Carbon isotope signatures from land snail shells: Implications for palaeovegetation reconstruction in the eastern Mediterranean

Abstract

In this study we compare carbon isotope values in modern Helix melanostoma shell carbonate (δ¹³C_shell) from the Gebel al-Akhdar region of Libya with carbon isotope values in H. melanostoma body tissue (δ¹³C_body), local vegetation (δ¹³C_plant) and soil (δ¹³C_soil). All vegetation in the study area followed the C₃ photosynthetic pathway. However, the δ¹³C_plant values of different species formed two distinct isotopic groups. This can be best explained by different water use efficiencies with arid adapted species having significantly more positive δ¹³C_plant values than less water efficient species. The ranges and means of δ¹³C_body and δ¹³C_plant were statistically indistinguishable from one another suggesting that δ¹³C_body was primarily a function of local vegetation composition. H. melanostoma δ¹³C_shell reflected the δ¹³C_plant of local vegetation with a positive offset between body/diet and shell of 14.5 ± 1.4‰. Therefore, in the Gebel al-Akhdar where only C₃ plants are present, higher mean δ¹³C_shell values likely reflect greater abundances of water-efficient C₃ plants in the snails diet and therefore in the landscape, whilst lower mean δ¹³C_shell values likely reflect the consumption of less water-efficient C₃ plants. The distribution of these plants is in turn affected by environmental factors such as rainfall. These findings can be applied to archaeological and geological shell deposits to reconstruct late Pleistocene to Holocene vegetation change in the southeast Mediterranean.

Introduction

Terrestrial snail shells are some of the most commonly preserved biological remains in archaeological deposits (Evans, 1972, Prendergast and Stevens, 2014) and geological deposits such as loess sequences (Rousseau, 1990), paleosols (Marcolini et al., 2003), and tufa (Preece and Day, 1994). They can provide a record of continental climatic conditions both through their biology and their shell chemistry (Goodfriend, 1999). The application of stable isotopes from land snail carbonate (δ¹⁸O_shell, δ¹³C_shell) for palaeoclimate reconstruction has steadily increased over the last decade (e.g. Balakrishnan et al., 2005a, Colonese et al., 2007, Colonese et al., 2011, Yanes et al., 2011, Yanes et al., 2011, Stevens et al., 2012). However, the interpretation of δ¹³C_shell as a palaeovegetation proxy can be complicated by factors such as water stress in plants, the ingestion of foreign carbonates, and uptake of atmospheric CO₂. These effects differ from species to species. It is therefore important for any species of interest in any new region of study, to test the empirical relationships between terrestrial shell stable isotopes and environmental parameters with a modern dataset before attempting to use δ¹³C_shell for palaeoenvironmental reconstruction. This study presents a modern carbon isotope calibration on a previously unstudied species, Helix melanostoma. This species is one of the most common land snails in the Mediterranean and northern Africa and one of the most abundant species in Mediterranean archaeological sites from the Middle Palaeolithic to the Roman period (Lubell, 2004, Barker et al., 2010, Lubell and Barton, 2011). Therefore, if this species can be validated as a useful palaeoenvironmental proxy, it can provide valuable information on late Pleistocene to Holocene past environments, in a region with comparatively little palaeoenvironmental data.

Terrestrial mollusc shell carbon is derived from the bicarbonate in snail body fluids. The CO₂ in these body fluids has three major carbon sources: metabolic CO₂, atmospheric CO₂, and CO₂ generated via the reaction of acid with ingested carbonates including limestone rock and recycled snail shell in the snail's stomach (Goodfriend and Ellis, 2002). Therefore, the carbon isotope ratios in land snail shells (δ¹³C_shell) are affected by three main factors: diet; atmospheric carbon; and ingested carbonates (Goodfriend and Ellis, 2002).

Previous studies have found that the primary variable affecting δ¹³C in terrestrial snail shells is diet. Therefore, in herbivorous species, the carbon isotope ratios may provide a proxy for palaeovegetation patterns in terms of the distribution of C₃ and C₄ plants and any changes due to water stress (Yapp, 1979, Francey, 1983, Goodfriend and Ellis, 2002, Stott, 2002, Metref et al., 2003, Balakrishnan et al., 2005a, Balakrishnan et al., 2005b, Baldini et al., 2007).

There are three metabolic pathways in plants: C₃ (Calvin–Benson), C₄ (Hatch–Slack), and CAM (Crassulacean Acid Metabolism). These pathways isotopically fractionate atmospheric CO₂ differently due to the presence of different carboxylating enzymes (Farquhar et al., 1989). Such differences can be discriminated isotopically. The majority of terrestrial plants including temperate grasses, all trees and shrubs use the C₃ metabolic pathway and have a δ¹³C range between −33‰ and −21‰ (Cerling and Quade, 1993). The δ¹³C in C₃ plants (δ¹³C_plant) is a result of the δ¹³C of atmospheric CO₂ (δ¹³C_atmosphere) and the fractionation processes that occur as CO₂ is incorporated into the plant. Atmospheric CO₂ diffuses into the gas space of the leaf through stomatal openings. CO₂ is then combined with the enzyme ribulose bisphosphate in a reaction called carboxylation (Farquhar et al., 1989).

