Two distinct age groups of melilitites, foidites, and basanites from the southern Central European Volcanic Province reflect lithospheric heterogeneity

Petrographic observations and in situ U–Pb ages of melilitites, foidites, basanites, phonolites, and trachytes from the southern part of the Central European Volcanic Province (CEVP) and related plutonic inclusions therein reveal two distinct age groups separated by a gap of ~ 20 Myr. A late Cretaceous to early Eocene group (~ 73–47 Ma; Taunus, Lower Main plain, Odenwald and Kraichgau area, Bonndorfer Graben and Freiburger Bucht area, Vosges and Pfälzerwald) is characterized by nephelinites and basanites mostly devoid of melilite and perovskite, and by rare haüynites, and trachytes. In contrast, a late Oligocene to late Miocene group (~ 27–9 Ma; Lorraine, southern Upper Rhine Graben, Urach, Hegau area) is dominated by melilitites, melilite-bearing nephelinites (both carrying perovskite), and phonolites. Both magmatic episodes are related to domal topographic uplift, erosion, and formation of major angular unconformities in the Upper Rhine Graben, suggesting an association with dynamic topography interrupted by phases of subsidence (or abatements of uplift). The investigated rocks in the southern CEVP (south of a line Eifel–Vogelsberg–Rhön–Heldburg), except for the Kaiserstuhl volcanic complex, mostly comprise small and isolated occurrences or monogenetic volcanic fields, whereas the northern CEVP is dominated by large volcanic complexes and dyke swarms, which are mostly SiO2-saturated to weakly SiO2-undersaturated. In the northern CEVP, evidence of spatially varying but recurrent volcanic activity exists since the Eocene, lacking the distinct 20 Myr gap as documented from the southern CEVP. While the temporal and spatial distribution of volcanism are a result of the Cretaceous to Miocene tectonic evolution in Central Europe, further studies are needed to explain the petrographic differences between the two age groups in the south.


Introduction
The late Cretaceous-Holocene Central European Volcanic Province (CEVP; Fig. 1, Table 1) is part of the Circum-Mediterranean Anorogenic Cenozoic Igneous province (CiMACI; Lustrino and Wilson 2007) and comprises several volcanic regions in eastern France, western, central, and southern Germany, western Poland, and the Czech Republic (e.g., Lustrino and Wilson 2007;Schmitt et al. 2007;Sirocko et al. 2013). Many of these regions are characterized by polygenetic, others by monogenetic igneous activities. Some of them comprise large-scale lava sheets or pyroclastic deposits, while others consist of accumulations of vents, stocks, or dykes. The volcanic products investigated in the studied area (Table 1) comprise basalts, basanites, tephrites,  latites, trachytes, phonolites, nephelinites, haüynites, melilitites, and carbonatites (e.g., Abratis et al. 2009;Bogaard and Wörner 2003;Braunger et al. 2018;Büchner et al. 2015;Dunworth and Wilson 1998;Jung and Hoernes 2000;Jung et al. 2005Jung et al. , 2006Jung et al. , 2011Kolb et al. 2012;Kramm and Wedepohl 1990;Schubert et al. 2015;Skála et al. 2015;Ulrych et al. 2016;Wörner et al. 1986). The volcanic activity in the CEVP is closely related to the evolution of the European Cenozoic Rift System (ECRiS; e.g., Dèzes et al. 2004;Merle 2011), in particular to the changes in the Alpine foreland stress field and the local reactivation of Variscan-age lithosphere-scale faults (Dèzes et al. 2004;Goes et al. 1999). The relationship between uplift, rifting, and volcanism in Central Europe is controversial, and several competing and complementary models of melt generation and melt sources have been proposed (e.g., Goes et al. 1999;Lustrino and Wilson 2007;Pfänder et al. 2018 and references therein). Explanations range from asthenospheric mantle upwelling caused by thermal instabilities (active upwelling) to melt formation by lithospheric deformation and thinning (passive rifting), both resulting in decompression melting (Lustrino and Wilson 2007;Pfänder et al. 2018;Ulrych et al. 2013). For the former case, an active large-scale mantle plume system, anchored near the core-mantle boundary layer (e.g., Albers and Christensen 1996;Goes et al. 1999), or several smaller (finger-like) plumes sourced in the asthenosphere or the transition zone have been proposed as the cause for magmatism (e.g., Dèzes et al. 2004;Goes et al. 1999;Granet et al. 1995;Haase et al. 2004;Hegner et al. 1995;Mertz et al. 2015;Ritter et al. 2001;Wedepohl and Baumann 1999). Opposing concepts that reject the dominant influence of mantle plumes invoke a partly metasomatized upper mantle beneath the CEVP that interacted with fluids and/or melts derived from subducted ancient oceanic and continental lithosphere, resulting in the presence of (low-melting) hydrous and partly carbonatebearing mantle domains (Blusztajn and Hegner 2002;Jung et al. 2005;Pfänder et al. 2018;Ulrych et al. 2008). Such models assume that decompression and subsequent melt formation are primarily caused by passive asthenospheric upwelling as a result of lithospheric thinning, uplift, and erosion (Eynatten et al. 2021;Fichtner and Villaseñor 2015;Lustrino and Carminati 2007).
Considering the variable composition, the different ages and magma volumes, and the occasional lack of spatial correlation between volcanism, graben structures, and thinned lithosphere for CEVP volcanic activity (Dunworth and Wilson 1998;Goes et al. 1999;Pfänder et al. 2018), it seems likely that a single general model alone cannot explain the petrogenesis of all the occurrences (Eynatten et al. 2021;Lustrino and Wilson 2007;Mertz et al. 2015). Rather, the spatial chemical heterogeneity of the lower lithospheric mantle, the formation depth and origin of the magmas, and the varying position of the lithosphere-asthenosphere boundary must be integrated into a comprehensive geodynamic scenario (Lustrino and Carminati 2007;Lustrino and Wilson 2007;Puziewicz et al. 2020).
