Bulk and compound-specific δ13C and n-alkane indices in a palustrine intermontane record for assessing environmental changes over the past 320 ka: the Padul Basin (Southwestern Mediterranean realm)

Here we provide valuable information about the palaeoenvironmental evolution of Southwestern Mediterranean region during the last ca. 320 ka through a biomarker-based study of the longest continuous continental Quaternary record in the Iberian Peninsula. The n-alkane content and δ13C values of these lipids were measured in 300 samples taken from the uppermost 55 m of the Padul Basin (PB) record. The δ13C signal of long-chain n-alkanes was a reliable proxy for C4/C3 terrestrial vegetation composition in the basin, as emergent macrophytes made a minor contribution to these homologues. In contrast, the δ13C values of C23 and C25 alkanes reflected mainly phases of increasing water level of the lacustrine/palustrine water body since aquatic macrophytes contain a large proportion of these compounds. Low δ13C values were attributed to a marked contribution of plants using the C3 photosynthetic pathway. Intervals with the lowest δ13C values were attributed to an important input of angiosperms, although they could also be explained by changing environmental conditions or environmental stress, as large shifts in δ13C occurred in long-chain homologues typically abundant in terrestrial plants. Shifts in δ13C of medium-chain homologues reflected limited CO2 availability induced by water temperature, salinity, pH, enhanced productivity, low atmospheric pCO2, or stagnant barriers, rather than the abundance of aquatic macrophytes. Our results also suggest enhanced isotopic fractionation during lipid synthesis by aquatic macrophytes within MIS 7 and the Holocene, leading to increased δ13C values of bulk OM and of long-chain n-alkanes. Hence, the δ13C logs were ideal for studying the contribution of aquatic macrophytes to the lipid and isotopic composition of sediments and for the reconstruction of palaeoenvironmental conditions. These results confirmed that C4 plants had a low presence in the PB. Comparison with biomarker analysis and pollen data of the PB and other records of the Southwestern Mediterranean revealed that δ13C values of bulk OM and of long-chain n-alkanes reflected global climatic oscillations during MIS 7 and the episodes Heinrich Events 3, 2, 1 and Younger Dryas.


