Miocene Stratigraphy: An Integrated Approach A. Montanari, C.S. Odin and R. Coccioni, eds. © 1997. Elsevier Science B. V. All rights reserved. Chapter C5
BIOSTRATIGRAPHY AND GEOCHRONOLOGY OF A MIOCENE CONTINENTAL VOLCANICLASTIC LAYER FROM THE EBRO BASIN, SPAIN
G.S. Odin, G. Cuenca Bescos, 1.1. Canudo, M. Cosca and M. Lago
INTRODUCTION
The Tertiary Ebro Basin in Spain is a large continental basin filled with Miocene fluvial to lacustrine sediments. In the Lanaja-Pefialba area (Zaragoza Province) a volcaniclastic layer has been found in 1989 within these Miocene sediments. The biostratigraphic age has been characterized using Charophyta and micromammals. This makes this material promising for correlation of continental and marine biostratigraphic and chronostratigraphic scales. This study reports new findings related to the presence of this unique volcaniclastic layer in five sections. A new biostratigraphic mammal site immediately overlying the layer has also been investigated.
LITHOSTRATIGRAPHY
Geology and sediments
The study area is located in the Ebro Basin (Aragon Province) in the northeastern part of the Iberian Peninsula (Fig. 1). This basin is limited by the Pyrenean Range to the north and the Iberian Range to the south. The nearest continental basin is the well-known Catalayud-Teruel Basin in the Iberian Range. The triangular Ebro Basin is one of the largest continental Cenozoic basins of western Europe, extending over nearly 45,000 krn2 and containing sediments of Middle Eocene to Quaternary age. The infilling is asymmetrical, with the northern part locally reaching a thickness of 5 krn. Overthrusted marine Mesozoic and Palaeocene nappes cover the area. In the southern part, the continental sediments unconformably overlap the margin of the basin; in the central part, the sediments are well-exposed and nearly horizontal (Riba et aI., 1983). Extensively studied by the Spanish geologists, the Ebro Basin has generated a number of unit names partly depending on the methodological approach. The unit containing the studied layer has been named: (1) as a lithologic unit: Formacion Sarifiena-Calizas de Pefialba (Quirantes, 1978); (2) a facies unit: Huesca fluvial system (Hirst, 1989); (3) a sequential analysis unit: Sistema Deposicional de los Monegros (Cabrera, 1983); (4) a tectonosedimentary unit: Nl (Arenas and Pardo, 1991); and (5) a genetic sedimentary unit: Unidad de Zuera (Hernandez et aI., 1991). The deposits of the Huesca fluvial system and the Formacion Sarifiena consist of geometrically distinct channel sandstone 298 G.s. Odin et al.
i N pRANCE
...... -, '\,'.'---.
~-~-) ...... , .....,' ' .. -~
Barcelona
,===~ __..;.100'
Fig. I. Schematic map of the lithological facies of the Tertiary Ebro Basin (after Riba et aI., 1983, modified). 1 = Pefialba, 2 = Sigena, 3 = Lanaja, 4 = EI Tejar, 5 = Tardienta. bodies enclosed within fine-grained floodplain deposits. Flow was mainly through a network of well defined channels within which a bulk of the sandy bedload material was deposited (Hirst, 1989). The Calizas de Pefialba and Sistema Deposicional de los Monegros were deposited mainly in lake to lake-margin environments. Lacustrine sediments predominated in the centre of the Oligocene-Miocene Ebro Basin.