Here, we present a detailed study on primitive to evolved, mostly SiO 2 -undersaturated rocks of the southern CEVP applying in situ U-Pb dating of the magmatic phases perovskite, apatite, titanite, zircon, and pyrochlore using LA-ICP-MS analyses on samples carefully studied petrographically. Considering the observed spatial and compositional variations, the present study not only contributes to understanding the chronology of magmatism in the CEVP, but provides also inferences on the driving forces responsible for the volcanic activity, regarding its close link to the preceding melting processes in the upper mantle.

3
Most of the studied occurrences are tied to obvious tectonic structures (Fig. 4): the Kaiserstuhl volcanic complex and several isolated stocks and diatremes are located within the URG between the Variscan massifs of the Vosges and the Schwarzwald. The volcanic dykes and vents in these massifs occur on the flanks of the URG (e.g., Freiburger Bucht) and in the WNW-trending Bonndorfer Graben, merging eastwards into the Hegau Graben that hosts additional dykes, plugs, domes, and larger tuff deposits. The Albstadt shear zone is a seismically active, few km wide, deformation zone Fig. 2 Map of SW Germany and E France showing the study areas of this work as circular diagrams representing the semi-quantitative distribution of groups of SiO 2 -undersaturated rock types in each region determined by number of known occurrences. The circle size roughly correlates with the number of described occurrences (Lorraine: n = 2, Bonndorfer Graben and Freiburger Bucht area: n = 31), except for the Kaiserstuhl, for which surface coverage percentages are used. The rock types were determined by microscopy of 215 thin sections from approx. 120 localities in the southern CEVP. For regions that also contain strongly evolved rocks, these are additionally stated (Mader et al. 2021) that trends parallel to the URG and connects the Hegau Graben with the westernmost margin of the Urach volcanic field, which is delimited by several minor faults (Fig. 4;Ring and Bolhar 2020). The volcanic rocks of the Lower Main plain occur along the eastern flank of the URG (Sprendlinger Horst), those from Forst (Pfälzerwald) and the more southerly Vosges along the western flank. The basanites and nephelinites in the Taunus trend parallel to the South Hunsrück-Taunus border fault, while the volcanic rocks in the Odenwald and Kraichgau area and Lorraine are not associated with any obvious structure.
The study area comprises some of the most primitive melilitites and nephelinites of the entire CEVP (Fig. 3). However, while several studies on the petrogenesis of such rocks exist for the northern part of the CEVP (e.g., Duda and Schmincke 1985;Harmon et al. 1987;Jung et al. 2006;Pfänder et al. 2018;Skála et al. 2015;Ulrych et al. 2008Ulrych et al. , 2011Ulrych et al. , 2013Ulrych et al. , 2016, similar occurrences in the southern CEVP have been much less investigated (Blusztajn and Hegner 2002;Dunworth and Wilson 1998;Hegner et al. 1995;Hegner and Vennemann 1997) and comparative work involves SiO 2 -saturated, basaltic and evolved rocks (e.g., Thirty-four new analyses performed according to the XRF method reported in Braunger et al. (2018) were supplemented by literature data from Dunworth andWilson (1998), Frenzel (1975), Hegner et al. (1995), Hurrle (1976), Keller et al. (1990), Krause and Weiskirchner (1981), Neumann et al. (1992), Staesche (1995), Stähle and Koch (2003), and Stellrecht and Emmermann (1970). All data have been renormalized to a volatile-free composition. Note that the naming of the rocks in the petrography chapter was done based on the mineral composition, which is why the position in the TAS diagram may differ from the assigned rock name. No whole-rock analyses are available for the trachytes of the Lower Main plain 1 3  Lustrino and Wilson 2007;Wedepohl et al. 1994;Wedepohl and Baumann 1999;Wilson and Downes 1992;Wörner et al. 1986).

Petrogenetic models explaining the diversity of volcanic rocks in the CEVP
Previous studies from the Urach and Hegau areas suggested that melilititic rocks crystallized from very low-degree, volatile-and trace-element-rich partial melts that originated in the sub-lithospheric mantle (> 100 km depth; dolomite-garnet peridotite stability field) and infiltrated and metasomatized the overlying lithospheric mantle (~ 75 km; Dunworth and Wilson 1998), or in some cases directly erupted at the surface. Other models, however, locate the melt generation of melilite-bearing rocks in the CEVP at the base of the lithosphere (Blusztajn and Hegner 2002), possibly in the thermal boundary layer (TBL; Heldburg region; Pfänder et al. 2018). Despite the timing, origin, and processes of metasomatic overprint are not yet well constrained, previous modifications to ancient subduction processes are widely accepted as the major forces for metasomatic overprinting of the mantle (e.g., Blusztajn and Hegner 2002;Hegner and Vennemann 1997;Lustrino and Wilson 2007;Puziewicz et al. 2020;Ulrych et al. 2011;Witt-Eickschen and Kramm 1998). According to Pfänder et al. (2018), prolonged subduction of oceanic and continental lithosphere with successive dehydration during Variscan Orogeny generates aqueous to supercritical fluids enriched in volatiles and other trace elements, causing metasomatism at great depths (> 125 km) in the thickened lithosphere. This process results in the formation of a carbonated, phlogopite-bearing garnet lherzolite that is later partially melted during evolution of the ECRiS in response to perturbation of the TBL and decompression associated with lithospheric thinning (< 85 km). For instance, Hegner and Vennemann (1997) identified an influence of recycled seawater that overprinted the lithospheric mantle prior to melt formation at Urach and in the Hegau area. However, the asthenospheric mantle is often considered as another major melt source component in melilite-bearing rocks, interacting and mixing with lithospheric domains (Jung et al. 2006;Lustrino and Wilson 2007;Skála et al. 2015;Ulrych et al. 2008Ulrych et al. , 2013Ulrych et al. , 2016. For melilite-free nephelinitic to basanitic rocks, slightly higher degrees of partial melting and shallower depths of extraction (spinel peridotite stability field or transition zone to garnet stability field) are assumed Pfänder et al. 2018;Schubert et al. 2015;Ulrych et al. 2013). Further, Ulrych et al. (2013) and Pfänder et al. (2018) suppose a less metasomatized lithospheric source for basanites than for the melilite-bearing rocks of the Heldburg region and the Eger Graben, without excluding an asthenospheric component. For basanites and nephelinites from the Westeifel, Vogelsberg, and Westerwald, Jung et al. (2011, and Schubert et al. (2015) favor an asthenospheric source coupled with thermal erosion of strongly metasomatized lithosphere and/or intense asthenosphere-lithosphere interaction. The less primitive nature of some of these rocks was explained by temporary stagnation and early polybaric fractionation of olivine, clinopyroxene, and chromium spinel combined with contamination and mixing processes in the upper lithosphere, apparently consistent with frequent occurrence of resorbed green-core pyroxenes in the Westeifel and Hocheifel (Duda and Schmicke 1985;Jung et al. 2006).