The fractionation of δ¹³C in C₃ plants can be affected by processes that alter the rate of photosynthesis or limit the CO₂ that passes through the stomatal pores. These factors include water availability, the canopy effect, air temperature, light availability, salinity, and the concentration of atmospheric CO₂. The most important factor to consider is water availability. When more soil water is available, the plant's stomata open wider, allowing more CO₂ into the plant, which causes ¹³C-depletion. Conversely, when less water is available, the plant's stomata close, causing ¹³C-enrichment (Farquhar et al., 1989, Condon et al., 1992). Water availability is a function of rainfall amount, relative humidity and aspects of the physiology of the plant. Plants that are more water-efficient tend to modulate their stomatal conductance, thus have more positive δ¹³C_plant ratios (Farquhar and Sharkey, 1982, O'Leary, 1988). Within the range of C₃ variation, there are marked inter and intra-specific δ¹³C ratios which are likely due to physiological factors based on the above mentioned mechanisms as well as potentially genetic variation (Handley et al., 1994).

Modern field studies in various regions of the world have found significant correlations between the carbon isotopic composition of local vegetation (δ¹³C_plant) and modern terrestrial snail δ¹³C_shell (e.g. Francey, 1983, Goodfriend and Ellis, 2002, Baldini et al., 2007, Yanes et al., 2008, Colonese et al., 2014). Snails in regions of predominantly C₃ vegetation have significantly more negative δ¹³C_shell ratios than snails found in regions with predominantly C₄ vegetation, whilst snails that eat a mixed diet will have intermediate δ¹³C_shell values. For example, Balakrishnan et al. (2005b) measured δ¹³C_shell ratios from −10.1 to −8.8‰ in areas of C₃ vegetation, and δ¹³C_shell ratios from −4.3 to +1.9‰ in areas of C₄ vegetation. Similar field correlations were achieved between the acid insoluble organic matrix of terrestrial shell carbonates and local vegetation, although this proxy is less commonly used than terrestrial snail carbonate (Goodfriend, 1988, Goodfriend, 1990).

Laboratory controlled feeding experiments on the land snail Helix aspersa showed that diet-derived metabolic CO₂ was the primary influence on δ¹³C_shell (Stott, 2002, Metref et al., 2003). Snails fed on an entirely C₃ diet (δ¹³C_diet = −27‰) had significantly higher δ¹³C_shell ratios than snails fed on entirely C₄ diets (δ¹³C_diet = −11.7‰). Animals fed on mixed diets had intermediate values, but they were closer to the C₃ range suggesting a preference for C₃ diet in this species. Likewise, the offsets between δ¹³C_diet and δ¹³C_shell was greater for the C₃ diet (Δδ¹³C = 13.75 ± 0.52‰) than for the C₄ diet (Δδ¹³C = 4.89 ± 0.87‰), suggesting that the animals may metabolise these plants in different ways (Stott, 2002).

The correlation between δ¹³C_plant and δ¹³C_shell has implications for palaeoclimate reconstruction as the distribution of vegetation types is related to other environmental variables such as rainfall. For example, Goodfriend, 1988, Goodfriend, 1990 used δ¹³C from the organic component of land snail shells to map the shifting distribution of C₃ and C₄ vegetation during the Holocene in the Negev Desert and by inference the shift in the desert boundary due to changing rainfall patterns. The effects of aridity may also be seen within entirely C₃ plant communities as water stress causes a ¹³C-enrichment in δ¹³C_plant ratios. Therefore, areas with high rainfall, tend to have more negative δ¹³C_plant ratios (Goodfriend and Magaritz, 1987, Goodfriend and Ellis, 2000, Goodfriend and Ellis, 2002). Such changes will be recorded in land snail δ¹³C_shell ratios.

Some studies have suggested that atmospheric CO₂ from gas exchange across the snail's body surface is an important source of δ¹³C_shell (Magaritz and Heller, 1980, Magaritz and Heller, 1983a, Magaritz and Heller, 1983b). The carbon isotopic composition of the local atmosphere can vary microgeographically due to the output of plant respiration and soil CO₂, so more heavily vegetated regions have more ¹³C-depleted δ¹³C_atmosphere values at the ground surface, therefore more ¹³C-depleted shell ratios (Magaritz et al., 1981, Magaritz and Heller, 1983b). However, it is likely that the effect of exchange with atmospheric CO₂ on δ¹³C_shell is more important for smaller snail species that have more of their surface area in contact with the atmosphere (Goodfriend and Hood, 2006). Regardless of the direct influence of atmospheric CO₂ on land snail δ¹³C_shell, plants derive their CO₂ from the atmosphere so changing concentrations of CO₂ in the atmosphere over glacial/interglacial timescales (e.g. Arens et al., 2000), and as a result of the input of depleted CO₂ into the atmosphere since the industrial revolution (Suess effect, e.g. Stuiver et al., 1984, Friedli et al., 1986) should be accounted for when reconstructing palaeoenvironment from land snail carbonates.

