Relief and calcium from gypsum as key factors for net inorganic carbon accumulation in soils of a semiarid Mediterranean environment

Geoderma 398 (2021) 115115 Available online 5 April 2021 0016-7061/© 2021 Elsevier B.V. All rights reserved. Relief and calcium from gypsum as key factors for net inorganic carbon accumulation in soils of a semiarid Mediterranean environment Vito Armando Laudicina a , Carmelo Dazzi a , * , Antonio Delgado b , Haydn Barros c , Riccardo Scalenghe a a Universit` a degli Studi di Palermo, Dipartimento Scienze Agrarie, Alimentari e Forestali, viale delle Scienze, 90128 Palermo, Italy b Universidad de Sevilla, ETSIA, Departamento Ciencias Agroforestales, Ctra. de Utrera, Km. 1, 41013Sevilla, Spain c Universidad Sim´ on Bolívar, Departamento de Física, Apartado Postal 89000, Caracas 1080-A, Venezuela ARTICLE INFO Handling Editor: Ingrid K¨ ogel-Knabner Keywords: Strontium isotope ratio Soil carbonates Atmogenic inorganic C Eucalyptus camaldulensis Toposequence Gypsiferous soils ABSTRACT In semiarid environments, the total inorganic carbon (TIC) in soil may contribute to the total carbon (C) pool more than the total organic C pool (TOC), thus playing a key role in storing atmospheric CO 2 . However, due to the different origin pathways of soil carbonates, not all of the TIC pool can be accounted for CO 2 sequestration. Indeed, the inorganic C can be accounted for a net sink of CO 2 only when calcium (Ca 2+ ) forming carbonates originate from non-carbonate minerals (atmogenic inorganic C, AIC). The aim of this study carried out in a gypsiferous area is to investigate the dissolution of Ca 2+ that comes from gypsum (CaSO 4 ⋅2H 2 O) in the formation of soil atmogenic carbonates and to quantify the different forms of inorganic C pools to understand their role in storing CO 2 in comparison to the TOC pool. To this end, five soil profiles were studied along a hillslope in a gypsiferous afforested area. Soil samples were analysed to determine their main chemical properties, as well as Ca 2+ , Sr 2+ , and the 87/86 Sr isotopes ratios in different soil fractions. Ca 2+ that comes from gypsum, which contributes to the formation of atmogenic carbonates, ranged from 0% to 63%; the remaining percentage resulted from the parent material or other sources (e.g., aeolian dust). The distribution of Ca 2+ in soils depended on the relief and the distance of the soils from gypsum outcrops. The accumulation of AIC in soil developed on Holocene deposits reached a maximum accumulation of 16 kg m −3 in the first meter of depth. The average Sr 2+ / Ca 2+ ratio in primary carbonates was 0.333, whereas it was 0.192 in secondary carbonates, thus suggesting that different sources contribute to Ca 2+ in both carbonate types. Consequently, the Sr 2+ /Ca 2+ ratio of soil carbonates could be a useful indicator of the presence of secondary carbonates. Overall, results suggest that gypsum plays a key role in the net accumulation of inorganic C in the soil, contributing to store atmospheric CO 2 . Finally, considering that the AIC pool was lower than the organic C pool, the latter was by far the most important element in fixing atmospheric CO 2 in the soils of the semiarid Mediterranean environment that were surveyed. 1. Introduction Soils contain 2500 GT of carbon (C) in the first meter of depth, accounting for the main C stock on Earth (Lal, 2004). Soil organic C is the major component of the global C cycle, accounting for more carbon than the terrestrial biomass and atmospheric pools combined (Amundson, 2001). A negligible shift of this huge pool may have substantial consequences on the total environment (Minasny et al., 2017). Soil C pools involved in the interactions between soil and atmosphere are inorganic (TIC) and organic (TOC). TIC, the most stable pool, depends on soil forming factors and especially parent material; TOC, the more labile of the two, depends on both biotic and abiotic factors (Badagliacca et al., 2020). In soils of semiarid environments, the TIC pool is often greater than the TOC pool (Lapenis et al., 2008; Lal, 2009; Camilli et al., 2016). However, in terms of atmospheric CO 2 storage, the relative abundance of the two pools may be inverted. These soils usually contain large amounts of pedogenic carbonates, especially when pedogenesis occurs from calcareous parent materials (Van der Hoven and Quade, 2002). In this case, dissolution and re-precipitation processes merely redistribute inorganic C along the soil profile and do not cause a net sink of atmospheric CO 2 (Monger et al., 2015; Bughio et al., 2017). When carbonates re-precipitate, the leaching of Ca 2+ and HCO 3 – results in a CO 2 flux * Corresponding author. E-mail address: [email protected] (C. Dazzi). Contents lists available at ScienceDirect Geoderma journal homepage: www.elsevier.com/locate/geoderma https://doi.org/10.1016/j.geoderma.2021.115115 Received 28 January 2021; Received in revised form 25 March 2021; Accepted 25 March 2021 Geoderma 398 (2021) 115115 2 towards the soil’s surface (Walvoord et al., 2005), according to the following reaction: Ca2++2HCO− 3→ CaCO3↓+H2O+CO2↑(1) This process, however, must be inferred with care, as Ca 2+ from other sources such as Ca-bearing silicate (e.g., wollastonite and montmorillonite) could combine with CO 2 to form pedogenic carbonates (Carmi et al., 2019). On the other hand, semiarid soils do not seem to be a major sink of atmospheric CO 2 when a soil CO 2 efflux comes from both biological and pedological processes (Rey, 2015; Schlesinger, 2017). In such soils, CO 2 fixation is assumed to come through an abiotic mechanism that includes carbonate dissolution and a CO 2 transport to groundwater (Schlesinger, 2017). Soil carbonates can be inherited from a calcareous parent material (primary or lithogenic carbonates) or formed during pedogenesis (secondary or pedogenic carbonates; Figure S1) (Monger et al., 2015). Schlesinger (1985) in the Mojave Desert in California and Laudicina et al. (2013) in Sicily, Italy estimated <3.5 g m −3 y −1 as the rate of pedogenic inorganic C (PIC) accumulation. Ryskov et al. (2008) reported that Russian soils’ fixed atmospheric CO 2 in pedogenic carbonates during the arid periods had a rate range of 0.9 to 2.2 kg C m −2 in the course of soil formation throughout the last 5000 years. However, according to Liu (2011) and Monger et al. (2015), pedogenic carbonates may not be an important sink of atmospheric CO 2 . Indeed, depending on the source of Ca 2+ , pedogenic carbonates are split into calcitic and silicatic carbonates (Monger et al., 2015). Calcitic carbonates do not store atmospheric CO 2 because they originated directly from primary carbonates or pre-existing carbonates through dissolution and reprecipitation processes (Fig. S1). During the dissolution of pre-existing CaCO 3 by carbonic acid (H 2 CO 3 ) that derived from the reaction of H 2 O and CO 2 , the produced Ca 2+ and 2HCO 3 – re-precipitate as a