The study area is located between the central Lhasa subterrane (CLS) and the southern Lhasa subterrane (SLS), which are separated by the Luobadui-Milashan fault zone (LMFZ) in southwestern Tibet (Figs. 1a–1b; Pan et al., 2006). There is an unconformable contact with Permian–Carboniferous metamorphic sedimentary rocks and Paleogene volcano-sedimentary rocks (Li et al., 2019; Zhang et al., 2018; Zhu et al., 2011). The structures are primarily EW-oriented faults of different scales and sequences, followed by genetically related NE-trending and NNW-trending faults, and a few NS-oriented faults (Fig. 1c). The exposed intrusions are mainly quartz syenite, granodiorite porphyry, monzonitic granite, and granodiorite, with zircon U-Pb ages ranging from 61–72 Ma (Liu et al., 2019; Zhang et al., 2017; Duan et al., 2014; Gao et al., 2012; Wang et al., 2012). These intrusions and the Paleogene volcanics (Fig. 1c) are closely spatially and genetically related to the known mineral deposits (Beina Pb-Zn-Ag deposit, Longgen Pb-Zn-Ag deposit, Chagele Pb-Zn-Ag-Cu deposit, Chazi Pb-Zn deposit, Gergeng Pb-Zn deposit, Dajiacuoxi Pb-Zn deposit) (Fig. 1c; Jiang et al., 2019; Liu et al., 2019; Zhang Y C et al., 2018; Zhang S Z et al., 2017; Gao et al., 2012).
The regional stream sediment geochemical dataset consisted of 758 stream sediment samples (Fig. 2) and 39 major and minor trace elements were analyzed at a density of one sample per 15 km2. Detection limits and analytical methods for each of the 39 elements are shown in Table 1.
Element Detection limit Analytical method Ag 20 ES As 1 HG-AFS Au 0.2 GF-AAS B 5 ES Ba 50 ICP-AES Be 0.5 ICP-AES Bi 0.05 ICP-MS Cd 30 ICP-MS Co 1 ICP-MS Cr 5 XRF Cu 1 ICP-MS F 100 ISE Hg 0.5 CV-AFS La 18 ICP-MS Li 5 ICP-AES Mn 30 ICP-AES Mo 0.2 ICP-MS Nb 5 ICP-MS Ni 2 ICP-AES P 100 XRF Pb 2 ICP-MS Sb 0.1 HG-AFS Sn 1 ES Sr 5 ICP-AES Th 4 ICP-MS Ti 100 XRF U 0.5 ICP-MS V 5 ICP-AES W 0.5 ICP-MS Y 1 XRF Zn 10 ICP-AES Zr 10 XRF SiO2 0.10 wt.% XRF Al2O3 0.10 wt.% XRF TFe2O3 0.05 wt.% XRF MgO 0.05 wt.% ICP-AES CaO 0.05 wt.% ICP-AES Na2O 0.05 wt.% ICP-AES K2O 0.05 wt.% XRF Note: the units of Ag, Au, Cd, and Hg are ppb, and the units of other elements are ppm, apart from oxides (wt.%). ES. Emission spectrometry; HG-AFS. hydride generation-atomic fluorescence spectrometry; GF-AAS. graphite furnace atomic absorption spectrometry; CV-AFS. cold vapor atomic fluorescence spectrometry; ISE. ion-selective electrode; ICP-MS. inductively coupled plasma mass spectrometer; XRF. X-ray fluorescence; ICP-AES. inductively coupled plasma atomic emission spectroscopy.
The DEM dataset was downloaded from publicly available data on the Geospatial Data Cloud Site, the Chinese Academy of Sciences (http://www.gscloud.cn). This dataset was processed using ASTER GDEM first edition (V1) and divided into four image files with a global spatial resolution of 30 m.
