The chemical composition of river waters gives a measure of the atmospheric CO2 fixed by chemical weathering processes. Since the dominating factors controlling these processes are lithology and runoff, as well as uplift and erosion, we introduce a new simplified geo-lithological map of the Alps (Alpine-Geo-LiM) that adopted a lithological classification compliant with the methods most used in literature for estimating the consumption of atmospheric CO2 by chemical weathering. The map was used together with published alkalinity data of the 33 main Alpine rivers (1) to investigate the relationship between bicarbonate concentration in the sampled waters and the lithologies of the corresponding drained basins, and (2) to quantify the atmospheric CO2 consumed by chemical weathering. The analyses confirm (as known by the literature) that carbonates are lithologies highly prone to consuming atmospheric CO2. Moreover, the analyses show that sandstone (which could have a nonnegligible carbonate component) plays an important role in consuming atmospheric CO2. Another result is that in multilithological basins containing lithologies more prone to consuming atmospheric CO2, the contribution of igneous rocks to the atmospheric CO2 consumption is negligible. Alpine-Geo-LiM has several novel features when compared with published global lithological maps. One novel feature is due to the attention paid in discriminating metamorphic rocks, which were classified according to the chemistry of protoliths. The second novel feature is that the procedure used for the definition of the map was made available on the Web to allow the replicability and reproducibility of the product.Carbon is the fourth most abundant element in the universe (Morgan and Anders, 1980; Anders and Ebihara, 1982), and it plays a vital role in Earth’s environment. This element migrates continuously among four sinks: oceans, atmosphere, ecosystems, and geosphere (Holland, 1978; Berner, 2003; Kump et al., 2009). Considering the time scale of the phenomena, the “short-term” carbon cycle (shorter than 1 m.y.) is distinguished from the “long-term” carbon cycle (longer than 1 m.y.). The 1 m.y. threshold is assumed in literature to be coherent with the residence time of Ca2+ in the ocean system (Donnini et al., 2016). In the “short-term” carbon cycle, carbon is rapidly exchanged within surficial systems, such as oceans, biosphere, soil, and atmosphere, where the anthropogenic CO2 production is also taken into account. In the “long-term” carbon cycle, carbon is slowly exchanged between the geosphere and the ocean-atmosphere system. Here, the concentration of atmospheric CO2 mainly derives from the balance between the CO2 produced by both volcanism and metamorphism, and the atmospheric CO2 consumed by weathering of silicates and carbonates (Berner et al., 1983; Berner, 1991, 1994, 2004, 2006; Berner and Kothavala, 2001; Gislason and Oelkers, 2011; Li and Elderfield, 2013).Because the solutes produced by chemical weathering enrich the river dissolved load, the composition of river waters can be considered as a good indicator of chemical weathering processes (Mackenzie and Garrels, 1966; Garrels and Mackenzie, 1971; Meybeck, 1987; Tardy, 1986; Probst, 1992; Gaillardet et al., 1999; Viers et al., 2007; Berner and Berner, 2012). Starting with knowledge of both the chemical compositions and flow rates of river waters, as well as of the lithologies of their basins, two different methods can be used to calculate the atmospheric CO2 consumed by chemical weathering (Hartmann, 2009; Hartmann et al., 2009): (1) the reverse and the (2) forward methods. Both methods assume that the only reactions occurring within the river basins are the alteration of silicates and the alteration of carbonates due to the presence of carbonic acid.The reverse method uses mass balance equations to discriminate weathering products by considering specific lithological end members (Garrels and Mackenzie, 1967; Meybeck, 1987; Gaillardet et al., 1999). Consequently, the stoichiometric relationships between the cations dissolved in the fluvial waters give, with good approximation, an estimate of the moles of atmospheric CO2 involved in the alteration processes (Probst et al., 1994; Amiotte-Suchet, 1995; Amiotte-Suchet and Probst, 1996; Boeglin and Probst, 1998; Mortatti and Probst, 2003; Donnini et al., 2016). Depending on the time scales, different reactions have to be considered in order to quantify the atmospheric CO2 consumed by chemical weathering (Huh, 2010; Donnini et al., 2016).The forward method assumes that lithology and runoff (i.e., the discharge per unit area) are the predominant controlling factors of bicarbonate concentration in river waters, which is a measure of the atmospheric CO2 consumed by chemical weathering. For specific lithologies, the runoff is linked to the atmospheric CO2 consumed by chemical weathering through empirical relationships. In this way, it is possible to quantify the atmospheric CO2 consumed by chemical weathering (Bluth and Kump, 1994; Amiotte-Suchet and Probst, 1993a, 1993b, 1995; Probst et al., 1994; Amiotte-Suchet et al., 2003; Hartmann, 2009; Hartmann et al., 2009).A good understanding of the nature of the rocks is fundamental for building the empirical relationships between CO2 consumption and lithology. As highlighted by Amiotte-Suchet et al. (2003) and by Moosdorf et al. (2010), geological maps often give scarce information regarding the chemical and physical nature of the rocks, focusing on the age of rocks, their deformation, their stratigraphy, and their structural position. This lack of information is problematic, especially for sedimentary rocks, which are very abundant in orogens (Doglioni, 1994; Einsele et al., 1996; Clift et al., 2001) and which have a highly variable chemical composition (Amiotte-Suchet et al., 2003). Moreover, it is often not simple to obtain information about the protoliths of metamorphic rocks.In the literature, a few lithological maps have been published at the global scale and are illustrated in the following. Gibbs and Kump (1994) presented a 2° × 2° global lithological map classified into the six following rock types: (1) carbonates, (2) shales, (3) sandstones, (4) extrusive igneous rocks, (5) shield areas (including both intrusive igneous rocks and metamorphic rocks), and (6) “complicated lithology” (where it was difficult to discern a single rock type within the 2° × 2° grid cell). That lithological map was used together with a derived 7.5° × 4.5° global runoff map to calculate the global riverine bicarbonate flux by using the relationships between runoff and bicarbonate flux from Bluth and Kump (1994).Amiotte-Suchet and Probst (1995) elaborated a 1° × 1° global map of CO2 consumption (Global Erosion Model for CO2 fluxes [GEM-CO2]) starting from the simplified lithological and soil maps published by the Food and Agriculture Organization (FAO) and United Nations Educational, Scientific and Cultural Organization (UNESCO) (FAO-UNESCO, 1971, 1975, 1976, 1978, 1979, 1981) and exploiting the relationships estimated by Meybeck (1986) considering more than 200 French monolithological basins (Amiotte-Suchet and Probst, 1993a, 1993b). Amiotte-Suchet and Probst (1995) defined the total atmospheric/soil CO2 flux consumed by rock weathering, ϕ(CO2)short, as the CO2 moles consumed per area unit in a given period of time. In the map, the following seven lithologies were considered: (1) plutonic and metamorphic rocks, (2) sand and sandstone, (3) acid volcanic rocks, (4) evaporite rocks, (5) basalts, (6) shales, and (7) carbonate rocks.Subsequently, Amiotte-Suchet et al. (2003) elaborated a 1° × 1° global lithological map considering six rock categories: (1) sands and sandstone, (2) shales, (3) carbonate