pedogenic carbonate. However, the latter does not sequester atmospheric C because the source of Ca 2+ is from a pre-existing carbonate and the CO 2 that is consumed during the reaction to form carbonic acid is released upon the re-precipitation of CaCO 3 (Drees et al., 2001; Monger et al., 2015). On the contrary, silicatic carbonates do store CO 2 since they derive from the precipitation of biogenic CO 2 (i.e., CO 2 from soil organic matter mineralisation and root respiration) and Ca 2+ from silicate rocks or other non-carbonate minerals (Monger et al., 2015). For this type of carbonate, Monger and Gallegos (2000; p. 285) proposed the term “atmogenic carbonates” to reflect the sequestration of atmospheric CO 2 . Monger and Gallegos (2000) have suggested six sources of Ca 2+ that form pedogenic carbonates: 1) Ca 2+ from in situ weathering of primary carbonate minerals; 2) Ca 2+ from illuvial CaCO 3 dust; 3) Ca 2+ in rain from dissolved CaCO 3 dust; 4) Ca 2+ from in situ weathering of noncarbonate minerals; 5) Ca 2+ in rain from sea-spray or dissolved noncarbonate minerals; and 6) Ca 2+ adsorbed to dust or in water-soluble, non-carbonate dust minerals. By using the 13 C/ 12 C ratio of soil carbonates, it is possible to discriminate lithogenic from pedogenic carbonates, but it is not possible to distinguish between calcitic and atmogenic soil carbonates since the source of Ca 2+ is unknown. The similar chemical behaviour of Sr and Ca allows 87/86 Sr ratios to be used as a proxy for Ca sources on an ecosystem scale and as a tracer of the origin of Ca 2+ in soil carbonates (Capo and Chadwick, 1999; Violette et al., 2010; Milli` ere et al., 2019). Sr 2+ has an ionic radius of 1.12 Å, very close to that of Ca 2+ (0.99 Å), which explains its role as a Ca 2+ analogue, replacing Ca 2+ in the composition of minerals (Faure and Powell, 1972). The 87 / 86 Sr ratio is not significantly changed during biological processes and mineral precipitation because all mass-dependent isotope fractionation effects are erased by normalisation to the natural 88 / 86 Sr ratio. Thus, the 87 / 86 Sr ratio of pedogenic CaCO 3 depends only on the mixing of available sources (Milli` ere et al., 2019). When the two main Ca inputs in terrestrial ecosystems – namely the weathering of parent rock as well as gypsum – are known and sufficiently contrasted, the origin of Ca included in pedogenic CaCO 3 can be traced and quantified using mass balance techniques (Capo et al., 1998). Moreover, since Sr 2+ can substitute Ca 2+ in the formation of carbonates during their dissolution/reprecipitation process, the Sr 2+ /Ca 2+ ratio may provide useful information about the primary or secondary origin of soil carbonates (Laudicina et al., 2013). By using the 87 Sr/ 86 Sr ratio, Quade et al. (1995) calculated the contribution of different sources in supplying Ca 2+ for the formation of soil carbonates. The authors also demonstrated that marine aerosol is the most significant source of Ca 2+ in soil carbonates in both the coastal and inland areas of southern Australia. In contrast, van der Hoven and Quade (2002) used the same ratio to assess the spatiotemporal variations of Ca 2+ sources in pedogenic carbonates in soils developed in six Quaternary basalt flows and a cinder deposit in the Zuni–Bandera volcanic field near Grants (NM), and Milli` ere et al. (2019) used it to assess the Ca sources in pedogenic needle fibre calcite from the Swiss Jura Mountains in Switzerland. In Sicily, region with Mediterranean semiarid climate, a geological formation called “Serie Gessoso-Solfifera” with different lithology, being gypsum the most extended parent material, accounts for an important surface (about 147,000 ha) (Fierotti et al., 1988). A survey carried out on soils developed on this lithological formation (Dazzi and Monteleone, 2002) confirms that gypsum dissolution and reprecipitation contributes to the formation of gypsiferous soils. Moreover, Laudicina et al. (2003) surveying an area of 1,280 ha, in the central-southern belt of Sicily, mapped 553 ha (43.2%) of gypsiferous soils developed on gypsum and also 260 ha (20.3%) of soils affected by gypsum but not developing on gypsiferous parent materials. This is due to surface and subsurface water dynamics and landscape morphology that promote dissolution of gypsum outcrops and its following re-precipitation and accumulation in lowlands (Lo Papa et al., 2003). Based on the above considerations, therefore, in semiarid environment Ca 2+ solubilised from gypsum may react and precipitate in soil according to eq. (1) thus storing atmospheric CO 2 . The aims of the current study, carried out in a semiarid Mediterranean environment, are: i) to assess the role of gypsum in supplying Ca 2+ for carbonate formation in soils developed on different parent materials in a gypsiferous area and ii) to evaluate the role of soil relief in the distribution of Ca 2+ in soil profiles along a toposequence, and iii) to quantify the atmospheric inorganic C pool to assess its role in storing atmospheric CO 2 . 2. Materials and methods 2.1. Study area The study was carried out in an area with a complex lithology comprised of diatomites, evaporitic limestone, gypsum, gypsarenites, and halite with clay, carbonate, and marl intercalation lithotypes of the Tortonian (>7.2 Ma BP), Messinian (7.2–5.3 Ma BP), and Holocene (<12,000 years BP) time scales (Dazzi and Monteleone, 2002). In that area, a hillslope located within the Mustigarufi wood (37◦31′21.93′′ N; 13◦55′29.14′′ E; Sicily, Italy) was selected (Fig. 1). From the top to the bottom, the slope begins with a doline (sinkhole) made up of clays, marly clays, and silts (Holocene in age, approximately 10,000 years old) accumulated in a continental environment by washing waters flowing into a scarcely drained area. It is followed by clays, sands, and conglomerate deposits belonging to the so-called “Terravecchia” formation (Grasso and Pedley, 1988), and by orographic terraces caused by the Salito River (Laudicina et al., 2013). The upper and the central part of the toposequence is constituted by steep slopes that become moderately sloping towards the bottom part of the hill where slope becomes nearly level. The hydrological setting is conditioned by the different amounts of water infiltrating into the soil profiles, which is decidedly scarce in the first two-thirds of the slope, but greater at the foot of the slope. Such different water infiltration capacities also depend on soil texture: the soil is decidedly finer in the upper and central part of the slope, yet coarser towards the bottom (Table 1). According to K¨ oppen–Geiger’s classification, the climate of the area is classified as V.A. Laudicina et al. Geoderma 398 (2021) 115115 3 Csa (Mediterranean hot summer climates). Climatic data recorded from 1985 to 2011 showed a mean annual air temperature ranging from 16.1 to 18.6 ◦C and a mean annual precipitation ranging from 440 to 495 mm (Sicilian