Evolved phonolites from the Hegau have been suggested to represent differentiates of primitive foiditic magmas (Mahfoud and Beck 1989). Abratis et al. (2015) showed that (tephri)phonolites in the Heldburg dyke swarm were derived from a nephelinitic parental magma, but subsequently modified during stagnation and fractionation by mixing and mingling with basanitic melts, producing intermediate compositions. For trachytes and phonolites from the Rhön and the Lower Main Plain, assimilation of crustal rocks is suggested as another contribution to their formation (Jung et al. 2013;Schmitt 2006a).
In summary, some models for the CEVP invoke the involvement of mantle plumes or at least active diapiric upwelling (Dunworth and Wilson 1998;Haase et al. 2004;Hegner et al. 1995;Jung et al. 2006;Mertz et al. 2015;Schubert et al. 2015), whereas other studies doubt the necessity of mantle plumes and explain partial melting by earlier metasomatism in the mantle source (Blusztajn and Hegner 2002;Jung et al. 2005Jung et al. , 2011Pfänder et al. 2018;Ulrych et al. 2008). Given the widely varying definitions of mantle plumes in terms of formation depth and thermodynamic properties, many of the modern models can be considered hybrid concepts. These do not attribute melting and volcanism exclusively to thermal or compositional anomalies of an uplifted asthenosphere, nor alone to metasomatism followed later by extensional-transtensional lithospheric thinning and uplift (Lustrino and Wilson 2007). However, deep-seated active asthenospheric upwelling causing a thermal anomaly below or in the TBL (mantle plumes in the narrower sense) is meanwhile mostly rejected for the formation of the CEVP (e.g., Fichtner and Villaseñor 2015).

Materials and methods
For this study, thin sections of 232 rock samples from approximately 130 localities in the CEVP were prepared for petrographic examination (Fig. 2), with a special focus on mineralogy, textures, and alteration state to classify them according to Le Maitre et al. (2002). Most samples originate from various collections in Germany (University of Tübingen, Goethe University Frankfurt, University of Mainz, LGBR Freiburg, HLNUG Wiesbaden) and France (Terrae Genesis, Le Syndicat), supplemented by resampling of several localities. Based on the presence of datable minerals (perovskite, apatite, titanite, zircon, and/or pyrochlore) of adequate grain size, 56 thin sections from 45 localities were selected for U-Pb dating and supplemented by mounted heavy mineral concentrates from five localities (Online Resource 1).
U-Pb age data, supported by ~ 3000 spot analyses, were collected on a Thermo Scientific Element XR sector field ICP-MS coupled with a Resonetics RESOLution 193 nm ArF Excimer laser (CompexPro 102) at FIERCE (Goethe-Universität Frankfurt am Main) using a slightly modified method as previously described in Zeh (2006, 2009). The ICP-MS was tuned for maximum sensitivity while keeping the oxide formation (UO/U < 0.2%) and element fractionation (i.e., Th/U ~ 1) low. Analytical details are reported in the metadata tables of each sequence (Online Resource 2-13). Raw data were corrected offline using an inhouse VBA Microsoft Excel© spreadsheet program Zeh 2006, 2009). The 207 Pb/ 206 Pb and 206 Pb/ 238 U ratios were corrected for mass biases, including drift over time for each of the 11 sequences measured, using the primary reference material SRM-NIST 612. An additional correction was applied on the 206 Pb/ 238 U ratios to account for compositional differences using Durango apatite (31.44 Ma;McDowell et al. 2005), in-house Nama titanite (999 Ma; TIMS data, Wolfgang Dörr, pers. comm.), GJ-1 zircon (Jackson et al. 2004; 603 Ma; TIMS data, Wolfgang Dörr, pers. comm.), and Ice River perovskite (361 Ma; Tappe and Simonetti 2012). Ages were calculated using Isoplot 4.15 (Ludwig 2009) and are plotted within the Tera-Wasserburg space (Tera and Wasserburg 1972). All uncertainties are reported at the 2σ level and were calculated following the guidelines provided by Horstwood et al. (2016).