rocks, (4) combined intrusive igneous rocks and metamorphic rocks (i.e., shield rocks), (5) acid volcanic rocks, and (6) basalts. Compared with the map presented by Gibbs and Kump (1994), the map of Amiotte-Suchet et al. (2003) has a greater resolution (1° × 1° vs. 2° × 2°), and it is more informative, since ∼27% of the total exposures are “complicated lithology” in the map of Gibbs and Kump (1994), and, as such, they are not precisely characterized (Amiotte-Suchet et al., 2003). Similar to Amiotte-Suchet and Probst (1995), ϕ(CO2)short was estimated by Amiotte-Suchet at al. (2003) through the relationship between ϕ(CO2)short and runoff published by Meybeck (1986).A more detailed global lithological map was published by Dürr et al. (2005) at 1:25,000,000 scale. In contrast to the maps published by Gibbs and Kump (1994) and by Amiotte-Suchet et al. (2003), which are two grid-based raster maps, the map of Dürr et al. (2005) is in vector format and includes 8300 polygons. The map considers 15 rock categories (excluding water and ice): (1) acid volcanic rocks, (2) basic volcanic rocks, (3) acid plutonic rocks, (4) basic plutonic rocks, (5) Precambrian basement, (6) metamorphic rocks, (7) consolidated siliciclastic rocks, (8) mixed sedimentary rocks, (9) carbonates, (10) semi- to unconsolidated sedimentary rocks, (11) alluvial deposits, (12) loess, (13) dunes, (14) evaporites, and (15) complex lithology (where sediments, volcanic, and metamorphic rocks are mixed together). Together with outcropping lithologies, the map contains three other thematic layers containing other geological information (major subsurface evaporite occurrences, geology, and limits of maximum Quaternary glaciation extent).Another global lithological map (named GLiM), in vector format, was presented by Hartmann and Moosdorf (2012). The map includes 1,235,400 polygons at 1:1,000,000 scale. Following Moosdorf et al. (2010), the map contains three levels of information (layers). The first one is mandatory and represents the general lithology. It considers 15 lithologies (excluding water and ice): (1) evaporites, (2) metamorphics, (3) acid plutonic rocks, (4) basic plutonic rocks, (5) intermediate plutonic rocks, (6) pyroclastics, (7) carbonate sedimentary rocks, (8) mixed sedimentary rocks, (9) siliciclastic sedimentary rocks, (10) unconsolidated sediments, (11) acid volcanic rocks, (12) basic volcanic rocks, (13) intermediate volcanic rocks, (14) Precambrian rocks, and (15) complex lithologies. The second and the third layers optionally contain information on the specific rock attributes.At regional scale, Donnini et al. (2016) presented a lithological map of the Alps. The map was used, together with the major-element concentrations of the 33 main Alpine river waters, to estimate the atmospheric CO2 consumption by chemical weathering in the Alpine region by applying the MEGA geochemical code (Amiotte-Suchet, 1995; Amiotte-Suchet and Probst, 1996), which implements the reverse method. This map was elaborated at 1:1,000,000 scale and considers eight lithological classes: (1) acid igneous rocks, (2) mixed carbonate, (3) clay and claystone, (4) debris, (5) mafic rocks, (6) metamorphic rocks, (7) pure carbonate rocks, and (8) sandstone.In this paper, we introduce a new high-resolution (1:1,000,000 scale) simplified geo-lithological map of the Alps (named Alpine-Geo-LiM) that adopted a lithological classification (10 lithological classes: (1) “pure carbonate,” (2) “mixed carbonate,” (3) “gypsum evaporite,” (4) “acid rocks,” (5) “mafic rocks,” (6) “intermediate rocks,” (7) “sandstone,” (8) “claystone,” (9) “metamorphic rocks,” and (10) “peats”), compliant with the reverse and the forward methods. Alpine-Geo-LiM was derived from the national geological maps of Italy, France, Germany, Switzerland, Austria, and Slovenia, and it represents an implementation of the map previously published in Donnini et al. (2016). Moreover, it is released together with the code adopted for building the map (Donnini et al., 2018). Although we used the same input data as in Donnini et al. (2016), Alpine-Geo-LiM differs from the map published in Donnini et al. (2016) in the lithological classification, i.e., eight lithological classes of Donnini et al. (2016) versus 10 lithological classes for Alpine-Geo-LiM, as well as in a more accurate analysis of the protoliths of metamorphic rocks. Moreover, unlike Donnini et al. (2016), Alpine-Geo-LiM is released in vector format together with both the informatic procedures used to elaborate the map and the original data (see Donnini et al., 2018). We define Alpine-Geo-LiM as a geo-lithological map since we provide the lithological map, but we also provide the original layers and procedure used to create the map. Moreover, we release, in the attribute table, the original geological information (Appendixes A, B, and C1).Alpine-Geo-LiM, together with the alkalinity of the 33 main Alpine rivers sampled in 2011 and 2012 (Donnini et al., 2016), was used: (1) to investigate the relationship between HCO3− concentration in the sampled river waters and the lithologies of the corresponding drainage basins, and, applying the forward method, (2) to quantify the atmospheric CO2 consumed by chemical weathering.The Alps (south-central Europe; Fig. 1) are a collisional belt generated by the Cretaceous to present convergence of the European and African (also named Adriatic or Apulian) continental margins, which caused the closure of the ocean located in the Mediterranean region (Trümpy, 1960; Frisch, 1979; Tricart, 1984; Haas et al., 1995; Stampfli et al., 2001; Dal Piaz et al., 2003; Schmid et al., 2004; Pfiffner, 2014).The Alps have an arc shape and can be roughly subdivided into the following different geological domains (Dal Piaz et al., 2003; Schmid et al., 2004; Pfiffner, 2014) shown in Figure 1: the eastern Alps, the Northern Calcareous Alps, the southeastern Eoalpine Calcareous Alps, and the western Alps. The Alps are partially continuous to the northwest with the Apennine chain and to the east with the Dinarides. The Pannonian basin bounds the Alps to the east, the Molasse Basin bounds the Alps to the north, and the Po Valley and Adriatic foreland bound the chain to the south. The Jura Mountains define the northwestern boundary of Alps. External to the Alps, in the north, there is the European foreland. The polygon in Figure 1 represents the study area, corresponding to the subdivision of the Alps into the 33 main Alpine river basins used by Donnini et al. (2016).The geology of the Alps can be roughly schematized using the following geological domains (Rossi and Donnini, 2018): (1) Austroalpine crystalline rocks in the eastern Alps; (2) carbonate rocks in the Jura Mountains, in the Northern Calcareous Alps, and in the southeastern Eoalpine Calcareous Alps; and (3) Helvetic calcareous units mixed with crystalline massifs and Penninic metamorphic-ophiolitic units in the western Alps. Outside of the Alpine chain, (1) the Molasse basin in the north is filled by Tertiary successions having several kilometers of thickness, and (2) the Po Valley and Adriatic foreland in the south mainly consist of alluvial deposits, as do the Pannonian basin and the European foreland.From a geomorphologic point of view, the Alps are characterized by altitudes ranging between 1200 and 1300 m above sea level (m.a.s.l.), extensive lowlands, deeply incised valleys, and mountains higher than 4000 m.a.s.l. (the highest peak is Monte Bianco at 4888 m.a.s.l.; Dal Piaz et al., 2003), leading to a strong topographic variability (Carraro and Giardino, 2004; Gobiet et al., 2014).Temperature extremes and annual precipitation are related to the physiography of the Alps. The valley bottoms are generally warmer and