Agrometeorological Information Service, 2011). The soil moisture regime is xeric, border with the aridic one, and the temperature regime is thermic. Vegetation largely consists of Eucalyptus camaldulensis Dehnh, planted about 60 years ago; soils were previously used for durum wheat cultivation (Laudicina et al. (2013); Scalenghe et al. (2015, 2016)). To the aims of this study, the same soil samples collected from the benchmark pedons reported in Laudicina et al. (2013) were used. From the top to the bottom (Fig. 1, Table 2), the soil samples include: pedon P3, located in the centre of the doline, developed on Colluvial deposits and classified as Calcic Pellic Vertisol (Hypereutric) (IUSS Working Group WRB, 2015), with an A-ABkss-Bkss sequence of horizons; pedon P4, located on the edge of the doline, developed on Colluvial deposits and classified as Gypsic Calcic Chernozem (Clayic), with an A-AB-Bk- 2By-2BCy sequence; pedon P22, located on the shoulder of the slope, Fig. 1. In the upper part of the figure, the toposequence is represented with the location of the five benchmark pedons (P3, P4, P22, P64, P1) along the slope (37 ◦31′21.93′′ N; 13◦55′29.14′′ E; Sicily, Italy). The substrata composing the slope are drawn out in the lower part of the figure. From the top to the bottom, the slope started with a doline made up of clays, marly clays, and silts, followed by the Terravecchia formation deposits, and finally by orographic terraces modelled by the Salito River (cf. Dazzi and Monteleone, 2002). Table 1 Main physical and chemical properties of soils along the slope. Pedon numbers follow their position along the slope from the top to the bottom. Numbers are means ± standard deviations. Soil texture data (sand ø 2–0.02 mm; silt 0.02<ø <0.002 mm; clay ø <0.002 mm) had a standard deviations below 5%. LC, loam clay (estimated in field due to the high gypsum content). Pedon Horizon Depth Bulk density Clay Silt Sand Total carbonates Gypsum Ca 2+ cm kg m −3 g kg −1 g kg −1 g kg −1 g kg −1 g kg −1 mol kg −1 3 A 0–12 1750 595 241 164 112 ±16 – 0.8 ±0.2 ABkss 12–50 1860 646 293 61 124 ±19 – 0.6 ±0.2 Bkss1 50–90 1900 591 332 77 162 ±20 24 ±4 8.6 ±1.2 Bkss2 90–150 1900 602 352 46 178 ±17 28 ±4 4.3 ±0.8 4 A 0–7 1650 552 312 136 162 ±31 – 1.3 ±0.2 AB 7–20 1750 573 341 86 188 ±22 21 ±4 0.7 ±0.1 Bk 20–40 1750 568 359 73 224 ±55 47 ±4 5.5 ±1.1 2By 40–65 1670 LC 124 ±18 391 ±35 5.4 ±0.9 2BCy 65–95 1670 LC 104 ±33 447 ±24 4.6 ±1.0 2Cy >95 1670 LC 124 ±39 276 ±31 4.3 ±0.5 22 A 0–7 1750 602 319 79 197 ±30 28 ±6 3.2 ±0.3 Bw 7–45 1800 625 346 29 157 ±9 41 ±7 3.4 ±0.3 Byss 45–75 1800 645 323 32 182 ±37 71 ±11 2.6 ±0.4 BCy 75–115 1800 615 364 21 93 ±33 54 ±8 2.8 ±0.2 C >115 1850 615 372 13 155 ±39 21 ±7 0.5 ±0.1 64 A 0–10 1750 497 227 276 15 ±6 21 ±4 1.4 ±0.2 AB 10–20 1800 534 285 181 37 ±3 34 ±6 0.3 ±0.1 Bw 20–50 1850 615 164 221 50 ±24 36 ±8 1.0 ±0.3 Byss 50–80 1900 636 163 201 40 ±17 70 ±7 2.8 ±0.6 BCy 80–105 1880 574 136 290 50 ±27 84 ±10 1.8 ±0.2 C >105 1930 584 147 269 43 ±14 34 ±6 3.7 ±0.6 1 A 0–12 1750 546 224 230 167 ±24 – 1.3 ±0.2 Bw 12–50 1850 632 129 239 185 ±19 – 1.4 ±0.2 Bss 50–100 1890 612 154 234 171 ±16 – 4.9 ±1.0 BC 100–130 1870 411 255 334 139 ±26 18 ±3 1.4 ±0.3 C >130 1690 328 351 321 147 ±16 16 ±3 4.6 ±0.9 V.A. Laudicina et al. Geoderma 398 (2021) 115115 4 developed on Clays and/or marly clays and classified as Endosodic Gypsic Vertisol (Calcaric), with an A-Bw-Byss-BCy sequence; pedon P64, located on the back slope, developed on Sandy clays and classified as Endosodic Gypsic Vertisol (Calcaric, Hypereutric), with an A-AB-Bw- Byss-BCy sequence; and pedon P1, located at the foot of the slope, developed on Marly and/or sandy clays and classified as Vertic Cambisol (Protocalcic, Clayic), with an A-Bw-Bss-BC sequence. 2.2. Laboratory analyses Bulk density was determined according to Blake and Hartge (1986) from undisturbed samples taken using a steel cylinder with a volume of 100 cm 3 (5 cm in diameter and 5 cm in height). Total carbonates were determined gas-volumetrically using the Dietrich-Frühling calcimeter (Williams, 1948) and the inorganic C was calculated stoichiometrically. Gypsum (CaSO 4 ⋅2H 2 O) was determined by the thermogravimetric method (Arteida et al., 2006). The total amount of Ca 2+ and Sr 2+ , and the 87/86 Sr in soil carbonates and silicates were determined as suggested by Capo et al. (1998) and hereafter reported. Briefly, the soil samples (0.25 g) were leached by shaking and centrifuging them five times with 40 mL of MilliQ grade water to remove soluble salts and gypsum and then three times with 40 mL of ammonium chloride ultrapure buffered to pH 8 to remove exchange cations. Leached samples were reacted with ultra-pure 1 M acetic acid in an ultrasonic bath for 30 min (De Paolo et al., 1983) to selectively dissolve carbonates without a substantial dissolution of silicate minerals (Asahara et al., 1995; Quade et al., 1995). Acetic acid solutions were evaporated to dryness and the precipitates dissolved in 7 mL of 3 M HNO 3 . The acetic acid solution was used to determine Ca 2+ , Sr 2+ , and 87/86 Sr ratios of soil carbonates (Capo et al., 1998). Carbonate-free soil samples were digested in closed Teflon containers with a mixture of ultrapure concentrated HF (1.5 mL) and HNO 3 (5 mL) to determine the amount of Ca 2+ , Sr 2+ , and 87/86 Sr ratios of silicates (Capo et al., 1998). Digested samples were evaporated to dryness on a heating plate and the precipitates dissolved in 7 mL of 3 M Table 2 Field description of pedons located along the slope (after Laudicina et al., 2013). Pedon Horizon Depth cm Boundary a Moist color Structure b Roots c Stages of gypsum morphology d Stages of carbonate morphology e P3 (◆) Calcic Vertisol (Hypereutric, Pellic) A 0–12 C 2.5Y 3/ 0 gr, f, 1 f, m, 3 – I, dispersed in soil matrix ABkss 12–50 G 2.5Y 3/ 0 abk, m, c, 2 m, 2 – II+, many soft nodules Bkss1 50–90 G 2.5Y 4/ 0 pr, c, 2 f, m, 1 I, common snowballs and filaments II+, many soft nodules Bkss2 90–150 – 2.5Y 4/ 0 pr, c, 2 – I, common snowballs and filaments II+, many soft nodules P4 (▴) Gypsic Calcic Chernozem (Clayic) A 0–7 C 5Y 3/2 sbk, f, m, 1 f, m, 3 – I, dispersed in soil matrix AB 7–20 C 5Y 4/2 abk, m, 2 f, m, 3 I, common snowballs and filaments II, common pseudomycelia and powdery pockets Bk 20–40 C 5Y 5/2 abk, m, 2 f, m, 2 I, many snowballs and filaments III, abundant soft nodules and carbonate coats 2By 40–65 G 5Y 6/4 pr, m, c, 2 m, 1 II, III, many snowballs and nodules II+, many soft nodules 2BCy 65–95 D 5Y 6/4 pr, ma, c, 2 m, 1 II, III, many snowballs and nodules II, common pseudomycelia and powdery pockets 2Cy >95 – 5Y 7/3 pr, ma, c, 2 m, 1 II, III, many snowballs and nodules II, common pseudomycelia and powdery pockets P22 (▾) Endosodic Gypsic Vertisol (Calcaric) A 0–7 C 5Y 4/3 sbk, f, m, 1 f, m, c, 3 I, common snowballs and filaments – Bw 7–45 C 5Y 5/4 abk, f, m, 2 m, c, 2 I, common snowballs and filaments I+, common pseudomycelia Byss 45–75 C 5Y 5/4 pr, m, 2 