith, in order to trace the migration of elements Keywords: Geochemical patterns and to understand the dispersion mechanisms. The 3D distribution maps of elements show that (1) coherent Concealed deposit anomalies occur at different depths of transported cover over the ore body, (2) Au tends to be enriched in the Overburden drilling top and bottom horizons and depleted in the middle horizon in the vertical direction, (3) the anomalous Desert terrains distribution of Au at the bottom is restricted to places at the interface of sediments and bedrock, and (4) the anomaly in the bottom sediments is conﬁned to a width of tens of metres, whereas that in top soils is much wider and can extend up to several kilometres. In addition, close positive correlations were found between the As, Hg, and Au distributions. © 2015 Elsevier B.V. All rights reserved.

1. Introduction 2007). However, there is still a critical need to study the three- dimensional (3D) distribution of elements in regolith. Such information Desert terrains are widely distributed in northwestern and northern is important for further elucidating the potential mechanisms for the China, and these areas are covered by widespread transported materials transfer of elements from the ore body upwards through the regolith that can mask geochemical signals from ore bodies; such geomorpho- cover to the surface and for understanding how to conduct successful logical structures present major obstacles to mineral exploration explorations in regolith-dominated terrains, whether for deposits (Wang et al., 2007). Over the past 20 years, partial extraction techniques concealed by the regolith or for those hosted within it. have been developed and proven effective in the search for concealed In this study, we used air reverse circulation (ARC) drilling technol- deposits in certain terrains (Antropova et al., 1992; Bajc, 1998; ogy over the Jinwozi gold deposit in China, which is covered by several Cameron et al., 2004; Clark et al., 1997; Cohen et al., 1998; El-Makky to tens of metres of transported materials, to determine the 3D distribu- and Sediek, 2012; Hamilton et al., 2004a,b; Kelley et al., 2003; Mann tion patterns of ore elements in regolith and to investigate the migration et al., 1995, 1998; Noble and Stanley, 2009; Wang, 1998; Wang et al., mechanisms. 2007; Williams and Gunn, 2002; Xie and Wang, 2003; Xie et al., 2011; Yeager et al., 1998), while at the same time, some migration models 2. Study area have been constructed and employed to explain the formation mechanisms of geochemical anomalies (Anand and Robertson, 2012; The Jinwozi gold ﬁeld is located 200 km southeast of Hami city at the Aspandiar et al., 2008; Cameron et al., 2004; Garnett, 2005; Hamilton, boundary of the Xinjiang and Gansu provinces in northwestern China 1998; Hamilton et al., 2004a,b; Kelley et al., 2003; Lintern, 2007; Luz (Fig. 1). There are two NE-trending mineralized zones in the Jinwozi et al., 2014; Mann et al., 2005; Smee, 1998; Wang, 2005; Wang et al., gold ﬁeld (Fig. 1). In the northern zone, the mineralization is characterized by an epithermal quartz-vein type. The auriferous quartz veins occur at the contact between porphyry and Devonian sequences. In ⁎ Corresponding author at: School of Earth Sciences and Resources, China University of Geosciences, Beijing 100083, China. the southern zone, the mineralization is characterized by tectonic E-mail address: [email protected] (B. Zhang). alterations. The ore bodies occur in a structural shear zone, which is

Fig. 1. Location and geology of the study area with drill hole sites. Coordinates are UTM Zone 46. mainly controlled by a NE trending fault that lies within the Devonian interbedded gravels, brown yellow sands, purple red sands, eluvium, sequence. The ores are mainly composed of pyrite, galena, and chalco- and bedrock. The lags are always covered by a dark and shiny substance pyrite. The pyrite is the primary Au-bearing mineral. The average Au called desert varnish. Such coatings represent a ﬁne mixture of clay grades of the two mineralized zones are c. 7 g/t and 4 g/t (Wang et al., minerals and Fe–Mn oxyhydroxides, which form micrometre-thick 2007). The proven total Au reserve of the gold ﬁeld is c.10t. The northern mineralized zone is located in an outcropping area and has a relatively high relief. The southern mineralized zone is situated in an area that is covered by the Gobi Desert with depths of a few metres to tens of metres. The regolith is composed of windblown sand, alluvium, colluvium, and residuum. The typical zonal structure of the regolith cover is illustrated in Fig. 2. The sequence of regolith materials from top to bottom: black gravels (lag), desert crusts, brown sands with