Geochemical statistical parameters like the skewness of the original dataset revealed the non-normal distribution of all 39 elements, especially Ag, Pb, Zn, Cu, Hg, and Au (Table 2). Based on the types of mineral combinations in the Ag-Pb-Zn deposit in the study area, Ag, Pb, Zn, and Cd were selected to generate logarithmic transformation histograms and Q-Q diagrams. The histograms showed an approximately log-normal distribution pattern (Fig. 3). However, the Q-Q plots revealed the existence of multiple populations in the single dataset (Fig. 4), indicating the presence of mixed origins due to the impact of complex geological processes (Zuo et al., 2011, 2009). They also clarified that these elements did not obey the central limit theorem in their spatial distributions, and element dispersion and enrichment may have had some singularity in space.
Element N Minimum Maximum Mean Standard deviation Skewness Kurtosis Ag 758 28 3 600 107.39 149.51 17.93 399.36 As 758 1.64 218.94 23.26 20.78 3.92 22.3 Au 758 0.27 59.2 1.02 2.15 26.06 703.91 B 758 6.3 233.3 40.11 18.33 2.85 19.44 Ba 758 131.15 1 278 512.56 129.52 2.11 10.92 Be 758 1.52 16.12 3.05 1.42 5.48 36.33 Bi 758 0.09 8.32 0.45 0.43 9.93 154.35 Cd 758 30 2 470 183.67 221.68 6.41 54.23 Co 758 1.6 22.07 6.28 2.59 1.89 6.85 Cr 758 4.22 221.96 25.82 19.51 5.16 44.78 Cu 758 2.18 292.48 12.18 15.08 13.94 236.34 F 758 189.75 1 569.11 468.9 173.83 2.26 8.11 Hg 758 1 1 440 16.42 54.73 23.58 607.51 La 758 17.93 92.28 33.92 7.93 1.13 4.86 Li 758 14.94 101.93 43.49 14.47 1.58 3.26 Mn 758 112.19 2 291 456.82 249.63 2.79 11.92 Mo 758 0.2 6.54 0.8 0.76 4.2 22.28 Nb 758 5.71 28.8 12.92 3.19 0.53 1.43 Ni 758 2.89 140.49 13.76 11.28 7.33 76.76 P 758 134.5 1 465 397.44 168.45 2.53 10.4 Pb 758 13.71 8 881 44.47 321.82 27.42 753.87 Sb 758 0.1 10.71 0.95 0.76 6.75 73.36 Sn 758 0.55 22.82 3.5 2.46 3.55 18.09 Sr 758 48.58 698.77 159.18 79.52 2.56 10.22 Th 758 4.02 73.26 15.86 8.57 3.36 16.53 Ti 758 672.09 5 089.42 2 106.49 654.74 0.71 0.66 U 758 1.02 191.3 5.36 14.74 9.83 103.07 V 758 8.77 132.06 39.63 17.52 1.1 2.2 W 758 1.06 28.54 3.77 2.7 4.09 21.11 Y 758 1.91 1.91 1.91 0 Zn 758 29.01 1 667 72.19 69.08 16.82 377.1 Zr 758 69.61 568.91 208.17 68.59 0.75 1.76 Al2O3 758 5.62 17.24 12.12 1.27 0.1 1.63 CaO 758 0.2 23.77 1.09 1.47 8.31 97.82 Fe2O3 758 0.9 7.14 2.76 0.85 1.12 2.14 K2O 758 1.73 6.4 3.61 0.57 1.23 5.99 MgO 758 0.13 4.18 0.74 0.44 2.99 15.39 Na2O 758 0.19 3.96 1.87 0.51 0.15 1.32 SiO2 758 34.45 86.67 74.96 4.69 -1.98 10.99 Note: The units of Ag, Au, Cd, and Hg are ppb and the units of other elements are ppm, apart from oxides (wt.%).