Olivine melilitites and melilite-bearing olivine nephelinites predominate at Urach, in the Hegau area, and in Lorraine and represent half of the occurrences in the southern URG but occur only rarely and often strongly altered in the Bonndorfer Graben and the Freiburger Bucht area and in the Vosges (Fig. 2). These are porphyritic rocks characterized by macrocrysts of melilite (< 2 mm; av. ~ 0.5 mm), olivine (< 10 mm; av. ~ 1 mm), and occasional clinopyroxene (< 2.5 mm; av. ~ 0.5 mm) set in a fine-grained groundmass of clinopyroxene, opaque phases, interstitial nepheline, melilite, biotite, betimes haüyne, and small amounts of subhedral to euhedral perovskite and apatite (Fig. 5a, b).
Olivine nephelinites and nepheline basanites, in contrast to the latter group, are melilite-and mostly perovskite-free. They are composed of macrocrysts of olivine (< 12 mm; av. ~ 1 mm) and clinopyroxene (< 10 mm; av. 1 mm) embedded in a very fine-grained or hyaline groundmass containing clinopyroxene, nepheline, and occasional haüyne. Such rocks are typical of the Vosges and Pfälzerwald, the Bonndorfer Graben and Freiburger Bucht area, the Taunus, the Lower Main plain, and parts of the Kaiserstuhl (Fig. 2). Two textural types of clinopyroxene macrocrysts can be distinguished: Type I is represented by beige to yellowish or pale brownish euhedral to subhedral crystals under polarized light. Type II (so-called green-core pyroxene, e.g., Duda and Schmincke 1985) consists of rounded greenish cores, surrounded by a fringe of Type I clinopyroxene. Some of these green-core pyroxenes enclose anhedral, stocky rounded apatite (Fig. 5c). Several samples with groundmass plagioclase and minor alkali feldspar are classified as basanitic nephelinites or nepheline basanites. Odenwald and Kraichgau samples may contain high amounts of biotite (Fig. 5d). In the Hegau area and the southern URG, some melilite-free olivine nephelinites occur as well (Fig. 2), but they lack greencore pyroxenes, contain betimes perovskite, and their texture is similar to the melilite-bearing rocks from these regions.

Timing and duration of volcanism in the southern CEVP
Using in situ U-Pb geochronology, we have dated numerous volcanic rocks of the southern CEVP for the first time, giving precise age constrains for several localities previously dated mostly by the K-Ar method on whole-rock samples (Baranyi et al. 1976;Horn et al. 1972;Kraml et al. 1995Kraml et al. , 1999Lippolt et al. 1963Lippolt et al. , 1973Lippolt et al. , 1975Lippolt 1983). In comparison with previous literature data, the results of this dating approach, supplemented with published ages applying different modern methods, show two age groups in the southern CEVP (Figs. 7 and 8). Discrepancies and former uncertainties result from known problems of the K-Ar Microstructure and mineral textures in melilitites, foidites, basanites, and phonolites of the southern CEVP. a Olivine melilitite with olivine (ol) macrocrysts and melilite (mll) phenocrysts, crossed polarization (Hohenbol, Urach). b Nepheline-bearing olivine melilitite with olivine macrocrysts and agglomerations of clinopyroxene (cpx) in a groundmass of fine-grained clinopyroxene, melilite, nepheline, magnetite (mag), and little biotite (bt) (Hohenstoffeln, Hegau area). c Olivine nephelinite with olivine macrocrysts, clinopyroxene phenocrysts and resorbed green-core pyroxenes, containing an inclusion of anhedral apatite (ap), embedded in a cryptocrystalline groundmass (Rotteckruhe, Freiburger Bucht). d Phonolitic phlogopite-nepheline basanite with lath-shaped, large phlogopite and medium-grained olivine crystals in a groundmass of clinopyroxene, magnetite, nepheline, plagioclase, and minor alkali feldspar (Steinsberg, Kraichgau). e Olivine-bearing nepheline haüynite with poikiloblastic clinopyroxene enclosing euhedral apatite and haüyne (hyn), magnetite, and olivine phenocrysts in a groundmass of nepheline, clinopyroxene, and minor alkali feldspar (Neckarbischofsheim, Kraichgau). f Phonolite with phenocrysts of alkali feldspar (afs), a strongly altered sodalite group mineral (sdl-hyn-nsn), titanite (ttn) and aegirine-augite (aeg-aug) in a groundmass of alkali feldspar, aegirine-augite, and nepheline (Hohentwiel, Hegau area). g Phonolite with altered sodalite group minerals showing exsolution textures in a groundmass of aegirine-augite, alkali feldspar, nepheline, and sodalite group minerals (Hohenkrähen, Hegau area). h Phonolite with a phenocrystal of titanite and small apatite in a fine-grained groundmass of aegirine-augite, alkali feldspar, and nepheline (Hohentwiel, Hegau area) ◂ 1 3 method (e.g., Baranyi et al. 1976;Horn et al. 1972), such as loss of 40 Ar due to devitrification of rocks and alteration, particularly on small grains, or excess 40 Ar caused by inheritance during magma formation and/or ascent (e.g., wall rock contamination, presence of xenocrysts). Low closure temperatures for certain K-rich minerals and alteration-related K loss can also lead to erroneous age determinations. The old K-Ar ages are discussed in Online Resource 14.
Crystallization and cooling during ascent and emplacement of subvolcanic to volcanic rocks have no significant influence on the determined mineral ages, since 2σ uncertainties are generally larger than the duration of such processes (Schmitt 2006b;Schmitt et al. 2010;Sundermeyer et al. 2020). Thus, resolvable age differences between texturally distinct minerals in the same sample, between different samples of similar rock types from the same locality, and between plutonic cumulates and their volcanic host rocks allow for deciphering time spans of several Myr relevant for transcrustal plumbing systems at local and regional scales.