Fig. 3. Gold distribution in different fractions of soils along the traverse line that crosses the Fig. 2. Sketch illustrating the vertical regolith proﬁles. two mineralized zones (Wang et al., 2007). 124 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

of 50–100 m. At the bore holes, samples were collected continuously every metre from the ground surface to the bedrock. Every sample was well mixed and sieved in the field through a b100 mesh, and the finer fraction (150 μm) passing through the mesh was retained for further analyses. The choice of fraction size was based on the experimental findings and previous geochemical work in the region. In total, 1046 soil samples were collected from 63 bore holes. It is undeniable that mining contamination occurs in this study area. Hence, drilling sites were selected very carefully to avoid drilling dust and mining contamination. In consideration of the possibility of cross-hole contamination from the drilling activities, drilling equipment and the sampling system were cleaned after use at every hole. The cleaning procedure involved the use of high-pressure air generated by an air pump, which was blown onto the drilling equipment and sampling tools to dislodge any residual debris. After that, we used a clean cloth to wipe off all of the equipment. In the southern mineralized zone, four soil samples were collected and combined to form a composite sample at a depth of 15–30 cm above the buried deposit to confirm the use of the appropriate sampling fraction in the drilling work. The choice of soil sample depth was based Fig. 4. Drill hole sites of the ARC drilling located in the southern mineralized zone. on considerations of the sandy clay enrichment in the study area and previous sampling depths used in this region (Wang et al., 2007). The composite sample was sieved into seven fractions in the field with laminations that parallel the topography of the rock substrate (Potter the following grain size fractions: 830–1700 μm, 380–830 μm, and Rossman, 1977). It is still debated whether such varnish is formed 250–380 μm, 180–250 μm, 150–180 μm, 120–150 μm, and b120 μm. by (Broecker and Liu, 2001; Krinsley, 1998): (a) slow diagenesis of During the process of drilling, the different regolith types (Quaternary dust particles deposited on rock surfaces, (b) leaching from the underly- alluvium, Tertiary red strata, eluvium, and bedrock) could be recognized ing rock substrate, or (c) direct deposition of dissolved constituents in on the basis of the bore hole cutting colour, granularity, and mineral the atmosphere. The desert crust is created by the breakdown of soil composition. Regolith profileswereconstructedforeachboreholethat structural units by flowing water, raindrops, and subsequent evapora- was sampled, and the results are shown in Fig. 5. tion. Influenced by the undulating relief of the palaeotopography, Tertiary purple red sand layers are absent in some places. 4.2. Sample preparation and analyses The area is characterized by an arid climate, low rainfall, high evaporation, vast temperature differences between day and night (the All the samples were ground to less than 200 mesh (75 μm) for temperature difference can reach 40 °C), and long hours of sunshine analyses. A 0.25-g sample was digested in a hot mixture of acids (HCl,

(16–17 h in summer). The annual rainfall is less than 250 mm, and HF, HNO3, and HClO4). Inductively coupled plasma-mass spectrometry the potential annual evaporation is c. 1500 mm. July is the hottest (ICP-MS) was used for the determination of Ag, Cu, La, Pb, Sb, Sr, Th, month of the year (mean daily temperatures range from 25 to 40 °C in U, and Zn concentrations. Additionally, a 10-g sample was digested in July), and January is the coldest month of the year (mean daily temper- aqua regia and analysed by graphite furnace atomic absorption ature range from −20 to 5 °C). Because of the harsh natural conditions, spectrometry (GF-AAS) to obtain the Au concentration. Furthermore, a this area often has sparse vegetation and small amounts of biomass. 0.5-g sample was subjected to an aqua regia digest and analysed by hydride generation atomic fluorescence spectrometry (HG-AFS) to determine the As and Hg concentrations. 3. Previous geochemical work In addition to the above analyses, 24 soil and sediment samples collected at different depths from one of the drill holes over the miner- In the study area, a regional scale geochemical survey over an area of alization were analysed for SiO ,Al O ,Fe O ,MgO,CaO,Na O, and K O 1200 km2 was conducted by Ye et al. (2004); their results showed that 2 2 3 2 3 2 2 by X-ray fluorescence (XRF). the fine fraction of soil samples (b96 μm) delineates reasonably well the Analytical accuracy and precision for the laboratory quality were location of the regional tectonic metallogenic belt and mineralization strictly controlled by laboratory replicate samples and Standard Refer- (Ye et al., 2004). Subsequently, Wang et al. (2007) employed a variety ence Materials (SRMs). Five laboratory replicates were inserted into of sample media to a prospect along a transverse line that ran across each batch of 50 samples for precision control. The relative deviation the two mineralized zones; their results demonstrated that the greatest (RD%) of determination values of the replicates was listed in Table 1 contrast between anomalous and background concentrations is in the and calculated based on the equation: fine fraction (b96 μm) of soils (Figs. 1 and 3). They advocated that the use of fine fraction samples from clay-rich horizons or selective leaching RD% ¼jC1−C2j=½ðÞC1 þ C2 =2 100% of elements adsorbed on clays or oxide coatings would be effective for locating buried deposits. where C1 is the first determination and C2 is the second determination. Four standard reference materials (GAU9aGSS1, GAU10aGSS2, 4. Methods GAU11GSS3, GAU12GSD1a) were inserted blindly into each batch of 50 samples and analysed simultaneously with the samples for accuracy 4.1. Sampling control. Accuracy was controlled by the logarithmic difference (ΔlgC) of the determination value (lgCi) and standard reference value (lgCs) for An ARC drilling programme was systematically conducted along each standard (Table 1). four transverse lines that ran across the southern mineralized zone The mineralogy of seven samples in different fractions and 24 (Figs. 1 and 4). The intervals between the transverse lines were c. samples from one of the drilling holes were further analysed semi- 250 m. Each line had 13–20 bore holes that were spaced at an interval quantitatively by X-ray diffraction (XRD). B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 125