drier than the surrounding mountains. In winter, nearly all precipitation above 1500 m.a.s.l. is in the form of snow. Snow cover lasts from approximately mid-November to the end of May at 2000 m.a.s.l (Diem et al., 2019). The precipitation stored as snow and ice in the winter season is released in the following months after their fusion (European Environmental Agency, 2010). The water in the Alpine region is in the form of lakes, aquifers, and glaciers, which feed many basins in Europe, including the Rhine, Danube, Po, and Rhone (Weingartner et al., 2007), which are the biggest European rivers in terms of flow rate and basin area. Glaciers cover an area of ∼2050 km2 (Paul et al., 2011), representing 1% of the area of the 33 main Alpine basins (Donnini et al., 2016).The following sections introduce the reader to (1) the basic equations governing atmospheric CO2 consumption and (2) the new Alpine-Geo-LiM.The chemical composition of river waters is an indicator of weathering processes (Mackenzie and Garrels, 1966; Garrels and Mackenzie, 1971; Meybeck, 1987; Tardy, 1986; Probst, 1992; Gaillardet et al., 1999; Viers et al., 2007; Berner and Berner, 2012), which contribute, together with atmospheric input (rain), pollution, biota, and evaporite dissolution, to the dissolved load (e.g., Gaillardet et al., 1999; Galy and France-Lanord, 1999; Roy et al., 1999; Moon et al., 2007; Wu et al., 2008; Gao et al., 2009; Jha et al., 2009; Donnini et al., 2016). Weathering reactions of silicate minerals, hydrolysis, and carbonate dissolution consume atmospheric/soil CO2 and produce an increase in the solution alkalinity. The forward reactions of silicate and carbonate alteration by carbonic acid are the only reactions that occur within the river basins in the “short-term” and are described by the following equations (Mortatti and Probst, 2003; Donnini et al., 2016):In the “long-term” period, not all the atmospheric CO2 is permanently removed by weathering reactions because some of the carbon is returned to the atmosphere (Huh, 2010; Donnini et al., 2016). In particular, the rivers’ dissolved load released to the oceans is partially precipitated as carbonate (mainly reverse reaction in Eq. 5) or authigenic clays (reverse reactions in Eq. 1 and Eq. 2). In particular, Equation 5 shows that weathering of CaCO3 consumes one unit of CO2 (forward reaction, “short-term”), and that the same amount of CO2 is returned to the atmosphere upon precipitation of CaCO3 in the seas and/or oceans (reverse reaction, “long-term”).A similar behavior might be due to the weathering of CaMg(CO3)2 (see Eq. 6), where two units of CO2 consumed by weathering (forward reaction, “short-term”) return to the atmosphere upon precipitation of CaMg(CO3)2 in the seas and/or oceans (reverse reaction, “long-term”). However, the direct precipitation of dolomite at ambient temperature from aqueous solution is prevented (e.g., Frondini et al., 2014, and references therein) by a strong solvation shell of aqueous Mg2+ as well as by crystallization barriers that inhibit the formation of Ca-Mg ordered dolomite (Montes-Hernandez et al., 2016). Therefore, the reverse reaction of Equation 6 is just theoretically considered.Weathering reactions of silicate minerals containing Na and K (albite and K-feldspar) consume atmospheric CO2 (Eq. 1 and Eq. 2 forward reactions, “short-term”). When Na+ and K+ ions are transported by rivers to the oceans and/or seas, they could be subjected to reverse weathering, forming authigenic clays and releasing atmospheric CO2 (Eq. 1 and Eq. 2 reverse reactions, “long-term”; Huh, 2010).Equation 3 and Equation 4 are not reversible, and the products of the forward reactions (HCO3−, Ca2+, and Mg2+) are involved in the reverse reactions described by Equation 5. In particular, half of the units of atmospheric CO2 consumed during weathering of Ca-plagioclase (Eq. 3) are precipitated in the oceans as CaCO3 (according to Eq. 5, reverse reaction). The same might also occur for the product of the forward reaction of olivine (Eq. 4), when considering the reverse reaction of Equation 6. However, the last speculation is just theoretically given, as previously explained. The assumption that Equations 1–6 are the only reactions that occur in the river basins is valid (a) if carbonic acid (H2CO3, derived from the interaction between river water and atmospheric CO2) is the only source of protons in weathering reactions (the contribution of other acids [HNO3, H2SO4] is negligible; Mortatti and Probst, 2003; Donnini et al., 2016), and (b) if pyrite (FeS2), gypsum (CaSO4•2H2O), and halite (NaCl) percentages are negligible in the river basins, and their dissolution is not considered in the model calculation (Perrin et al., 2008). Overall, the described conditions are valid in nonpolluted areas, for temperate climates, and for lithologies without pyrite.As previously stated, two different methods were used to calculate the atmospheric CO2 consumed by chemical weathering (Hartmann, 2009; Hartmann et al., 2009): (1) the reverse and the (2) forward method. In the following, a brief description of both approaches is reported.In the reverse method, the moles of atmospheric/soil CO2 consumed by chemical weathering are computed from Equations 1–6, considering the forward (“short-term”) and the reverse (“long-term”) reactions. In practice, the measured dissolved cations in river waters are used to estimate the moles of consumed atmospheric/soil CO2.In Equation 7 and Equation 8, ϕ(X + Y) is the sum of the fluxes of two generic chemical species X and Y in river waters, given by its molar concentration multiplied by the runoff, while the suffixes “sil” and “carb” indicate the considered chemical species derived from either silicate or carbonate weathering (Huh, 2010; Donnini et al., 2016). Starting from the measured concentration of Ca2+ and Mg2+ in river waters, the contributions of silicate weathering, (Ca + Mg)sil, and carbonate dissolution, (Ca + Mg)carb, to the total riverine fluxes can be distinguished by using specific ionic ratios of water drained from monolithological basins (e.g., Meybeck, 1986, 1987).The reverse method has been already applied by many authors to estimate the atmospheric CO2 consumption in the Congo, Amazon, and Niger watersheds and in the 33 main Alpine river basins (Probst et al., 1994; Amiotte-Suchet, 1995; Amiotte-Suchet and Probst, 1996; Boeglin and Probst, 1998; Gaillardet et al., 1999; Mortatti and Probst, 2003; Moon et al., 2007; Donnini et al., 2016).In the forward method, the moles of atmospheric/soil CO2 consumed by chemical weathering are computed from Equations 1–6 considering only the “short-term” forward reactions.Similar to Equation 7 and Equation 8, in Equation 9, ϕ(HCO3) is the flux of moles of HCO3−, given by its molar concentration multiplied by the runoff. The suffixes “sil” and “carb” indicate whether the considered chemical species (HCO3−) derives from silicate or carbonate weathering.In the literature, a set of empirical relationships links, for different lithologies, the flux of atmospheric/soil CO2 consumed by chemical weathering on the “short-term,” ϕ(CO2)short, to the runoff. Amiotte-Suchet and Probst (1993a, 1993b, 1995), Probst et al. (1994), and Amiotte-Suchet et al. (2003) estimated the relationship between ϕ(CO2)short and runoff from the dissolved load and the runoff of more than 200 French monolithological river basins (Meybeck, 1986, 1987). Similar relationships were estimated by Bluth and Kump (1994) from the dissolved load and the runoff of ∼100 monolithological catchments across the United States, Puerto Rico, and Iceland. In Hartmann (2009) and Hartmann et al. (2009), for the first time, the relationship between ϕ(CO2)short and runoff was estimated through a