Mineral ages in xenoliths may potentially suffer from thermal overprint by the host magma during ascent, which would result in partial reset of the original age if closure temperatures of the dated minerals are below those of the xenolithtransporting magma. Closure temperatures for the U-Pb system vary considerably between minerals, depending on grain size and cooling rate and increase from ~ 350-550 °C for apatite, via ~ 570-700 °C for titanite and ~ 660-950 °C for perovskite, to ~ 740-1010 °C for zircon (Carlson 2019;Chew and Spikings 2015;Flowers et al. 2007;Heaman and Parrish 1991;Spear and Parrish 1996;Wu et al. 2010). In the present study, several xenoliths of the same occurrence have been dated with different crystals and several different mineral phases (apatite, pyrochlore, titanite, zircon), with ages consistent within analytical uncertainty (Table 4). Thus, we consider them to represent crystallization ages of the xenoliths.
Our data (Tables 3 and 4) indicate that the oldest volcanic rocks in the southern CEVP formed in the late Cretaceous and anhedral perovskite (prv) (Sternberg, Urach). c Ijolite with skeletal magnetite and perovskite surrounded by subhedral to euhedral clinopyroxene, nepheline, zeolite, and minor apatite (Hohenstoffeln, Hegau area). d Nepheline syenite xenolith with an anhedral zircon grain next to a tabular biotite crystal surrounded by alkali feldspar and zeolite (2 km E Weil, Hegau area) and represent strongly evolved trachytes and haüyne-rich rocks from the Lower Main plain (~ 73-65 Ma) and the Odenwald and Kraichgau area (~ 70-62 Ma) in the east of the northern URG, consistent with U-Pb zircon ages in the same region for the Sporneiche trachyte (68.6 ± 1.9 Ma; Schmitt 2006a) and the Katzenbuckel phonolite (69.6 ± 1.9 Ma; Schmitt et al. 2007). A similar age range (~ 67-58 Ma) determined for the Freiburger Bucht and the Bonndorfer Graben area on the eastern flank of the URG coincides with the eruption of the Trois Épis olivine melilitite in the Vosges on the opposite graben shoulder with an amphibole 40 Ar/ 39 Ar age of 60.9 ± 0.6 Ma . For the Taunus, a new titanite age for Naurod (55.4 ± 0.9 Ma) is very similar to a 40 Ar/ 39 Ar age for Rabenkopf (58.7 ± 0.4 Ma; Fekiacova et al. 2007). However, significantly older ages for apatite inclusions in titanite (65.7 ± 2.4 Ma) and apatite in a second sample from Naurod (68.1 ± 2.4 Ma) reveal prolonged magmatic activity in this area started at least 10 Myr earlier, potentially revealing discontinuous magma replenishment events. During the early Eocene (~ 56-47 Ma), volcanism in the southern CEVP was basaltic to nephelinitic and restricted to the Lower Main plain and the Pfälzerwald (Forstberg, Messel maar, Stetteritz, Forst, and Kisselwörth; Fekiacova et al. 2007;Lenz et al. 2015;Lutz et al. 2013).
After a volcanic gap of ~ 20 Ma, the oldest known olivine melilitites of the younger series erupted in the late Oligocene (~ 26-25 Ma) near Essey-la-Côte in Lorraine and at the Buggingen potash salt deposit in the southern URG (Table 4). Subsequently, no magmatism is documented until a phase of intense volcanism represented at Urach (~ 19-12 Ma), in the Hegau area (~ 15-9 Ma), and the remaining southern URG (~ 17-16 Ma; Mahlberg, Herbolzheim) including the Kaiserstuhl volcanic complex (~ 19-15 Ma), whose activity is well constrained by recent dating of numerous rock types using U-Pb, 40 Ar/ 39 Ar, Rb-Sr, (U-Th)/He, and fission track (Ghobadi et al. 2021 and references therein; Kraml et al. 1995Kraml et al. , 1999, confirmed by our apatite and perovskite U-Pb ages for a haüyne melilitite from the Kaiserstuhl. The   Kröchert et al. (2009) and Ring and Bolhar (2020). Towards the end of volcanism at Urach, relocation of magmatic activity from the northern to the southern tip of the Albstadt shear zone (Fig. 4) is expressed by phonolitic domes in the eastern Hegau (~ 14-11 Ma) and primitive melilititic to nephelinitic eruptions in the western Hegau (~ 12-9 Ma), representing the last magmatic signal in the southern CEVP. Age data from evolved nepheline syenitic xenoliths in tuffs and melilititic to nephelinitic rocks (~ 15-11 Ma) match those of the Hegau phonolites and syenitic autoliths therein. Thus, there is no age gap between the emplacement of primitive and evolved rocks, suggesting a close genetic relationship between both rock types and differentiation and storage processes over such time scales in the upper crust. The spatial separation of melilite-bearing nephelinites in the west and phonolites in the east may suggest that minor structural differences in the nature of the upper lithosphere led in some cases to prolonged stagnation and evolution in a magma chamber and in other cases to direct eruption. The ijolitic schlieren from the Hohenstoffeln are interpreted as in situ differentiation products accumulating and crystallizing from the surrounding magma, consistent with their identical age within the uncertainty compared to that of the nephelinitic host rock.  (Abratis et al. 2007(Abratis et al. , 2009Büchner et al. 2015;Jung et al. 2006;Kramm and Wedepohl 1990;Mertz et al. 2015;Pfänder et al. 2018;Skála et al. 2015;Ulrych et al. 2016). Conversely, evolved rocks are rare in the study area, comprising phonolites in the eastern Hegau area, diverse Kaiserstuhl rocks, and trachytes in the Lower Main plain. The age distribution for the northern and southern CEVP based on literature and our new data, disregarding K-Ar whole-rock ages (Fig. 8), indicates continuous volcanism since the middle Eocene at varying locations. In contrast to the southern part (Fig. 2), however, no obvious correlation between rock types and ages in the northern CEVP exists. The only evidence for late Cretaceous igneous activity in the northern CEVP are trachytic xenoliths and camptonitic dykes at the Vogelsberg, (~ 74-66 Ma; Bogaard and Wörner 2003;Martha et al. 2014;Schmitt et al. 2007). This volcanic activity overlaps with the formation of trachytes in the Lower Main plain, haüyne-rich foidites in the Odenwald and Kraichgau area (Fig. 8), and an apatite inclusion age at Naurod (Taunus). The basaltic to nephelinitic rocks in the Lower Main plain and Pfälzerwald (~ 56-47 Ma) indicate the end of early magmatic activity in the southern CEVP.