Table 2. Descriptive statistics for original stream sediment geochemical data
1.1. Geological Setting
1.2. Data Sources
1.3. Geochemical Parameters and Distribution Statistics
As applied in the current investigation, the general procedure for SFA consisted of two main phases: the first phase involved extracting 'clean' factors and the second phase extracted the remarkable multi-element anomalous signature for mineralization to calculate reasonable factor scores (FSs) and loadings (Yousefi et al., 2014). Each of the main SFA phases was divided into sub-phases (called stages) depending on the geochemical dataset feature and type of known mineral deposits (Yousefi, 2017; Yousefi et al., 2014). To avoid significant bias in multivariate statistical analysis (e.g., SFA), we performed an ilr-transformation on the dataset (Chen et al., 2018, 2016; Yousefi et al., 2014). In addition, the ilr-transformation dataset was tested using the Kaiser-Meyer-Olkin (KMO > 0.7). Test and significance of Bartlett's test of sphericity (Sig. < 0.01) to determine its factorability (Kaiser and Rice, 1974). Then we used SFA until the third stage (Table 3) to extract remarkable multi- element anomalous geochemical signatures for Ag-Pb-Zn deposits using Geochem Studio software (Gao et al., 2016). To improve the interpretability of the data, we retained factors with eigenvalues > 1 and used thresholds with loadings higher than 0.6 (Treiblmaier and Filzmoser, 2010; van Helvoort et al., 2005). Hence, the factors were regarded as 'clean' factors, and the elements (threshold value ≤0.6) were 'noisy' elements in the 'clean' factors (Yousefi et al., 2014).
First main phase Second main phase First stage (FSFA) First stage (FSFA) Second stage (SSFA) Third stage (TSFA) Element FSF1 FSF2 FSF3 FSF4 Element FSF1 FSF2 FSF3 FSF4 Element SSF1 SSF2 SSF3 SSF4 Element TSF1 Ag -0.42 0.62 -0.12 -0.06 Sn 0.04 0.56 0.42 -0.04 Ag -0.44 -0.13 -0.7 0.01 Ag 0.76 As -0.01 0.18 -0.57 0.02 Sr -0.36 -0.66 -0.67 0.06 Be -0.63 -0.19 0.38 0.16 Cd 0.89 Au -0.04 -0.35 -0.07 0.06 Th -0.36 0.33 0.06 -0.11 Cd -0.11 0.04 -0.8 0.08 Pb 0.84 B 0.33 -0.01 -0.02 0.05 Ti 0.90 -0.10 0.03 -0.07 Co 0.85 -0.06 0.08 0.04 Zn 0.90 Ba -0.54 0.01 -0.01 -0.05 U -0.41 0.00 -0.01 -0.06 Cr 0.84 -0.03 0.10 -0.04 Eigenvalue 2.87 Be -0.64 -0.34 -0.08 0.00 V 0.90 -0.10 0.03 -0.07 La -0.30 0.71 0.22 0.00 Var.% 71.8 Bi 0.03 0.55 -0.04 0.00 W -0.26 0.02 0.01 -0.04 Nb -0.21 0.84 0.00 0.21 Cum. Var.% 71.8 Cd 0.03 0.82 -0.16 0.01 Y -0.57 -0.14 0.07 -0.60 Ni 0.89 -0.11 0.03 0.02 KMO 0.75 Co 0.84 -0.09 0.03 -0.07 Zn 0.03 0.82 -0.16 0.01 Pb -0.24 0.15 -0.7 0.13 Bartlett 0.00 Cr 0.83 -0.07 0.02 0.02 Zr 0.07 0.44 0.67 -0.09 Sr -0.36 -0.68 0.29 -0.15 Cu 0.57 0.50 -0.03 0.05 Al2O3 -0.33 0.27 0.10 -0.72 Ti 0.88 0.12 0.17 0.14 F -0.28 0.27 0.06 -0.09 CaO 0.05 -0.26 -0.20 0.86 V 0.93 -0.10 0.19 0.01 Hg 0.34 -0.01 -0.02 0.05 Fe2O3 0.78 0.38 0.03 0.02 Zn 0.06 0.26 -0.9 0.19 La -0.33 -0.03 0.65 -0.10 K2O -0.60 0.43 0.01 -0.10 Zr 0.06 0.81 -0.09 0.18 Li -0.52 -0.02 -0.01 -0.05 MgO 0.76 0.03 -0.04 0.41 Al2O3 -0.29 0.37 -0.24 0.76 Mn 0.04 0.40 0.08 -0.07 Na2O -0.48 -0.42 -0.02 -0.03 CaO -0.01 -0.24 0.16 -0.92 Mo -0.27 0.33 0.07 -0.10 Eigenvalue 9.02 5.73 2.68 1.42 Fe2O3 0.79 0.30 -0.28 -0.01 Nb -0.29 0.40 0.72 -0.15 Var.% 23.75 15.09 7.05 3.74 MgO 0.80 -0.02 0.02 -0.41 Ni 0.86 -0.09 0.03 -0.07 Cum. Var.% 23.75 38.84 45.89 49.63 Eigenvalue 6.30 4.11 2.34 1.15 P 0.48 0.01 0.01 0.05 KMO 0.77 Bartlett 0.00 Var.% 35.03 22.86 13.00 6.38 Pb -0.28 0.74 -0.12 -0.04 Cum. Var.% 35.03 57.88 70.89 77.26 Sb -0.01 0.15 -0.48 0.02 KMO 0.80 Bartlett 0.00 Note: Loadings in bold represent the selected elements based on a threshold of 0.6 (the absolute threshold value) for each stage. KMO. Kaiser-Meyer-Olkin measure of sampling adequacy; Bartlett. significance of Bartlett's test of sphericity.