multivariate nonlinear regression analysis starting from 382 Japanese river basins draining more than one lithology.The forward method considers that lithology and runoff are the dominating factors controlling the atmospheric CO2 consumption processes, and that other factors, such as relief or land cover, are less important at both regional and global scale (Hartmann, 2009; Hartmann et al., 2009). A temperature dependence of the atmospheric CO2 consumption is implemented only for the global basalt-weathering model (Dessert et al., 2003).Starting from the empirical relationships between ϕ(CO2)short and runoff and knowing only the outcropping lithology and the runoff within a given territory, it is possible to estimate the moles of atmospheric CO2 consumption. Data on the chemical composition of river water are not needed for this estimation. The forward model has been applied at basin scale in the Garonne, Congo, and Amazon River basins (Amiotte-Suchet and Probst, 1993a, 1993b, 1995; Probst et al., 1994), at regional scale in the Japanese Archipelago (Hartmann, 2009), and at a global scale (Amiotte-Suchet and Probst, 1995; Amiotte-Suchet et al., 2003; Hartmann et al., 2009).For the elaboration of our geographic information system (GIS)–based simplified geo-lithological map (1:1,000,000 scale) of the Alps, we took advantage of the geological layers, in vector format, extracted from (1) the geological map of Italy at 1:500,000 scale (Bonomo et al., 2006) released by the Italian Institute for Environmental Protection and Research (ISPRA; http://www.isprambiente.gov.it), (2) the geological map of Switzerland at 1:500,000 scale (Bundesamt für Landestopografie, 2005) released by the Swiss Federal Office of Topography (Swisstopo; http://www.swisstopo.admin.ch), (3) the geological map of Germany at 1:1,000,000 scale (BGR, 2011), (4) the geological map of Austria at 1:500,000 scale (Egger et al., 1999) released by the Geological Survey of Austria (GBA; http://www.geologie.ac.at), (5) the geological map of France at 1:1,000,000 scale (BRGM, 2003), and (6) the geological map of Slovenia at 1:250,000 scale (Buser, 2010). These two last maps were obtained from the European Geological Data Infrastructure (EGDI; http://www.europe-geology.eu/metadata). The six maps are released in ESRI shapefile formats having different coordinate reference systems and different accuracy and information quality. The layers of France, Germany, and Slovenia contained several topological errors (e.g., gaps between polygon borders, overlapping polygon borders, etc.) and were corrected by removing duplicate boundaries and areas smaller than, respectively, 1 m2, 600 m2, and 50 m2 (the longest boundary with adjacent area was removed).The attribute tables of the vector maps contain different attribute fields where the description of the geological information is stored. Those fields are listed in the Appendix B (see footnote 1).According to Moosdorf et al. (2010, p. 2), “classification is a constant compromise between exactness and simplicity.” The lithological classification used in Alpine-Geo-LiM is a compromise among the 6–7 rock categories used by Gibbs and Kump (1994), Amiotte-Suchet and Probst (1995), and Amiotte-Suchet at al. (2003), and the 15 rock categories used by Dürr et al. (2005), Hartmann and Moosdorf (2012), and Moosdorf et al. (2010), since we consider the first classification too simplified and the second one too detailed. Ten lithologies were taken into account for Alpine-Geo-LiM: (1) “pure carbonate,” (2) “mixed carbonate,” (3) “gypsum evaporite,” (4) “acid rocks,” (5) “mafic rocks,” (6) “intermediate rocks,” (7) “sandstone,” (8) “claystone,” (9) “metamorphic rocks,” and (10) “peats.”The “pure carbonate” category includes rocks composed mainly of calcite, aragonite (CaCO3), and dolomite [MgCa(CO3)2], such as limestone, dolomite, and travertine, as well as marble, for which the protolith is composed by carbonate rock (Pettijohn, 1957; Garrels and Mackenzie, 1971; Boggs and Boggs, 2009).The “mixed carbonate” category includes rocks composed of carbonate minerals mixed with noncarbonate minerals. In this category, there are impure carbonate rocks, calcarenites, and marls (Pettijohn, 1957).In the “gypsum evaporite” category, we include gypsum and anhydrite. We know that, generally, the term evaporites refers to anhydrite, gypsum, and halite (Garrels and Mackenzie, 1971). However, since the analyzed bibliographic sources (see Appendixes A and C, and references therein [footnote 1]) excluded the presence of halite in the Alps, in this work only gypsum and anhydrite were included in the “gypsum evaporite” group.The subdivision among “acid rocks,” “mafic rocks,” and “intermediate rocks” was done according to (1) the total-alkali-silica (TAS) diagram (Le Bas et al., 1986; Middlemost, 1994), which classifies many common types of volcanic rocks starting from the relationships between the combined alkali (Na2O + K2O) and silica (SiO2) contents, and (2) an adaptation of the same diagram for plutonic rocks (Le Bas et al., 1986; Middlemost, 1994). Thus, we considered “mafic rocks” those with less than 50%–52% of SiO2, “intermediate rocks” the rocks with SiO2 content between 50%–52% and 60%–62%, and “acid rocks” those with more than 60%–62% of SiO2. The metamorphic rocks, which are not included in the two TAS diagrams, were classified according to Mottana et al. (2009). Based on the compositions of the protoliths, an orthogneiss was considered as “acid rocks” (assuming a granitic protolith composition), and a serpentinite was considered as “mafic rocks” (Mottana et al., 2009).In the “sandstone” category, we included arkose, graywacke (Garrels and Mackenzie, 1971), and conglomerate, the last one being similar to sandstone in terms of origin and depositional mechanisms (Boggs and Boggs, 2009). Moreover, the metamorphic rock quartzite falls in the “sandstone” category, as its protolith (Mottana et al., 2009).In the “claystone” category, we included shale, argillite, siltstone, and mudstone (Garrels and Mackenzie, 1971; Boggs and Boggs, 2009), as well as, again considering the protoliths (Mottana et al., 2009), the metamorphic phyllite, schists, and paragneiss.The generic “metamorphic rocks” category was used only when information on protoliths was unavailable or unclear (e.g., in the case of migmatite, mylonite, and metasediments).Finally, we introduced the further lithology “peat,” due to the presence of these types of deposits in the Alps.Rocks composed by more than one lithotype posed some problems for their classification. Goldich (1938) introduced a weathering series of silicates that was further modified by Railsback (2006), who added some nonsilicate minerals. According to this weathering series, carbonate dissolution is considerably higher than silicate dissolution, whereas gypsum/anhydrite dissolution is higher than carbonate dissolution. For this reason, (1) in the “gypsum evaporite” category, we also included rocks composed of a mix of carbonate and gypsum/anhydrite, and (2) in the “mixed carbonate rocks” category, we also included the rocks composed of more than one lithotype, where at least one of them was composed of “mixed carbonate rocks” (e.g., a lithotype composed by sandstone, graywacke, and marl was considered “mixed carbonate rocks”).For the silicate rocks composed by more than one lithotype, we adopted the “principle of prevalence.” We classified these rock types according to the most abundant lithologies among those listed in the different fields of the reference map attribute tables. For example, an outcrop (a polygon of the vector map) where the different field attributes reported the presence of basalt, trachybasalt, and andesite was included in the “basic rocks” class, since, following