Pliocene volcanic rocks in Germany are only reported for the Westerwald, with a basanite dated to 5.0 ± 0.2 Ma (Schubert et al. 2015). Furthermore, a new phase of volcanic activity occurred in Lusatia in the Pleistocene (~ 370-170 ka; Mrlina et al. 2007;Wagner et al. 1998), coinciding with the Eifel volcanism. In the French Massif Central (~ 14 Ma-5 ka) and the Pannonian Basin (~ 18.5 Ma-8 ka), on the other hand, the main phase of volcanism dates back to the middle Miocene and has continued with short interruptions until today, the spectrum of erupted magmas ranging from basanites to rhyolites (Harangi et al. 2010(Harangi et al. , 2015Hurai et al. 2013;Lexa et al. 2010;Lukács et al. 2018;Molnár et al. 2019;Nowell et al. 2006;Riisager et al. 2000;Seghedi and Downes 2011;Wijbrans et al. 2007).
Considering the entire CEVP, there are no obvious correlations between spatial, compositional, and temporal evolution of volcanism like propagating hot-spot tracks that would support a plume-like mantle anomaly (Lustrino and Wilson 2007). Although thermal asthenospheric upwelling cannot be ruled out as driving force, the most striking differences of the southern CEVP compared to the northern one, i.e., the ~ 20 Ma age gap, the more pronounced SiO 2 -undersaturation of the rocks, and the usually monogenetic character, indicate a structural control on the occurrence of volcanism linked to the geodynamic evolution of Europe. However, there are some spatial variations at same times and temporal variations within same areas. For instance, in some regions with bimodal or polygenetic volcanism, primitive rocks follow more differentiated ones and/or strongly SiO 2 -undersaturated rocks follow less SiO 2 -undersaturated or SiO 2 -saturated ones (e.g., Hegau, Vogelsberg; this study; Bogaard and Wörner 2003), sometimes vice versa (e.g., Heldburg; Abratis et al. 2015;Fig. 7 Overview of all new in situ U-Pb age data amended by existing 40 Ar/ 39 Ar, fission track, (U-Th)/He, Rb-Sr, and biostratigraphical results. The previous age distributions for the different regions based on old K-Ar age data are shown in light grey (Baranyi et al. 1976;Horn et al. 1972;Lippolt et al. 1963Lippolt et al. , 1973Lippolt et al. , 1975Lippolt 1983). The current data indicates that there was no continuous primitive SiO 2 -undersaturated volcanism in SW Germany and E France since the late Cretaceous, but two phases of activity separated by a middle Eocene to early Oligocene gap (shown in pink). This contrasts with previous K-Ar data. Age data from this work were supplemented with results from Aziz et al. (2010), Fekiacova et al. (2007), Ghobadi et al. (2021), Gregor (2003), Keller et al. (2002), Koban and Schweigert (1993), Kraml et al. (1995Kraml et al. ( , 1999, Kraml et al. (2006), Lenz et al. (2015), Lutz et al. (2013), Rasser et al. (2013), Schleicher et al. (1990), Schmitt (2006a), Schmitt et al. (2007), and Wagner (1976). Note that for some 40 Ar/ 39 Ar and U-Pb ages, the published error range is smaller than the symbol size. gm groundmass, WR wholerock, am amphibole, phl phlogopite, sa sanidine, bt biotite, czt calzirtite, zrl zirconolite ◂ Pfänder et al. 2018), and sometimes with age gaps as in the Eifel (Lustrino and Wilson 2007). In other regions, primitive rocks follow more differentiated ones, in turn succeeding primitive rocks (Büchner et al. 2015;Przybyla et al. 2018;Ulrych et al. 2011Ulrych et al. , 2013Ulrych et al. , 2016. Such compositional sequences have been attributed to pre-, syn-, and post-rift stages (e.g., Eger Graben; Ulrych et al. 2011), differentiation due to magmatic plumbing systems (e.g., Siebengebirge; Przybyla et al. 2018), and/or different sources and depths of melt generation (e.g., Heldburg region, Kaiserstuhl, Eger Graben; Braunger et al. 2018;Ghobadi et al. 2021;Pfänder et al. 2018;Ulrych et al. 2016). Likewise in the southern CEVP, the high temporal, spatial, and compositional variability of the individual regions on the one hand, and the compositional similarities of coeval occurrences on the other hand, refer to heterogeneities in the sub-lithospheric mantle beneath Central Europe (Puziewicz et al. 2020). The temporal trend for primitive rocks in the southern CEVP from melilite-free nephelinitic-basanitic rocks to melilitebearing olivine nephelinites and olivine melilitites may indicate decreasing partial melting degrees, greater formation depths, a thicker lithosphere, a more carbonate-accentuated source, and/or a reduced influence of assimilation and fractional crystallization for the latter (e.g., Bogaard and Wörner Fig. 8 Age distribution of igneous rocks in the Central European Volcanic Province. The data of this work are amended by literature data (except for K-Ar data; Abratis et al. 2007;Aziz et al. 2010;Büchner et al. 2015;Fekiacova et al. 2007;Ghobadi et al. 2021;Gregor 2003;Harangi et al. 2010;Harangi et al. 2015;Hurai et al. 2013;Keller et al. 2002;Koban and Schweigert 1993;Kraml et al. 1995Kraml et al. , 1999Kraml et al. 2006;Lenz et al. 2015;Lexa et al. 2010;Linthout et al. 2009;Lukács et al. 2018;Lutz et al. 2013;Mayer et al. 2014;Mertz et al. 2000;Mertz et al. 2007;Mertz et al. 2015;Molnár et al. 2019;Nowell et al. 2006;Pfänder et al. 2018;Przybyla et al. 2018;Rasser et al. 2013;Riisager et al. 2000;Schleicher et al. 1990;Schmitt 2006a;Schmitt et al. 2007;Schubert et al. 2015;Shaw et al. 2010;Singer et al. 2008;van den Bogaard 1995;Wagner 1976;Wagner et al. 1998;Wijbrans et al. 2007). Continuous magmatism throughout the province can be traced back to the late Cretaceous. However, primitive, strongly SiO 2 -undersaturated volcanism in the southern CEVP is restricted to two periods: from the late Cretaceous to the early Eocene and from the late Oligocene to the late Miocene. mel. melilitites, neph. nephelinites, bas. basanites, phon. phonolites 2003;Dunworth and Wilson 1998;Jung et al. 2006;Pfänder et al. 2018).