m, 2 I, II, abundant snowballs, filaments, some nodules I+, common pseudomycelia BCy 75–115 G 5Y 6/4 pr, m, c, 2 – I, abundant snowballs and filaments – C >115 – 5Y 6/4 ma – I, common snowballs and filaments – P64 (■) Endosodic Gypsic Vertisol (Calcaric, Hypereutric) A 0–10 C 5Y 4/3 sbk, f, m, 1 f, m, 3 I, common snowballs and filaments – AB 10–20 C 5Y 5/4 sbk, f, m, 1 f, m, 3 I, common snowballs and filaments – Bw 20–50 G 5Y 5/4 pr, m, c, 2 m, 2 I, abundant snowballs and filaments II, common pseudomycelia and powdery pockets Byss 50–80 G 5Y 6/4 pr, m, c, 2 m, 2 II, many snowballs, nodules II, common pseudomycelia and powdery pockets BCy 80–105 D 5Y 6/4 ma – II, many snowballs, nodules – C >105 – 5Y 6/4 ma – I, abundant snowballs and filaments – P1 (●) Vertic Cambisol (Calcaric, Eutric, Clayic) A 0–12 C 5Y 4/3 gr, sbk, f, 2 f, m, 3 – – Bw 12–50 C 5Y 4/3 abk, m, 2 f, m, 1 – – Bss 50–100 G 5Y 5/3 pr, c, 2 m, 2 – II, common pseudomycelia and powdery pockets BC 100–130 D 5Y 5/3 pr, c, 2 m, 2 I, common snowballs and filaments I+, common pseudomycelia C >130 – 5Y 5/4 pr, c, 2 m, 2 I, common snowballs and filaments I+, common pseudomycelia a Boundary: C =clear, G =gradual, D =diffuse. b Structure: gr =granular, pr =prismatic, abk =angular blocky, sbk =subangular blocky, ma =massive; 1 = moderate, 2 =strong; f =fine, me =medium, c =coarse. c Roots: 1 =few, 2 =common, 3 =many; f =fine, m =medium, c =coarse. dStages of gypsum morphology according to Buck and Van Hoesen, 2002; e Stages of carbonate morphology according to Gile et al., 1981. V.A. Laudicina et al. Geoderma 398 (2021) 115115 5 HNO 3 . In addition, samples of gypsum-rock outcrops (sampled at the top of the slope) were dissolved in Milli-Q water and then re-precipitated with acetone ultrapure. Finally, the acetone was evaporated and samples were re-dissolved with 7 mL of 3 M HNO 3 . All of the 3 M HNO 3 solutions were passed through cation exchange columns packed with 2–3-butyl-cyclohexane-18-heterocyclic-6 (DtBuCH18C6) in 1-octanol absorbed on an inert support (Amberchrom CG-71 ms) and supplied as resin “Sr.Spec” (size particles 50–100 μ m) in polyethylene cartridges (5 cm ×5 mm) ready for use (Eichrom Industries, France). This procedure was carried out to separate Sr 2+ from Rb + and other cations to reduce inductively coupled plasma-mass spectrometry (ICP-MS) interferences (Chassery et al., 1998). To perform a chromatographic separation, once the resin was activated with 5 mL of 3 M HNO 3 , 5 mL of the HNO 3 solution containing Sr 2+ to be separated was first injected into the cartridge containing the activated resin. Then, two washings of 5 mL +5 mL of the resin with 3 M HNO 3 were carried out and, finally, the strontium was extracted by passing 0.05 M of the HNO 3 solution through the resin. The flow of solutions passing through the resin was kept constant at a rate of 1 mL per minute. It was not necessary to recover 100% of the Sr 2+ since the goal was to evaluate the ratio between the two isotopes of the same element rather than the total quantity of each isotope. The total amounts of Ca 2+ and Sr 2+ were determined on aliquots of the 3 M HNO 3 solution through atomic absorption spectroscopy (AAS) and ICP-MS (X7 SERIES ICPMS, Thermo Elemental, UK), respectively. The 87/86 Sr isotopic ratios were measured after passing the solutions through the resin by the same ICP-MS equipment but using a different set of measuring parameters (peristaltic pump, plasma coil power, and main argon, auxiliary, and coolant fluxes) specially tuned to maximise the sensitivity and mass resolution for the elements in the mass region of 87/86 Sr (Barros et al., 2005, pp. 181). At the same time, the possible presence of 85 Rb (natural isotopic abundance =72.17%) was also investigated, as it is the most abundant of the rubidium isotopes and, if absent, excludes the presence of the 87 Rb isotope (natural isotopic abundance =27.83%), a possible interference of 87 Sr. To use Sr isotopes as Ca 2+ tracers, the 87/86 Sr and Sr 2+ /Ca 2+ ratios of possible Ca 2+ sources must be known (Milli` ere et al., 2019). Knowing such variables, the amount of Ca coming from a source A and making its contribution in a Ca “mix” in a two-component system is calculated as follows (Capo et al., 1998): X A =[( 87/86 Sr mix - 87/86 Sr B ) ×R B ] / {[( 87/86 Sr mix - 87/86 Sr B ) ×R B ] +[( 87/86 Sr A - 87/86 Sr mix ) ×R A ]} (2) where A, B, and mix indicate the values of the two sources (gypsum and silicates) and their mix (carbonates), whereas R is the Sr 2+ /Ca 2+ ratio. The 87/86 Sr and Sr 2+ /Ca 2+ ratios of gypsum were those determined on gypsum rock outcrop, of silicates on the HF-HNO 3 solution, and of mix on acetic acid solution as previously described. It is important to mention that the 87/86 Sr and Sr 2+ /Ca 2+ ratios determined on the HFHNO 3 and acetic acid solutions include both autochthonous and, eventually, allochthonous silicates and carbonates (i.e., aeolian dust) potentially acting as a source of Ca 2+ . By knowing the amount of Ca 2+ coming from the gypsum and pedogenic carbonates (Laudicina et al., 2013), the amount of atmogenic inorganic C (AIC) was determined. The total Ca 2+ in Eucalyptus camaldulensis leaves and the litter fall collected within a radius of 10 m around each soil profile was determined by AAS after a HCl-HNO 3 digestion of 0.1 g of pulverised leaves previously dried at 70 ◦C for 48 h (Benton and Vernon, 1990). The Ca 2+ in the saturated soil paste extract was also determined by AAS. The saturated paste extracts were prepared according to Rhoades (1982). Bartlett’s test for the homogeneity of variances was used to compare the Ca 2+ in the saturated paste extract and in the plants (Dytham, 2003; p. 192). The K sp (theoretic solubility product) for the different minerals was obtained from their predicted Gibbs free energies ΔGf◦according to Vieillard (2000). The Log Q/K ratio was calculated to characterise the real equilibrium condition (Territo et al., 2006). The kinetic and thermodynamic geochemical model KINDISP (KINetics of DISsolution and Precipitation) was used to describe the thermodynamic equilibrium between minerals and aqueous solutions. We also calculated the ionic activities in soil solutions using the EQUIL Code as described by Mad´ e et al. (1994). We were thus able to obtain, from their predicted Gibbs free energies ΔGf◦, the theoretic solubility product for the different minerals. We then calculated the individual log Q/K ratio to estimate the real equilibrium condition. 