Rock alteration at the post-Variscan nonconformity: implications for Carboniferous–Permian surface weathering versus burial diagenesis and paleoclimate evaluation

A nonconformity refers to a hiatal surface located between metamorphic or igneous rocks and overlying sedimentary or volcanic rocks. These surfaces are key features with respect to understanding the relations among climate, lithosphere and tectonic movements during ancient times. In this study, the petrological, mineralogical and geochemical characteristics of Variscan basement rock as well as its overlying Permian volcano-sedimentary succession from a drill core in the Sprendlinger Horst, Germany, are analyzed by means of polarization microscopy, and environmental scanning electron microscope, X-Ray diffraction, X-ray fluorescence and inductively coupled plasma mass spectrometry analyses. In the gabbroic diorite of the basement, the intensity of microand macro-fractures increases towards the top, indicating an intense physical weathering. The overlying Permian volcanic rock is a basaltic andesite that shows less intense physical weathering compared with the gabbroic diorite. In both segments, secondary minerals are dominated by illite and a mixed-layer phase of illite and smectite (I– S). The corrected chemical index of alteration (CIA) and the plagioclase index of alteration (PIA) indicate an intermediate to unweathered degree in the gabbroic diorite and an extreme to unweathered degree in the basaltic andesite. The τ values for both basaltic andesite and gabbroic diorite indicate an abnormal enrichment of K, Rb and Cs that cannot be observed in the overlying Permian sedimentary rocks. Accompanying minerals such as adularia suggest subsequent overprint by (K-rich) fluids during burial diagenesis which promoted the conversion from smectite to illite. The overall order of element depletion in both basaltic andesite and gabbroic diorite during the weathering process is as follows: large-ion lithophile elements (LILEs)> rare earth elements (REEs)> high-field-strength elements (HFSEs). Concerning the REEs, heavy rare earth elements (HREEs) are less depleted than light rare earth elements (LREEs). Our study shows that features of supergene physical and chemical paleo-weathering are well conserved at the post-Variscan nonconformity despite hypogene alteration. Both can be distinguished by characteristic minerals and geochemical indices. Based on these results, a new workflow to eliminate distractions for paleoclimate evaluation and evolution is developed.