Introduction
Lacustrine and palustrine sediments provide excellent palaeoclimatic and palaeoenvironmental reconstructions from the study of a variety of proxies, including elemental, isotopic and molecular organic geochemical ones. Among these, biomarkers are commonly used to reconstruct past environmental conditions (Meyers 1997(Meyers , 2003Meyers and Ishiwatari 1993;Ortiz et al. 2010;Xie et al. 2004;Zhou et al. 2005, among others). In particular, n-alkanes have been extensively used for this purpose. The sources of n-alkanes in geological records can be inferred from their molecular distributions. The n-alkanes of the leaf wax of terrestrial plants are enriched in C 27 to C 31 , whereas aquatic macrophytes are characterized by the predominance of C 23 and C 25 homologues Hamilton 1963, 1967). Short-chain n-alkanes (C 17 and C 19 ) are attributed mainly to algae and cyanobacteria (Cranwell et al. 1987). Given these considerations, indexes like the average chain length (ACL) (Poynter 1989) have been developed to estimate the contribution of different types of vegetation to sediments. In this regard, the aquatic macrophyte proxy (Paq) (Ficken et al. 2000) provides information about the input of aquatic macrophytes to marine, lacustrine and palustrine environments.
The carbon isotopic composition of organic matter (OM) in lake sediments is widely used in palaeolimnology. δ 13 C values can be used for studying changes in bulk OM sources (mainly C 3 and C 4 vegetation), for reconstructing palaeoenvironmental conditions (Farquhar et al. 1989;Ficken et al. 1998;Street-Perrot et al. 2004;Eley et al. 2016, among others) and past productivity rates (e.g., Brenner et al. 1999;Hodell and Schelske 1998;Hollander and McKenzie 1991;Hollander et al. 1992;Meyers 2003), and for identifying changes in nutrient availability in surface waters (Bernasconi et al. 1997;Teranes and Bernasconi 2005). In this regard, the distinct CO 2 fixation pathways of plants produce characteristic differences in the stable carbon isotope composition of leaf waxes, with C 3 plants usually having more depleted δ 13 C values than C 4 plants, which in turn are related to climatic characteristics, as the abundance of the latter shows good correspondence with arid conditions. Similarly, phytoplankton (C 3 algae) preferentially utilize 12 C to produce OM, which is approximately 20 0 / 00 lighter than the 13 C/ 12 C ratio of their dissolved inorganic carbon (DIC) source (cf., O'Leary 1988;Wolfe et al. 2001), and this proxy can be used to estimate productivity.
The δ 13 C composition of specific compounds has been employed in recent years for palaeoenvironmental interpretations. The δ 13 C composition of individual n-alkanes has been used to infer biosynthetic processes, source inputs, and palaeoenvironmental conditions, and it provides more specific data on past changes in C 3 and C 4 plant abundance than the δ 13 C of bulk OM (Street-Perrot et al. 1997;Ficken et al. 1998;Huang et al. 2000Huang et al. , 2001Sinninghe Damsté et al. 2011;Sun et al. 2013). Thus, the δ 13 C values of long-chain n-alkanes, which are derived mainly from terrestrial vegetation, are used to reconstruct the input of OM from C 3 and C 4 plants to sediments (Ficken et al. 1998;Huang et al. 2001Huang et al. , 2006Liu et al. 2015;Sinninghe Damsté et al. 2011;Yamamoto et al. 2010). Changes in the contribution of C 3 and C 4 plants are controlled mainly by climatic and environmental factors, such as temperature, aridity and atmospheric CO 2 levels (Cerling et al. 1997;Huang et al. 2001), although other factors (e.g. pH, salinity, productivity) are likely to affect both the bulk OM and compound-specific δ 13 C signals (Hollander and McKenzie 1991;Hollander et al. 1992;Meyers 2003;Street-Perrot et al. 2004;Eley et al. 2016).
Thus, the δ 13 C composition of bulk OM and of long-chain n-alkanes in lake sediments has been traditionally considered a reliable way to track changes in the contribution of C 3 and C 4 plants. A key premise is that long-chain leaf wax components are derived primarily from land plants. However, the input of aquatic macrophytes can be underestimated (Liu et al. 2015). Moreover, the wide range of δ 13 C values of medium-and some long-chain n-alkanes may bias the interpretation towards more C 3 or C 4 plants, as aquatic macrophytes show large variations in δ 13 C values (Aichner et al. 2010;Liu et al. 2015).
Located in the south of the Iberian Peninsula, the Padul Basin (PB) has one of the longest terrestrial records in the world, the main one in the Southern Mediterranean realm, with more than 100 m covering ca. 1 Ma (Ortiz et al. 2004. The importance of the PB was first recognised in the pollen studies of Menéndez Amor and Florschütz (1962Florschütz ( , 1964 and Florschütz et al. (1971) performed in a ca. 100-m deep core, and of Pons and Reille (1988) in a composite record of 22.8 m. In 1997, a 107 m-long core (SPD) was drilled where the basin reaches its maximum depth near its western edge (Nestares and Torres 1998). High resolution pollen and organic geochemistry studies from this watershed reveal changes in vegetation that can be attributed to alterations in climatic and environmental conditions in the Southern Mediterranean realm over ca. 1 Ma (Flörschutz et al. 1971;Pons et al. 1988;Ortiz et al. 2004Ortiz et al. , 2010Torres et al. 2020). The record is characterized by an overwhelming presence of Pinus. The abundance of the typical steppe taxa Artemisia and Chenopodiaceae are linked to arid phases, while an increase in Mediterranean vegetation, deciduous trees and Poaceae plants is interpreted as the development of a Mediterranean-like climate (Torres et al. 2020). In contrast, periods with predominance of pteridophytes (ferns) and deciduous trees, which both arise in more humid conditions, are scarce in the record. Ramos-Román et al. (2018a) also provided a climate reconstruction of the PB for the last 4700 cal yr BP based of the pollen record of a core drilled in the western part of the basin. Ramos-Román et al. (2018b) reconstructed the vegetation and environmental conditions of the last ca. 11.6 cal ka BP by integrating pollen, inorganic and organic geochemical and sedimentological analyses. Also, Camuera et al. (2019) provided the pollen record of the last 200 ka. However, the chronological model used by these authors should be considered with caution, as it was constructed using only 2 ages older than 50 ka (ca. 65 ka and ca. 120 ka, after the rejection of 3 ages), and using linear extrapolation with two different sediment rates beyond ca. 120 ka (Camuera et al. 2018).
Lithology and organic geochemical proxies reveal the existence of two main phases in the record: from 107 to 60 m (1 Ma to 400 ka), with lacustrine conditions, and in the upper 60 m (last 400 ka), with palustrine conditions (Ortiz et al. 2004). The relative abundance of high molecular weight (MW) n-alkanes shows good correspondence with these large palaeoclimatic periods.
Research using other palaeoclimatological proxies has continued over the years. In this regard, the concentration of organic carbon, the atomic H/C and C/N ratios, and δ 13 C and CPI values of bulk OM have proven to be excellent proxies for the study of the palaeoclimatological and palaeohydrological evolution of the Padul peat bog (Ortiz et al. 2004). However, global climatic changes (glacials and interglacials), such as those occurring from ca. 170 to 25 ka B.P. (from meter 33.6 to 7), did not greatly affect these palaeoenvironmental proxies, the values of which show little variation. Ortiz et al. (2010) demonstrated that the relative percentages of high MW alkanes (C 27 , C 29 and C 31 ) are useful for reconstructing the palaeoenvironmental evolution of the PB.
In the present study, we re-examined the last ca. 320 ka (upper 55 m) of the palaeoclimatological record of the PB, focusing on new data, namely the stable carbon isotope composition of n-alkanes, and various indexes related to the n-alkane content (mainly the n-alkane predominant chain, ACL and Paq indexes). In fact, ACL, Paq and δ 13 C have been used together to reconstruct the palaeoenvironmental conditions of lacustrine environments Sun et al. 2013).
The objectives of this study were (1) to provide new insights into the vegetation and climate change in the Southern Mediterranean realm during the last 320 ka based on stable carbon isotopic measurements of biomarkers in the PB record, (2) to examine the variability in the carbon isotopic composition of the Padul record: bulk and compoundspecific δ 13 C signals, focusing on n-alkanes, and (3) to combine the information provided by δ 13 C values and n-alkane proxies.