3. Results Total carbonates ranged from 15 to 224 g kg −1 (Table 1). P64 showed the lowest content of carbonates, whereas P4 and 22 had the highest contents. Total carbonates occurred everywhere (Table 2), as dispersed in the soil matrix (carbonates at stage I), as filaments and powdery pockets with diameters of 1–2 mm (carbonates at stage II), and as filaments and powdery pockets with diameters of 1–2 mm (carbonates at stage III). They did not show any relation to depth among the profiles. Gypsum occurred in all soil profiles (Table 2) in the form of snowballs, filaments (stage I), or nodules (stage II), and, in general, increased with soil depth (Table 1). P3, P4, and P1 did not have gypsum in their topsoil. P4, P22, and P64 showed a gypsic horizon where gypsum was accumulated up to stage II. The theoretic solubility products, calculated to describe the real equilibrium condition, indicate that calcites and smectites reached thermodynamic equilibrium with a slight undersaturation of gypsum and the illitic phase in all soil profiles (Fig. 2). The Ca 2+ in the saturated paste extract ranged from 0.3 to 8.6 mol kg −1 and increased with soil depth, except for P22, in which it decreased (Table 1). P3 and P4 showed values of Ca 2+ in the saturated paste extract that were higher than that of P22, P64, and P1. The concentration of Ca 2+ determined in the 1 M acetic acid soil leachate (i.e., coming from the selective dissolution of carbonates) ranged from 2 to 59 g kg −1 (Table 3), and was significantly correlated to total carbonates (R 2 = Fig. 2. Theoretical solubility product for gypsum (empty circles, dotted lines) and calcite (closed circles, continuous lines) in the five benchmark pedons (P3, P4, P22, P64, P1) along the slope. V.A. Laudicina et al. Geoderma 398 (2021) 115115 6 0.697; P <0.001; Fig. 3), thus confirming the effectiveness of the procedure to selectively dissolve carbonates within the soil. The Sr 2+ in the same leachate ranged from 1 to 22 mg kg −1 (Table 3). By sorting carbonates by their origin (i.e., primary and secondary [Laudicina et al., 2013]), significantly different Sr 2+ /Ca 2+ ratios in the acetic acid soil leachate were found. In primary carbonates, the ratio was 0.333, whereas in secondary carbonates it was 0.192 (Fig. 4), which means that different sources of Ca contributed to both carbonate types. The 87/86 Sr ratio ranged from 0.690 to 0.712 in the acid acetic fraction and from 0.701 to 0.714 in the hydrofluoric fraction (Table 3). The 87/86 Sr isotope ratio determined in the acetic acid fraction behaved differently for each soil profile (Fig. 5). In P3 and P64, it was mostly constant along the profile, showing a preponderance of a particular Ca source. In P4 and P22, the isotope ratio decreased with depth, highlighting the importance of the superficial and sub-superficial sources of Ca 2+ for the formation of carbonates. In P1, no relation with depth occurred, possibly due to the variations in the relative contributions that arrived at the lower part of the slope at different times in recent history. The 87/86 Sr ratio determined by the gypsum outcropping at the top of the slope was 0.692 ±0.002 (Fig. 5). The percentage of Ca 2+ from gypsum, calculated according to Equation 2 for each soil horizon that contributed to the formation of atmogenic carbonates, ranged from 0% to 63% (Fig. 6). The highest percentage of Ca 2+ from gypsum was found in P3 (48% on average of all P3 horizons), whereas the lowest percentage was in P64 (10% on average of all P64 horizons). The content of Ca 2+ that was determined in the fresh leaves and litter fall did not show significant differences among the sampling areas: it ranged from 8.5 to 17.4 g Ca kg −1 in the fresh leaves and from 8.4 to 16.6 g Ca kg −1 in the litter fall (Table 4). The total Ca 2+ accumulated in the forest floor ranged from 50 to 140 kg Ca ha −1 . The AIC ranged from 0.2 to 11.1 g kg −1 (Table 5). In P3, the AIC was distributed along the entire profile down to 150 cm, and was 7.6 g kg −1 on average. In P4, the AIC was 65 cm at its lowest and 3.1 g kg −1 on average. P22 had AIC only in the topsoil (0.9 g kg −1 ), whereas P1 had it Table 3 Calcium, strontium, and 87/86 Sr ratio determined on acetic (AA) and hydrofluoric acid (HF) soil extracts for the quantification of Ca 2+ sources in soil carbonates formation. Pedon numbers follow their position along the slope from the top to the bottom. Pedon Horizon Depth Ca 2+ Sr 2+87/86 Sr Ca 2+ Sr 2+87/86 Sr cm g kg −1 mg kg −1 mg kg −1 mg kg −1 AA AA AA HF HF HF 3 A 0–12 22.0 ±2 4.9 ±2 0.698 ±0.001 13.0 ±0.8 0.69 ±0.16 0.705 ±0.002 ABkss 12–50 22.2 ±1 4.6 ±0 0.697 ±0.001 11.2 ±0.4 0.75 ±0.05 0.706 ±0.001 Bkss1 50–90 28.9 ±3 5.7 ±1 0.697 ±0.001 14.1 ±1.3 0.76 ±0.11 0.714 ±0.002 Bkss2 90–150 28.1 ±3 5.2 ±1 0.697 ±0.001 12.9 ±1.2 0.79 ±0.11 0.703 ±0.001 4 A 0–7 31.8 ±3 7.3 ±3 0.702 ±0.001 16.7 ±1.8 0.80 ±0.18 0.702 ±0.001 AB 7–20 33.1 ±3 7.7 ±2 0.701 ±0.001 16.6 ±1.4 0.76 ±0.10 0.702 ±0.001 Bk 20–40 35.5 ±4 7.7 ±2 0.699 ±0.001 18.9 ±1.5 0.83 ±0.17 0.703 ±0.002 2By 40–65 – – 0.698 ±0.002 14.8 ±3.1 1.29 ±0.15 0.701 ±0.001 2BCy 65–95 – – 0.697 ±0.001 – – – 2Cy >95 – – 0.692 ±0.003 – – – 22 A 0–7 57.4 ±9 21.5 ±3 0.703 ±0.002 20.7 ±3.7 0.82 ±0.08 0.712 ±0.001 Bw 7–45 37.2 ±7 12.7 ±2 0.692 ±0.001 – – – Byss 45–75 35.0 ±5 14.1 ±2 0.690 ±0.003 – – – BCy 75–115 21.1 ±2 8.3 ±1 0.708 ±0.001 12.4 ±0.8 0.88 ±0.11 0.711 ±0.001 C >115 30.0 ±3 11.6 ±1 0.690 ±0.002 – – – 64 A 0–10 2.1 ±0.2 1.1 ±0.2 0.709 ±0.001 7.6 ±0.5 0.90 ±0.13 0.712 ±0.001 AB 10–20 11.1 ±1.1 2.2 ±0.1 0.708 ±0.001 7.5 ±0.2 0.92 ±0.12 0.717 ±0.002 Bw 20–50 6.3 ±1 2.4 ±0.2 0.709 ±0.001 7.0 ±0.5 0.85 ±0.07 0.717 ±0.001 Byss 50–80 5.3 ±0.6 2.2 ±0.1 0.709 ±0.002 6.9 ±0.3 0.80 ±0.06 0.711 ±0.001 BCy 80–105 8.5 ±0.4 2.4 ±0.3 0.712 ±0.001 8.8 ±0.4 0.79 ±0.05 0.712 ±0.001 C >105 4.6 ±0.3 2.4 ±0.3 0.712 ±0.002 12.2 ±0.6 0.75 ±0.05 0.713 ±0.001 1 A 0–12 53.0 ±7 18.7 ±2 0.700 ±0.001 – – – Bw 12–50 58.6 ±8 21.3 ±3 0.702 ±0.001 – – – Bss 50–100 32.4 ±5 12.2 ±1 0.691 ±0.002 – – – BC 100–130 27.3 ±3 9.8 ±1 0.699 ±0.001 15.1 ±1.6 0.68 ±0.03 0.701 ±0.001 C >130 27.4 ±2 9.6 ±1 0.696 ±0.001 13.0 ±0.8 0.73 ±0.02 0.703 ±0.001 Fig. 3. Correlation between Ca 2+ coming from the selective dissolution of soil carbonates by a 1 M acetic acid fraction (De Paolo et al., 1983) and total soil carbonates. Fig. 4. Correlation between Ca 2+ and Sr 2+ in soil carbonates selectively dissolved by a 1 M acetic acid solution (De Paolo et al., 1983). Different coloured points indicate the Sr 2+ /Ca 2+ in primary (blue squares) and secondary (red circles) soil carbonates as assessed by δ 13 C isotopes analyses and observed during field observation (Laudicina et al., 2013). V.A. Laudicina et al. Geoderma 398 (2021) 115115 7 only in the subsoil (on average 1.85 g kg −1 ). P64 showed the lowest AIC (on average 0.6 g kg −1 ). Four C pools were sorted: total organic C (TOC), total inorganic C (TIC), pedogenic inorganic C (PIC), and atmogenic inorganic C (AIC). The amount of TOC stored in the soils (0–100 cm depth) ranged from 17 to 42 kg (Table S1) and decreased according to the following sequence: P3 >P4 >P1 >P64 >P22. Total inorganic C was of the same order of magnitude of TOC and ranged from 11 to 39 kg m −3 (Table S1). Along the soil profiles, the TIC was always lower than the TOC in the topsoils and higher in the subsoils. The PIC ranged from 3 to 31 kg m −3 (Table S1); it