widely covered by only a thin layer of Cisuralian volcano-sedimentary rocks and the contact surface has been penetrated by 65 numerous drillings (Kirsch et al., 1988).
For this study, we selected a representative drill core reaching from unweathered basement rock into the volcanicsedimentary cover which was analyzed at high resolution in particular near the nonconformity. This drill core allowed not only to study alteration in Variscan basement rocks, but also subsequent alteration of overlying early Permian basalt. We carried out a detailed petrographical, mineralogical, and geochemical study and applied a new workflow to distinguish the 70 supergene and hypogene alteration processes within a first-order nonconformity. The workflow is based on normalizing mineral types and geochemical weathering indices to un-overprinted conditions and to quantify the observed deviations.
With corrected geochemical and mineral information of the weathered profile, the weathering and palaeo-climatic condition and the alteration scenario were addressed. wackes and siltstones. Both the top of the basement rock and the top of the volcanic lava constitute palaeosurfaces which faced intense alteration throughout their exposure which is supposed to be significantly shorter for the volcanic rocks. The macroscopic alteration underneath these surfaces is intense which is expressed by a higher degree of fracturing, bleaching and grain disaggregation. The degree of macroscopic alteration decreases with increasing depth in these parts. The core offers the unique opportunity to study subsequent weathering intervals and to compare typical endmembers of rock types 150 with weak (granodiorite) and strong (andesitic basalt) vulnerability to chemical weathering.
In total 24 samples were extracted from the GA1 drill core from which 11 samples belong to the basement, 6 samples to the overlying volcanic rock (Fig. 1) and 7 samples to the sedimentary rocks of the Lower Langen Formation. In order to capture the small-scale petrographic and geochemical variations in the weathering zones appropriately, we reduced the sampling interval in the topmost part of the volcanic rock and in the basement to 40 cm, and for the rest part, to avoid the fractures, the 155 interval is around 1 m.
The samples were used to prepare thin sections which have been analyzed by polarization microscopy and SEM/EDX for their petrographic characteristics. For mineral composition, trace element and major element analyses all samples were crushed and milled into a powder with a diameter of less than 63 microns. One part of the powder was examined by X-ray diffraction (XRD) at Goethe University Frankfurt for whole rock mineral composition. The preparations were poor in 160 orientation by carefully backside filling of the powder. Using a PANalytical X'Pert diffractometer equipped with a braggbrentano goniometer (Copper beam), each sample was scanned under 40 kV and 30 mA for 2 hours. The start angle was 2.5°, the end angle was 70.0° and a step size of 0.008° was applied. The time for each step was 50 s. The mineral phase proportions were estimated by weighted XRD peak intensities after conversion with their typical Reference Intensity Ratios (RIR) as found in the powder diffraction file (PDF-2 and PDF-4 of the International Centre of Diffraction Data: 165 www.icdd.com, see Tab. 1) with the software MacDiff (Petschick et al., 1996).
The powder samples were sent to State Key Laboratory of Isotope Geochemistry, Guangzhou Institute of Geochemistry, Chinese Academy of Science for examination of major elements by X-ray fluorescence (XRF) and examination of trace elements by Inductively Coupled Plasma Mass Spectrometry (ICP-MS). Before XRF analysis, samples were roasted under 900 ℃ for 3 hours and weighed before and after heating to measure the loss on ignition (LOI). Subsequently, 0.51-0.53 g of 170 each sample was mixed with a the ratio of 1:8 with Li2B4O7 and fused at 1150 ℃ in a Pt-crucible to make a glass disk for XRF analysis.
For trace and rare earth elements analysis, firstly, 40 mg sample powder was weighed and placed into high-pressure-resistant Savillex Teflon beakers to which 0.8 ml 1:1 HNO3, 0.8 ml HF and 0.5 ml 3N HClO4 was added. The mixture was heated for 48 hours under 100 ℃ and then evaporated. Secondly, 0.8 ml 1:1 HNO3 was added and heated under 100 ℃ for 12 hours. 175 Thirdly, 0.8ml HF and 0.5ml 3N HClO4 was added and the beaker was sealed and moved into an oven with a temperature of https://doi.org/10.5194/se-2020-221 Preprint. Discussion started: 25 January 2021 c Author(s) 2021. CC BY 4.0 License.
190 ℃ for 48 hours to make sure the sample was completely dissolved. Fourthly, the beakers were opened, the solution was evaporated and 4 ml 4N HNO3 were added. After that, the beakers were sealed and moved into the oven with a temperature of 170 ℃ for 4 hours. Lastly, the solution was diluted with 3% HNO3 until the weight of the solution was 250 times the weight of the sample. 0.25 g of the solution was taken and diluted with 3% HNO3 to 2.00 g, mixed with 2.00 g Rh-Re 180 internal standard solution and examined by ICP-MS. To monitor the analytical quality, international standards of GSR-1(granite), GSR-2(andesite), GSR-3(basaltic andesite) were applied.

Lava 195
The lava has a phaneritic, amygdaloidal texture. Fractures in this part are very limited and occur between 13.8 and 14.7 m with high angle to horizon. The width is less than 1 mm and the fractures are also filled with secondary minerals (Fig. S1).
The fresh part of the volcanic rock mainly consists of plagioclase crystals and amygdaloid bodies which consist of calcite and zeolite locally accompanied with chalcedony ( Fig. 2F). With a decrease in depth, the content of plagioclase gradually decreases under the microscope. However, the thin section in the topmost part of the lava (13.9 m) conduct a sudden change, 200 compared with the samples from lower parts. Here nearly all primary minerals were altered to secondary clay minerals but the primary grain shapes are still relatively intact (Fig. 2G). Some residual grains consist of a kaolinite rim and acore of illite or illite-smectite (I-S) mixed layers (Fig. 2H). In the void, adularia with kaolinite can be observed (Fig. 2I).