Implications for the geological evolution in the southern CEVP area
The older volcanic rocks of the southern CEVP (~ 73-47 Ma) postdate the late Cretaceous (~ 95-75 Ma) inversion tectonics in Central Europe (Voigt et al. 2021) and predate Upper Rhine Graben formation (< 47 Ma; Grimmer et al. 2017). Late Cretaceous to early Eocene volcanic activity in the southern CEVP overlaps in time with the ~ 75-55 Ma topographic uplift in Central Germany (Eynatten et al. 2021). The late Cretaceous inversion and the late Cretaceous to Eocene doming and respective magnitudes of shortening and uplift in the southern CEVP area are yet-due to the lack of Cretaceous sedimentary rocks-only inferred from apatite fission track and U-Th/He thermochronological data from the rift flanks of the URG (Link 2010). Time-temperature modeling of apatite fission track length distributions indicates late Cretaceous rapid cooling of crystalline basement rocks along both the western and eastern rift flanks of the URG (Link 2010), similarly to what documented in the Odenwald, Harz, and Thüringer Wald (Wagner et al. 1989;Eynatten et al. 2019Eynatten et al. , 2021Thomson and Zeh 2000). A major portion of apatite fission track data in the Schwarzwald and Odenwald displays late Cretaceous to Eocene ages predating rift-related Cenozoic uplift and cooling (e.g., Link 2010; Wagner et al. 1989). A well-defined erosional angular unconformity recording a > 100 Myr hiatus in the subcrop of the URG documents this pre-Eocene uplift and erosion (e.g., Grimmer et al., 2017). In the northern URG, removal of the entire Mesozoic succession indicates that a zone of major uplift was probably located in the Taunus-Odenwald area, where numerous late Cretaceous to early Eocene volcanic rocks are documented as well. The dynamic topography concept of Eynatten et al. (2021) considers that thermally induced uplift of the lithosphere-asthenosphere boundary (LAB) and regional-scale doming terminated the late Cretaceous inversion at ~ 75 Ma in Central Germany. On the other hand, SW-NE shortening was probably ongoing until ~ 42 Ma along the western European continental margin (> 1000 km to the NW in Ireland), as indicated by calcite twinning analysis (Craddock et al. 2022). Eynatten et al. (2021) suggest that late Cretaceous to Eocene LAB uplift caused a dynamic topography of ~ 1 km and concomitant erosion of up to ~ 4 km. The formation of low-percentage melts that translocated to upper crustal levels appears to be controlled by interfering processes such as LAB shallowing and decompression, metasomatism, and deep-rooted faults such as the URG boundary fault system and the Albstadt shear zone. Dynamic topography can be induced by either convection-induced LAB uplift and subsequent cooling (or an abatement of LAB shallowing) or by translocation of continental lithosphere over upward convecting asthenospheric domains (e.g., Braun et al. 2013). Domal uplift terminated late Cretaceous shortening at ~ 75 Ma in western Central Europe. Late Cretaceous to early Eocene volcanic rocks occur around and possibly in the subcrop of the northern URG, where major uplift is considered for the Odenwald and southern Taunus, but also in the southern Schwarzwald (Bonndorfer Graben and Freiburger Bucht; Figs. 1 and 8; Table 3). Dykes strike NW, N, and NE as documented from maps, geophysical anomalies, and field observations (e.g., Mäussnest 1975;Stellrecht and Emmermann 1970). The ~ 47-27 Ma volcanic age gap temporarily overlaps with the early Eocene to Oligocene syn-rift phase of the URG during (W)NW-extension (e.g., Schumacher 2002).
