Étienne Juvigné, André Pouclet, Stéphane Pirson & Jacques-Marie Bardintzeff

Further Considerations of the Provenance of the Rocourt Tephra: Volcanic Mafic Minerals and Age

Résumé

Le Téphra de Rocourt est une retombée de cendre volcanique qui a été découverte en Belgique, il y a trois quarts de siècle. Depuis lors, toutes les recherches destinées à découvrir son volcan émetteur ont été vaines. Deux volcans de l’Eifel occidental ont été cités, à savoir le Dreiser Weiher et le Pulvermaar, mais il a été démontré par la composition chimique des pyroxènes qu’ils sont incompatibles avec le Téphra de Rocourt (Juvigné et al., 2024, DOI: 10.1007/s00445-024-01756-2). Le présent travail présente des arguments supplémentaires qui supportent cette conclusion ; ils relèvent de la comparaison des associations de minéraux mafiques volcaniques des trois parties impliquées. Par ailleurs, les détails de l’étude stratigraphique qui ont conduit à estimer l’âge du Téphra de Rocourt entre 70 et 80 ka sont rapportés, de façon à élargir la tolérance avec d’éventuelles datations de volcans de l’Eifel ou d’ailleurs.

Abstract

The Rocourt Tephra (RT) is a pyroclastic fallout deposit that was discovered in Belgium three-quarters of a century ago. Since then, the search for its source volcano has been in vain. Recently, however, two volcanoes of the West Eifel Volcanic Field have been put forward, namely the Dreiser Weiher and the Pulvermaar, but it has been shown by the geochemical fingerprints of associated pyroxenes that neither of them is compatible with the RT (Juvigné et al., 2024, DOI: 10.1007/s00445-024-01756-2). We present here additional arguments that support this conclusion. They arise from the comparison of volcanic mafic mineral associations of the RT and the Dreiser Weiher and Pulvermaar tephras. Furthermore, details of the stratigraphic study that led to the age estimate of between 70 and 80 ka for RT are reported, in order to widen the scope for determining the ages of volcanoes or tephras in the Eifel Volcanic Field (or elsewhere).

Table des matières

Received: 24 December 2024 – Accepted 20 March 2025

This work is distributed under the Creative Commons CC BY 4.0 Licence.

1. Introduction

1The enstatite bearing volcanic ash deposit (now known as Rocourt Tephra, RT, see below) was discovered in a dispersed state (i.e., as a cryptotephra) in reworked deposits of Upper Belgium (Gullentops, 1952; Tavernier and Laruelle, 1952), then in a paleosol of the Eemian–Weichselian transition of the Rocourt loess stratotype (Gullentops, 1954). Since then, it has been found in about 35 sites in Belgium, the Grand-Duchy of Luxembourg, the Netherlands and the Lower Rhine Bay/Germany (Fig. 1; Table 1).

Figure 1: Location of sites where the volcanic mafic minerals (vmm) of the Rocourt Tephra (RT) have been identified as well as some localities where enstatite has been recognized as traces in various associations. Explanation: the numbers referring to localities correspond to the relevant literature references listed in Table 1 ; sites in bold and red font are of major interest, providing mineralogical and/or chemical and/or stratigraphical data. Abbreviations: WEVF = West Eifel Volcanic Field; PvM = Pulvermaar; DrW = Dreiser Weiher; Jfw = Jungferweiher; HlM = Hoherlist Maar; EiM = Eigelbach Maar; EEVF = East Eifel Volcanic Field; B-M = Baraque Michel; Eu = Eupen; Lg = Liège; Ma = Malmedy; Mo = Monjoie; Re = Remouchamps; St = Stavelot; StV = Saint-Vith; Ve = Verviers.

Table 1: For the Rocourt Tephra, articles reporting the presence of the RT in deposits at the sites shown in Fig. 1.