Plutonic rock 205
The XRD results are listed in the supplementary material (Table S1) and plotted in Fig. 3. The plutonic basement part is composed of Ca-poor plagioclase, K-feldspar (only in one sample), quartz, amphibole (mainly Mg-hornblende), and mica phases, with secondary minerals of illite (not separable from micas), vermiculite, I-S mixed layers, minor kaolinite, anatase, hematite, calcite and dolomite. Amphibole abundance decreases in the middle part and disappears in the uppermost part. In the topmost part (20.6-23.5 m) amphibole is not found. Also, plagioclase decreases in abundance from around 40% at the 210 bottom to about 8% at the top. Considering the mineralogical composition of the fresher parts of the plutonic rock, based on the QAPF diagram for plutonic rocks, the protolith of the basement rock can be classified as a quartz diorite/tonalite ( Fig.   S2), which corresponds well with Mezger (2013).

Lava
The fresh samples of the volcanic rock are composed of augitic pyroxene and plagioclase. In the weathered part, most of the 215 phases appear as secondary minerals, such as quartz, hematite and anatase, and clay minerals such as illite, mixed layer illitesmectite (I-S), vermiculite, and kaolinite, as well as carbonate minerals like calcite and minor dolomite. The uppermost part of the lava is dominated by I/S mixed-layer minerals while the plagioclase content is less than 5% and pyroxene is absent.
The abundance of plagioclase (and pyroxene) gradually decreases from bottom to top, and illite and vermiculite exhibit an increasing tendency ( fig. 3). Based on the QAPF diagram for volcanic rocks, the protolith of the weathered volcanic rock can 220 be classified as a basalt or andesite (Fig. S2).

Geochemical rock classification
In order to further verify the lithological type of the plutonic and volcanic rock in the GA1 drill core, for the basement part, the geochemical data from comparatively fresh samples (23.5-56.5 m ) are plotted in the TAS-diagram (Middlemost, 1994) 225 ( Fig. 4). Here, the results are mainly plotting in the gabbroic diorite field (Fig. 4A), which is grossly consistent with the results from the petrographic classification (Fig. S2). For the volcanic rocks, due to the unaltered samples is very limited, to minimize the effects of alteration, the Revised Winchester-Floyd diagram is used which is based on immobile trace elements (Pearce, 1996). Most lava samples fall into the andesite/basaltic andesite field (Fig. 4B), which is in accordance with the petrographic classification result as well. In the following, we use the result of chemical classification and term the two 230 protolith rocks as gabbroic diorite and basaltic andesite.

Major elements
The concentrations of major elements are listed in supplementary material (Table S2)

Trace elements 245
Trace element data is also given in the supplementary material (Table S3) Concerning the basaltic andesite part, HFSE such as Zr, Hf, Nb, Ta and Th (in ppm) all exhibit an increasing tendency from bottom to the top with a sharp increase at the topmost sample. LILE such as Rb and Cs also show an increasing tendency, whereas Sr reveals an opposite trend with a decrease from 123 ppm at the bottom to 47 ppm in the topmost part. The tendency for Ba is irregular compared to the other elements, but the overall trend is decreasing.

Rare Earth Elements (REEs) 255
REE concentrations are listed in supplementary material (Table S4) and are shown as chondrite-normalized patterns (McDonough and Sun, 1995) in Fig. 7. The calculation for the anomalies of cerium (Ce) and europium (Eu) are defined as follows: and 260 The Ce* and Eu* are the hypothetical concentrations of trivalent Ce and Eu, XN represents the normalized value of the element X. The distribution patterns of both gabbroic diorite and basaltic andesite are nearly parallel in different depths and exhibit decreasing values from bottom to the top (Fig. 7). All samples are moderately enriched in light rare earth elements (LREEs) and have gently right-dipping REE patterns. They exhibit no Ce anomalies and slightly negative Eu anomalies. 265