Table 3. Rotated factor matrix for the third-stages of staged factor analysis
A logistic function can be used to transform boundless FSs into values in the range [0, 1]. In recent logistic function association studies, this was called the geochemical mineralization probability index (GMPI) to eliminate the negative impact on the fractal model and obtain fuzzy weights for geochemical mapping (Yousefi et al., 2014). To convert the FS values into geochemical mineralization prospectivity indices (GMPI) in the range of 0 and 1, we applied the following logistic function (Yousefi et al., 2015, 2012)
where FS is the factor score of each sample obtained from the SFA. Thus, based on the third stage factor one (TSF1, Ag-Cd-Pb-Zn) acquired by third stage factor analysis (TSFA), the GMPI value for each geochemical sample was
where FSAg-Cd-Pb-Zn is a factor score describing a multi-element ore-bearing indicator in the study area. Non-linear transformation of TSF1 into the range of 0 and 1 using the logistic function (Eq. (2); Yousefi et al., 2014) has resulted in stronger discrimination between anomalous and background values.
S-A fractal model is a generalization of the concentration- area fractal model (C-A) applying a Fourier Transform to the frequency domain to achieve the separation of background and anomalies (Cheng, 1999). Anomalies are varied, and their intensity, size, and background values are also diverse, which can be determined by self-similarity routines in the frequency domain (Cheng, 2004, 2000). Based on self-similarity routines (Cheng, 2004), the "spectral energy density-area" (S-A) fractal model of multifractal fields was established
where A(≥S) represents the area that exceeds the threshold of spectral energy density, $ \propto $ represents the proportion, and β is the fractal dimension. The power-law spectral energy density (S) was determined on the double logarithmic plot, according to the relationship S=(|X (ω)|2). The scatter plot was fitted to a three-segment straight line using the least square method to find the appropriate cut-off points S1, S2 (S1 < S2) to construct a fractal filter (Liu et al., 2020). Using this fractal filter, the combined map of geochemical elements was decomposed into three different components, namely noise, background, and anomaly (Cheng et al., 2009). These two intersections were used to define the cutoff value of the filters, with S < S1 representing noise, S > S2 describing the background, and S1 < S < S2 representing anomalies. Filters were then determined as follows.
where G1(ω) is the background filter, and G2(ω) is the anomaly filter.
However, considering spatial correlations, the multifractal inverse distance weighting (MIDW) method (Cheng, 2000) incorporates a local singularity into the moving average interpolation model. Compared with the traditional moving average interpolation methods, the MIDW process has the advantages of improving interpolation accuracy and retaining the local structure of the interpolation surface (Zuo, 2011; Cheng, 2000, 1999). Because of these advantages, MIDW is more suitable for the interpolation of S-A fractal model than other interpolation methods. Anomalies and background from the S-A method could be calculated using the GeoDAS software system (Cheng, 2000, 1999).