was higher in soils developed on non-calcareous parent material (P3 and P4; centre and edge of the doline) than in those developed on the calcareous Terravecchia formation (clay/marly/sandy clays: P1, P22 and P64). P3 had the highest content of AIC (from 4.4 to 11.1 g kg −1 ), followed by P4 (from 1.6 to 4.3 g kg −1 ) and then by P1 (from 0.6 to 3.1 g kg −1 ); P22 and P64 had the lowest content of AIC (<1 g kg −1 ). In the first cubic meter of soil, AIC was absent in P1, 0.1 kg m −3 in P22, 1 kg m −3 in P64, 4 kg m −3 in P4, and 16 kg m −3 in P3. Fig. 5. Patterns of the 87/86 Sr isotope ratios determined in soil carbonates and selectively dissolved by 1 M acetic acid solution (De Paolo et al., 1983) in the five benchmark pedons (P3, P4, P22, P64, P1) along the slope. Fig. 6. Percentage of Ca 2+ coming from gypsum and contributing to the formation of atmogenic inorganic carbonates in the five benchmark pedons (P3, P4, P22, P64, P1) along the slope. Table 4 Total calcium in fresh leaves and in litterfall determined in the different pedons along the slope. Pedon numbers follow their position along the slope from the top to the bottom. Pedon Fresh leaves Litterfall g Ca kg −1 g Ca kg −1 3 12.8 ±2.2 10.5 ±1.1 4 8.5 ±1.4 8.4 ±0.7 22 11.5 ±2.1 9.7 ±1.5 64 15.3 ±1.9 16.8 ±1.7 1 17.4 ±2.2 14.6 ±1.3 Table 5 Content of the different forms of C in the soil profiles. TOC, total organic C; TIC, total inorganic C; PIC pedogenic inorganic C; AIC, atmogenic inorganic C (LIC can be computed by the difference between TIC and PIC). Pedon numbers follow their position along the slope from the top to the bottom. Pedon Horizon Depth TOC TIC PIC AIC cm g kg −1 g kg −1 g kg −1 g kg −1 3 A 0–12 43.8 ± 3.2 13.5 ± 1.9 12.2 ± 1.2 4.4 ±2.1 ABkss 12–50 21.9 ± 3.0 14.9 ± 2.3 13.5 ± 1.7 7.3 ±0.7 Bkss1 50–90 19.1 ± 2.4 19.4 ± 2.4 17.6 ± 1.6 11.1 ± 1.5 Bkss2 90–150 16.0 ± 2.8 21.4 ± 2.0 19.6 ± 2.1 7.7 ±1.8 4 A 0–7 35.5 ± 3.4 19.4 ± 3.7 16.4 ± 3.1 1.6 ±0.5 AB 7–20 29.1 ± 4.1 22.5 ± 2.6 20.6 ± 1.7 2.2 ±0.8 Bk 20–40 25.9 ± 3.0 26.9 ± 6.6 23.8 ± 6.0 4.2 ±1.0 2By 40–65 6.4 ±2.0 14.9 ± 2.1 13.3 ± 1.3 4.3 ±1.4 2BCy 65–95 5.9 ±1.4 12.5 ± 3.9 n.d. n.d. 2Cy >95 2.1 ±1.1 14.8 ± 4.7 n.d. n.d. 22 A 0–7 33.5 ± 4.1 23.7 ± 3.5 3.5 ±0.4 0.9 ±0.3 Bw 7–45 12.5 ± 3.2 18.9 ± 1.1 n.d. n.d. Byss 45–75 5.9 ±1.5 21.9 ± 4.4 n.d. n.d. BCy 75–115 2.9 ±1.1 11.2 ± 3.9 0.4 ±0.1 n.d. C >115 1.2 ±0.5 18.6 ± 4.6 n.d. n.d. 64 A 0–10 39.3 ± 3.3 1.8 ±0.7 1.6 ±0.6 0.2 ±0.1 AB 10–20 17.7 ± 2.1 4.5 ±0.3 4.3 ±2.4 0.9 ±0.4 Bw 20–50 13.5 ± 1.1 6.0 ±2.9 5.0 ±2.1 0.7 ±0.2 Byss 50–80 10.3 ± 3.1 4.8 ±2.1 4.5 ±2.1 0.5 ±0.1 BCy 80–105 5.6 ±1.4 6.0 ±3.2 2.4 ±0.3 n.d. C >105 4.3 ±1.3 5.2 ±1.6 0.8 ±0.4 n.d. 1 A 0–12 43.2 ± 2.9 20.0 ± 2.9 n.d. n.d. Bw 12–50 12.5 ± 2.3 22.2 ± 2.3 n.d. n.d. Bss 50–100 12.5 ± 1.8 20.5 ± 1.9 n.d. n.d. BC 100–130 10.5 ± 2.1 16.7 ± 3.1 5.0 ±0.6 0.6 ±0.2 C >130 8.4 ±1.6 17.6 ± 1.9 6.6 ±0.6 3.1 ±0.3 V.A. Laudicina et al. Geoderma 398 (2021) 115115 8 4. Discussion The presence of carbonates in the topsoils of the investigated soil pedons suggests a limited translocation of carbonates from eluvial to illuvial horizons, which is likely due to the low annual precipitation. In addition, some deposition of carbonates in soil surfaces associated with dust inputs from southern areas is another possibility that cannot be excluded (Guerzoni and Molinaroli, 2005). Moreover, a prevalent transport and a subsequent deposition of silicates with respect to carbonates is highlighted by Castorina and Masi (2015), Gobbi et al. (2019), Varrica et al. (2019), and Rodríguez et al. (2020). However, considering that our aim was to investigate the role of gypsum in the calcium and carbon cycles, we only compared five soils that should receive the same amount of dust regardless of the external input. From a mineralogical point of view, carbonates are mainly represented by calcite that reaches thermodynamic equilibrium along the entire soil profiles, whereas a slight undersaturation for gypsum markedly occurs in the first 60 cm of the soil profiles, as previously reported by Territo et al. (2006). 4.1. Ca 2+ sources for carbonate formation The highest values of Ca 2+ in the saturated paste extract of P3 and P4 suggest that the relief and distance from gypsum outcrops (150 m and 100 m, respectively) were the driving factors affecting its distribution. Indeed, because gypsum is a soluble salt (2.4 g L -1 ), the Ca 2+ -gypsum that comes from the dissolution by rainwater is then addressed towards soils according to their position along a slope (Birkeland, 1999, p. 261). This determined that values of Ca 2+ in the saturated paste extract from soils situated in the doline and nearest the gypsum outcrops (P3 and P4) were higher than those in the other soils (P22, P64, and P1; 250 m, 900 m, and 1500 m, respectively). Moreover, this Ca 2+ redistribution, along with the SO 4 2- , contributed to the formation of lenticular secondary gypsum in the soils developed on non-gypsiferous parent materials (Dazzi and Monteleone, 2002; Poch et al., 2010). The 87/86 Sr ratio values suggest that gypsum, in addition to silicates, provides Ca 2+ for the formation of soil carbonates. Similar studies have reported silicates (Naiman et al., 2000; Van der Hoven and Quade, 2002), rainfall (Hamidi et al., 2001), oceans and volcanic ash (Quade et al., 1995), and limestone and aeolian dust (Naiman et al., 2000; Rodríguez et al., 2020) as Ca 2+ sources, but not gypsum. Furthermore, these studies have not investigated all of the soil profile horizons. We have found that the amount of Ca 2+ from gypsum depends on the hydrological setting of the slope, as well as on the distance of the soils from the gypsum source. This in turn determines the extent of water infiltration and hence the depth reached by Ca 2+ from gypsum. In the P3 and P4 profiles, the amount of Ca 2+ from gypsum was the highest (P3 > P4) compared to the other profiles, probably due to their proximity to gypsum outcrops. Furthermore, the amount of Ca 2+ coming from gypsum increased with depth, which is probably due to the relief that favoured the flow of rainwater rich in Ca 2+ towards the centre of the doline and leached the Ca 2+ in depth. On the contrary, in P22, which is situated on the shoulder of the slope, water infiltration is limited due to the hydrological setting, so Ca 2+ from gypsum decreases in the subsoil. In P64, which is located in the backslope, the amount of Ca 2+ from gypsum was lower than P3 and P4 but higher than P22, and it also decreased with soil depth, which suggests a small amount of water infiltration along the profile. The presence of Ca 2+ in only the deepest horizons of P1 could be attributed to its position along the slope (foot slope) that allows it to receive a greater amount of water as compared to the other soils, thereby moving Ca 2+ towards the deepest horizons. This factor, coupled with P1′s greatest distance from gypsum outcrops, explains why Ca 2+ from gypsum was found only in the two deepest P1 horizons. Such Ca 2+ patterns from gypsum within and among soil profiles confirmed what was theoretically interpreted by Birkeland (1999): that local relief controls most of the distribution of soil in the landscape, and the slopes of the semiarid Mediterranean environment are characterised by a lateral translocation of the more soluble compounds such as gypsum. 