5. Results amounts of clay minerals. The fine fraction (b150 μm) consisted largely of calcite and clay minerals including illite, kaolinite, and chlorite. 5.1. Concentrations and variations of elements in different soil fractions The analytical results of different soil fraction samples sieved from the composite sample are shown in Fig. 6.Itisobviousthat The distributions of elements in different soil fractions are controlled Au-concentrations of the fine fractions (b150 μm) were higher than by mineralogy. The mineralogy of the study area is shown in Table 2. those of the other fractions (N150 μm). Some other ore elements (Ag, The coarser fraction (N830 μm) consisted largely of quartz and feldspar, As, Hg, Pb, Zn, Ni, Co, Cr) displayed the same distribution. Considering followed by gypsum and calcite. The intermediate fraction (150–830 μm) the weight percentage of fine fractions in soils (Table 2), the proportion was dominated by quartz, gypsum, feldspar, and calcite, with small of the total Au in the fine fractions reached up to 48.93%.

Fig. 5. Structure of the regolith proﬁles presented by drilling. 126 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

Table 1 Calculated values of relative deviation and logarithmic difference for precision and accuracy control.

Parameters Analytical method Unit Detection limit RD% ΔlgC

Au GF-AAS ppb 0.2 24.53 0.05 Ag ICP-MS ppb 20 13.35 0.04 As HG-AFS ppm 1 7.03 0.03 Cu ICP-MS ppm 1 9.19 0.02 Hg HG-AFS ppb 2 7.17 0.02 La ICP-MS ppm 1 7.06 0.03 Pb ICP-MS ppm 2 5.99 0.02 Sb ICP-MS ppm 0.05 6.33 0.03 Sr ICP-MS ppm 5 2.21 0.02 Th ICP-MS ppm 1 8.57 0.03 U ICP-MS ppm 0.2 6.19 0.03 Zn ICP-MS ppm 2 7.02 0.02

SiO2 XRF % 0.1 1.86 0.005

Al2O3 XRF % 0.1 2.21 0.01

Fe2O3 XRF % 0.1 0.65 0.01 MgO XRF % 0.05 1.96 0.02 CaO XRF % 0.05 1.05 0.02

Na2O XRF % 0.1 1.79 0.03

K2O XRF % 0.05 0.95 0.01 Notation: GF-AAS: graphite furnace atomic absorption spectrometry; HG-AFS: atomic ﬂuorescence spectrometry; ICP-MS: inductively coupled plasma mass spectrometry; XRF: X-ray ﬂuorescence spectrometry; RD: relative deviation; ΔlgC: logarithmic difference.

Table 2 Mineral composition of different soil fraction samples sieved from the composite sample.

Fraction Quartz Feldspar Calcite Dolomite Hematite Hornblende Gypsum Illite Kaolinite Chlorite Weight percentage (μm) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) (%)

830–1700 32.4 33.0 8.3 – 3.1 0.9 14.6 5.4 1.9 0.4 21.9 380–830 19.7 24.2 15.5 – 4.8 – 24.3 7.6 3.3 0.6 4.1 250–380 18.7 24.4 9.4 – 5.7 – 30.6 7.1 3.5 0.6 18.5 180–250 23.9 13.5 11.4 – 5.0 – 31.0 10.6 3.7 0.9 14.1 150–180 15.4 13.3 15.3 4.3 5.3 1.5 24.7 13.8 5.2 1.3 15.9 120–150 13.9 14.1 21.9 – 3.5 0.7 16.4 18.8 8.6 2.2 6.0 b120 13.6 9.4 19.8 4.8 2.3 0.8 9.0 26.2 11.2 2.9 19.5

Fig. 6. Au, Ag, As, and Hg concentrations in different sieve fractions in the composite soil sample from the southern mineralized zone. B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 127

Fig. 7. Scatterplots of Au, Ag, As, and Hg concentrations versus clay content fractions in the composite soil sample from the southern mineralized zone. r, correlation coefﬁcient.