our classification, basalt and trachybasalt can be considered as “basic rocks” and only andesite as is classified as “intermediate rocks.”In the rare occasions when an outcrop is composed by “intermediate rocks” and “acid rocks” (or “mafic rocks”) in the same proportions, it was considered as “acid rocks” (or “mafic rocks”). This is the case, as an example, of an outcrop composed by monzonite (“intermediate rocks”) and granite (“acid rocks”); it was considered as “acid rocks.” An in-depth study of alpine geology, described in the Appendix C (see footnote 1), was carried out to classify specific geological units.The new Alpine-Geo-LiM is a portion of the “Geo-Lithological Map of Central Europe” (Geo-LiM; Donnini et al., 2018), which was released in vector format, and which is freely downloadable at the Web address (https://doi.org/10.5281/zenodo.3530257, Donnini et al., 2018). The map is composed by 12,001 polygons. Some very small polygons exist in the map (due to the cut of the map along the boundaries of the studied area). The biggest polygon has an area of ∼11,197.5 km2, and the average polygons size is ∼16.5 km2.The preprocessing (cleaning of topological errors) and processing (unions, intersections, and classifications) steps to build the map were performed using GRASS GIS (Neteler and Mitasova, 2008; Neteler et al., 2012), an open-source GIS software, and PostgreSQL (PostgreSQL - http://www.postgresql.org), an open-source relational database management system (RDBMS), with its PostGIS spatial extension (PostGIS; http://www.postgis.org).The attribute table of the resulting map is composed by the following 10 fields: litho_irpi, rsil_mm, orig1, orig2, orig3, orig4, orig5, orig6, orig7, and country. The litho_irpi field was compiled with one of the 10 aforementioned lithological classes (“acid rocks,” “mafic rocks,” “intermediate rocks,” “metamorphic rocks,” “sandstone,” “claystone,” “pure carbonate rocks,” “mixed carbonate rocks,” “gypsum evaporite,” and “peat”). The orig1, orig2, orig3, orig4, orig5, orig6, and orig7 fields contain the original geological information derived from the six original geological maps (Italy, Switzerland, Germany, Austria, France, Slovenia; see Appendix B [footnote 1]), and the country field contains the name of the country.The command lines and queries used for building the map based on the original data are provided together with the geo-lithological map herein.Alpine-Geo-LiM is shown in Figure 2. The colors used to distinguish the different lithologies in Alpine-Geo-LiM were derived from the lithologic legend adopted by the U.S. Geological Survey (USGS) for the geologic maps of the United States. The legend and the red-green-blue (RGB) codes are made available by the USGS on the Web (https://mrdata.usgs.gov/catalog/lithclass-color.php).The abundance of rock types outcropping in the Alpine region is shown in Table 1, and it was estimated within an area of 197,773 km2, corresponding to the area of the main Alpine river basins (Fig. 3) defined in Donnini et al. (2016). The table shows that carbonate rocks are the most abundant type in Alpine region, with 23.75% of “mixed carbonate” and 20.82% of “pure carbonate,” for a total of 44.57%. They are followed by “sandstone” (26.99%), “claystone” (12.87%), and volcanic rocks (with 7.38% of “acid rocks,” 2.69% of “mafic rocks,” and 0.43% of “intermediate rocks,” for a total of 10.50%). “Metamorphic rocks” represent 1.81% of the study area, while “peats” and “gypsum evaporite” represent less than 1% of the study area (respectively 0.48% and 0.08%). A small area (2.69%) is covered by “water” in the form of lakes and glaciers. The data about the abundance of each outcropping rocks type (% Area in Table 1) in the Alpine region are quite similar to the percentages computed by Donnini et al. (2016), which, however, underestimated the percentage of claystone and did not fully differentiate the metamorphic rocks. Looking at the elevation and slope values reported in Table 1, and based on the 25-m-resolution European digital elevation model (EU-DEM; Bashfield and Keim, 2011; https://www.eea.europa.eu/data-and-maps/data/copernicus-land-monitoring-service-eu-dem), we find that “claystone,” “acid rocks,” “mafic rocks,” “metamorphic rocks,” and “intermediate rocks” have the highest mean elevation (from 1664 m to 1993 m) and slope values (from 23.95° to 28.32°) associated with relatively low standard deviation values. This is due to the fact that these lithologies compose the crystalline massifs that are the tallest mountains of the Alps (e.g., Monte Bianco, 4808 m.a.s.l.; Monte Rosa, 4634 m.a.s.l.; Dent Blanche, 4357 m.a.s.l.).High mean elevation values (1662 m) are associated also with “water” (lakes and glaciers), with a high standard deviation (1301 m); moreover, a medium slope value (12.55°), with a relatively high standard deviation (14.72°), is associated with “water.” This big variation is related to the fact that the “water” class includes lakes (which usually are located in the valley) and glaciers (which are located at high altitude).The mean elevation and the mean slope of “pure carbonate rocks” are equal to 1279 m and 22.55°, respectively. This is confirmed by the fact that the highest calcareous mountains of the Alps have quite high elevations (e.g., Ortles, at 3905 m.a.s.l.; Gran Zebrù, at 3857 m.a.s.l.; and Marmolada, at 3343 m.a.s.l., in the southeastern Eoalpine Calcareous Alps; Parseierspitze, at 3036 m.a.s.l., in the Northern Calcareous Alps; and Crêt de la Neige, at 1720 m.a.s.l. in the Jura Mountains). Table 1 shows that “mixed carbonate rocks,” mainly composed by impure carbonate rock, calcarenite, and marl, have mean elevation and slope values, respectively, of 1157 m and 18.38°.Elevation and slope of the rocks classified within the “sandstone” class are relatively low with respect to the other lithologies (respectively 698 m and 8.50°, with associated standard deviation of 519 m. and 10.27°, respectively). This is due to the fact that we included, in this lithology, conglomerates and uncemented sediments produced by the erosion of the massifs.“Gypsum evaporite” presents low mean elevation (603 m) and slope (11.50°) values, as does “peat” (555 m; 2.14°). This last observation is not surprising, since “peat” is formed in vegetated and flattened wetlands by the degradation of the vegetation (Bracco et al., 2004).With the aim to quantify the atmospheric CO2 consumed by chemical weathering using the forward method, the proportions of the outcropping rocks (Table 2) were estimated for each river basin. We considered all the 33 river basins (Fig. 3) defined in Donnini et al. (2016), including the four Rhine subbasins (Linth, Reuss, Alpine Rhine, and Aare).Glaciers and lakes pose some uncertainty in estimating atmospheric CO2 consumption by chemical weathering. Anderson et al. (1997) and Anderson (2005) showed that glaciers increase runoff within basins, but they do not enhance silicate weathering processes. Regarding lakes, Cole et al. (2007) and Tranvik et al. (2009) showed that freshwaters (lakes, rivers, and reservoirs) act both in transporting and in producing the atmospheric carbon (CO2 and CH4), confirming the atmospheric CO2 and CH4 production in lakes as highlighted by Huttunen et al. (2003), Del Sontro et al. (2010), Diem et al. (2012), and Pighini et al. (2018). In this work, we decided to simplify our approach by considering glaciers and lakes just like parts of the hydrographic network.In Alpine-Geo-LiM, the sediments of the alluvial valleys and terraces were included in the “sandstone” class. Here, conglomerates and uncemented sediments have been put in place by the erosion, transport, and deposition of the rocks constituting the upper part of the watersheds. As