The volcaniclastic layer
The volcaniclastic layer studied in this work extends over nearly 100 km. From northwest to southeast, the layer has been observed in five sites at Tardienta, Lanaja, E1 Tejar, Pefialba, and Sigena (Fig. 2). The first report of volcanogenic sediments in the Ebro Basin is recent (Hirst, 1989). This author interpreted a horizon at Tardienta as a volcanic ash layer. The layer consists of a lower white, sanidine-rich sandstone, about 2 cm thick, and overlain by about 10 cm of pure smectite clay. The clay is pinkish with a soapy texture. Canudo et al. (1993) first reported the ash layer as a 45 km wide more or less continuous horizon in the Lanaja-Pefialba area. The wide distribution, independent of the facies, is a good criterion for truly aerial transportation of the volcanic material. The layer displays two levels of different colour. The base is characterized by a lithological change between fine-grained sedimentary material below: clay in Lanaja or limestone in Pefialba and white volcaniclastic material above for the lower level. The colour of the upper level varies from red in Lanaja to green in Pefialba. The thickness generally varies between 0.15 m to the northwest and 0.30 m to the southeast. The white level is present alone in Sigena but absent in El Tejar. Biostratigraphy and geochronology of a Miocene continental volcaniclastic layer 299
0: C- 00 00 00 >- C- c- ~ ~ J: oj o: g; oc " "- "'"t: ~ '" ~ ~ " Fig. 2. Lithologic succession in five sections from the Ebro Basin where a volcaniclastic layer has been observed. BIOSTRATIGRAPHY Biozonation with mammals Miocene continental biostratigraphy is primarily based on the evolution of mammals. Limits of large units (continental stages) are traditionally marked with large mammals, and smaller units (biozones) commonly use micromammals. For instance, the Aragonian unit (of stage or superstage ranking) was originally identified as the interval between the first appearance of the horse Anchiterium below and the first appearance of the horse Hipparion above (Daams et aI., 1977). The unit was formally established and characterized for sediments and mammals in the Catalayud-Teruel Basin; it comprises two land mammal stages: Orleanian and Astaracian which would cover MN zones 3 to 8 of Mein (1989). The approximate marine equivalent would be Early (pp) and Middle Miocene (Table 1). Subsequent work (Daams and Freudenthal, 1981, 1989) improved the biostratigraphy but revealed that the base of the Aragonian unit (first appearance of Anchiterium) could not be identified confidently in the type area and the event could not serve as a marker horizon in absence of the key fossil. Daams et al. (1987) redefined the base of 300 C.S. Odin et al. Table I Schematic relation between continental and marine subdivisions (partly after Mein, 1989, and Steininger et aI., 1989) OLIGOCENE AQUITAN. BURDIGALIAN LANGH. SERRA V ALLIAN AGENIAN I ORLEANIAN I ASTARACIAN 2a I 2b I 3 I 4 I 5 6 7 8 Y2 Z A B C 0 E F G "AGENIAN pp" RAMBLIAN ARAGONIAN (2)• (1)• (1) initially presumed location of the dated layer; (2) newly established location according to the fauna discovered in the section at Tardienta. the Aragonian in correspondence with the first appearance of modem cricetid rodents and created the Ramblian regional stage for strata underlying the Aragonian ones. This revision places the 'biozone A' of their previous study (Daams and Freudenthal, 1981) in the Ramblian. They also define the base of their Ramblian stage in correspondence with extinction of the Rhodanomys-Ritteneria group of eomyd rodents. However, this level is not concretely represented in the type area of the Ramblian (Calamocha-Navarrete del Rio in the Catalayud-Teruel Basin). Daams et aI. (1987) identified two local biozones in the Ramblian stage: (1) zone Z would be characterized by the presence of cricetids and large number of glirids (Rodentia); and (2) zone A would show the predominance of the Eomyidae (an extinct family of rodents) with genera Ligerimys and Pseudotheridomys. However, the zonal limit is difficult to identify both in the Catalayud-Teruel and the Ebro basins (Hernandez et aI., 1987; Cuenca et aI., 1993). The distribution of the