Palaeoenvironmental reconstruction of a Pleistocene lacustrine sequence from faunal assemblages and ostracode shell geochemistry, Baza Basin, SE Spain

ECOLOGY

ELSEVIER

Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 191-205

Palaeoenvironmental reconstruction of a Pleistocene lacustrine sequence from faunal assemblages and ostracode shell geochemistry, Baza Basin, SE Spain P. Anad6n, R. Utrilla, R. Julih lnstitut de Cibncies de la Terra (Jaume Almera) CSIC, C. Martl i Franquks sn, 08028 Barcelona, Spain

Received 7 June 1993; revised and accepted 19 January 1994

Abstract

We have reconstructed the Early Pleistocene palaeoenvironmental record for a shallow lacustrine sequence from the Baza Basin (Southern Spain), using faunal assemblages and the shell chemistry of two ostracode species (Candona sp. and Cyprideis torosa). In the NE basin sector (Orce area), a Lower Pleistocene lacustrine sequence, up to 10 m thick, contains alternating phases of two fossil assemblages which differ in their salinity requirements. The faunal assemblages record phases of (1) slightly saline, Ca + bicarbonate-rich water, when freshwater organisms predominated, and (2) saline, NaCl-dominant water in which marine-like organisms lived. The ostracode shells from intervals with a saline fauna have higher 6180 values than those from intervals with a freshwater fauna. This feature corresponds to a parallel alternation of positive and negative hydrologic balances. 6180 and 613C values from ostracode calcite display a covariant trend which indicates that the ostracodes lived in a closed lacustrine system. The isotopic data show a better agreement with the inferences from faunal assemblages than with trace element trends. The trace element data, although locally fitting well with the isotope and faunal interpretations, suggest overall that trace element contents in ostracodes are not consistent indicators of salinity and/or temperature variations, but depend on a set of complex factors, which may not be easily deciphered. The alternation of concentration/dilution phases recorded in the studied section can be correlated with climatic cycles described from synchronous ocean basin records from the Late Matuyama chron.

1. Introduction

Lacustrine carbonates may contain a high resolution record of variations in biotic/abiotic interactions in lakes that result from changes in the lacustrine environment, which in turn depend on both, local and global factors (Kelts and Talbot, 1990). Biogenic carbonate from limnic organisms, such as molluscs and ostracodes, provide two types of information. First, the presence of fossil provides information about the depositional environment through knowledge of its environmental tolerances. These environmental conditions include 0031-0182/94/$7.00 © 1994 Elsevier Science B.V. All rights reserved SSDI 0031-0182(94)00009-W

physico-chemical properties of the lake water (salinity, ionic composition, temperature,...) which comprise the species tolerance range (e.g. Delorme, 1969; Forester, 1983, 1986; Anad6n, 1992). Other ecological data, concerning biotic interactions (predation, competition, etc.), may also be inferred in some cases. The second type of information is derived from the trace element and stable isotope geochemistry of the shells. Geochemical studies can also inform us about the physico-chemical conditions of the lake waters, such as temperature, salinity and hydrological budget. Nevertheless it is not always easy to decode geochemical data and

192

P. Anad6n et aL/Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 191-205

understand correctly the biochemical processes involved in the uptake of trace elements and stable isotopes by a particular species. Among limnic organisms, ostracodes are quite selective with respect to their environmental tolerances (De Deckker and Forester, 1988). Ostracode shells are commonly found in lacustrine sediments and, in some cases, they form ostracode coquinas. Ostracodes moult several times (usually 8 to 9) before reaching maturity and, as ostracode life and moulting process are short, a single ostracode shell represents the environmental conditions of the lake over a time interval of a few days to weeks (Chivas et al., 1983). Nevertheless, the moult itself is even more rapid and calcification (chemistry) of the valve after the moult is fixed more or less in hours (Turpen and Angell, 1971). Over the last decade considerable attention has been given to inferring the palaeohydrochemistry of ancient lakes using the chemical composition of ostracode shells (trace elements and stable isotopes of O and C). After the early work on ostracode shell composition by Cadot et al. (1972), and Bodergat and Andreani (1981), Chivas et al. (1983) cultivated ostracodes under controlled conditions (temperature, salinity, pH, water composition) and analyzed the ostracode shells with inductively coupled plasma-argon emission spectrometry (ICP-AES). From this and subsequent works on laboratory cultures and field collections of ostracodes, Chivas et al. (1985, 1986a,b) established that the Sr/Ca in the calcite shells of ostracodes is related to the Sr/Ca of the water. This relationship, or partition coefficient, varies among ostracode genera and seems to be little affected by water temperature. K o (Me) = Meost . . . . de/Mewater

As the trace element content is commonly expressed as the molar metal/Ca ratio the KD is: Ko (Me) = ( Me/Ca)ost.... de/(Me/Ca)water Morse and Bender (1990) pointed out that this type of relationship is actually a partition coefficient, and not a distribution coefficient, because it is based on empirical data, rather than theoretical considerations of thermodynamic equilibrium. Chivas et al. (1986a,b) proposed that the Mg/Ca of the ostracode shell is controlled by water tern-