Geographical, geological and bioclimatological settings
The PB catchment has a total area of 44 km 2 , while the present peat bog covers 4.6 km 2 and is located 720 m a.s.l. (Fig. 1). The western boundary of the basin is formed by the elevations of the Albuñuelas range (1000 m a.s.l), while the eastern one is the bald and craggy Sierra Nevada, which reaches its highest point at the Mulhacén peak (3482 m a.s.l.), only 20 km east of Padul village (Fig. 1). The northern boundary is a small elevation (Suspiro del Moro, 830 m a.s.l.) that divides the PB and the Granada depression. The southern boundary is defined by the Dúrcal river gorge. The PB resembles a typical endorheic basin that is tectonically controlled (Alfaro et al. 2001): an alluvial fan system developed along its eastern border, grading into the lacustrine-palustrine environment. The still active Padul fault provides continuous subsidence at the western footwall of the basin, and the alluvial fan apices are aligned along the fault trace at 800-900 m a.s.l. (Alfaro et al. 2001) (Fig. 2). A maximum thickness of ca. 110 m is reached towards the eastern boundary of the basin, with Pleistocene sediments overlying Miocene marine ones. A dominant groundwater regime controls the hydrology of the basin, with a hydrogeological connection between the karst rock of Sierra Nevada and the PB itself, the run-off input of water being estimated to contribute only about 8% of the total (Cañada 1984).
The climate in the basin is Mediterranean with a continental influence. Winters are cold and dry, while summers are extremely hot, with maximum temperatures over 40 ℃. The average annual rainfall ranges from 400 to 450 mm yr −1 , evapotranspiration is 700-900 mm yr −1 , and the mean annual temperature is between 15 and 16 ℃ (Rivas-Martínez and Rivas-Sáenz 2009).
The chronology of the PB record was calculated through consistent numerical datings ( 14 C, Th/U, amino acid racemization and palaeomagnetism) of SPD core (Ortiz et al. 2004;Torres et al. 2020) and using Bayesian age-depth modelling created with the Bayesian R-code package "Bacon 2.3.7" (Blaauw and Christen 2011).

Materials and methods
The drilling was performed by the Geological Service of Spain, using a ring with a conventional drill pipe and direct flow of bentonite-water mud. The stratigraphy of the Padul SPD core (Latitude: 37º01' 01'' N; Longitude: 3º36' 07'' W; Elevation: 714.20 m) is shown in Fig. 2. For more details see Ortiz et al. (2004) and Ortiz et al. (2010). Coarse detrital material was predominant in the lower part of the core (from 107 to 70 m), with some scarce peaty intervals, whereas peaty lutites were dominant between 70 and 50 m. In the uppermost 50 m, peat was abundant, with some marly (37-33 m) and sandy interbeds (7-5 m).
A total of 330 samples for lipid fraction and 270 samples for OM and n-alkane stable carbon isotope analysis were taken at approximately 15 to 20 cm intervals along the upper 55 m of the core.

Lipid extraction and analysis (biomarker analysis)
About 5-10 g of sediment was ground, and biomarkers were extracted following the LEB protocol (Lucini et al. 2000). Briefly, this procedure consists of a 24-h soxhlet extraction with dichloromethane and methanol 2:1 (suprasolv Merck) and concentration of the isolated bitumen using a rotor-vapor device. Three bitumen fractions were extracted through liquid chromatography in a silica-alumina glass column using solvents of distinct polarity, namely n-hexane, dichloromethane/n-hexane 4:1 and methanol. Samples were injected into an HP 6890 gas chromatograph equipped with a selective mass detector (HP 5973) and an ATM-5 column (250 × 0.25 mm; 0.20 μm). Helium was used as the carrier gas. The oven temperature was programmed from 60 to 300 ℃ at 6ºC/min (holding time 20 min). The injector was programmed at 275ºC. The compounds were identified using the Data Analysis Program and the Wiley Library. n-Alkanes were calculated from the GC/MS chromatograms of mass/ charge m/z 57 from the first bitumen fraction (extracted with n-hexane), and decafluorobiphenyl was used as an internal standard.

Bulk organic matter δ 13 C analysis
Carbon isotope ratios of bulk OM were calculated at the "Estación Experimental El Zaidín". After carbonate removal with 1:1 HCl, the δ 13 C values of OM were measured in selected samples by means of an EA-IRMS elemental analyser connected to a Finnigan MAT 251 mass spectrometer.

Compound-specific δ 13 C analysis
Compound-specific δ 13 C analyses were performed on the aliphatic fraction (extracted with hexane) using a TraceGC Ultra (Thermo-Finnigan) gas chromatograph coupled to a Delta XP isotope ratio mass spectrometer (IRMS) at the "Estación Experimental El Zaidín". We used an ATM-5 column (250 × 0.25 mm; 0.20 μm). Helium was the carrier gas. The oven temperature was programmed from 60 to 300 ℃ (held 20 min) at 3ºC/min. Carbon isotope ratios are expressed relative to the PDB standard. The reproducibility for replicate analysis is 0.1 0 / 00 .