In this study, the instrument used for detecting the trace element contents of samples was a Niton XL3t Goldd (Thermo Fisher Scientific). During the test, the system automatically corrected instrument parameters after a certain period to ensure the accuracy of the data (Lemière, 2018). Depending on the sample type being analyzed, three test modes (mining mode, soil mode, and all-around mode) can be adopted. The mining model is appropriate for analyzing elements of higher concentrations (element content > 0.2%–0.5%), the soil mode is suitable for minor and trace elements, and the all-around mode obtains comprehensive results faster than other modes (Burley et al., 2019). In this study, the all-around model was used to test the ore-related rock concentrations and improve fieldwork efficiency.
The pXRFA can effectively detect many elements (e.g., Ag, As, Cu, Cd, Pb, Sb, Zn, etc.), directly or indirectly (Druzbicka and Craw, 2013; Kovács et al., 2006). Field profile measurements of newly discovered ore-bearing rocks in the Dajiacuo region were carried out using pXRFA.
2.1. Staged Factor Analysis (SFA) and Logistic Function
2.2. S-A Fractal Model and Multifractal Inverse Distance Weighting (MIDW)
2.3. Portable X-Ray Fluorescence Analysis
We used SFA up to the third stage (Table 3): the first and second stages were used to remove the 'noisy' elements, and the third stage was used to extract the essential composite elements indicating geochemical anomalies (Yousefi et al., 2014). It can be seen from Table 3 that KMO values for each stage were > 0.75 and Sig. were all < 0.01, indicating that this dataset was suitable for factor analysis. For the dataset in this paper, the first stage factor one (FSF1) represented a Be-Co-Cr-Ni-Ti-V- Fe2O3-MgO association, FSF2 represented an Ag-Cd-Pb-Sr-Zn association, FSF3 represented a La-Nb-Sr-Zr association, and FSF4 represented an Al2O3-CaO association (Table 3).
Because of the existence of indicator elements (i.e., Ag, Pb, Zn, Cd, Sr), the FSF2 signature reflected the existence of Ag-Pb-Zn deposits. However, the 20 elements in the first stage factor analysis (FSFA) such as As, Au, did not contribute importantly to all factors; hence, they were not used in the second stage factor analysis (SSFA). After removing the 20 elements, the contributions of Ag, Cd, Pb, and Zn to the second stage factor three (SSF3) were enhanced in the SSFA (Table 3), and the negative correlation of Sr was eliminated. Factors SSF1, SSF2, and SSF4 (Table 3) were considered 'clean' because there were no 'noisy' elements.
A small number of extracted factors provide the best results with respect to interpretability (Reimann et al., 2002), and SSF1, SSF2, and SSF4 are not appropriate multi-element features. Therefore, even though each of these factors was positively correlated with elements that had a small positive or negative influence in the SSF3, the SSF3 would still include "noisy" elements for every sample (Yousefi et al., 2014). To further filter out geochemical noise, only elements with absolute loading values > 0.6 on SSF3 were used in the third stage factor analysis (TSFA) (Table 3). In the TSFA, we obtained the combination of Ag-Cd-Pb-Zn, which means that it was related to the Ag-Pb-Zn mineralization. Hence, the third stage factor one (TSF1) was selected as a factor correlated with Ag-Pb-Zn mineralization
A Fourier transform was used to transform the GMPI datasets into the frequency domain. The abscissa and ordinate were spectral energy density (S) and area (A), respectively, and plotted on a double logarithmic plot (Fig. 5). Three straight lines were fitted to the data, separated with the threshold values log S1=4.33 (S1=76.03) and log S2=8.61 (S2=5 511.45). The three lines defined three filters: S≤76.03 was the noise filter, 76.03≤S≤5 511.45 was the anomaly filter, and S≥5 511.45 was the background filter. The left-line represented high-frequency noise in the geochemical progress (Cheng et al., 2010). The midline indicated the middle-frequency anomaly, which is usually connected to complex geological processes and mineralization (volcanic rock and intersections of faults) that are favorable areas for Ag-Pb-Zn deposits. The right-line displayed the low-frequency background, which may have reflected beneficial rock types (Cheng et al., 2010). Anomalies and background maps for geochemical combination elements were obtained by transforming the three intervals of S from the frequency domain to the spatial domain using the inverse Fourier transform (Figs. 6a–6b).