Positive correlations (generally r N 0.8) were found between the clay greater contrast between anomalous and background concentrations content and element concentrations (Fig. 7). Based on the above results than coarse grain-sized fractions. Thus, the selection of the b150 μm and previous research, it can be concluded that ﬁne grain-sized fraction from the drilling samples for analysis and research was fractions have higher concentrations of target elements and show a appropriate.

Fig. 8. Vertical variation of mineral contents in the drill proﬁle over the mineralization hole (Hole A shown in Fig. 4). 128 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

Table 3 Mean values for the mineralogical composition of samples from different lithological units (Quaternary alluvium, Tertiary red strata, eluvium, and bedrock); these samples were collected from one drill hole over the mineralization.

Unit Clay (%) Quartz (%) Feldspar (%) Calcite (%) Dolomite (%) Hematite (%) Hornblende (%) Gypsum (%)

Quaternary alluvium 27.4 19.1 5.9 10.2 26.1 1.6 2.6 7.1 Tertiary red strata 21.5 42.9 5.6 9.1 19.7 –– 1.2 Eluvium and bedrock 22.9 45.9 4.2 7.5 19.2 –– 0.3

Fig. 9. Vertical variation of element concentrations, pH, and conductivity of soils in the drill proﬁle over the mineralization hole (Hole A shown in Fig. 4).

5.2. Vertical variation of the mineral composition and element concentra- In addition, hematite and hornblende are found to occur only in tions in a drill hole Quaternary alluvium. The concentrations of elements also vary with depth. Gold concen- Soil and sediment samples (b150 μm) at different depths from one trations are high in the top parts of the bore profile, then they gradually of the drill holes (Hole A in Fig. 4) over the mineralization were analysed decrease to low levels in Quaternary alluvium and in Tertiary red strata. to determine their mineral compositions (Fig. 8)(Table 3) and elemen- Close to the eluvium, Au concentrations are high. In the eluvium, the Au tal concentrations (Fig. 9)(Table 4). The structure of the regolith in this concentrations reach 2593 ppb. The distribution pattern of Au in the drill hole is shown in Fig. 2. regolith profile is crescent-shaped, i.e., it has a tendency to be enriched Although the major minerals of the soils and sediments from the in the lower and upper parts of the profile. The Ag, As, Hg, Zn, and a few ground surface to the site of mineralization are similar, i.e., quartz other elements (e.g., W and Bi) have distribution patterns similar to Au. (14–53%), dolomite (15–34%), illite (13–23%), and calcite (5–15%), This kind of distribution pattern is similar to the vertical variation of clay certain variations were found in regard to mineral composition minerals, and it is indicative of the positive correlations between these with depth; these variations were especially noticeable in areas components. In contrast, Cu is high in Quaternary alluvium and low in where characteristic changes in the regolith occurred. Clay minerals Tertiary red strata and eluvium. For the major elements, Quaternary (illite, kaolinite, and chlorite) are more abundant in Quaternary alluvium has less Si and more Fe and Ca than Tertiary red strata and alluvium and low in Tertiary red strata, but are present in higher eluvium, which coincides with the vertical variation of quartz, hematite, amounts in weathered eluvium close to the mineralization. Quartz and gypsum. Furthermore, the pH levels of the upper half horizons of is much lower in Quaternary alluvium than in Tertiary red strata Quaternary alluvium are greater than those in the other horizons. This and eluvium. Conversely, gypsum is much higher in Quaternary is due to the precipitation of some secondary minerals such as gypsum, alluvium than in Tertiary red strata and eluvium. Feldspar is higher which is related to long-term evapotranspiration processes. In addition, in Quaternary alluvium than in Tertiary red strata and eluvium; but electrical conductivity is high in Quaternary alluvium and low in at depths of 17–18 m, feldspar is much higher than at other depths. Tertiary red strata and eluvium.

Table 4 Mean values for the element concentrations, pH, and conductivity of soils from different lithological units (Quaternary alluvium, Tertiary red strata, eluvium, and bedrock); these samples were collected from one drill hole over the mineralization.