Authors	Year	Site #		Host sediment
Tavernier and Laruelle	1952	n/a		Alluvial plain
Gullentops	1952	1		Soil
Gullentops	1954	2		Loess section
		3 & 4		Slope deposits
		5		Slope deposits & terrace
Bourguignon	1955	1		Soil
Hermans	1955	6		Slope deposits
Bastin et al.	1972	7		Valley bottom: periglacial deposits
Juvigné and Mullenders	1972	8		Loess section
Bustamante Santa Cruz	1973, 1974	9		Terrace
Juvigné	1973	10		Loess section
		11		Terrace
Juvigné	1974	12		Loes section
Pissart	1974	13		Terrace
Ek	1974	14		Cave deposits
Bastin et al.	1974	15		Rampart of lithalsa
Pissart et al.	1975	7		Valley bottom: periglacial deposits
Juvigné	1976	2		Loess section
		12		Loess section
		10		Loess section
		16		Loess section
		16		Loess section
		16		Short term excavation
		17		Short term excavation
Juvigné	1977b	n/a		Varia
Juvigné	1977a	18		Loess & eolian sand
Bastin and Juvigné	1978	19		Valley bottom: periglacial deposits
Juvigné	1979a	20		Loess section
Juvigné	1979b	s.o.		Terrace
Juvigné and Pissart	1979	15		Slope deposits
Quinif et al.	1979	21		Cave deposits
Bolline et al.	1980	22		Loess section
Pissart and Juvigné	1980		15	Rampart of lithalsa
Ballmann et al.	1980		23	Cave deposits
Haesaerts et al.	1981		n/a	Loess section
Pissart and Juvigné	1982		24	Valley bottom: periglacial deposits
Meijs and de Lang	1983		25	Loess section
			26	Loess section
Mees and Meijs	1984		25	Loess section
Juvigné and Mörner	1984		n/a	Palaeolake
Juvigné	1985		27	Terrace
Juvigné	1990		n/a	Varia
Pouclet and Juvigné	1993		27	Loess section
Lacroix	1993		28	Cave deposits
Juvigné	1993		n/a	Varia
Juvigné et al.	1996		12	Loess section
Gullentops and von der Hocht	1998		29 & 30	Loess section
Meijs and Groenendijk	1999		12	Loess section
Bringmans et al.	1999-2000		n/a	Loess section
Renson et al.	2002		28	Cave deposits
Meijs	2002		12 & 25	Loess section
Pirson et al.	2004		28	Cave deposits
Juvigné et al.	2008		31	Loess section
Pouclet et al.	2008		2, 12, 32	Loess section
			28	Cave deposits
			27	Terrace
Rixhon and Juvigné	2010		33	Valley bottom: periglacial deposits
Meijs	2011		10	Loess section
Pirson and Juvigné	2011		28	Cave deposits
Juvigné et al.	2013		32	Loess section
Juvigné	2016		34	Rampart of lithalsa
Jouannic et al.	2016		2	Loess section
Haesaerts et al.	2016		16	Loess section
Juvigné et al.	2022		35	Valley bottom: periglacial deposits

2In all cases, its identification is based on the volcanic mafic mineral (vmm) assemblage as determined using a polarizing microscope, and more particularly by a high frequency of enstatite and brown amphibole. It became the Tuf de Rocourt (Juvigné, 1977b) then the Rocourt Tephra (Juvigné, 1991). The optical determination of the enstatite was confirmed by microprobe analyses by Bustamante-Santa Cruz (1973) and that of the clinopyroxenes and the amphibole by Juvigné (1990).