Figure 6. Geochemical background (a) and anomalies (b) map of GMPI in the study area, compiled using the spectrum-area model. Dajiacuo: Ag-Pb-Zn deposit; Dajiacuoxi: Pb-Zn deposit; Beina: Pb-Zn-Ag deposit; Chagele: Pb-Zn-Ag-Cu deposit; Longgen: Pb-Zn-Ag deposit; Chazi: Pb-Zn deposit; Gergeng: Pb-Zn deposit.
The anomalies map shows that high-value areas (close to red) appear to have SE-NW trend (main) and NE-SW trend (subordinate), and occurred around intrusions and along SE-NW trending faults, or intersections of faults (Fig. 6b). They were favorable areas for the formation, preservation, and prospecting of ore deposits and indicated the potential for the discovery of Ag-Pb-Zn deposits in the study area. However, the high-value areas (near red) in the background map had an E-W trend and were mainly located in the middle of the study area (Fig. 6a). The high-value regions were firmly related to volcanic rocks and intrusions, which suggested a source of Ag-Cd-Pb-Zn in the region. The S-A method not only effectively narrowed the exploration area (Zuo, 2014, 2011) but also separated the mineralization-derived geochemical anomalies from the background.
In the attempt of any exploration target, the exploration scale and the corresponding metallogenic sub-system must be determined. For this reason, the ore-forming process is divided into three sub-systems: pre- syn-, and post-mineralization sub- systems, which is helpful to better understand their operation mode at different scales (Yousefi et al., 2019). However, the relative importance of these three subsystems depends on the scale of key ore-forming geological processes, and the improper combination of relevant target standards of different sizes will inevitably lead to biased and ineffective target exploration models (Yousefi et al., 2019). Therefore, the exploration targets from regional to deposit scale are outcrop mineralization, partial outcrop mineralization, concealed orebody and blind orebody in each subsystem. Based on the sampling density and the phenomenon of oxidation and secondary geochemical element diffusion under superficial conditions, it is inferred that the stream sediments involved in this paper belong to the regional scale and the post-mineralization sub-systems. Therefore, according to the modeling idea of the exploration information system, the measurement of the stream sediments can achieve the purpose of pointing to the potential deposit location (outcrop mineralization) (Yousefi et al., 2019, 2013, 2012).
Figure 7. DEM-based geochemical model (DGM). (a) Geochemical anomalies and (b) geochemical background map of the study area. 0. Dajiacuo Ag-Pb-Zn deposit; 1. Dajiacuoxi Pb-Zn deposit; 2. Beina Pb-Zn-Ag deposit; 3. Chagele Pb-Zn-Ag-Cu deposit; 4. Longgen Pb-Zn-Ag deposit; 5. Chazi Pb-Zn deposit; 6. Gergeng Pb-Zn deposit.
From the DGM, it appeared that the anomaly to the east of the Dajiacuoxi (No. 1) deposit was generated by mineralization. An anomalous area also exists to the west of the Beina (No. 2) deposit, which appeared to originate from it at a higher elevation, but the area was relatively small. Chagele (No. 3) and Longgen (No. 4) Deposits were the two most significant deposits in the study area, with Pb-Zn reserves of more than 1.1 million tons (Zhang et al., 2018; Gao et al., 2012). Therefore, the area of geochemically anomalous responses to the two deposits was more extensive than that of other deposits. The anomalous area to the east of the Gergeng (No. 6) deposit was presumed to be a geochemical halo produced by weathering and leaching of the deposit. At the same time, the anomaly to the north was not its down-drainage, as indicated by the DGM.