Unit Au (ppb) As (ppm) Hg (ppb) Ag (ppb) Cu (ppm) Pb (ppm) Zn (ppm) SiO2 (%) Fe2O3 (%) CaO (%) pH EC (s/m) Quaternary alluvium 7.8 18.5 8.8 202 28.8 31.3 66.5 40.2 4.9 14.9 8 6.2 Tertiary red strata 3.1 10.9 7.6 225 20.7 19.9 52.3 50.8 3.7 11.6 7.8 2.5 Eluvium and bedrock 671.8 12.7 10.5 302 18.8 34.2 63.9 55.9 3.6 9.9 8 2.3 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 129

Fig. 10. Three-dimensional spatial distribution of Au.

5.3. Three-dimensional geochemical patterns of ore-forming elements Quaternary alluvium, Tertiary red strata, eluvium, and bedrock) are presented in Table 5. The distribution of elements in different transported overburden horizons is of special significance for geochemical research, especially the patterns in the top and bottom layers. The top layer participates 5.3.1. Gold directly in the process of hypergenesis and is the highest horizon that At the interface between overburden and basement, regolith Au elements can reach during the process of vertical migration. At the concentrations were elevated in the mineralized zone (Fig. 10a). Gold same time, the soil of the top layer offers an economical sampling medi- concentrations in the anomalous centre reach up to 2.59 ppm. At the um for geochemical exploration. The bottom layer is next to bedrock positions of 0–1m,1–2m,and2–3 m, anomalies were found above and contains weathering products from bedrock. Accordingly, some di- the mineralized zone and in its dip direction (NW trending). The area rect mineralization information can be exposed by studying this layer. of anomalous distribution is larger in the shallower position. The surface In order to present the 3D geochemical patterns of ore-forming soil (c.10–30 cm) has the largest enrichment area, and it is related to elements in transported overburden material, the spatial distributions the direction of the tectonic alteration zone and seasonal rainwater of Au, As, Hg, Ag, and Cu are shown in Figs. 10–14. The contour plots flow. It is obvious that Au tends to be enriched in the lower and upper were made by Surfer software, and the kriging method was used to parts and depleted in the middle parts of the profile in the vertical represent the distribution of these elements in the regolith. The statisti- direction (Fig. 10b and c). Furthermore, the anomalous distribution at cal values (mean, standard deviation, minimum, and maximum) for the bottom is restricted in the lowest places at the interface of the the elemental ARC data from the various lithological units (surface, overburden and basement. 130 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

Fig. 11. Three-dimensional spatial distribution of As.

5.3.2. Arsenic, mercury, and silver positive correlations with As and Hg, a very weak correlation with Ag, The As and Hg distributions were similar to that of Au (Figs. 11–12). and no correlation with Cu. Anomalies occur continuously in the different depths of overburden over the ore body. There are close positive correlations among the As, Hg, and Au distributions. The Ag distribution showed slight 6. Discussion and conclusions resemblance to that of Au, especially the Au distribution at the surface (Fig. 13a). Fig. 5 shows that palaeotopographic relief is high in the research area. The depth of transport overburden in the northwest area is about 2–12 m. In the middle area, the depth of regolith becomes 5.3.3. Copper shallower, but abruptly, the terrain slopes down southeastward along The Cu data displayed an irregularly distributed profile (Fig. 14). the transverse lines. Because of the descent of the ancient landform, a There is no obvious enrichment of Cu related to Au mineralization or thick-layer of Tertiary red strata has been deposited over the bedrocks. any close correlation between Au and Cu. In regard to the distribution The thicknesses of the Tertiary red strata range from several metres to of Cu and the structure of the regolith profiles, strong Cu anomalies more than 50 m, over which there is still transport overburden. mainly occur in the lower segment of the transport overburden lying High-relief palaeotopographic features may influence the mechanical to the southeast of the mineralization. This portion of Cu-bearing dispersion of weathering products from mineralization so that materials may have been from past river transportation events. Copper transported regolith containing anomalous trace element concentra- anomalies also occur in the bedrock lying to the northwest of the tions can be found in the low-lying areas of ancient landforms. In mineralization. addition, the depth of the regolith cover across the mineralized zone In addition, correlation coefficients among elements (Au, As, Hg, Ag, in the middle of lines 1 and 2 is shallow (1–6 m deep), which makes it and Cu) in all soil samples were calculated (Table 6). Gold showed close easier for the element to migrate. B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 131

Fig. 12. Three-dimensional spatial distribution of Hg.