Chemical alteration
During chemical weathering alkalis and alkaline earth elements contained in silicates such as feldspar, mica minerals, pyroxene and amphibole will be gradually depleted while aluminum tends to remain in situ generating clay minerals (Clift et al., 2014;Nesbitt and Young, 1982;Vázquez et al., 2016). Based on this mechanism, different types of weathering indices 270 were developed to evaluate the weathering intensity. Among those, the Chemical Index of Alteration (CIA) and Plagioclase Index of Alteration (PIA) were proposed by Nesbitt and Young (1982) and Fedo et al (1995), respectively. They are defined as: and 275 All portions are given in molecular weight and in both equations CaO* is the Ca content within silicate minerals only. For the correction of CaO* the approach proposed by Mclennan (1993) is applied as follows: CaOrest= CaO -P2O5*10/3, whereas the P2O5 is related to apatite, CaO* = CaOrest when CaOrest<Na2O, otherwise, CaO* = Na2O. With this calculation, the CIA value for fresh feldspar is about 50, unaltered basaltic andesite is between 30 and 45, granitoids range between 45 and 55, 280 illite is from 75 to 85, muscovite yields a value of 75 and kaolinite and chlorite have the highest value of nearly 100 (Fedo et al., 1995). As a modification of the CIA, the PIA value for fresh rock is around 50, and for clay minerals such as kaolinite, illite, and gibbsite it is close to 100 (Fedo et al., 1995;Patino et al., 2003). 74% of the constituent is I/S. This is well expressed by the PIA, which yields a value of 98, hence, the PIA is more consistent with the mineral and petrological character compared to the CIA.  (Fedo et al., 1995;Nesbitt and Markovics, 1997).
To better evaluate the weathering intensity, the A-CN-K ternary diagram is applied (Fedo et al., 1995;Nesbitt and Young, 1984). The letter A stands for Al2O3, CN for (CaO*+Na2O) and K is the content of K2O all in molecular proportions. The 290 ideal weathering trend for different types of parent rocks in the upper continental crust should be parallel to the A-CN axis, but due to diagenetic alteration the original data tend to deviate from the theoretical weathering trend (Babechuk et al., 2015;Fedo et al., 1995;Zhou et al., 2017). The trends for both basaltic andesite and gabbroic diorite samples uniformly deviate from the ideal weathering tendency and excurse to the K apex (Fig. 8A), which clearly indicates a relative K enrichment. The enrichment of K is always interpreted as K metasomatism due to conversion among clay minerals such as the transformation 295 from kaolinite to illite, or from plagioclase to K-feldspar (Fedo et al., 1995;Nesbitt and Young, 1984;Zhou et al., 2017). K metasomatism results in a lower CIA value relative to the actual weathering intensity and can also explain the deviation from the PIA. To address this problem, Fedo et al. (1995) suggested that the proportion of "pre-metasomatic" compositions of the weathering products could be determined by correcting each point on the A-CN-K diagram back to its predicted position.
The method proposed by Panahi et al. (2000) for the K correction is applied: 300  (Fig. 8B). These values are much lower than values from the topmost part and suggest an incipient weathering degree. Based on the XRD results, the plagioclase content at the bottom is 77% and declines gradually upsection to 54% in 320 14.5 m before it suddenly drops to 0% in 13.9 m. In 14.5 m depth, the plagioclase grains are still mostly fresh, which is clearly different from the uppermost half meter in which they are completely altered.