Results
The logs of the palaeoenvironmental proxies used in this study, i.e. concentration of total organic carbon (%TOC), diverse n-alkane indices, bulk δ 13 C values, and δ 13 C values of some n-alkanes, are shown in Figs. 2, 3.
The %TOC values were between 52.38 and 0.06. From a general point of view, the %TOC values showed oscillations, although most were between 30 and 55%. There were, however, some minima in the %TOC, which correlated with horizons consisting of marls (37-34 m) and sands (4 beds between 8 and 4.5 m). Of note, two other minima occurred at 52-51 m and 43-42 m.

n-Alkanes
All the samples showed an odd-over-even carbon number predominance (CPI values were higher than 2; Ortiz et al. 2004), with a chain-length distribution ranging mainly from C 15 or C 17 to C 33 or C 35 , maximizing either at low MW (C 17 or C 19 ; Fig. 3) or long-chain alkanes (C 27 , C 29 and C 31 ), the former group predominant in only a few samples (Fig. 3).
The profiles of the n-alkane predominant chain, ACL and Paq indexes were considered. The Paq index, calculated as the (C 23 + C 25 )/(C 23 + C 25 + C 29 + C 31 ) ratio (Ficken et al. 2000), was postulated to reflect the relative contribution of emergent and submerged/floating aquatic macrophytes, which typically maximize at C 23 and C 25 , and of terrigenous plants. Paq values ranged between 0.08 (2.27 m) and 0.83 (51.50 m).

δ 13 C values
The δ 13 C values of bulk OM ranged between − 31.46 and − 17.83 0 / 00 They did not show marked oscillations, with most of the samples ranging between − 27.0 and − 29.0 0 / 00 , with the exception between 39 and 34 m (220-180 ka), which showed a shift to less negative values, and in the upper 7 m, in which some samples provided values over − 26.0 0 / 00 . The stable carbon isotopic composition of n-alkanes varied substantially. C 31 and C 33 n-alkanes were typically the most depleted and they ranged from − 26.2 0 / 00 to as much as − 46.6 0 / 00, with the latter showing the most depleted values. δ 13 C values for C 29 oscillated between − 27.7 and − 39.6 0 / 00 and for C 27 between − 23.5 and − 39.6 0 / 00 .

Origin of n-alkanes
n-Alkane profiles were used to distinguish the diverse sources of OM, namely algal, aquatic or terrigenous. Each sample can be characterized by the predominant n-alkane chain length. Phytoplankton and algae are dominated by low MW n-alkanes, maximizing at C 17 (Blumer et al. 1971;Fig. 3 Profiles of δ 13 C values of bulk and even n-alkanes from C 23 to C 33 in the uppermost 55 m of the Padul record, together with the stratigraphy, chronology (Torres et al. 2020) and MIS Cranwell et al. 1987;Gelpi et al. 1970). Submerged/floating macrophytes maximize at C 21 , C 23 and C 25 (Cranwell 1984;Ogura et al. 1990;Viso et al. 1993), while emergent macrophytes have a composition similar to that of terrestrial plants, peaking at C 27 and C 29 (Cranwell 1984). Land plants contain a high proportion of higher MW n-alkanes (C 27 , C 29 and C 31 ) in their epicuticular wax (Cranwell et al. 1987;Eglinton and Calvin 1967;Hamilton 1963, 1967;Nott et al. 2000;Pancost et al. 2002;Rieley et al. 1991). Deciduous trees typically maximize at C 27 , whereas C 31 is dominant in marsh plants and possibly grasses (Cranwell et al. 1987;Ortiz et al. 2004Ortiz et al. , 2010Schwark et al. 2002). However, data from a broad survey of modern plants show that n-alkane chain-length distributions are highly variable within plant groups, and chemotaxonomic distinctions between grasses and woody plants are difficult to make, with the exception of aquatic plants and Sphagnum moss (Bush and McInerney 2013). In contrast, changes in chain length distribution are likely to be a result of temperature and/or humidity conditions (Bush and McInerney 2015).
In the PB record, most of the samples from the upper part of the core maximized at C 27 , C 29 or C 31 n-alkanes, while the C 17 and C 23 were predominant in some samples between 54 and 51 m (320-300 ka), 39 and 37 m (320-300 ka) and the upper 2.5 m (7.5 cal ka BP). Therefore, the n-alkanes present in the sediment probably derived from terrestrial plants, with some exceptions in which the C 17 and C 23 homologues were predominant, indicating a major input from algae and aquatic macrophytes, respectively. Samples in which C 27 was the most abundant n-alkane might be also linked to aquatic macrophytes. Of note, a bimodal distribution was detected in some samples, with maxima at medium-and long-chain n-alkanes, thereby indicating the contribution of distinct types of OM: land plants and aquatic macrophytes.
The ACL index did not show significant oscillations, with values ranging between 29.5 and 27.0, confirming the major input of terrestrial plants. Only between 54 and 51 m (320-300 ka) and in the uppermost 2.5 m did ACL values fall below 25.0. This observation is attributed to an important contribution of algae and/or aquatic macrophytes. Of note, between 39 and 34 m (220-180 ka), and 7 and 4 m (25-10 cal ka BP), ACL values were between 27.0 and 25.0, indicating a significant input of aquatic macrophytes.
The Paq index reflected the abundance of emergent and submerged/floating aquatic macrophytes. According to Ficken et al. (2000), values < 0.1 are linked to a dominant contribution from land plants, while those between 0.1 and 0.4 indicate a significant input from emergent macrophytes. Values > 0.4 are typical in sediments with a major n-alkane input from submerged/floating macrophytes.
In the Padul record, most Paq values fell between 0.1 and 0.4. Therefore, although the predominant n-alkane chain and ACL values indicated a major input of land plants, the Paq index pointed to a considerable contribution of emergent macrophytes. Therefore, a mixed contribution of terrestrial plants and aquatic macrophytes occurred. It must also be highlighted that samples from between 54 and 51 m, and a few between 39 and 34 m (220-180 ka), 7 and 4 m (25-10 cal ka BP) and the upper 2.5 m (7.5 cal ka BP) registered Paq values over 0.4, thereby indicating a major input of submerged macrophytes, which is consistent with lower ACL values and the predominance of short-chain n-alkanes.