significant taxa is shown in Fig. 3. Biostratigraphic location of the volcaniclastic layers Our first study of the ash layer (Canudo et aI., 1993) used the information available from Hernandez et aI. (1991) who located the corresponding beds within the Zuera Formation accepted to be Aragonian in age. According to Hernandez et aI. (1991) the Zuera Formation is stratigraphically above the Calizas de Pefialba unit which contains a diagnostic 'pre-Ramblian' fauna including the eomyd Ritteneria (Cuenca et aI., 1989). However, following our recent studies, the Calizas de Pefialba unit appears to be at the very bottom of the Zuera unit. In addition, the age of the ash has been documented with a newly discovered fauna in the Tardienta section (Fig. 2) and this new fauna is pre-Aragonian. The fauna comes from a fluvial silt bed overlaying the ash bed in Tardienta. The age is considered Ramblian based on (1) the presence of the glirids Peridyromys murinus and P. turbatus, Pseudodryomys ibericus and Quercomys daamsi, and the lagomorph Prolagus aff. vasconiensis fortis, and (2) the absence of Ritteneria and cricetids. The 302 C.S. Odin et al. glirids are the most abundant mammal group in the locality. They are usually considered poor for correlation (Alvarez-Sierra et aI. , 1990; Daams, 1990). Nevertheless, because other small mammals are scarce in the Ebro Basin (Cuenca et aI., 1993), glirids are used in this study. Quercomys daamsi represents a more primitive morphological stage compared to Ramblian representatives such as Quercomys parsani; this would suggest a pre-Ram blian age. The Tardienta fauna is almost identical to the close San Juan and La Galocha (1, 2, 3) faunas (Alvarez-Sierra et aI., 1990). The San Juan fauna contains a tooth of Rhodanomys or Ritteneria which may be a vestigial form of the pre-Ramblian (Agenian) assemblage. For this reason, the authors included the San Juan locality in the Y2 mam mal biozone. About 10 and 20 m above are the localities La Galocha 1 and La Galocha 2 and 3, respectively. Alvarez-Sierra et aI. (1990) group the three localities within the same mammal zone (Y2 = MN2b) of the Agenian stage because they share Peridyromys turbatus which is also present in Tardienta. The lack of Rhodanomys or Ritteneria in Tardienta could be caused either by local ecological or by regional biostratigraphical factors. In short, Tardienta should be included in biozone Y2 based on glirid assemblage, or in the Z zone (Ramblian) based on the absence of Ritteneria and cricetids. Lagomorpha could be very useful for local correlation; unfortunately, the material collected from Tardienta is poor. Nevertheless, the taxon identified is closer to Prolagus vasconiensis fortis from San Juan and La Galocha than to any other Prolagus species. The reference level of Mein's zone MN2a in Spain is Cetina de Arag6n (Zaragoza, Duero Basin; Mein, 1989) which contains Ritteneria molinae (Alvarez-Sierra, 1990) and Quercomys bijmai (Daams, 1990) among other rodents. They represent more primitive morphological stages than Ritteneria manca and Quercomys daamsi, respectively. The reference level for Mein's zone MN2b in Spain is Navarrete del Rio in the Catalayud Teruel Basin. It contains no Ritteneria nor Rhodanomys being thus a Ramblian locality and no hypsodont glirids as Quercomys. In the correlative Ramblar 1 locality, Quercomys parsani is the hypsodont glirid present according to Daams (1990). In summary, the presently known constraints allow us to place the Tardienta locality either in mammal biozone Y2 or in biozone Z and the immediately underlying ash layer is at the limit Agenian/Ramblian or at the very bottom of the Ramblian. GEOCHRONOLOGY Petrography and mineralogy of the geochronometers Study of the ash layer was undertaken using microscopy, microchemistry (micro probe), and X-ray diffraction. The layer shows sporadic clinopyroxene (0.01 to 1.0 mm) and abundant sanidine and plagioclase as the main crystallized phases. The chemical composition of clinopyroxene (between E142.5 W044 FS 13 .5 and En16.5 W040.5 Fs43 ) suggests a tholeiitic affinity. Plagioclase composition is near the anorthite end member. These minerals have a planar orientation and follow continuous