3When the volcanic ash was first discovered by Gullentops (1952), its origin from the West Eifel Volcanic Field (WEVF) was suspected by the author, but, to date, the emitting volcano has never been identified. Following the identification of the RT in a loess section of the Lower Rhine Bay, Gullentops and von der Hocht (1998) pointed out that the Dreiser Weiher volcano was a possible source on the sole basis of its large size and its relative proximity in the northern part of the WEVF. This hypothesis has not received any further attention. Lenaz et al. (2010) identified a tephra containing traces of enstatite in the Jungferweiher lake core (WEVF), and they clearly correlated it with the RT. Förster et al. (2020) took up the above correlation by integrating a tephra also containing traces of enstatite in two other lacustrine sequences of the WEVF, namely Eigelbach and Hoherlist maare. By designating the Pulvermaar as the source volcano for the RT occurrences in these three maare (Jungferweiher, Eigelbach, Hoherlist), the authors provoked a comparative study of the RT with the proximal tephra of Pulvermaar (Juvigné et al., 2024) This study also included the Dreiser Weiher. Juvigné et al. (2024) demonstrated by the study of the chemical composition of the pyroxenes that neither the Pulvermaar nor the Dreiser Weiher could be the source of the RT. In their paper, the optical determinations of the vmm had little weight because of the variations inherent in the complexity of the factors that can modify, at least quantitatively, the mineralogical association of the same tephra. Those factors can include the heterogeneity of the magma chamber, the alteration in the host terrains, as well as air transport and laboratory practices. Nevertheless, the results obtained by Juvigné et al. (2024) support the conclusion obtained from the chemical composition of the pyroxenes. Furthermore, in tephra correlation research, the age of a deposit constitutes a valued guide, and it has been used by Lenaz et al. (2010) and Förster et al. (2020) for the RT. However, the RT has never been found in a primary position. Despite the progress of loess stratigraphy in Belgium (Haesaerts et al., 2016), the vertical distribution of its vmm in the Lower Weichselian loess sequences weakens the narrowness of the estimate currently proposed between 78 and 80 ka. Detailed data on this subject will be provided below to illustrate this reservation.

2. Analytical Procedure

4Dense minerals (δ > 2.8) were separated as follows: boiling in HCl_10%vol; sieving by 355/75 µm; extraction of dense minerals with bromoform (δ = 2.8) in a separating funnel by repeating agitation–decantation–harvest cycles, until no more harvest was obtained (generally three to five cycles). Aliquots were examined under the binocular magnifier, and smear slides were prepared in Canada balsam for identification with a petrological microscope.

3. Volcanogenic Mafic Minerals of the Rocourt Tephra

5The association consists of orthopyroxenes, clinopyroxenes, brown amphiboles, titaniferous magnetite and Cr-spinel. Due to several factors including the zonation of magma and the fallout (natural factors), as well as from sampling to optical determination (technical factors), the frequency ranges of the individual minerals are somewhat large. In the sites with high frequencies of vmm, a characteristic mineral suite was calculated (Table 2 ).

6Table 2: Percentage ranges of the vmm of the Rocourt Tephra. Explanation: Cpx = clinopyroxene; Ens = enstatite; Amp = amphibole; n = the number of grains counted; n.a. = not available.

Locality	Determinator	Ref.	Cpx	Ens	Amp	Spinel	n
Eupen, Soor valley, one section	Juvigné	[1]	53	27	20		n.a.
Eupen, Soor valley, various sections	Juvigné	[2]	50	25	25		n.a.
Rocourt, loess section	Juvigné	[3]	58.7	9.4	31.8		388
Kesselt, loess section	Juvigné	[3]	65.5	13.4	21.1		739
Tongrinne, loess section	Juvigné	[3]	49.7	25.6	24.6		107
Wanlin, loess section	Juvigné	[4]	28.9	43.1	28		2888
Liernu, loess section	Juvigné	[5]	22.9	34.9	42.9		73
Hautes Fagnes, Trô Maret, terrace	Juvigné	[5]	46.8	22	31.2		n.a.
Vroenhoven, loess section	Meijs	[6]	76.7	8.7	13.8	0.8	n.a.
Hautes Fagnes, Trô Maret, terrace	Juvigné	[7]	45.3	25.5	29.1		1300
Trooz, Walou, cave	Pirson & Juvigné	[8]	45.4	40.1	14.3		1279
Bassenge, Romont, loess section	Juvigné	[9]	90.9	3.8	5.2		10550
Remicourt, loess section	Juvigné	[10]	67.2	10	22.7		3735
Min.			22.9	3.8	5.2
Max.			90.9	43.1	42.9
References: [1] – Bastin et al. (1972); [2] – Pissart et al. (1975); [3] – Juvigné (1976); [4] – Juvigné (1979a); [5] – Bolline et al. (1980); [6] – Meijs and de Lang (1983); [7] – Juvigné (1985); [8] – Pirson and Juvigné (2011); [9] – Juvigné et al. (2008); [10] – Juvigné et al. (2013).