Three-dimensional distribution maps of elements that were alluvium relative to the Ternary red strata in the profiles. The mineralized generated in this study for the Jinwozi gold deposit in China show that zone is completely covered by the Gobi Desert, although the covering (1) coherent anomalies occur at different depths of transported cover layer is very shallow in some places. Hence, anomalies in top sediments over the ore body, (2) Au tends to be enriched in the lower and upper formed by vertical migration of elements. In addition, broad patterns in parts and depleted in the middle parts of the profile in the vertical direc- surface soils were driven by continuous vertical migration of elements tion, and the distribution has a ‘C-shaped’ pattern, (3) the anomalous from the ore body to the surface and further lateral diffusion by wind distribution of Au in the bottom horizon is restricted to places at the and seasonal rainwater in all directions, providing the theoretical basis interface of the sediments and bedrock, and (4) the anomaly in the for the use of low-density geochemical mapping when prospecting for bottom sediments is confined to a width of tens of metres, whereas concealed deposits. The possibility may still exist that some anomalous that in top soils is much wider and can extend to several kilometres. high surface values could have resulted from mining activity contamina- In addition, close correlations were found between the As, Hg, and Au tion (despite efforts taken by us to avoid this). distributions, but no obvious enrichment of Cu related to Au mineraliza- The possible modes of vertical migration in desert terrains have been tion was observed. described by Wang et al. (2007) (Fig. 15). Elements are released from Enrichment of elements in the bottom materials was due to the the ore body and associated altered rocks during weathering. Because weathering of the ore body, and the elements were mechanically the area is so arid, the water-table is hundreds of metres deep and transported and deposited in the lowest places by the process of vegetation is sparse. Water and vegetation would therefore seem to pediplanation. The contact between the Quaternary alluvium and play a very limited role in the transport of elements upwards to the Tertiary strata formed an unconformity along the palaeo-landform surface, although evaporation, capillary action, or uptake by plants surface, which allows for the possibility that mechanical dispersion of may occur during rainfall events. Hence, gas is regarded as the main weathering products from the ore zone onto this surface may have medium for vertical migration of the elements. Ultra-fine Au particles contributed to some of the higher concentrations present in Quaternary can be adsorbed onto the surfaces of gas bubbles and migrate with the 132 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

Fig. 13. Three-dimensional spatial distribution of Ag. bubbles upwards to the surface. Such gases may be derived from the In the future, we plan to conduct additional geochemical mapping atmosphere and driven to the surface by barometric pumping research on some known concealed deposits in desert terrains and ap- (Cameron et al., 2004), or they may be released from ore minerals or plying our mapping technology to other areas of potential geological derived from mantle degassing (Gold and Soter, 1980). Nanoscale Au significance. In addition, further studies on the migration mechanisms particles have been observed in gas over Jinwozi gold deposits using a of elements are needed to gain better understanding of the genesis of transmission electron microscope (TEM) equipped with an energy ore element anomalies over concealed deposits. dispersive spectroscope (Wang and Ye, 2012); these data provide direct evidence of the migration mechanism. Three-dimensional geochemical patterns of regolith presented by Acknowledgments overburden drilling also provide good evidence for the transport mode of vertical migration. ‘C-shaped’ patterns have been observed in This paper presents the results of the “Deep-Penetration Geochemi- regolith of other landscapes, for example, over ore bodies in the Yilgarn cal Detection Technology Project” (201011055, SinoProbe-04-03), craton of Australia (Gray et al., 2008). This pattern is thought to be the which was supported by China's Ministry of Land and Resources. This result of depletion of elements in the middle of the regolith due to saline study was also financially supported by the Natural Science Foundation groundwater under an arid climate. In our study area, the groundwater of China (41203038) and the China Geological Survey (1212011120206, table is below 200 m and the rain water, with an annual rainfall of 12120113100900). The authors thank Dr. Ravi R. Anand, who is affiliat- b250 mm, does not extend down 20 cm from the soil surface (since ed with Australia's Commonwealth Scientific and Industrial Research the beginning of the Quaternary). Thus, it is impossible for elements Organization (CSIRO), for providing us with a critical review of an to have been depleted by saline groundwater in this very young earlier draft of the manuscript. We also thank the reviewers and editors wind-blown soil regolith terrain. for their thorough work and very helpful comments on prior versions of B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 133

Fig. 14. Three-dimensional spatial distribution of Cu.

Table 5 Statistical parameters for elemental (Au, As, Hg, Ag, and Cu) ARC data from the various lithological units (surface, Quaternary alluvium, Tertiary red strata, eluvium, and bedrock).