levels with respect to the partly devitrified matrix. SEM analysis of pyroxene and feldspars shows fragmentation structures. Zircon and apatite crystals are present, idiomorphic (Fig. 4) but scarce in Tardienta. X-ray diffraction of the matrix shows the presence of analcime, rodocrosite (?) and clay minerals: corrensite, illite, smectite, and saponite. Two samples Biostratigraphy and geochronology of a Miocene continental volcaniclastic layer 303 Fig. 4. Scanning electron microscopic views of idiomorphic pyroclastic heavy minerals; apatite crystal (top, length is 200 /Lm) and zircon crystal (bottom, length is 140 /Lm) (pictures: Dto Ciencias de la Tierra, Zaragoza). collected from Lanaja (TI57) and Pefialba (TI72) were used for separation of datable minerals. The Lanaja sample is pink in colour, and binocular microscope examination shows essentially white milky minerals, some transparent, and common oxy-hydroxides which give the colour to the rock. The sample was treated following the usual procedure for separation of pyroclastic components using washing, sieving, magnetic fractionation, and heavy liquids (bromoform, mixtures bromoform-acetone, methylene iodide). The single magnetic mineral observed is bottle-green, transparent pyroxene (about I %0 versus whole-rock = IWR). The non-magnetic fraction includes feldspar (about 3% IWR) and quartz including common bipyramidal idiomorphic crystals. Apatite is rare and pink zircon crystals (>0.05%0 IWR) are short, 0.15 to 0.3 mm long, idiomorphic, commonly poorly shaped, and contain black inclusions. The array of pyroclastic minerals confirms the volcanic origin of the sample. 304 G.S. Odin et al. SANIDINE Q -1 PLAG I OC LASE ...... 1M.... M Q 30° Fig. 5. X-ray diffraction patterns of the feldspars separated from the Lanaja volcaniclastic sample. The interval (131)-(131) is emphasized; quartz (Q) is present. Diagrams obtained in the Laboratoire de Geologie of the Museum National d'Histoire Naturelle, Paris using Ka Cu ray; abscissa units are in "2e). Biostratigraphy and geochronology of a Miocene continental volcaniclastic layer 307 Table 2 Results of the 40 Ar/39 Ar incremental heating measurements on sanidine from Lanaja °C Apparent age Ma ±2a %39Ar % Rad. KlCa 650 18.!1 0.28 0.1 63 1.4 800 18.80 0.24 0.3 85 1.3 950 19.19 0.22 1.7 96 1.4 1000 19.34 0.22 1.8 99 1.8 1050 19.28 0.22 2.2 99 2.3 1100 19.19 0.22 10.5 99 3.0 1150 19.28 0.22 16.2 98 3.2 1200 19.37 0.22 24.0 98 5.5 1250 19.38 0.22 24.5 97 6.0 1300 19.46 0.22 12.6 96 5.3 1375 19.52 0.24 4.6 91 4.4 1450 19.34 0.24 1.3 85 2.5 1600 19.69 0.28 0.2 71 4.4 °C 40* 39 38 37 36 40/39 650 732.4 ± 1.5 117.9 ± 0.1 2.059 ± 0.005 42.49 ± 0.28 0.936 ±0.05 3.887 ± 0.02 800 283.4 ± 0.6 59.64 ± 0.12 0.883 ± 0.002 21.68 ± 0.08 0.149 ± 0.002 4.037 ± 0.02 950 6426± 13 1490± 3 20.79 ±0.04 521.5 ± 1.0 1.074 ± 0.005 4.122 ±0.01 1000 6243 ± 12 1485 ± 3 20.06 ±0.04 403 .9 ±0.8 0.337 ± 0.005 4.153 ± 0.01 1050 7667 ± 15 1831 ±4 24.44 ±0.05 385 .0 ±0.8 0.361 ± 0.005 4.141 ±0.01 1100 171400 ± 300 41100 ± 100 544.8 ± l.l 6659 ± 13 7.81 ±0.03 4.122 ± 0.01 1150 988200 ± 2000 234600 ± 500 3196 ± 6 35480 ± 70 62.35 ± 0.!2 4.1 41 ±0.01 1200 1479500 ± 3000 347900 ± 700 4697 ± 9 30730 ± 60 109.4 ± 0.2 4.161 ±0.01 1250 1520400 ± 3000 354600 ± 700 4778 ± 10 28750 ± 60 152.4 ±0.3 4.162±0.01 1300 791600 ± 1600 181900±400 2490 ± 5 16720 ± 30 106.3 ±0.2 4.180±0.01 1375 82500± 165 17910 ±40 250.5 ± 0.5 2532 ± 5 25.45 ±0.05 4.192±0.01 1450 25220± 50 5170 ± 10 78.12±0.16 995.5 ± 2.0 12.84 ±0.Q3 4.153 ±0.01 1600 220.2 ±0.4 36.98 ±0.07 0.596 ± 0.001 4.14 ± 0.08 0.216 ± 0.002 4.230 ± 0.02 HD-BI = 24.21 Ma; error on ] curve = 0.5%; T (total fusion age) : 19.36 (± 0.21) Ma; (plateau age 1000- 1300°C): 19.33 ± 0.21 Ma. ] = 0.00259; 40/ 39 =40Ar radiogenic/39ArK; gas in 10- 15 moles; analytical precision la in lower table. Sample weight: 125 mg. (Analytical data G.S .O. in Lausanne.) mean plateau age is 19.33 ± 0.21 Ma (2a intralaboratory analytical uncertainty). The plateau age is similar to the total age; this is a priori an excellent age result which is both precise and well qualified by the obtained spectrum. The isochron plots of the step-heating analytical results were then calculated. If one uses the