7Photographs of the three most common volcanogenic mafic minerals of the Rocourt Tephra are presented in Figure 2.

Figure 2: Photographs of the three most common volcanogenic mafic minerals of the Rocourt Tephra. Rows: E = enstatite; A = amphibole; C = clinopyroxene. Columns: (1) and (2) orthogonal positions in plane-polarized light to highlight pleochroism; (3), (4), (5) maximum illumination between cross-polarized light to highlight birefringence with intercalated gypsum (4) and (5).

4. The Dreiser Weiher Tephra

8The location of the volcano and the detailed positions of the samples stratigraphically are available in Juvigné et al. (2024, Fig. 3). Two populations share some 90% of the entire population: very dark green to black euhedral minerals coated with dark magmatic glass and colorless to greenish grains which are not coated with magmatic glass. The magmatic coating impairs the transparency during microscope examination. Nevertheless, numerous dark green euhedral minerals could be identified as clinopyroxene as well as agglomerates of small euhedral clinopyroxenes. The colorless crystals consist of olivine. Traces of another two minerals were found: anhedral lawn-green crystals with serrated edges (clinopyroxene) and prismatic brown minerals (amphiboles). No significant stratigraphic variation was observed in the sequence (Table 3). Photographs of the three most common volcanogenic mafic minerals of the Dreiser Weiher Tephra are presented in Figure 3.

Table 3: Frequency (%) of the volcanogenic mafic mineral suite of the Dreiser Weiher Tephra after optical determinations (magnifier and microscope). Explanation: n = number of mineral grains counted.

Label	Clinopyroxene	Enstatite +amphibole	Olivine	n
Dreiser Weiher 2	38.5	0	61.5	117
Dreiser Weiher 3	40.0	2.9	57.1	105
Dreiser Weiher 4	46.7	0	53.3	112

Figure 3: Photographs of the two most common volcanogenic mafic minerals of the Dreiser Weiher Tephra. C = cpx; O = olivine.
(A) Under magnifier: dark grains are clinopyroxenes; colorless to pale greenish grains are olivines.
(B) Rows: C = clinopyroxene; O = olivine. Columns: (1) and (2) orthogonal positions in plane-polarized light to highlight pleochroism; (3), (4), (5) maximum illumination between cross-polarized light to highlight birefringence with intercalated gypsum in (4) and (5).

5. The Pulvermaar Tephra

9The location of the volcano and the detailed position of the samples are given in Juvigné et al. (2024, Fig. 4).

10Two populations make up most of the mass: very dark green to black grains mainly coated with magmatic glass are largely dominant over transparent colorless grains. Under the microscope, magmatic glass coating impairs the transparency of the former minerals. Nevertheless, several of them could be identified as subhedral to euhedral clinopyroxene as well as agglomerates of small euhedral clinopyroxenes. The transparent colorless minerals are likely to be olivine. Otherwise, a few anhedral lawn-green minerals with serrated edges are present (clinopyroxene) as well as traces of prismatic brown grains (amphibole). The latter two minerals and the olivines are not coated with magmatic glass and so are grains of coarser crystals. Qualitative examination of the eighteen samples has been done and no significant stratigraphical variation was observed (Table 4 ).

Table 4: Frequency (%) of the volcanogenic mafic mineral suite of the Pulvermaar Tephra after optical determinations (magnifier and microscope). Explanation: n = number of mineral grains counted.

Label	Clinopyroxene	Enstatite + amphibole	Olivine	n
Pulvermaar 1	88.7	0	11.3	106
Pulvermaar 2	76.1	2.2	21.7	105
Pulvermaar 3	92.7	0	7.3	123
Pulvermaar 4	95.3	0	4.7	128
Pulvermaar 5	96.2	1.9	1.9	106
Pulvermaar 6	97.1	0	2.9	134
Pulvermaar 7	98.3	0	1.7	116
Pulvermaar 8	93.4	1.6	4.9	122
Pulvermaar 9	91.8	2	6.1	105
Pulvermaar 10	88.4	11.6	0	109
Pulvermaar 11	81.8	18.2	0	110
Pulvermaar 12	94.8	5.2	0	144
Pulvermaar 13	88.6	8.5	2.9	105
Pulvermaar 14	87.5	12.5	0	112
Pulvermaar 15	96.8	13.2	0	126
Pulvermaar 16	94.9	5.1	0	118
Pulvermaar 17	89.6	10.4	0	134