However, none of the deposits in the study area were located in the center of the geochemical anomalous halo. Due to the critical influence of topography and drainage, the elements migrated under supergene conditions. The flow rate of water increased as relief increased, and weathering and denudation also accelerated, which was conducive to the migration of elements (Fig. 8). The downstream dilution of stream sediment anomalies was derived from the erosion and dissolution of water (Jiang et al., 2006; Hawkes 1976; Figs. 8a–8b). Steep terrain is prone to landslides and collapse, resulting in the displacement of an anomaly from its source (Zhang 2007; Figs. 8c–8d). Furthermore, the elements Ag, Cd, Pb, and Zn are chalcophile elements, and their chemical properties are very similar (Goldschmidt, 1937). Cadmium is generally present in sphalerite in the form of isomorphism, and silver is present in galena in the form of isomorphism or mechanical mixing (Levinson, 1980). According to the DGM and the migration behavior of the elements Ag, Cd, Pb, and Zn, we speculate that it was likely that there was an Ag-Pb-Zn deposit in the Dajiacuo region (Fig. 9). The DGM not only intuitively produces three-dimensional (3D) visualization of anomalies and backgrounds (Zuo et al., 2016), but also provides an excellent method to identify sources of anomalies.
Figure 8. Schematic diagram of dispersion from mineralization. (a) Definition of drainage parameters. (b) The downstream dilution of stream sediment anomalies. (c) Separation anomaly caused by the landslide. (d) Buried halo formed by collapse. The dashed and solid lines represent the state before and after the collapse and landslide, respectively, in (c) and (d).
According to the geochemical anomalies in the Dajiacuo region, we measured the cross-section of the mineralized zone at intervals of about 50 m (Fig. 9). pXRFA was used to analyze samples along two transects, with measurements taken at intervals of approximately 50 m. The first transect (PM01) consisted of 14 points and the second (PM02) had 35 points (Fig. 10). One of the applications of this method is to measure ore grade with reliable accuracy (Lemière, 2018; Bosco, 2013). The statistical parameters for the grade values of the samples measured by pXRFA are shown in Table 4.
Figure 10. Section diagrams of element grades of Ag, Cd, Pb and Zn on sections 1 (PM01) and 2 (PM02) at Dajiacuo using pXRFA.
Element Cutoff grade Highest grade The arithmetic mean of exceeding the cutoff grade Ag 40 ppm 147.9 ppm 62.6 ppm Pb 0.30 wt.% 4.20 wt.% 1.70 wt.% Zn 0.50 wt.% 24.20 wt.% 6.00 wt.%
Table 4. Statistical parameters for the grades in the Dajiacuo region.
The highest grades of Pb and Zn were found at the center of the mineralized zone, and Ag also showed a good relationship with Pb and Zn in the profile. The principal host minerals for Pb and Zn are galena and sphalerite, respectively, which can be seen with the naked eye. However, it is challenging to identify scarce ore minerals such as argentite in the field. Argentite is easily hosted in the galena crystal lattice, which can be identified by pXRFA (Fig. 10). With the assistance of the pXRFA, we verified the presence of Ag, Pb, and Zn above the cutoff grades in the anomalous area.
The new exploration was used to verify that the geochemical anomaly area of Dajiacuo (No. 0) (Fig. 7b) is a newly discovered deposit in the southwestern part of the study area. The skarn mineralization alteration zone (length: 500 m, width: 300 m) hosting Pb-Zn mineralization (length: 200 m, width: 100 m) was found successfully, and the skarn and the NW-trending faults strictly controlled mineralization. The alteration characteristics showed strong skarnization and epidotization (Figs. 11a–11b), and the mineralization consisted of galena and sphalerite (Figs. 11e–11f), accompanied by malachite and pyrite (Figs. 11c–11d).
Figure 11. Representative photographs of rocks from geochemically anomalous areas in the Dajiacuo deposit. (a) Epidote skarn; (b) garnet skarn; (c) massive pyrite hosted in tuff; (d) malachite staining on epidote skarn; (e), (f) sphalerite and galena hosted in the tuff. Mineral symbols: Grt. garnet; Ep. epidote; Mlc. malachite; Gn. galena; Sph. sphalerite; Py. pyrite.