δ 13 C values
The isotopic proxies and n-alkane ratios provide a general overview of the palaeovegetation and palaeoenvironmental changes in the PB record. δ 13 C values of bulk OM have been examined previously (Ortiz et al. 2004), but here we widened the study by addressing compound-specific δ 13 C signals and exploring the relationships with n-alkane proxies.

δ 13 C values of bulk OM
The carbon isotopic composition of OM (δ 13 C) in lake sediments is frequently used to distinguish the source of this material, especially between terrestrial and aquatic plants. It can further differentiate different types of land plants (C 3 and C 4 ) and can be used to constrain algal productivity (Talbot and Johannessen 1992).
Most C 4 species are grasses, belonging to Poaceae (40%) and sedges included in Cyperaceae family (15-20%) and are associated with warm and arid climates. Likewise, most Chenopodiaceae are C 4 plants. Moreover, gymnosperms do not use the C 4 photosynthetic pathway, and only 3% of angiosperms use this cycle for CO 2 fixation.

3
Crassulacean acid metabolism (CAM) is a less common C 3 pathway (Winter el al. 2005) used by plants mostly under certain environmental niches, e.g., desert succulents and tropical epiphytes (Keeley and Rundel 2003). Thus, CAM plants make a very low contribution to biomass and provide intermediate δ 13 C values between those of C 3 and C 4 plants.
The rapid decomposition of algae in comparison with land plants implies that there is little preservation of OM derived from phytoplankton and algae and that this material contributes only a low concentration of n-alkanes (Cranwell et al. 1987;Meyers 2003;Montagna and Ruber 1980). Nevertheless, several additional factors can determine δ 13 C values, namely the geochemistry of the water, the exchange of CO 2 with the atmosphere, the dissolved inorganic matter reservoir, and the carbon acquisition mechanisms (Håkansson 1985;McKenzie 1985). In fact, phytoplankton and dominant aquatic plants can sometimes enrich the 13 C of biomass when photosynthesis occurs under CO 2 limiting conditions (Hodell and Schelske 1998;Sharkey and Berry 1985). Thus, an active uptake of HCO 3 − rather than CO 2 can lead to a 13 C-enrichment of biomass (Bernasconi et al. 1997;Espie et al. 1991;Goericke et al. 1994;Hollander and Mckenzie 1991;Laws et al. 1998).
Aquatic macrophytes can be especially abundant in some lacustrine environments (Aichner et al. 2010;Ficken et al. 2000;Liu et al. 2013) and can therefore contribute a significant amount of organic carbon and medium-chain n-alkanes to sediments.
The bulk δ 13 C values of aquatic macrophytes are usually within the range of C 3 plants, although the characteristics of the water mass (high pH, alkalinity, high CO 2 consumption, producing limited CO 2 availability) can lead to less negative δ 13 C values (because of HCO 3 − assimilation), similar to those of C 4 plants (Prins and Elzenga 1989;Keely and Sandquist 1992). Aichner et al. (2010) reported bulk δ 13 C values ranging from − 5.8 to − 18.1 0 / 00 in submerged aquatic macrophytes (in the range of δ 13 C values for C 4 plants), and between − 25.7 and − 26.4 0 / 00 for emergent macrophytes. Similarly, the aquatic macrophytes studied by Chikaraishi and Naraoka (2003) showed bulk δ 13 C values of between − 14.6 and − 16.5 0 / 00 . The small oscillations of δ 13 C values, ranging between − 27.0 and − 29.0 0 / 00, observed in most of the samples from the PB core may indicate a major input from C 3 plants. Such plants may have been either grasses or trees characteristic of areas with distinct precipitation rates. Furthermore, C 3 trees include a huge variety of species with diverse ecological characteristics, although it appeared that most of the OM derived from angiosperms, as they usually have more depleted δ 13 C than gymnosperms. Nevertheless, according to pollen data, there is an overwhelming presence of Pinus along the record (Torres et al. 2020), although there is evidence that gymnosperms produce very few n-alkanes relative to angiosperms (Diefendorf et al. 2011), and Pinus pollen can be largely overrepresented due to the air and water buoyancy linked to its enlarged sac wings. Moreover, in the PB, gymnosperms were located at higher elevations, far from the lacustrine/palustrine water body, with run-off input of water estimated to contribute only about 8% of the total (Cañada 1984). Consequently, changes in the water table in the peat bog were controlled indirectly by infiltration of water from the surrounding mountains (artesian waters), where accumulated snow during colder phases subsequently melted.
However, these δ 13 C values may also be characteristic of aquatic macrophytes (Müller and Mathesius 1999). Therefore, samples of this group cannot be ascribed to a specific palaeoenvironmental event.
In the uppermost ca. 7 m, many samples indicated that the OM derived from algae and/or aquatic macrophytes and land plants, as ACL values were < 26, and Paq ranged mainly between 0.6 and 0.3. However, some variations of the δ 13 C org values between 7.8 and 4.5 m (30-11.5 cal ka BP), with a maximum difference of 5 0 / 00 (− 23.3 to − 28.3 0 / 00 ), were especially marked. This finding may indicate a change in vegetation, but within C 3 plant species. However, variations in macrophytes could also explain this increase.
Samples with lower δ 13 C values of bulk OM [(7.8-7.7 m (30-29 cal ka BP); 6.9-6.7 m (22-20 cal ka BP); 5.9-5.6 m (18-17 cal ka BP) and 4.6 m (12 cal ka BP)] coincided with predominant n-alkane chains of 31 carbon atoms, which are typical of grasses (Cranwell 1973), and C 27 or C 29 predominant n-alkanes, common to trees (Cranwell 1973). Of note, these intervals can be correlated with a marked increase in steppe plants and a significant decrease in xerophilous taxa (Torres et al. 2020) and a significant reduction of the percentage of C 27 with respect to the sum C 27 + C 29 + C 31 , thereby indicating the occurrence of cold-dry phases.