11Photographs of the two most common volcanogenic mafic minerals of the Pulvermaar Tephra are presented in Fig. 4

Figure 4: Photographs of the two most common volcanogenic mafic minerals of the Pulvermaar Tephra. (A) Under magnifier: dark grains are clinopyroxenes; colorless to pale greenish grains are olivines. (B) Rows: (1) Cpx = euhedral clinopyroxene (single mineral or agglomerate); O = olivine. Columns: (1) and (2) orthogonal positions in plane polarized light to highlight pleochroism; (3), (4), (5) maximum illumination between cross polarized light to highlight birefringence with intercalated gypsum in (4) and (5).

6. Stratigraphical Distribution of the Rocourt Tephra

12In sites where the Upper Pleistocene lithostratigraphic units are sufficiently developed, the vertical distribution of the vmm was investigated, with the aim of finding the RT layer visible to the naked eye at a peak of concentration. However, the objective of identifying the RT in its primary position has never been achieved. In some cases, secondary concentration peaks have been highlighted (Fig. 5). The part above the main peak comes from sedimentary reworking and the lower part from bioturbation (Juvigné, 1977b).

Figure 5: Vertical distribution of vmm of the RT in Upper Pleistocene sequences in Belgium. Legend: 1 = Humiferous Complex of Remicourt (HCR); 2 = palaeosol; 3 = frequency of vmm of the RT; 4 = unconformity. Horizontal scale = quantity of vmm for varying weights of sediment from one site to another (see the original publications).

13In the loess region of Middle Belgium (Rocourt, Kesselt, Romont, Veldwezelt, Remicourt, Tongrinne – Fig. 5), the tephra is systematically associated with a humiferous pedocomplex known as the Humiferous Complex of Remicourt (HCR; Haesaerts et al., 1997, 2016). In High Belgium, two sites yielded detailed data for the RT stratigraphic distribution. In the Walou Cave, the distribution peak is located on top of a humiferous horizon (unit CV-1), overlying a complex sequence including palaeosols (units DI-BT, CV-3 and CV-2; Pirson and Juvigné, 2011). In the Wanlin brickyard, the peak of the RT was observed at the boundary between two palaeosols separated by a stony layer (Juvigné, 1979a).

14Hence, it seems that the fallout of the RT has occurred during a period of (relatively strong) soil formation, with sufficiently intense biological activity to disseminate the RT into the underlying units. In all the cases, the peak is situated above the Rocourt Pedocomplex or its equivalent. In the most complete loessic sequences, the vmm concentration peak is generally in the Humiferous Complex of Remicourt (see Juvigné et al., 2013; Haesaerts et al., 2016) There are a few exceptions, however. In one of the sequences studied at Rocourt, the RT peak was found above the HCR, but in other sequences, it was observed inside the HCR. In two sequences from Kesselt, the peak is below the HCR, but on other sequences from the same site, the highest concentration of RT-vmm was found inside the HCR. At Tongrinne, the peak is located at the contact between the Rocourt Pedocomplex and the overlying HCR.

15Regarding the age of the RT, according to Haesaerts et al. (2016), the Rocourt Pedocomplex is attributed to the Eemian interglacial and to the main part of the Weichselian Early Glacial (GS 25 to the lower half of GI 21, sensu Rasmussen et al., 2014). Therefore, again according to Haesaerts et al. (2016), the overlying Humiferous Complex of Remicourt bearing the RT is attributed to the end of GI 21 (ca. 78–80 ka; Juvigné et al., 2013). It is worth mentioning here that following Antoine et al. (2016), the equivalent of HCR is correlated with GI 20 and GI 19. Based on this viewpoint, the age of the RT would be slightly younger, 76.5–70 ka, following the Rasmussen et al. (2014) chronology. We therefore suggest here to consider the range 80–70 ka for the age of the fallout in order to widen the scope for determining the ages of volcanoes or tephras in the Eifel Volcanic Field.