4.2. Ca 2+ in the soil–plant system As emphasised by Milli` ere et al. (2019), calcium in soil undergoes several interactions related to its mineral phases, the soil solution, and the vegetation. Originating from the soil solution, the elements excreted by leaves and suction move into the roots and then flow through the plant. Thus, vegetation plays a key role in spatially homogenising Ca 2+ in soils. Along the slope, the Ca 2+ content of the plant materials tends to increase towards the bottom of the slope and in the pedon located in the centre of the doline. Given the variability of total carbonates and gypsum, a general spatial redistribution of Ca 2+ was evident where the main soil minerals potentially providing Ca 2+ were greater in quantity at the top part of the slope, with an overall mean of 12 ±3 g Ca 2+ kg −1 litter. Litter accumulation on the soil surface accounted for 13 Mg ha −1 , which corresponded to 150 kg Ca 2+ ha −1 (Table 4). These results are in agreement with a study by Turpault et al. (2019), which measured a Ca 2+ uptake up to 200 kg ha −1 within a stand, with Ca 2+ returns to the soil from 20 kg ha −1 (litter) to 100 kg ha −1 (fine root decomposition). The forest floor below Eucalyptus camaldulensis accumulated at rates around 13 Mg ha −1 yr −1 , where Ca 2+ ranged from 50 kg ha −1 to 140 kg ha −1 . By comparing Ca 2+ within soil solution verses Ca 2+ in plants (averaged by pedon) through a Bartlett’s test (K-squared =8.8624, df = 2, p-value =0.0119), it was determined that Ca 2+ contained in the soil varies three-dimensionally across the whole space of the slope, but trees return the Ca 2+ contained to the soil in the volume explored by the rhizosphere, homogenising it. Indeed, according to Scalenghe et al. (2015), the tree cover of the slope is homogeneous by species and age. The rhizosphere, given the planting scheme, occupies the entire volume of the soil explored by the roots, and it is interconnected. In general, the release of Ca 2+ from the soil is controlled by the pCO 2 of the soil, pH, kinetics at the mineral surface, and mass transport (Inskeep and Bloom, 1985; Morse et al., 2007). However, Ca 2+ is highly recycled within the system, most of which is annually absorbed by plants returned to the soil (Pritchett and Fisher, 1987; pp. 162-164). 4.3. Carbon pools The TOC and TIC pools were of the same order of magnitude. The TIC pool ranged from 30% to 66% of the total soil C pool. This falls within the range reported for similar environments by Eswaran et al. (2000) and You et al. (2020). The large variability of TIC within and among soil profiles on the slope can be ascribed mainly to the different lithological substrates, the soil age, and the active carbonation process that lead to the accumulation of secondary soil carbonates with different stages of carbonate morphology (Laudicina et al., 2013; Monger et al., 2015). The amount of AIC pool was always lower than that of the organic C pool. This suggests that the latter pool is the most important in storing CO 2 in the semiarid gypsiferous Mediterranean environment. However, considering that the turnover rate of the inorganic C (>85,000 years; Schlesinger, 1997) is far greater than that of the organic C (<1,000 years and <200 years in hot biomes; Luo et al., 2019), it is clear that a small amount of atmogenic inorganic C plays a key role in storing CO 2 in soils (Minasny et al., 2017). The distribution of AIC along the soil profiles aligned with that of Ca 2+ from gypsum, thereby suggesting that an external source of Ca 2+ is crucial for the net accumulation of new C within the soil (Monger et al., 2015). 4.4. The Sr 2+ /Ca 2+ ratio as a potential indicator of secondary carbonates Although some of the investigated pedons did not develop on calcareous or gypsiferous parent materials, P3 and P4 show a calcic horizon (Bk). Secondary carbonate formation, which is necessary to V.A. Laudicina et al. Geoderma 398 (2021) 115115 9 individuate the calcic horizon, occurred following both the dissolution and re-precipitation of primary soil carbonates, as well as new precipitation of carbonates from the combination of Ca 2+ from gypsum and biogenic CO 2 . According to both the World Reference Base for Soil Resources (IUSS WG WRB, 2015) and the Soil Taxonomy (Soil Survey Staff, 2014), a calcic horizon has to be ≥15-cm thick and shows accumulation of secondary carbonates. Both of these soil classification systems do not consider inputs from aeolian dust in the calcic horizon definition. During field observation, it is difficult to differentiate between primary and secondary carbonates, especially when they are dispersed in the soil matrix and concretions or nodules are not evident. To this purpose, a very useful technique is one based on the δ 13 C isotopic signatures (Laudicina et al., 2013; Monger et al., 2015). Unfortunately, such techniques are not fully available in many soil laboratories and are also time-consuming and expensive. By performing a selective dissolution of soil carbonates with 1 M acetic acid solution and by sorting the soil horizons according to the prevalence of primary and secondary carbonates (cf. Laudicina et al., 2013), different Sr 2+ /Ca 2+ ratios were found. The highest values of such ratios occurred in soil carbonates of primary origin, whereas the lowest ones were in those of secondary origin (Fig. 4). Such a finding is reasonable since Sr 2+ and Ca 2+ have a similar ionic radius (255 pm and 231 pm, respectively), therefore a substitution of the former with the latter may occur. For this reason, due to the higher abundance of Ca 2+ compared to that of Sr 2+ , the Sr/Ca ratio decreases during the dissolution and re-precipitation or precipitation of carbonates in the soil. Therefore, the Sr 2+ /Ca 2+ ratio can be proposed as an alternative and less expensive approach to distinguish pedogenic carbonates than the one based on δ 13 C isotopic signatures. Thus, an accurate distinction between primary and secondary soil carbonates could be achieved by the determination of the Sr 2+ to Ca 2+ ratio in the acetic acid soil leachate. 