Compound-specific δ 13 C values
Individual carbon isotopic compositions of n-alkanes are shown in Fig. 3, showing some oscillations that differed in medium-chain relative to long-chain n-alkanes, which complement the information on δ 13 C of bulk OM.
In general, the δ 13 C values of medium-and short-chain n-alkanes are more negative than those of long-chain homologues (Sinninghe Damsté et al. 2011;Sun et al. 2013). However, in the Padul record, the δ 13 C values of C 29 to C 33 n-alkanes were more depleted than those of the C 23 homologue.
Moreover, the δ 13 C values of a single n-alkane homologue of a plant can vary in a given location under changing environmental conditions, i.e. δ 13 C n-C 27 alkane of the birch Betula pendula shows seasonal variations of ca. 3 0 / 00 (Pedentchouk et al. 2008). Variations for n-C 27 , n-C 29 and n-C 31 alkanes have also been reported for some saltmarsh species.
The shifts in the δ 13 C values of sedimentary leaf waxes may result not only from climate variables but also from temporal shifts in the distribution of vegetation, as well as from changes in the isotopic signatures of higher plants undergoing metabolic processes or subjected to environmental stress (Pedentchouk et al. 2008), or as a result of adaptation to salinity (Eley et al. 2016).
Thus, the relatively narrow range of δ 13 C values limits the application of carbon isotope techniques. Moreover, given that plant species produce widely varying amounts of n-alkanes, further complexity is introduced when interpreting biomarker-based δ 13 C reconstructions, as species may bias the sedimentary record to a greater or lower extent depending on the amount of alkyl lipids they produce (Eley et al. 2016). Nevertheless, these uncertainties do not affect the interpretation of overall trends, where relatively enriched (depleted) bulk and/or n-alkane δ 13 C values indicate a decrease or increase in the input of C 3 / C 4 plants (Castañeda and Shouten 2011).
Here we observed a certain correlation between the ACL index and bulk δ 13 C logs of OM and C 23 and C 25 n-alkanes, with low oscillations between 34 and 7 m (180-25 ka), and generally lower ACL values when δ 13 C signals increased. These findings indicate that changes in vegetation may affect these proxies. However, no correspondence was observed between the Paq index and δ 13 C logs, thereby indicating considerable and almost constant presence of emergent macrophytes throughout the record and that other factors were responsible for changes in δ 13 C values. Therefore, the n-alkane indices revealed the contribution of vegetation to the sedimentology of the PB. However, specific considerations could be made on the basis of δ 13 C values.
Similarly, the bulk δ 13 C log was comparable to those of C 23 , C 25 and C 27 homologues, with lower correspondence with long-chain n-alkanes (Fig. 3). The multivariate analysis (Table 1) showed that the δ 13 C values of C 23 , C 25 and C 27 were the main contributors to the bulk δ 13 C signal, as they showed significant correlation coefficients. Of note, the C 27 δ 13 C signal was influenced mostly by the input of both aquatic macrophytes and land plants, as the correlation coefficient between δ 13 C 27 and δ 13 C 23 (0.66) and between δ 13 C 27 and δ 13 C 31 (0.43) was significant.
In general, compound-specific δ 13 C values reproduced the bulk δ 13 C signal. However, the range of values of each component differed. Furthermore, small differences were observed in the δ 13 C values of long-chain n-alkanes (C 31 and C 33 ) between the interval located at 39-34 m (220-180 ka) and the rest of the core. In contrast, large differences were detected in those of medium-chain homologues. Therefore, between 39 and 34 m (220-180 ka), the bulk δ 13 C signal was influenced mainly by medium-chain alkanes, typically abundant in aquatic macrophytes. However, it must be considered that, in this interval, the n-alkane predominant chain and ACL values were interpreted in terms of higher inputs of algae and aquatic plants, with terrestrial plants predominating in only few samples. Moreover, the Paq index did not show a significant increase, thereby indicating that the contribution of aquatic macrophytes was similar to that of the previous interval. In our view, the environmental conditions of the water body were responsible for these oscillations, rather than changes in the characteristics of the vegetation. Of note, some macrophytes take up isotopically heavier HCO 3 − when CO 2 is limited (Allen and Spence 1981;Keeley and Sandsquist 1992;Van der Berg et al. 2002). In this regard, the δ 13 C value of HCO 3 − is ca. 10 0 / 00 higher than that of dissolved CO 2 (Mook et al. 1974;Prins and Elzenga 1989), thus producing a shift towards less negative δ 13 C values. According to Aichner et al. (2010), limited CO 2 availability in alkaline and/or shallow lakes with a dense macrophyte population can lead to submerged macrophytes of a single species showing a wide range of bulk δ 13 C values. Moreover, the isotopic fractionation during n-alkane biosynthesis in plants is enhanced with increasing chain length and when more 13 C-enriched carbon sources (like HCO 3 ) are assimilated. Some submerged aquatic macrophytes tend to increase the production of long-chain alkanes, especially in very shallow lakes where macrophytes are exposed during dry episodes.