7. About Correlations with the Rocourt Tephra

16The identification of RT in some 35 sites in Belgium and in the neighboring regions of the Netherlands and the Lower Bay Rhine was based solely on determinations of vmm made under the microscope, and more particularly on the presence of enstatite as a marker mineral. The correlation of RT with products of the Dreiser Weiher (Gullentops and von der Hocht, 1998) or of the Pulvermaar (Förster et al., 2020) was rejected mainly by the geochemical fingerprints of pyroxenes (Juvigné et al., 2024). The detailed optical mineralogy data added in this paper allows us to argue in the same way. The photographs show fragments of megacrysts (without glass coatings) in the RT and euhedral clinopyroxenes with glass coating at Pulvermaar. Moreover, in a ternary diagram, there is no overlapping of the fields of the mineralogical associations of the three tephras (Fig. 6 ). It is also difficult to accept that weathering could have caused the disappearance of the amphiboles and enstatites from the Pulvermaar T. and the Dreiser Weiher T. and/or the olivines from the RT. Even if their total alteration were accepted, one could not obtain overlapping of the respective fields.

Figure 6: Comparisons of vmm associations of RT (black circles), Pulvermaar (black triangles) and Dreiser Weiher (black stars). Legend: 1 = possible distribution of the vmm association assuming the disappearance by alteration of brown amphiboles and enstatites in the proximal tephra of the Pulvermaar T.; 2 = possible distribution of the vmm association of the RT assuming the disappearance by alteration of olivines in all types of host sediments.

8. About the Presence of the RT at Schwalbenberg

17Fischer et al. (2021) report the presumed presence of the RT in the Schwalbenberg loess section (Middle Rhine valley). It is a centimeter-thick, coarse-grained tephra resting on a paleosol which could be the equivalent of the HCR in Belgium (see above). Unfortunately, the authors do not report any mineralogical or geochemical data concerning the RT in the Schwalbenberg loess. Their hypothesis calls for further investigations, because: (1) the site is on the road to the Inden and Garzweiler mines (Lower Bay Rhine) where Gullentops and von der Hocht (1998) found the RT; (2) the source volcano could be in the nearby East Eifel Volcanic Field, an unexpected hypothesis.

9. Conclusion

18Thus far, the volcano that has erupted the RT has not yet been identified. However, both the Pulvermaar and the Dreiser Weiher volcanoes can be ruled out, even though they have been suggested (erroneously) as sources in the literature. To propose the correlation of tephras with the RT, it is essential to refer not only to the chemical composition of the pyroxenes, but also to the association of volcanic mafic minerals which is characterised by high frequencies of megacrystal fragments of enstatite and brown amphibole in the absence of olivine. The earlier age estimate for the RT of 78 to 80 ka (Juvigné et al., 2024)) is dependent on the stratigraphy of the loess compared to the INTIMATE curve, because the original stratigraphical position of the tephra is unknown. However, another stratigraphic-based suggestion for the age of RT is from 76.5 to 70 ka. Therefore, evidence of the presence of RT should not be sought only in the narrow age range of 78 to 80 ka but in the range 80-70 ka.

Acknowledgments

19We warmly thank Dr. Andreas Schüller, coordinator of scientific research in UNESCO Natur- und Geopark Vulkaneifel, who agreed that we could take samples from the old quarry opening in the rampart of the Pulvermaar. David J. Lowe (emeritus professor, University of Waikato) is warmly thanked for improving the text and revising the English. Constructive review and comment by Pierre Antoine also improved the manuscript and are greatly appreciated.

Further Information

Author contributions

20Etienne Juvigné: conceptualization, formal analysis, methodology, project administration, original draft writing, writing revision; André Pouclet, Jacques-Marie Bardintzeff, Stéphane Pirson: formal analysis, methodology, original draft writing, writing revision.

Conflicts of interest

21The authors declare that there is no conflict of interest.

Bibliographie

Pour citer cet article

A propos de : Étienne Juvigné

A propos de : André Pouclet

A propos de : Stéphane Pirson

A propos de : Jacques-Marie Bardintzeff