Elements Sample types N Min. Max. Mean SD

Au (ppb) Surface soil 63 2.78 386.00 48.38 81.43 Quaternary alluvium 456 0.58 146.30 9.10 20.93 Tertiary red strata 372 0.19 19.07 2.38 2.78 Eluvium and bedrock 155 0.31 2593.70 56.99 169.37 As (ppm) Surface soil 63 3.24 64.03 20.89 8.34 Quaternary alluvium 456 6.77 78.90 20.12 7.24 Tertiary red strata 372 4.64 35.00 12.04 4.50 Eluvium and bedrock 155 1.90 213.99 22.41 23.76 Hg (ppb) Surface soil 63 6.00 84.50 21.43 17.81 Quaternary alluvium 456 2.50 69.50 8.46 6.17 Tertiary red strata 372 5.00 31.50 10.00 3.05 Eluvium and bedrock 155 2.50 90.50 12.40 15.75 Ag (ppb) Surface soil 63 58.97 1339.42 116.47 62.62 Quaternary alluvium 456 35.97 837.24 128.85 83.26 Tertiary red strata 372 45.38 909.72 135.13 106.69 Eluvium and bedrock 155 18.52 2239.87 142.89 240.77 Cu (ppm) Surface soil 63 17.73 32.34 25.00 3.18 Quaternary alluvium 456 5.96 485.41 32.17 31.31 Tertiary red strata 372 10.08 200.51 23.42 14.22 Eluvium and bedrock 155 3.96 57.74 23.27 12.32

Notation: N = number of samples; Min. = minimum; Max. = maximum; SD = standard deviation. 134 B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135

Table 6 Correlation coefﬁcients between the elements (Au, As, Hg, Ag, and Cu) in all soil samples.

Element Correlation coefﬁcients

Au As Hg Ag Cu

Au 1 –– –– As 0.71 1 ––– Hg 0.82 0.69 1 –– Ag 0.17 0.07 0.06 1 – Cu −0.02 0.05 −0.01 0.03 1

Fig. 15. Conceptual model of migration in desert regolith proﬁle (Wang et al., 2007).