In fact, it has been proposed that shifts towards higher δ 13 C values in C 23 and C 25 alkanes, typical of macrophytes, are caused by limited CO 2 availability induced by water temperature, salinity, pH, enhanced productivity, low atmospheric pCO2, or stagnant barriers (Street-Perrot et al. 2004). Moreover, when environmental conditions lead to the partial emersion of macrophytes (drop in water level due to dry episodes, thereby increasing exposure to sunlight and air), these plants increase the biosynthesis of C 25 and longchain n-alkanes in favour of the C 23 homologue (Aichner et al. 2010). Therefore, our results point to enhanced isotopic fractionation during lipid synthesis by aquatic macrophytes within MIS 7, leading to an increase in the bulk and n-alkane δ 13 C values. The presence of marls in this interval (39-34 m; 220-180 ka) reinforced this hypothesis. In this case, ACL values decreased, and there was a significant and positive shift of δ 13 C values in medium-chain homologues (C 23 and C 25 ), at which aquatic macrophytes maximize. In contrast, long-chain n-alkanes, especially C 29 , C 31 and C 33 , did not vary. In this regard, the pollen analysis of the PB record revealed that this stage was characterized, in general, by Mediterranean-like conditions with increasing moisture (Fig. 4): high taxa diversity, with an increase in Mediterranean, Riparian, Hygrophilous and Mesophilous, and pollen from these taxa predominated over that from aquatics (Torres et al. 2020). In addition, some phases of increasing aridity occurred. Futhermore, in other long Southwestern European pollen records within the same climatic region, for example those of Valle di Castiglione (Follieri et al. 1988), the marine core ODP-976 (Sánchez-Goñi et al. 2008), and Villarquemado (Valero-Garcés et al. 2019), MIS7 was also characterized by a marked expansion of deciduous forest, although some short phases of increasing aridity were also identified within this stage (Fig. 4).
As occurred with the bulk δ 13 C signal, the δ 13 C values of C 27 , C 29 , C 31 and C 33 n-alkanes fell within the range of C 3 plants, but showed greater variability. Episodes with lower δ 13 C values [(7.8-7.7 m (30-29 cal ka BP); 6.9-6.7 m (22-20 cal ka BP); 5.9-5.6 m (18-17 cal ka BP) and 4.6 m (12 cal ka BP)] could be attributed to an important input from angiosperms. We compared the palaeoenvironmental characteristics obtained from the profiles of δ 13 C values of bulk and of C 25 n-alkane in the PB with the relative percentage of C 27 with respect to the sum C 27 + C 29 + C 31 obtained in the PB ) and in the Fuentillejo maar record (Ortiz et al. 2013) (Fig. 4). The low % C 27 values, which are linked the retreat of deciduous forest and thus to drier conditions, are correlated with the maxima values of δ 13 C curves, which corresponded to colder and drier episodes. This increase in aridity during these events was also confirmed by pollen analysis data in Padul (Torres et al. 2020) and Fuentillejo (Vegas et al. 2008(Vegas et al. 2010, observing a considerable rise in steppe plants, whereas xerophilous plants decreased. Therefore, these episodes represented cold-dry phases, which caused the recession of temperate forests and the extension of grasses, which can be correlated with Heinrich Events 3, 2 and 1 (Last Glacial Maximum) and the Younger Dryas, respectively.
These findings are consistent with palynological data from other long Southwestern European pollen records such as those of marine cores drilled off the coast of Spain (ODP-976, Sánchez-Goñi et al. 2008), and the continental sequences of Villarquemado (Valero Garcés et al. 2019), and Valle di Castiglioni (Follieri et al. 1988), in which the Heinrich Events and the Younger Dryas are characterized by a contraction of temperate tree populations and the expansion of steppe vegetation. Likewise, low δ 13 C values occur in some lakes as a result of organisms feeding on isotopically light methane-derived carbon (Jones and Grey 2011;Kankaala et al. 2006). In anoxic conditions, anaerobic respiration produces 13 C-depleted methane, which can cause a decrease in δ 13 C of algae and aquatic macrophytes (Holander and Smith 2001). This phenomenon is likely to have occurred in the PB, as the medium-chain components also showed significant variations, in some cases coinciding with those of long-chain homologues. However, this variability can be attributed mainly to changing environmental conditions (Pedentchouk et al. 2008), temporal shifts in the distribution of vegetation or environmental stress (Pedentchouk et al. 2008), as large δ 13 C shifts occurred in long-chain homologues, which are typically abundant in terrestrial plants.
In the uppermost 7 m of the core (ca. 25 ka), the rise in δ 13 C values in medium-chain components (C 23 and C 25 ) could be attributed to an increase in the relative contribution of aquatic macrophytes to OM, as ACL values decreased and the Paq index increased. Likewise, the uptake of heavier CO 2 from the water body may have occurred, leading to higher δ 13 C values.
These results confirm that C 4 plants had a low presence in the PB during MIS7 as also observed in the pollen record (Torres et al. 2020). Therefore, the δ 13 C logs were ideal for testing the influence of aquatic macrophytes on the lipid and isotopic composition of sediments. Coinciding with Liu et al. (2015), the δ 13 C values of the C 31 alkane was a reliable proxy for C 4 /C 3 terrestrial vegetation composition in Lake Qinghai (Tibet Plateau), as emergent macrophytes made a minor contribution to this homologue. In contrast, the δ 13 C values of C 23 and C 25 alkanes reflected mainly ecological conditions due to the large portion of these compounds in the aquatic macrophytes of this lake. Fig. 4 Correlation between the profiles of δ 13 C values of bulk and C 25 n-alkane in the Padul Basin and the relative percentage of C 27 with respect to the sum C 27 + C 29 + C 31 (a proxy for increasing deciduous tres and humidity) in Padul ) and in Fuentillejo (Ortiz et al. 2013), the palaeoenvironmental conditions during MIS interpreted in Padul based on the pollen content (Torres et al. 2020) and those in the following long Southwestern Euro-pean pollen records: Valle di Castiglione (Follieri et al. 1988); ODP-976 (Sánchez-Goñi et al. 2008), andVillarquemado (Valero-Garcés et al. 2019). AP arboreal pollen, NAP non-arboreal pollen. The pollen records were plotted in age or depth according to their original published source, and the correlation was performed according to the description of environmental episodes and MIS found in the original papers 1 3