Quantification of elements transfer
For quantification of element transfer due to weathering and diagenesis the τ model is applied (Anderson et al., 2002;Nesbitt, 1979;Nesbitt and Markovics, 1997). The model uses the relation between the mobile element concentration in the 325 sample (Msample) vs. protolith (Mprotolith), and the immobile element concentration in the sample (Isample) vs. protolith (Iprotolith).
Among the immobile elements, Ti is widely used as a key element (Middelburg et al., 1988). Thus the model is defined as: When τM > 0, element M is enriched during alteration, when τM = 0, element M is immobile during alteration, when 0> τM > -1, element M is depleted, when τM = -1, element M is completely lost from the material. For the basaltic andesite and 330 gabbroic diorite, the samples from the bottom (19.3 m and 55.5 m, 56.5 m, respectively) were selected as protoliths to provide the lowest degree of alteration based on CIA values, petrographic features and XRD results.
The results for major and trace elements are listed in supplementary material (Table S5) and are plotted in Fig. 9. In the basaltic andesite part, both Ca and Na are strongly depleted in the topmost part (13.9 m) with τCa and τNa of -0.93 and -0.99 respectively. τCa gradually increases from -0.60 to -0.07 and τNa from -0.60 to -0.18 with increasing depth from 14.3 m down 335 to 17.8 m. The discontinuity between the topmost and the rest part correspond well with the CIA and PIA value. However, in the gabbroic diorite part, the τNa values from the top gradually increase from -0.70 (20.6 m) to -0.67 (21.5 m), followed by a sharp increase in the lower part with values between -0.12 and 0.06. This is consistent with the trends of both CIA and PIA. τCa values show a high variability compared to τNa. The sample from 20.6 m yields a τ value of -0.03 which indicates a slight depletion of Ca while the sample from 21.5 m shows a τ value of 0.17, which suggests a slight enrichment of Ca (Fig. 9B). A 340 similar enrichment also occurs in 23.5 m, which yields a value of 0.13 (Table S5).
The Sr/Ca ratio can be applied as a parameter to distinguish different phases of diagenetic fluids (Berndt et al., 1988;Brandstätter et al., 2018). To figure out the source of Ca, the Sr-CaO diagram is applied (Fig. 9A). The ratio of Sr/CaO at the top of the gabbroic diorite in 20.6 m and 21.5 m depth show a close relation with the Sr/CaO ratios of the basaltic andesite and clearly deviate from the general trend of the gabbroic diorite and the overlying sedimentary rocks. This hints to a 345 chemical overprint of the gabbroic diorite by the overlying basaltic andesite, whereas the Ca in the lower part appears to be primary. The spike of τCa in 23.5 m depth can also be explained by porosity data (Weinert et al., 2020) which can be considered being a measure of fracture density and grain disaggregation in igneous rocks. The porosity decreases sharply from 24% in 21.5 m to 3% in 23.6 m depth (Fig. 9B). This can be explained by the fact that before the Permian basaltic to andesitic lava flow flooded the basement, the fractures provided pathways for meteoric water. Ca was leached by meteoric 350 water and transferred downward through these fractures and accumulated around the interface, where the porosity sharply decreases.
In contrast to Ca and Na, elements closely related to clay formation such as K, Rb and Cs show significant enrichment (Fig.   6). In the gabbroic diorite part, the τ values of K, Rb and Cs decrease from 2.5 to 0.2, from 5.4 to 0.5 and from 19.8 to 2.3 from top to bottom, respectively. Similarly, they decrease in the basaltic andesite from 4.2 to -0.4, from 5.5 to -0.4, and from 355 8.4 to 0.2. To search for the origin of this enrichment, we plotted correlation diagrams for gabbroic diorite, basaltic andesite, and sediments (Fig. 9C). The linear and close relationships between Cs, Rb and K in the gabbroic diorite and basaltic andesite point to a joint alteration of both, whereas the overlying sediments can be excluded as a source. Palmer and Edmond (1989) claimed that mobile elements such as K, Rb and Cs are very easily being extracted by hydrothermal fluids and transferred during hydrothermal activity. In addition to clay transformation, typical hydrothermal minerals are observed in thin-sections and XRD such as dolomite with accompanied secondary quartz and adularia (Fig. 3). This observation supports the model of a second alteration which was pervasive through the nonconformity and must have happened during burial diagenesis by geothermal fluids.  In both gabbroic diorite and basaltic andesite, the τ values gradually increase from LREE to HREE (Fig. 10). This indicates that the depletion degree from LREE to HREE is reducing during the alteration process if the same conditions exist. In a 370 study of granodiorite, Nesbitt (1979) showed that REE are removed by acidic leaching of meteoric water which becomes buffered with depth and loses its etching effect due to rising pH. Moreover, Nesbitt (1979) proposed that the fractionation of LREE and HREE may be controlled by the mineral type. Kaolinite and illite are favorable for LREE while vermiculite, Fe-Ti-oxy-hydroxides, relict hornblende and biotite are more favorable for HREE. In our study, the reducing depletion degree from LREE to HREE in both weathering profiles also indicate that in the same acid weathering environment LREE are more 375 mobile than HREE. Also, REE fractionation of the gabbroic diorite is less systematic from top down compared with the basaltic andesite (Fig. 10B). By comparison with rock textures, we assume that this is due to physical fracturing and more heterogeneous chemical alteration in the basement. The strongly depleted samples at 20.6 m and 22.5 m depth are close to fracture zones and possibly more affected by leaching. In contrast, the enriched samples from 21.5 m and 23.5 m depth are without macro-fractures. This can be explained by acidic meteoric water which used the pathways provided by the macro-380 fractures in the topmost part of the gabbroic diorite. REEs in the fracture zone were leached and transported downward and accumulated around the interface of macro-fracture and macro-fracture-free zone. This is comparable to the behaviour of Ca.
The high field strength elements (HFSE) Zr, Nb, Hf, Ta and Th are expected to be immobile. In the gabbroic diorite the τ values scatter significantly for specific elements within single samples. The samples in 21.5 m and 23.5 m depth show mostly enriched values together with Pb and U, which is in line with REE. Other element shifts appear to be controlled by 385 heterogeneous conservation and alteration of specific minerals due to the fractured and granular texture of the rock. The depletion of Pb and U can be well explained by oxidation during weathering into the mobile species Pb 6+ and U 4+ , respectively and subsequent leached by meteoric water. In addition, from LILE to HFSE, the overall depletion degree decreased in both basaltic andesite and gabbroic diorite part during alteration process.