A major change in deformation characterized by (E)NEtrending extension and transtension is considered to have started ~ 18 Ma (Schumacher 2002) or earlier (~ 25 Ma; Grimmer et al. 2017), associated with major activity of the NNW-trending Ludwigshafen hinge zone and the onset of distinct early Miocene subsidence of the Heidelberg basin (e.g., Grimmer et al. 2017). U-Pb ages of calcite fibers from NNE-trending Alpenrhein graben boundary faults yield 25.3 ± 5.6 Ma to 21.8 ± 3.4 Ma, interpreted with the initial phase of graben formation postdating Helvetic nappe emplacement (Ring and Gerdes 2016). The NNE-trending phonolitic Heldburg dyke (Fig. 1), yielding 17-14 Ma ( 40 Ar/ 39 Ar-ages; Abratis et al. 2015), and U-Pb ages of NWstriking calcite veins of 15.3-2.3 Ma from Miocene domal uplift (Ring and Bolhar 2020) indicate that (E)NE-extension was initiated at that time and persisted during Miocene domal uplift, volcanism, and erosion, and hence throughout the Neogene. Apparent conformable sedimentation in the URG from ~ 47 Ma to ~ 18-16 Ma was interrupted by uplift and erosion, resulting in a rift-wide, angular unconformity that separates conformably deposited older graben filling sediments in the footwall from younger, i.e., Pliocene to Quaternary graben fillings (e.g., Grimm 2011). The late Oligocene to late Miocene volcanic activity (~ 27-9 Ma; Table 4) temporarily overlaps with middle to late Miocene domal uplift in the southern CEVP as documented by the Burdigalian cliff (~ 20-18 Ma) elevated to > 900 m a.s.l. in the WSW and to < 400 m a.s.l. in the ENE of the Swabian Alb (Hoffmann 2017;Fig. 4), and by Miocene apatite fission track data in the southern Schwarzwald (Timar-Geng et al. 2006). East of the southern Schwarzwald, the Upper Miocene younger Juranagelfluh in the Hegau area comprises a late Miocene unroofing sequence, documenting erosion of an uplifted area in the W to NW of the Hegau area (Schreiner 1965). Dykes and elongated plugs show variable trends in the Hegau and Urach areas (e.g., Mäussnest 1974), indicating that pre-existing, shallow-rooted structures were used for the emplacement of melts, either locally along pre-existing 1 3 structures or during variations between strike-slip and normal faulting in the northern foreland of the Alps (Egli et al. 2017;Grimmer et al. 2017;Ring & Bolhar 2020;Craddock et al. 2022). In summary, though several details and processes still need to be clarified and quantified, it appears that since ~ 75-70 Ma dynamic topography has exerted major control on deformation, structural, and sedimentological development of the URG area and correlates temporally with the volcanic activity in the southern CEVP until today.

Conclusion
Based on new U-Pb mineral ages (perovskite, apatite, titanite, zircon, pyrochlor) for volcanic rocks and related plutonic inclusions from nine regions in the southern CEVP and complemented by existing age data, two distinct periods of volcanism are identified, with an age gap of ~ 20 Ma in between: a late Cretaceous to early Eocene (~ 73-47 Ma) group and a late Oligocene to late Miocene (~ 27-9 Ma) group. These two groups also differ in terms of mineralogy and spatial distribution coinciding with topographically uplifted areas. The older group occurs in the Lower Main plain, the Odenwald and Kraichgau area, the Taunus, the Bonndorfer Graben and Freiburger Bucht area, the Vosges and the Pfälzerwald and mostly comprises nephelinites, basanites, haüynites, and trachytes. The young group is mainly represented by perovskite-bearing olivine melilitites, melilite-bearing olivine nephelinites, and phonolites of Lorraine, the southern URG, Urach, and the Hegau area. As an exception, the Kaiserstuhl comprises basanites, tephrites, melilitites, nephelinites, haüynites, carbonatites, phonolites, and noseanites, all of which are of Miocene age.
Magmatism in both periods is linked to dynamic topography and the tectonic evolution in Central Europe, expressed in its age and location, reflecting large-scale stress field changes in the interplay of Pyrenean and Alpine orogeny interrupted by phases of subsidence causing a volcanic hiatus. The older volcanic rocks tend to occur in or along NE-SW striking structures and/or in the periphery of or on the shoulders of present-day graben structures (Fig. 4). They postdate late Cretaceous inversion tectonics (NE shortening; > 73 Ma) and, together with erosion of Permo-Mesozoic units, possibly reflect late Cretaceous to Eocene doming, shortening and uplift in the southern CEVP. The temporal and spatial coincidence of the 47-27 Ma volcanic gap with the initial phase of the URG development suggests a lithosphere-scale event such as cooling-induced regional subsidence and/or an abatement of asthenospheric uplift following the 75-55 Ma period considered for dynamic topography in Central Germany. The younger volcanic rocks are bound to Cenozoic N-S striking fault zones that developed with the ongoing formation of the ECRiS during Neogene (E) NE-extension-transtension (< 25 Ma). Again, this partly coincides with doming, uplift, and erosion in the southern CEVP and the URG from the middle to late Miocene, followed by subsidence during the Pliocene and Quaternary.
Evolved phonolites and (nepheline) syenite xenoliths in the melilite-bearing nephelinites of the Hegau area show that felsic alkaline magmatism beneath the Hegau area was already active before ~ 15 Ma, predating the first melilititic to nephelinitic eruptions (~ 12 Ma ago). The onset of volcanism is possibly related to the simultaneous attenuation, extinction, and relocation of Urach volcanic activity  at the northern tip of the Albstadt shear zone. Volcanism at Urach was not caused by the Ries impact (14.6 ± 0.2 Ma).
Magmatism in the northern part of the CEVP differs from the igneous activity in the southern sectors studied here. The generally mildly SiO 2 -undersaturated volcanic rocks of the northern CEVP cover larger areas mostly forming voluminous, diversely composed complexes and are often more evolved than the isolated and low-volume, primitive nephelinites and melilitites in the southern CEVP. Even though volcanism in the northern CEVP has been recurrently active since the middle Eocene and the 20 Myr gap is absent, some temporal parallels with activity in the southern CEVP are apparent. Altogether, a spatial heterogeneous sub-lithospheric mantle source, varying lithospheric and crustal thicknesses, different depths of melt formation and degrees of partial melting, differentiation processes, and crustal contamination combined with the tectonic evolution in Central Europe can explain the different compositions of volcanic rocks and the magmatic evolution in the CEVP. The petrography and new age data do not provide evidence for a large-scale and deep-routed active asthenospheric upwelling beneath Central Europe. However, to delimit, determine, and describe these possible causes more precisely, detailed geochemical and petrological studies are in progress.

Conflict of interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/.