Conclusions
The n-alkane content, and δ 13 C values of these lipids and of bulk OM allowed us to infer the palaeoenvironmental conditions of the PB in the last 320 ka. The δ 13 C signal of long-chain n-alkanes was a reliable proxy for C 4 and C 3 plant content in the basin, as emergent macrophytes made a minor contribution to these homologues. Episodes with depleted δ 13 C values were linked to a significant contribution of plants using the C 3 photosynthetic pathway, with the lowest δ 13 C values indicating a considerable input of angiosperms, linked to cold and dry conditions. Of note, these values could also be attributed to changing environmental conditions or environmental stress, as important shifts in δ 13 C occurred in long-chain homologues, which are typically abundant in terrestrial plants.
In contrast, the δ 13 C values of medium-chain alkanes (C 23 and C 25 ) reflected mainly the ecological conditions of the lake since the aquatic macrophytes of this water body contained a large portion of these compounds. The shifts in δ 13 C of these homologues were attributed to limited CO 2 availability induced by water temperature, salinity, pH, enhanced productivity, low atmospheric pCO 2 , or stagnant barriers, rather than the abundance of aquatic macrophytes. The δ 13 C values of n-alkanes and of bulk OM did not show significant variations and therefore were not greatly affected by palaeoenvironmental variations. In contrast, our results point to enhanced isotopic fractionation during lipid synthesis by aquatic macrophytes (and/or algae) within MIS 7 and the Holocene, which led to increased bulk OM and n-alkane δ 13 C values. In coherence with biomarker analysis and pollen data of the PB and other Southwestern Mediterranean records, these proxies indicated that MIS 7 was characterized by increasing moisture. Furthermore, some episodes belonging to MIS 2 and 1 with lower n-alkane δ 13 C values were attributed to cold-dry episodes (Heinrich Events 3, 2, 1 and Younger Dryas). Therefore, general global climatic oscillations that occurred during these stages appear to have also influenced the environmental conditions in southern Spain. Similar findings were obtained for the biomarker analysis for the PB record , located in southern Spain. The δ 13 C logs were, therefore, ideal for testing the influence of aquatic macrophytes on the lipid and isotopic composition of the sediments. Our results confirm that C 4 plants were not abundant in the PB during MIS 7.