5. Conclusion In a semiarid Mediterranean gypsiferous environment, the availability of soluble Ca 2+ from gypsum and the relief that influences surface and sub-surface water flows are the main drivers of inorganic C accumulation in soils; therefore, an external source of Ca 2+ is crucial to accumulate new C within the soil. In this context, vegetation contributed to spatially homogenise Ca 2+ in soils. The soil organic C pool, more than the atmogenic inorganic C pool, acts as a C sink. However, atmogenic inorganic C may account for a significant fraction of the total inorganic C in soils, and even a small amount may play a key role in storing CO 2 in soil due to its longer turnover time. Finally, the expeditious Sr 2+ /Ca 2+ ratio is proposed as a new approach to identify primary and secondary carbonates in soils for diagnostic purposes. Declaration of Competing Interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. Appendix A. Supplementary data Supplementary data to this article can be found online at https://doi. org/10.1016/j.geoderma.2021.115115. References Amundson, R., 2001. The carbon budget in soils. Annu. Rev. Earth Planet. Sci. 29, 535–562. https://doi.org/10.1146/annurev.earth.29.1.535. Arteida, O., Herrero, J., Drohan, P.J., 2006. Refinement of the differential water loss method for gypsum determination in soils. Soil Sci. Soc. Am. J. 70, 1932–1935. Asahara, Y., Tanaka, T., Kamioka, H., Nishimura, A., 1995. Asian continental nature of 87 Sr/ 86 Sr ratios in north central Pacif sediments. Earth Planet. Sci. Lett. 133, 105–116. Badagliacca, G., Romeo, M., Presti, E.L., Gelsomino, A., Monti, M., 2020. Factors governing total and permanganate oxidizable C pools in agricultural soils from Southern Italy. Agriculture 10 (4), 99. Barros, H., Abril, J.M., Laudicina, V.A., Delgado, A., 2005. Application of ICP-MS for measuring soil metal cations from sequential extraction. In: M´ endez-Vilas, A. (Ed.), Proceedings of the first International Meeting on Applied Physics (APHYS-2003) on Recent advances in multidisciplinary applied physics, pp. 179-183. 13-18 October, 2003. Badajoz, Spain. pp. Birkeland, P.W., 1999. Soils and Geomorphology, 3rd edition. Oxford University Press, New York, p. 430p. Blake, G.R., Hartge, K.H., 1986. Bulk Density, In: Klute, A. (Ed.), Methods of Soil Analysis, Part I. Physical and Mineralogical Methods: Agronomy Monograph no. 9, 2nd ed. , pp. 363–375. Buck, B.J., Van Hoesen, G.J., 2002. Snowball morphology and SEM analysis of pedogenic gypsum, southern New Mexico, U.S.A. J. Arid Environ. 51, 469–487. Bughio, M.A., Wang, P.P., Meng, F., Chen, Q., Li, J., Shaikh, T.A., 2017. Neoformation of pedogenic carbonate and conservation of lithogenic carbonate by farming practices and their contribution to carbon sequestration in soil. J. Plant Nutr. Soil Sci. 180, 454–463. Camilli, B., Dell’Abate, M.T., Mocali, S., Fabiani, A., Dazzi, C., 2016. Evolution of organic carbon pools and microbial diversity in hyperarid anthropogenic soils. J. Arid Environ. 124, 318–331. https://doi.org/10.1016/j.jaridenv.2015.09.003. Capo, R.C., Chadwick, O.A., 1999. Sources of strontium and calcium in desert soil and calcrete. Earth Planet. Sci. Lett. 170, 61–72. https://doi.org/10.1016/S0012-821X (99)00090-4. Capo, R.C., Stewart, B.W., Chadwick, O.A., 1998. Strontium isotopes as tracers of ecosystem processes: theory and methods. Geoderma 82, 197–225. Carmi, I., Kronfeld, J., Moinester, M., 2019. Sequestration of atmospheric carbon dioxide as inorganic carbon in the unsaturated zone under semi-arid forests. Catena 173, 93–98. https://doi.org/10.1016/j.catena.2018.09.042. Castorina, F., Masi, U., 2015. Sr and Nd isotopes as tracers in pedogenic studies: Evidence for Saharan dust contribution to the soils of Muravera (Sardinia, Italy). Geochemistry 75, 301–315. Chassery, S., Grousset, F.E., Lavaux, G., Qu´ etel, C.R., 1998. 87 Sr/ 86 Sr measurements on marine sediments by inductively coupled plasma-mass spectrometry. Fresenius 360, 230–234. Dazzi, C., Monteleone, S., 2002. Soils and soil landform relationships along an elevational transect in a gypsiferous hilly area in central Sicily, Italy. In: Zdruli, P., et al. (Ed.), 7th International meeting on Soils with Mediterranean type of climate, pp. 73–86. CIHEAM-IAMB, Bari, Italy, (accessed on 2020, April 29) http://om. ciheam.org/om/pdf/a50/04002021.pdf. De Paolo, D.J., Kyte, F.T., Marshall, B.D., O’Neil, J.R., Smit, J., 1983. Rb-Sr, Sm-Nd, K- Ca, O and H isotopic study of Cretaceous-Tertiary boundary sediments, Caravaca, Spain: evidence for an oceanic impact site. Earth Planetary Science Letters 64, 356–373. Drees, L., Wilding, L., Nordt, L., 2001. Reconstruction of soil inorganic and organic carbon sequestration across broad geoclimatic regions. In: Lal, R. (Ed.), Soil Carbon Sequestration and the Greenhouse Effect. Soil Science Society of America Special Publication, pp. 155–172. Dytham, C., 2003. Choosing and using statistics: a biologist’s guide. Blackwell, UK, p. 248. Eswaran, H., Reich, P.F., Kimble, J.M., Beinroth, F.H., Padmanabhan, E., Moncharoen, P., 2000. Global carbon stocks. In: Lal, R., Kimble, J.M., Eswaran, H., Stewart, B.A. (Eds.), Global climate change and pedogenic carbonate. Lewis Publishers, Boca Raton, Florida. Faure, G., Powell, J.L., 1972. Strontium Isotope Geology. Springer, Berlin. Fierotti, G., Dazzi, C., Raimondi, S., 1988. A report on the soil map of Sicily. University of Palermo, Istituto di Agronomia. Gile, L.H., Hawley, J.W., Grossman, R.B., 1981. Soils and geomorphology in the Basin and Range area of southern New Mexico—Guidebook to the Desert Project. New Mexico Bureau of Mines and Mineral Resources, Socorro, New Mexico. Gobbi, G.P., 2019. An inclusive view of Saharan dust advections to Italy and the Central Mediterranean. Atmos. Environ. 201, 242–256. Grasso, M., Pedley, H.M., 1988. The sedimentology and development of Terravecchia Formation carbonates (Upper Miocene) of North Central Sicily: Possible eustatic influence on facies development. Sed. Geol. 57, 131–149. Guerzoni, S., Molinaroli, E., 2005. Input of various chemicals transported by Saharan dust and depositing at the sea surface in the Mediterranean Sea. In: Saliot, A. (Ed.), The Handbook of Environmental Chemistry. Springer, Berlin, pp. 237–268. https:// doi.org/10.1007/b107149. Hamidi, E.M., Colin, F., Michard, B., Boulang` e, B., Nahon, D., et al., 2001. Isotopic tracers of the origin of Ca in carbonate crust from the Middle Atlas, Morocco. Chem. Geol. 176, 93–104. Inskeep, W.S., Bloom, P.R., 1985. An evaluation of rate equations for calcite precipitation at p less than 0.01 atm and pH greater than 8. Geochim. Cosmochim. Acta 49, 2165–2180. IUSS Working Group WRB, 2015. World Reference Base for Soil Resources 2014, update 2015. International soil classification system for naming soils and creating legends for soil maps. World Soil Resources Reports No. 106. FAO, Rome. Lal, R., 2004. Soil carbon sequestration to mitigate climate change. Geoderma 123, 1–22. Lal, R., 2009. Sequestering carbon in soils of arid ecosystems. Land Degrad. Dev. 20, 441–454. https://doi.org/10.1002/ldr.934. Lapenis, A.G., Lawrence, G.B., Bailey, S.W., Aparin, B.F., Shiklomanov, A.I., Speranskaya, N.A., Torn, M.S., Calef, M., 2008. Climatically driven loss of calcium in steppe soil as a sink for atmospheric carbon. Global Biogeochem. Cycles 22, GB2010. https://doi.org/10.1029/2007GB003077. V.A. Laudicina et al.