a consequence, one can assume that the lithology of those materials should reflect that of the upstream massifs. It becomes relevant to check if, in the 33 river basins, the sandstones are mainly associated with the presence of carbonate rocks or silicate rocks. To better investigate the association of the “sandstone” class with the other lithologies, we performed a cluster analysis using the well-known unsupervised k-means algorithm (Hartigan and Wong, 1979) implemented in the R software (R Core Team, 2016) and imposing four clusters. The percentages of the lithologies representing the centers of the four clusters are shown in Table 3.Cluster 1 shows a percentage of “sandstone” equal to 19.72% associated with: (1) 73.80% of carbonate rocks (“mixed carbonate” and “pure carbonate”), (2) 6.24% of “claystone” and igneous and metamorphic rocks (“acid rocks,” “mafic rocks,” “metamorphic rocks,” and “intermediate rocks”), and (3) 0.26% of other lithologies (“peat” and “gypsum evaporite”).Cluster 2 shows a similar percentage of “sandstone” (15.02%) associated with a relatively low percentage of carbonate rocks (28.66%), and a high percentage of “claystone” and igneous and metamorphic rocks (56.32%). Other lithologies (“peat” and “gypsum evaporite”) are negligible (0.01%).In cluster 3, the “sandstone” percentage is equal to 20.74%, carbonate rocks are 64.46%, “claystone” and igneous and metamorphic rocks are 14.75%, and other rocks are 0.06%.Cluster 4 shows relatively high concentration of “sandstone” (40.79%), followed by carbonate rocks (37.86%) and by “claystone” and igneous and metamorphic rocks (20.29%). As for the other clusters, other lithologies have a low percentage (1.06%).The analysis shows that the “sandstone” class is associated with a high percentage of “claystone” and igneous and metamorphic rocks only in cluster 2, while in the other clusters (including cluster 4 characterized by high “sandstone” concentration), “sandstone” is associated with a high percentage of carbonate rocks. Since the “sandstone” lithology is produced by the erosion of the massifs, we maintain that in the studied area, “sandstone” is mostly composed of carbonate rocks.Figure 4 shows the spatial distribution of the four clusters, highlighting (1) the presence of an inner core mainly composed of crystalline silicate rocks (mainly “claystone”; cluster 2), (2) a western and eastern (mainly) bound with carbonate rocks (cluster 1 and cluster 3), and (3) “sandstone” rocks (cluster 4) in the northern and southern sectors of the Alps (Molasse Basin and Po Valley) composed of sandstone rich in carbonates (see Appendix A [text footnote 1]).To estimate the atmospheric CO2 consumed by chemical weathering in the Alpine region, we applied a revised version of the forward method (Hartmann, 2009; Hartmann et al., 2009).For the purpose, we considered the subdivision of the Alpine region in 33 river basins (Fig. 3) from Donnini et al. (2016) and the alkalinity of river waters sampled near the basin outlets during spring and winter seasons in the 2011–2012 hydrological year (Donnini et al., 2016).The data on discharges were obtained from different sources, including international, national, and local authorities (see references in Table 4). For most catchments, the data were available as flow rates (m3/s); for the catchments where only stage measurements (m) at gauging stations were available, the stage measurements were converted to estimated discharge (m3/s) by using rating curves provided by river basin authorities or derived using empirical data.Table 4 shows that data on flow rate markedly varied within the two sampling campaigns, with mean PD[Q](w/s) of 3.06%, and a coefficient of variation equal to 23.79%. Moreover, Table 4 shows that 22 rivers (almost 70% of considered rivers) have the lowest flows in winter season and the highest in spring season (PD[Q](w/s) > 0), which is typical of rivers with glacial- and snowmelt-dominated regimes. A comparison of Q(s) and Q(w) with Q(my) highlights the fact that almost all the flow rate values, measured at the time of the two sampling campaign, were lower than the mean annual discharge estimated considering the daily discharge in one hydrological year. On the contrary, Table 4 shows that alkalinity values are less variant across the seasonal measurements with respect to the flow rates, having mean PD[HCO3](s/w) of 17.38% and a variation coefficient equal to 1.40%. More generally, it seems that there is not a correlation between the flow rate and the alkalinity. The weak correlation between flow rate and alkalinity is also shown in Figure 5, where the alkalinity (HCO3) of the river waters sampled in the two campaigns (in spring season and in winter season) is plotted with respect to the flow rate (Q) of the river waters registered at the moment of the two sampling campaigns.We performed the analysis considering the 10 lithological classes of Table 2, and, due to the scarce presence of some lithologies in the study area, also considering four general lithological classes defined according to the following schema: (1) “sandstone,” (2) “claystone,” (3) “total carbonate” (including “pure carbonate,” “mixed carbonate,” “gypsum evaporite,” and “peat”), and (4) “igneous and metamorphic rocks” (including “acid rocks,” “mafic rocks,” “intermediate rocks,” and “metamorphic rocks”).The coefficients resulting from the linear multiple regression analysis performed considering 10 lithological classes are shown in Table 5, where b represents the estimated coefficient for each lithology; Std Error measures the standard error in b estimation; the P value expresses the probability that the b value is equal to 0 by chance; and Significance level is a literal classification of the P value expressing the reliability of the analysis.High b values are correlated with relevant HCO3− concentrations and therefore identify lithologies more prone to consuming atmospheric CO2 by chemical weathering. Conversely, a low value of b is indicative of a low atmospheric CO2 consumption by chemical weathering from the corresponding lithology. Small P values indicate weak correlations between the predictor ([HCO3]) and response (SR) variables. High P values correspond to low significance (Significance level in Table 5) of the analysis.Figure 6 shows the alkalinity values of the Alpine rivers measured during the two sampling campaigns (Alkalinity observed) versus the alkalinity values of the same Alpine rivers predicted by applying Equation 11 (Alkalinity predicted). The coefficient of determination (R2) of the linear fit (passing through the zero) between measured and fitted values is close to 1 (0.95), while the median, the mean, and the standard deviation of the residuals are –3.742 × 10−5, –1.92 × 10−9, and 5.335 × 10−4 respectively, highlighting the capability of the model in reproducing the observed data.In Table 5, the lithologies are ordered from the highest to the lowest b value. The table shows very high chemical alterability and related CO2 consumption rates for “gypsum evaporites” and “peat” (high b values, respectively, equal to 2.35 × 10−2 mol L−1 and 2.2 × 10−2 mol L−1), with very low significance for “gypsum evaporites” (P value = 0.51) and high significance for “peat” (P value = 0.0017).High b values are shown for “sandstone” (b = 4.13 × 10−3 mol L−1), with very high significance (P value = 1.77 × 10−10), and “pure carbonate” and “mixed carbonate” (b values, respectively, equal to 2.62 × 10−3 mol L−1 and 2.32 × 10−3 mol L−1; P values = 2.03 × 10−7 and 1.96 × 10−4, respectively).Very low b values together with low significance are shown for “metamorphic rocks” (b = 2.28 × 10−3 mol L−1; P value = 0.48) and “claystone” (b = 1.44 × 10−3 mol L−1; P value = 0.12).Negative b