this paper. All of these supportive efforts are acknowledged here with field investigation at the Marsh Zone gold property. Geochem. Explor. Environ. Anal. 4, 33–34. appreciation. Hamilton, S.M., Cameron, E.M., McClenaghan, M.B., Hall, G.E.M., 2004b. Redox, pH and SP variation over mineralization in thick glacial overburden. Part II: field References investigation at cross lake VMS property. Geochem. Explor. Environ. Anal. 4, 45–58. Anand, R.R., Robertson, I.D.M., 2012. The role of mineralogy and geochemistry in forming Kelley, D.L., Hall, G.E.M., Cross, L.G., Hamilton, I.C., McEwen, R.M., 2003. The use of partial anomalies on interfaces and in areas of deep basin cover: implications for explora- extraction geochemistry for copper exploration in northern Chile. Geochem. Explor. tion. Geochem. Explor. Environ. Anal. 12, 45–66. Environ. Anal. 3, 85–104. Antropova, L.V., Goldberg, I.S., Voroshilov, N.A., Ryss, Y.S., 1992. New methods of regional Krinsley, D., 1998. Models of rock varnish formation constrained by high resolution exploration for blind mineralization: application in the USSR. J. Geochem. Explor. 43, transmission electron microscopy. Sedimentology 45, 711–725. 157–166. Lintern, M.J., 2007. Vegetation controls on the formation of gold anomalies in calcrete and Aspandiar, M.F., Anand, R.R., Gray, D.J., 2008. Geochemical dispersion mechanisms other materials at the Barns Gold Prospect, Eyre Peninsula, South Australia. Geochem. through transported cover—implications for mineral exploration in Australia. CRC Explor. Environ. Anal. 7, 249–266. LEME Report 246. CRC LEME, Perth. Luz, F., Mateus, A., Matos, J.X., Gonçalves, M.A., 2014. Cu- and Zn-soil anomalies in the NE Bajc, A.F., 1998. A comparative analysis of enzyme leach and mobile metal ion selective border of the South Portuguese Zone (Iberian Variscides, Portugal) identified by extractions: case studies from glaciated terrain, northern Ontario. J. Geochem. Explor. multifractal and geostatistical analyses. Nat. Resour. Res. 23, 195–215. 61, 113–148. Mann, A.W., Birrell, R.D., Gay, L.M., 1995. Partial extractions and mobile metal ions. In: Broecker, W.S., Liu, T., 2001. Rock varnish: recorder of desert wetness? GSA Today 11, 4–10. Camuti, K. (Ed.), Extended Abstracts of the 17th IGES, pp. 31–34. Cameron, E.M., Hamilton, S.M., Leybourne, M.I., 2004. Finding deeply-buried deposits Mann, A.W., Birrell, R.D., Mann, A.T., Humphreys, D.B., Perdrix, J.L., 1998. Application of using geochemistry. Geochem. Explor. Environ. Anal. 4, 7–32. the mobile metal ion technique to routine geochemical exploration. J. Geochem. Clark, J.R., Yeager, J.R., Rogers, P., Hoffman, E., 1997. Innovative enzyme leach Explor. 61, 87–102. provides cost-effective overburden/bedrockpenetration.In:Gubins,A.G.(Ed.), Mann, A.W., Birrell, R.D., Fedikow, M.A.F., Souza, H.A.F., 2005. Vertical ionic migration: Geophysics and Geochemistry at the Millennium. Proceedings of Exploration mechanisms, soil anomalies, and sampling depth for mineral exploration. Geochem. 97, pp. 371–374. Explor. Environ. Anal. 5, 201–210. Cohen, D.R., Shen, X.C., Dunlop, A.C., Putherford, N.F., 1998. A comparison of selective Noble, R.R.P., Stanley, C.R., 2009. Traditional and novel geochemical extractions applied to extraction soil geochemistry and biogeochemistry in the Cobar area, New South aCu–Zn soil anomaly: a quantitative comparison of exploration accuracy and Wales. J. Geochem. Explor. 61, 173–189. precision. Geochem. Explor. Environ. Anal. 9, 159–172. El-Makky, A.M., Sediek, K.N., 2012. Stream sediments geochemical exploration in the Potter, R.M., Rossman, G.R., 1977. Desert varnish: the importance of clay minerals. Science Northwestern Part of Wadi Allaqi Area, South Eastern Desert, Egypt. Nat. Resour. 196, 1446–1448. Res. 21, 95–115. Smee, B.W., 1998. A new theory to explain the formation of soil geochemical responses Garnett, D., 2005. Element mobility in transported overburden—are we looking in the over deeply covered gold mineralisation in arid environments. J. Geochem. Explor. wrong direction? Assoc. Appl. Geochem. Explor 127, 3–5. 61, 149–172. Gold, T., Soter, S., 1980. The deep earth gas hypothesis. Sci. Am. 242, 130–138. Wang, X., 1998. Leaching of mobile forms of metals in overburden: development and Gray, D.J., Sergeev, N.B., Britt, A.F., Porto, C.G., 2008. Supergene Mobilization of Gold and applications. Geochem. Explor. 61, 39–55. Other Elements in the Yilgarn Craton-final Report. Cooperative Research Centre Wang, X., 2005. Conceptual model of deep-penetrating geochemical migration. Reg. Geol. for Landscape Environments and Mineral Exploration. CRC LEME, Perth (open file China 24, 892–896 (in Chinese with English abstract). report 228). Wang, X., Ye, R., 2012. Findings of nanoscale metal particles: evidence for Hamilton, S.M., 1998. Electrochemical mass-transport in overburden: a new model to deep-penetrating geochemistry. Acta Geosci. Sin. 32 (1), 7–12 (in Chinese with account for the formation of selective leach geochemical anomalies in glacial terrain. English abstract). J. Geochem. Explor. 63, 155–172. Wang, X., Wen, X., Ye, R., 2007. Vertical variations and dispersion of elements in arid Hamilton, S.M., Cameron, E.M., McClenaghan, M.B., Hall, G.E.M., 2004a. Redox, pH and SP desert regolith: a case study from the Jinwozi gold deposit, northwestern China. variation over mineralization in thick glacial overburden. Part I: methodologies and Geochem. Explor. Environ. Anal. 7, 163–171. B. Zhang et al. / Journal of Geochemical Exploration 164 (2016) 122–135 135

Williams, T.M., Gunn, A.G., 2002. Application of enzyme leach soil analysis for epithermal Ye, R., Wang, X., Zhao, L., 2004. Deep penetration geochemistry methods in gobi- gold exploration in the Andes of Ecuador. Appl. Geochem. 17, 367–385. overburden terrain of the Jinwozi metallogenic belt. Geophys.Geochem. Explor. 40 Xie, X., Wang, X., 2003. Recent developments on deep penetrating geochemistry. Earth (6), 65–70 (in Chinese with English abstract). Sci. Front. 10 (1), 225–238 (in Chinese with English abstract). Yeager, J.R., Clark, J.R., Mitchell, W., Renshaw, R., 1998. Enzyme leach anomalies associated Xie, X., Lu, Y., Yao, W., Bai, J., 2011. Further study on deep penetrating geochemistry over with deep Mississippi Valley-type zinc ore bodies at the Elmwood Mine, Tennessee. the Spence porphyry copper deposit, Chile. Geosci. Front. 2 (3), 303–311. J. Geochem. Explor. 61, 103–112.