perature and the Mg/Ca of the water. De Deckker et al. (1988a, b) have also applied this idea, in the case of Cyprideis, in palaeoenvironmental (ancient water chemistry) and palaeogeographical reconstructions. In recent years these methods have also been used by other authors apart from the Australian team (i.e. Anad6n and Juliet, 1990; Engstrom and Nelson, 1991; Holmes et al., 1992). The use of stable isotope (6180 and 613C) data from ostracode shells to reconstruct continental palaeoenvironments has been limited in comparison with the stable isotope data from nonmarine inorganic carbonates. However, various studies have shown that different ostracode taxa from the same sample may show differences in 6180 of as much as 1.3%o (Durazzi, 1977; Fritz et al., 1975; Lister, 1988; Lister et al., 1991; Benson et al., 1991; Von Grafenstein et al., 1992). Eyles and Schwarcz (1991) found consistent differences in isotope values for two Candona species from a 100 m sequence of Pleistocene glaciolacustrine sediments. In this case the isotope differences are interpreted as suggesting a thermally stratified lake and confirm modern ecological observations that the two species have different depth preferences. In order to improve the use of ostracode shell geochemistry as a tool for paleoenvironmental reconstructions, Chivas et al. (1986a) argued that a combination of 6180 and trace element measurements, in particular Mg content, on non-marine ostracodes should be used to resolve palaeotemperature and palaeosalinity variations. In this way, Chivas et al. (1993) have used a combination of trace element and stable isotope measurements of ostracode shells as palaeoclimatic indicators. They have made all the analyses on the same valve after phosphoric treatment. Simultaneously, Palacios et al. (1993) have made comparative interpretations from ostracodes using faunal assemblages and combined geochemical analyses on ostracode shells (trace elements and stable isotopes). The aim of the present paper is to reconstruct ancient environmental variations in a Pleistocene lacustrine sequence from the Baza Basin by analysing ostracode shells for trace elements and stable isotopes (O, C). An evaluation of the sensitivities of these techniques will be provided. We will

P. A nad6n et al./Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 191-205

compare the results of our analytical work with the changes in water chemistry and ecological conditions of the lacustrine environment deduced from simultaneous changes in faunal assemblages. We have used our reconstruction of palaeoenvironmental changes, as a further check on prior palaeoclimatic inferences made for this basin (Anad6n et al., 1986, 1987; Agusti and Julia, 1990).

2. Geological setting 2.1. The studied sequence The Cenozoic Guadix-Baza Intramontane Basin is located in the central area of the Alpine Betic Chain (Southern Spain), at the contact between two main structural units: the External and the Internal zones. The External Zone is mainly formed of Mesozoic carbonate rocks (limestones and dolostones) and Triassic evaporites (gypsum anhydrite and minor halite). The Internal Zone is formed of metamorphic rocks. The basin contains a sedimentary sequence ranging from Lower Miocene to Pleistocene. During the Pliocene and Pleistocene a lacustrine sequence over 100 m thick was deposited in the central part of the basin (Baza Fm; Vera, 1970). In the Orce-Venta Micena area (NE part of the basin) a lower Pleistocene shallow lacustrine sequence up to 25 m thick crops out extensively. This sequence comprises calcareous and dolomitic mudstones, sand, sandstones, limestones and dolostones. These lacustrine deposits overlie calcareous and dolomitic mudstones which were deposited in a nonmarine mudflat environment during a major water level recession (Anad6n et al., 1987). Preliminary paleomagnetic results from the Lower Pleistocene succession of the Baza Basin (J.M. Par6s and O. Oms, unpublished data) show that the Orce section records a reverse polarity event with a probable short normal event at the top. The Orce section deposits may have accumulated during a reversed part of the Matuyama chron, below the boundary of the Jaramillo normal event (J.M. Par6s and O. Oms, pers. comm.). Previous work on the Pleistocene lacustrine sequence in the Orce-Venta Micena area (NE Baza

193

Basin, Spain) based on palaeoecological studies of ostracodes and other fossils (Anad6n et al., 1986, 1987) has documented dramatic alternations of saline (Na-C1) dominant and slightly saline or fresh- (bicarbonate-rich) water phases. An early reconstruction of the palaeohydrochemistry of some shallow lacustrine sequences based on the ostracode shell chemistry was presented in Anad6n and Juli/t (1990). Major changes in the chemical composition of the water were deduced for the sequences that originated in shallow, littoral areas, where ostracode valves from the saline phases had lower Sr/Ca values than those from sediments which formed under lower salinity conditions. In this paper we present the results of a geochemical study of ostracode shells from the Orce section (11 m thick). This section (Fig. 1) is located 500 m to the N of the Barranco de Orce section described in Vera et al. (1984) and Anad6n et al. (1987). The studied section comprises three lowermiddle Pleistocene mammal sites: ORCE 2 (Mimomys ostramosensis zone of Agusti, 1986) at the base, ORCE 4 (Allophaiomys pliocaenicus zone of Agusti, 1986) and ORCE 3 (Mimomys savini zone of Agusti, 1986) near the top of the section (Fig. 2). The studied section is formed of an alternation of quartzose sands, sandy calcareous mudstones and limestones (Fig. 2). Dolomite is present in minor amounts and is probably detrital. Bioclastic debris accounts for a large proportion of grains in the calcareous mudstones. Some sand beds are formed by accumulations of skeletal carbonates (ostracode and mollusc shell fragments). In the studied section fine lamination is absent; most of the section is massive and bioturbated.

2.2. Fossil invertebrate content Two fossil invertebrate asemblages, distinguished in previous studies in the Orce-Venta Micena area (Anad6n et al., 1986, 1987; Anad6n and Juli/t, 1990), have also been recognized in the Orce section (Fig. 2). One assemblage, occurring between 2-2.5 m and again between 5.5-6 m, comprises an euryhaline fauna (extremely eurytopic organisms) of ostracodes (Cyprideis torosa, smooth valves, Loxoconcha sp.), foraminifers

P. Anadrn et aL/ Palaeogeography, Palaeoclimatology, Palaeoecology 111 (1994) 191-205

194

~

LEISTOCENE LACUSTRINE FACIES NEOGENE AND

.

r=~ ~

,..~,~.oc~.~

,,,,

MAINLY CARBONATE MESOZOIC ROCKS M - T ' M O - P H ' - ROCK-

~..

,~

~"~'~"'--