seven steps forming the plateau, the 40 Ar/36 Ar versus 39 ArP6 Ar plot gives an apparent age at 19.22 ± 0.21 Ma with an excellent mean standard weighted deviation (MSWD) of 0.7 and an initial 40 ArP6 Ar ratio at 383 ± 59; the 36 Ar/40 Ar versus 39 Arl 40 Ar plot gives similar results with an apparent age at 19.20 ± 0.21 Ma, a MSWD of 0.8, and an initial ratio at 397 ± 62 apparently higher than the atmospheric ratio. However, the isochron plot of all steps (see Fig. 7) would suggest a ratio smaller than atmospheric and it is assumed that the information given by the isochron plots is not significant from this point of view. 308 c.s. Odin et al. DISCUSSION AND CONCLUSIONS The peculiar layer recently discovered in several localities in the Ebro Basin is purely volcanic in origin in Lanaja and probably in other sections too. This is shown in the field and the laboratory by the sedimentologic and petrographic characteristics; the variety . of pyroclastic minerals and their shape also confirm the volcanic origin following air , transport in Lanaja. The biostratigraphic age has been improved during the study from mid-Aragonian to earliest Ramblian or pre-Ramblian following discovery of a new site for mammal fauna in a section presumably time-equivalent to the one dated (see Table 1). Geologically, the apparent age measured is most probably connected to the time of crystallization of the sanidine at the time of eruption of the volcanic material. As discussed above, the purely volcaniclastic nature of the layer which results from aerial transport of the erupted material and sub synchronous deposition indicates that the age of the sanidine is reasonably similar to that of the deposition. The age at 19.3 ± 0.2 Ma may be compared directly to the other ages obtained in this volume using the same reference material. For use by comparison to external data, the intralaboratory analytical error bar must be combined with the error on the known age of the monitor (±2.5% of the age, 2a) which should be propagated to the analytical age in order to derive a realistic geological age of 19.3 ± 0.7 Ma. This age corresponds to a moment within the first quarter of the Burdigalian global stage and may be used for connection between the continental and marine chronostratigraphic scales. The Aragonian continental stage is sometimes shown as Middle Miocene in age in the Spanish literature. In terms of Ma, this would correspond to an interval between 16 and 11 Ma. However, Hernandez et al. (1991), among others, locate the Zuera Formation in the Early Miocene; Steininger et al. (1985) show the Aragonian in correspondence to the entire Burdigalian to Serravallian Stage interval which would correspond to the interval between about 20.5 Ma and 11 Ma. Introduction of the Ramblian has modified this picture in Steininger et al. (1989). The present study reports data documenting the isotopic age of a level which is not Aragonian in age, as originally accepted, but practically coincident with the Ramblianl Agenian limit of the Ebro Basin (= within Mein's zone MN2b, = somewhere within local mammal zonal units Y2-Z). According to evolutionary criteria, the entire regional Ramblian 'stage' would comprise only a small interval of time (local zonal units Z + A or a portion of Mein's mammal zone MN2b). Compared to the Aragonian, this would probably represent about 1 Ma. The brevity of this interval may lead to question the necessity to create a 'stage' which is not (or difficult to be) subdivided on the basis of faunal units. The age of 19.3 Ma obtained in the present study suggests the pre-Ramblianl Ramblian limit about 0.5 Ma above the base of the Burdigalian. We have shown in Table 3 the influence of our new result on the previously proposed relations between chronostratigraphic units. In Table 3A, we show the proposal by Steininger et al. (1989). However, the numerical scale used by these authors is not correct as far as marine stages are concerned. As a first improvement, we have redrawn the columns keeping unchanged the Biostratigraphy and geochronology of a Miocene continental volcaniclastic layer 309 Table 3 Time relation between chronostratigraphic and biostratigraphic units around the Ramblian Spanish mammal stage 25 20 15 m OLIGOCENE AQU. BURDIGALIAN SERRA VALL. TO. co Ol ~ '-"' LL MN CIl (/) 0 AGEN. pp RAMBLIAN ARAGONIAN V. 25 20 15 IGOCENE I AQUITAN. BURDIGALIAN I LA. I SERRA V ALLIAN AGENIAN ORLEANIAN Iii ASTARACIAN 2a I 2b 3 I 4 I 5 6 7 8 Y2 Z A B C 0 E F G AGENIANpp RAMBLIAN ARAGONIAN .. 