Burial diagenesis 390
In addition to correct the alteration tendency, the A-CN-K diagram can also be used to kinetically predict the primary weathering products of plutonic and volcanic rocks (Nesbitt and Young, 1984;Panahi et al., 2000). In this case, this concept is applied to separate surface weathering and burial diagenesis by comparing the remaining secondary minerals in the profile with the theoretical weathering products. Fig. 11. Theoretical weathering tendency and retrograde tendency during burial diagenesis (Fedo et al., 1995), a-observed weathering trend for both basaltic andesite and gabbroic diorite; b-theoretical weathering trend for basaltic andesite; c-retrograde trend of topmost basaltic andesite; d-retrograde trend of the rest part of basaltic andesite; e-theoretical weathering trend for gabbroic diorite; f-retrograde trend for gabbroic diorite.

395
Rock types which contain only tiny amounts of K tend to be weathered directly by forming smectite and kaolinite instead of illite regardless under what kind of climate owing to the chemical composition restriction (Nesbitt and Young, 1989). In our case, both gabbroic diorite and basaltic andesite contain minor K. However, secondary minerals mainly consist of illite which makes metasomatic addition of this element during burial diagenesis highly probable (Fedo et al., 1995). As discussed in section 5.2, K-metasomatism is most likely promoted by hydrothermal activities. According to the A-CN-K diagram (   climate. Physical weathering in the basaltic andesite is much weaker compared to the gabbroic diorite. The reconstructed pristine clay mineral association is dominated by kaolinite which indicates a change of climate to humid conditions. An overprint by burial diagenesis is indicated by K metasomatism, clay mineral transformation and neoformation of minerals, as well as enrichment of K, Rb and Cs in the alteration zone. In both, gabbroic diorite and andesitic basalt, the primary 495 weathering products such as smectite and kaolinite were transformed into illite, and newly formed minerals are found, (e.g., adularia, dolomite, secondary quartz).
At both nonconformities, distinct gradients from the top downward demonstrate that surficial weathering is the major alteration process. During burial diagenesis, hydrothermal fluids preferentially percolated along the post-Variscan nonconformity and at the basalt/sediment boundary, due to a higher permeability. This led to clay mineral transformation and 500 some neoformation of minerals, but did not change the alteration pattern. Deeper parts of both parent rocks are pristine and not affected neither by surficial weathering nor by hydrothermal fluids during burial diagenesis.
Our case study shows that surface weathering in the past has a prime control on the petrography and geochemistry, and also guides fluids through the system during burial diagenesis. Moreover, we could demonstrate that formation of the saprolite zone depends on rock composition, climatic conditions, as well as the duration of the process. Our results have implications 505 for paleoclimatic studies. In order to separate hypergene and supergene alteration, we provide a workflow for nonconformities and shed light on the use of paleo-weathering surfaces for paleoclimate research.

Sample availability
All samples are available at the Institute of Applied Geoscience, TU Darmstadt and can be requested from liang@geo.tudarmstadt.de.