values were obtained for “mafic rocks,” “acid rocks,” and “intermediate rocks,” associated with very low significance and P values.These results show a surprising behavior of “peat,” with a positive and highly significant value of the calibration parameter b. Since peat is composed by at least 30% of organic matter (Joosten and Clarke, 2002), its dissolution leads to the release of organic carbon into the atmosphere (e.g., Chow et al., 2003; Bengtsson and Törneman, 2004; Schwalm and Zeitz, 2015; Selvam et al., 2017); for this reason, we would have expected a negative value of b. This discrepancy could be explained by the fact that the “peat” presence is concentrated in the Isar basin (6.84% of the basin), where it is associated with “sandstone” (41.6%), “pure carbonate” (24.12%), and “mixed carbonate” (16.76%), i.e., lithologies particularly effective at CO2 consumption.The relative high values of the Std Error values (and low P values) obtained for “mafic rocks,” “acid rocks,” and “intermediate rocks” demonstrate that the values of the b coefficients are not statistically different from 0, and therefore that the contribution of these lithologies to the CO2 consumption is negligible.Two other linear multiple regression analyses were performed using the linear multiple (lm) and the multiple nonnegative linear (nnl) regression analysis tools in the R software (R Core Team, 2016) considering the following four lithological classes: (1) “sandstone,” (2) “claystone,” (3) “total carbonate” (including “pure carbonate,” “mixed carbonate,” “gypsum evaporite,” and “peat”), and (4) “igneous and metamorphic rocks” (including “acid rocks,” “mafic rocks,” “intermediate rocks,” and “metamorphic rocks”). Table 6 shows the values of the b coefficients obtained by modeling the lithologies with the lm and nnl regression models (R Core Team, 2016). We observe that in the lm model, all the coefficients are significant (very high significance for “sandstone” and “total carbonate,” and medium significance for “claystone” and “igneous and metamorphic rocks”). Moreover, we note that the b value for the “igneous and metamorphic rocks” is still negative. This value does not agree with the assumptions that acid, mafic, intermediate, and metamorphic rocks are involved in a process of CO2 consumption (Eqs. 1–4). For this reason, in Table 6, we also show the values of the b coefficients obtained using a nonnegative linear (nnl) regression analysis. We observe that, as expected, the b value for “igneous and metamorphic rocks” becomes 0, whereas the coefficients of “sandstone” and “total carbonate” do not significantly change with respect to the values obtained with the lm regression. Conversely, the b coefficient of “claystone” obtained using the nnl model (6.30 × 10−4 mol L−1) is lower with respect to the values obtained using the lm regression (2.00 × 10−3 mol L−1). The determination coefficient of the fitting between the measured concentrations and the fitted values of the lm regression (R2) is again equal to 0.95, whereas the median, the mean, and the standard deviation of the residuals are –5.035 × 10−5, –1.65 × 10−9, and 5.935 × 10−4 respectively. Analogue values for the nnl model were obtained (R2: 0.94, residuals median: 5.943 × 10−5, residuals mean: 1.453 × 10−5, residuals standard deviation: 6.187 × 10−4).The linear model derived using four lithological classes was tested against a model built using literature values and in particular the b values estimated by Amiotte Suchet et al. (2003). In particular, we assumed b equal to (1) 1.52 × 10−4 for “sandstone,” (2) 6.27 × 10−4 for “claystone,” (3) 3.17 × 10−3 for “total carbonate” and (4) 9.50 × 10−5 for “igneous and metamorphic rocks.” Residuals between the linear model built using the Amiotte Suchet et al. (2003) coefficients and the measured alkalinity have values of median, mean, and the standard deviation respectively equal to –6.74 × 10−4, –7.86 × 10−4, and 9.06 × 10−4. As expected, the model built using the Amiotte Suchet et al. (2003) coefficients is less accurate (larger absolute values of mean and median values) and precise (larger standard deviation) in predicting the original alkalinity values with respect to the proposed model derived using four lithological classes.where RO is the runoff, SRi is the proportion (from 0 to 1) of the surface area covered by lithology i, bi is the calibration parameter for lithology i, and a is a parameter having value 1 in case of silicate rocks and value 0.5 in case of carbonate rocks (see Eq. 9).Table 7 shows the fluxes, ϕ(CO2)short, of atmospheric CO2 consumed by chemical weathering and estimated at basin scale by applying Equation 12. The values of b coefficients were derived by the lm regression analysis performed using 10 lithologies (see Table 5). Where the Significance level of the b values was very low (i.e., for “gypsum evaporites,” “metamorphic rocks,” “mafic rocks,” “acid rocks,” and “intermediate rocks,” as shown in Table 5), we considered b equal to 0. This choice is reinforced by the fact that (1) if the Significance level of b is very low, it means that b is statistically not different from 0, and (2) the negative b values for “mafic rocks,” “acid rocks,” and “intermediate rocks” do not agree with the assumption that these lithologies are involved in the CO2 consumption processes (see Eqs. 1–4). The CO2 fluxes were then calculated considering that “sandstone” is composed mainly either by silicate rocks (silicate-sandstone scenario) or by carbonate rocks (carbonate-sandstone scenario). In the silicate-sandstone scenario, the a parameter was considered equal to 0.5 only for pure carbonate and mixed carbonate categories and 1 for the remaining rock categories. In the carbonate-sandstone scenario, the a parameter was considered equal to 0.5 for pure carbonate, mixed carbonate, and sandstone categories and 1 for the remaining rock categories. Finally, RO values were calculated from Q(s), Q(w), and Q(my) (see Table 4), leading to three sets of ϕ(CO2)short for each scenario: ϕ(CO2)S(s), ϕ(CO2)S(w), and ϕ(CO2)S(my).Considering the silicate-sandstone scenario, Table 7 shows that during the spring season, the flux of atmospheric CO2 consumed by chemical weathering, ϕ(CO2)S(s), ranges from 3.93 × 104 mol km−2 yr−2 (Durance) to 3.71 × 106 mol km−2 y−2 (Livenza). Similar values were obtained by using both Q(w) and Q(my), since: (1) ϕ(CO2)S(w) ranges from 3.48 × 104 mol km−2 yr−2 (Durance) to 3.33 × 106 mol km−2 yr−2 (Lech), and (2) ϕ(CO2)S(my) ranges from 7.73 × 104 mol km−2 yr−2 (Durance) to 5.26 × 106 mol km−2 yr−2 (Isonzo). Also, the average values of ϕ(CO2)S(s), ϕ(CO2)S(w), and ϕ(CO2)S(my) were quite similar, being 9.43 × 105 mol km−2 yr−2, 9.81 × 105 mol km−2 yr−2, and 1.52 × 106 mol km−2 yr−2, respectively.The lowest ϕ(CO2)short values were systematically obtained considering the carbonate-sandstone scenario, since the a parameter for “sandstone” for this scenario was considered equal to 0.5, in contrast to the carbonate-sandstone scenario, where the a parameter was considered equal to 1. Considering the spring season, ϕ(CO2)S(s) ranged from 3.27 × 104 mol km−2 yr−2 (Durance) to 2.28 × 106 mol km−2 yr−2 (Livenza); considering the winter season, ϕ(CO2)S(w) varied from 2.89 × 104 mol km−2 yr−2 (Durance) to 2.33 × 106 mol km−2 yr−2 (Lech); and considering Q(my), the ϕ(CO2)S(my) ranged from 6.24 × 104 mol km−2 yr−2 (Durance) to 4.34 × 106 mol km−2 yr−2 (Isonzo). Quite similar values were obtained considering the average values of ϕ(CO2)S(s), ϕ(CO2)S(w), and ϕ(CO2)S(my), which were, respectively, 6.89 × 105 mol km−2 yr−2, 7.05 × 105 mol km−2 yr−2, and 1.09 × 106 mol km−2 yr−2.In Table 7, PD[ϕ(CO2)S(my)](carb/sil) represents the percentage difference, calculated following Equation 