25 20 15 SERRA V ALLIAN I QI t OL IGOCENE I AQUITAN· I BURDIGALIAN I LA. ] ~ I AGENIAN ORLEANIAN ! ASTARACIAN ¥I 2a I 2b 3 I 4 I 5 6 7 8 Y2 Z A B C 0 E F G AGENIAN pp RAMBL. ARAGONIAN 81@ I A) Scale proposed by Steininger et al. (1989)•; B) Scale A kept unchanged but for application of a better numerical scale to the marine Stages; C) Scale B improved to agree with an age of about 19.3 Ma for the basal boundary of the Ramblian stage. Note that Agenian stage does not go into the Oligocene (S. Sen: personal communication) relations between marine and continental units but using an up-to-date numerical scale (see Odin and Odin, 1990; Odin 1994). Proposal B is obtained; it is still inconsistent with our new geochronological study. In a second improvement, we have thus drawn a third proposal (Table 3C) by disconnecting the relationship between the continental units and the marine stage column. Relations within the continental units are kept unchanged because these units are interrelated by their comparable faunas, and thus the relations 310 C.S. Odin et al. are assumed to be correct. Compared to the initial scheme, the final picture shows: (1) a short 'Ramblian' unit; (2) MN zones of quite comparable apparent durations; (3) an Agenien/Orleanian boundary about 3 Ma younger than supposed before. The validity of these conclusions depends on the three hypotheses done (i) in changing the direct relation between time and continental units assumed to be poorly known; (ii) in changing the geometric relation between marine and continental units also known to be uncert~in; and (iii) in accepting that the correct relation between Spanish and European MN faunal zones is known. With our new geochronological results, we have established with an acceptable reliability that the base of the Ramblian unit in the Tertiary Ebro Basin, is located near and slightly younger than the Aquitanian/Burdigalian global stage boundary. SOMMAIRE - BIOSTRATIGRAPHIE ET GEOCHRONOLOGIE D'UN NIVEAU VOLCANOCLASTIQUE MIOCENE DU BASSIN CONTINENTAL DE L' EBRE, ESPAGNE (Manuscrit soumis: Decembre 1993; revise: Decembre 1994: redacteur responsable: GSO) Le Bassin tertiaire de L'Ebre est un grand bassin montrant une succession fossilirere continentale. Un niveau volcanoclastique a ete decouvert, caracterise et repere dans plusieurs sections ainsi qu'une faune de Mammireres tres proche. Ce niveau corre spondrait a un horizon voisin de la limite entre les etages continentaux Agenien et Rarnblien. Grace a la presence, notamment, de sanidine en excellent etat d'apres les analyses mineralogiques, la datation isotopique a ete realisee par la technique 39 Ar/40 Ar et chauffage par paliers. Le spectre d'age obtenu revele un plateau tres bien individualise dont l'age de 19,36 ± 0,21 (incertitude analytique interne, 20-) Ma est identique a rage total et aux ages deduits des isochrones. Cet age permet de revoir et de modifier de fac;on appreciable, les correlations entre la succession des unites biostratigraphiques continentales vis a vis de I' echelle chronostratigraphique. D'apres cette datation directe, la base de l"etage' Ramblien se situerait dans Ie Burdigalien inferieur et cette unite ramblienne serait tres courte. (Sommaire des auteurs) ACKNOWLEDGEMENTS Preliminary steps of mineral separation were undertaken by A. Amorosi (Univ. Bologna and Univ. P. & M. Curie, Paris) as part of his thesis. S. Sen (Museum, Paris) is thanked for his comment on this paper. Isotopic measurements for this study by G.S. Odin were achieved thanks to facilities made available by the Fonds National Suisse in Lausanne and permission of the responsible of the University. X-ray diffraction analyses were achieved in the Laboratoire de Geologie du Museum, Paris. Field and palaeontological studies were partially supported by the Spanish DGCYT Project No. PB 92-0013. No funding was provided by French Organizations for this research.