Various studies have shown that δ¹³C_shell ratios of land snails can be significantly affected by the ingestion of carbonates as a result of grazing on limestone rocks and carbonate-rich soils since these are usually much more positive in δ¹³C than plants (Goodfriend and Stipp, 1983, Goodfriend, 1999, Goodfriend and Hood, 2006, Goodfriend and Ellis, 2002, Yates et al., 2002, Li et al., 2007, Romaniello et al., 2008, Yanes et al., 2008). This leads to δ¹³C_shell ratios that are more positive than expected from diet alone. This effect seems to be highly species-specific and varies from location to location (Goodfriend and Ellis, 2000). Some species such as H. aspersa showed no detectable influence of foreign carbonate ingestion (e.g. Stott, 2002, Chiba and Davison, 2009), whilst others have shown that carbonate ingestion can account for up to 40% of the δ¹³C_shell ratios (Goodfriend and Ellis, 2000, Yanes et al., 2008). If a significant proportion of δ¹³C_shell is derived from carbonate ingestion, Goodfriend, 1988, Goodfriend, 1990 suggests analysing the acid insoluble organic matter within the shells for palaeoenvironmental reconstruction.

H. melanostoma (Draparnaud 1801) is a herbivorous air breathing land snail (terrestrial pulmonate gastropod) in the family Helicidae. This species has been reported across the Mediterranean in northern Algeria, Tunisia, Greece, Albania, Italy, and on the south coast of France (Baker, 1938, Morel, 1973, Lubell et al., 1975, Kerney et al., 1983). This study confirms its existence in Libya. Previous surveys of H. melanostoma in Algeria and Tunisia have shown that it inhabits humid shady areas in bush and parkland habitats, often favouring calcareous substrates (Baker, 1938, Morel, 1973, Lubell et al., 1975). During dry periods, the snail digs into the soil and aestivates (Baker, 1938, Lubell et al., 1975). In Tunisia, active H. melanostoma were observed from February until June, particularly following rainfall events. The snails probably aestivate from November to January as they were not seen in the landscape during those months (Morel, 1973). Lubell et al. (1975) noted that in the Algerian winter, the snails formed a thick epiphragm suggesting longer periods of inactivity whereas in summer they formed several thin epiphragms suggesting shorter periods of aestivation.

The Mediterranean is located in a transitional zone where tropical and mid-latitude systems both affect climate variability. Therefore, Mediterranean moisture sources are varied. The region is characterised by hot dry summers and cool wetter winters. During summer, a high pressure belt steers storm tracks away from the region. During winter, the sub-tropical high pressure belt moves southwards, bringing cold air with it. The mixing of cold air from the high pressure belt with relatively warmer sea water causes increased evaporation and cyclonogenesis (Gat and Carmi, 1970). Rainfall in the eastern Mediterranean primarily occurs in winter and is of cyclonic origin (Wigley and Farmer, 1982, Kostopoulou and Jones, 2007).

The Gebel al-Akhdar is a limestone massif in the southeastern Mediterranean that rises to a height of >850 m above sea level over 10 km (Fig. 1). It forms a condensed climatic gradient with sharp differences in temperature and rainfall amount over a short distance. At sea level, rainfall is around 250 mm/yr rising to >700 mm/yr on the upper escarpments. Rainfall falls sharply on the southern slopes before dropping to arid levels in the pre-desert (Libyan National Meteorological Center, 2012). McBurney and Hey (1955) note that the Gebel al-Akhdar is “the only area of high ground along 2500 km of flat coastline between Homs in Tripolitania and Mount Carmel in Palestine”. The high elevation of the Gebel al-Akhdar, which acts as a trap for rainfall from the westerlies, ensures that climate and vegetation patterns are distinct from the surrounding regions. Today, the area provides a refuge of fertile Mediterranean vegetation bounded by arid coastal corridors and the Sahara Desert. The region has the richest species diversity of any region in Libya (Hegazy et al., 2011). The vegetation assemblage is predominantly maquis scrubland dominated by Juniperus phoenecia, Quercus coccifera, Pistacia lentiscus, and Ceratonia siliqua with some areas of steppic vegetation including Sarcopterium spinosum and Artemisia (Al-Sodany et al., 2003, El-Darier and El-Mogaspi, 2009, Simpson and Hunt, 2009). Like the Magreb in northwestern Africa, it may have served as a refugium for vegetation, animals and human populations at times of climatic extremes during the late Quaternary (McBurney, 1967, Barker et al., 2010).

The Gebel al-Akhdar contains many important archaeological sites spanning from the last interglacial (c. 130,000 years ago) to present. Many of these sites, particularly the multi-period cave site of Haua Fteah, contain abundant land snail remains (McBurney, 1967, Barker et al., 2007, Barker et al., 2008, Barker et al., 2009, Barker et al., 2010, Barker et al., 2012). These sites therefore offer enormous potential for palaeoenvironmental reconstruction from land snail stable isotopes.