10, between the two values of ϕ(CO2)S(my) computed considering the carbonate-sandstone scenario and the silicate-sandstone scenario. The obtained percentage differences show that in the carbonate-sandstone scenario, the ϕ(CO2)S(my) value is on average –26.94% with respect to the fluxes estimated in the silicate-sandstone scenario, with the minimum value of –41.25% for Sesia and the maximum value of –13.74% for Var.The fluxes of atmospheric CO2 consumed by chemical weathering coming from silicates, ϕ(CO2)S(my)-sil, and from carbonates, ϕ(CO2)S(my)-carb, were estimated considering the two scenarios and are reported in Table 8. In the silicate-sandstone scenario, ϕ(CO2)S(my)-sil values were estimated considering the weathering of “sandstone” and “claystone,” while ϕ(CO2)S(my)-carb values were estimated considering the weathering of “pure carbonate” and “mixed carbonate.” In the carbonate-sandstone scenario, ϕ(CO2)S(my)-sil values were estimated considering only the weathering of “claystone,” while ϕ(CO2)S(my)-carb values were estimated considering the weathering of “pure carbonate,” “mixed carbonate,” and “sandstone.” The relative percentages of ϕ(CO2)S(my)-sil and of ϕ(CO2)S(my)-carb for the two scenarios were estimated with respect to the total ϕ(CO2)S(my). Table 8 shows that, considering the silicate-sandstone scenario, the contribution of silicate weathering is: (1) between 25% and 50% for eight river basins (Var, Brenta, Isonzo, Durance, Mincio, Tagliamento, Piave, and Sava), (2) between 50% and 75% for 14 river basins (Isar, Roia, Isere, Lech, Reuss, Alpine Rhine, Iller, Aare, Adige, Linth, Enns, Rhine, Drau, and Inn), and (3) between 75% to 100% for 11 river basins (Rhone, Livenza, Dora Baltea, Tanaro, Mella, Adda, Oglio, Mur, Po, Ticino, and Sesia).Considering the carbonate-sandstone scenario, the contribution of silicate weathering decreases significantly, being: (1) between 0% to 25% for 25 river basins (Brenta, Isonzo, Livenza, Tagliamento, Piave, Iller, Lech, Roia, Var, Isar, Sava, Tanaro, Mincio, Aare, Mella, Rhine, Linth, Reuss, Durance, Po, Alpine Rhine, Inn, Rhone, Isere, and Oglio), (2) between 25% to 50% for seven river basins (Sesia, Enns, Adige, Adda, Drau, Dora Baltea, and Ticino), and (3) between 50% and 75% for one river basin (Mur).The principal aims of this work were: (1) to investigate the relationship between the lithological composition of the main Alpine river basins and their water alkalinity, and (2) to provide generic mathematical parameters that link lithology to the moles of atmospheric CO2 consumed by chemical weathering.Assuming that the chemical reactions occurring within the basins are those reported in Equations 1–6, the lithological classes of Alpine-Geo-LiM were chosen by considering the mineralogic composition of outcropping rocks. For this reason, metamorphic rocks were classified according to the chemistry of protoliths, and all the rocks for which data on protoliths were unavailable or unclear (e.g., in the case of migmatite, mylonite, and metasediments) were included in the class “metamorphic rocks” (which occupies 1.81% of the whole study area). This criterion represents a novel feature when compared with other global lithologic maps (Gibbs and Kump, 1994; Amiotte-Suchet and Probst, 1995; Amiotte-Suchet at al., 2003; Dürr et al., 2005; Hartmann and Moosdorf, 2012; Moosdorf et al., 2010), where lithologies with very different behavior in the atmospheric CO2 consumption processes were included in the generic “metamorphic” class. This is the case, for example, of marble, a metamorphic rock composed of carbonate minerals highly prone to consuming atmospheric CO2. Marble has a very different behavior with respect to other metamorphic rocks, such as, for example, orthogneiss, which is a metamorphic rock derived from a granite/rhyolite protolith that is much less prone to consuming atmospheric CO2 with respect to marble. The classification of metamorphic rocks according to the chemistry of their protoliths is of particular interest as regards the Alpine chain, where, considering the global lithological map GLiM elaborated by Hartmann and Moosdorf (2012), metamorphic rocks are quite abundant, representing 25.84% of the whole area.Another novel feature of this work is the release of the map with the procedures (GIS commands and database queries) used to produce the map. We decided to share this information to allow reproducibility and replicability of the research and following the concept of open science (Nüst et al., 2018).Looking to Alpine-Geo-LiM, it shows that carbonate rocks are the most abundant type in the Alpine region (44.57%), followed by “sandstone” (26.99%), “claystone” (12.87%), “volcanic rocks” (10.50%), “metamorphic rocks” (1.81%), “peats” (0.48%), and “gypsum evaporite” (0.08%). A small area (2.69%) is covered by “water” in the form of lakes and glaciers. The effort in discriminating metamorphic rocks according to the chemistry of protoliths in Alpine-Geo-LiM is demonstrated by the fact that almost all the metamorphic rocks outcropping in the study area (25.84% according to Hartmann and Moosdorf, 2012) were assigned to a specific rock category, and only 1.81% of the study area remains in the general “metamorphic rock” category, used only when information on protoliths was unavailable or unclear.Overall, the map highlights the presence of an inner core mainly composed of crystalline silicate rocks, bounded to the north and south by rocks mainly composed of carbonates, and finally the presence of rocks composed of sandstones in the basins external to the Alpine chain, coherent with studies by Donnini et al. (2016) and Rossi and Donnini (2018).To investigate the relationship between lithological composition of river basins and their water alkalinity, three linear multiple regression analyses were performed following an approach derived from Hartmann et al. (2009); see Equation 11 herein. The first analysis used the linear multiple (lm) regression analysis tool (R Core Team, 2016) and considered the original 10 lithological classes of Alpine-Geo-LiM (Table 2). Due to the scarce presence of some lithologies, the other two analyses were performed using the linear multiple (lm) and the multiple nonnegative linear (nnl) regression analysis (R Core Team, 2016), and considering four lithological classes (“sandstone,” “claystone,” “total carbonate,” and “igneous and metamorphic rocks”).The b values obtained from the nnl regression analysis (Table 6) were compared with values from literature that considered monolithological basins (Bluth and Kump, 1994; Amiotte-Suchet et al., 2003), and that considered multilithological basins (Hartmann, 2009). The comparison shows that, in the present work, the calibration parameter b for “total carbonate” (2.45 × 10−3 mol l−1) is included among the range of literature values (1 × 10−3 to 8 × 10−4 mol L−1), as well as for “claystone,” with an estimated b value equal to 6.30 × 10−4 mol L−1, i.e., of the same order of magnitude of literature values ranging from 2 × 10−4 to 9 × 10−4 mol L−1. Conversely, the comparison also shows that the estimated b value for “sandstone” (4.50 × 10−3 mol L−1) is noticeably higher than literature values (6 × 10−4 to 6 × 10−5 mol L−1). Moreover, the b values obtained for the “sandstone” class are always larger than those calculated for “pure carbonate,” “mixed carbonate,” and “total carbonate” (see Tables 5 and 6). This large discrepancy is explained by the aforementioned cluster analysis (Table 3) results, highlighting that, in the study area, the “sandstone” class is probably composed by a relevant carbonate component. The presence of a relevant carbonate component in the Alpine forelands is explained in Appendix A, and it is well noted in the literature (see, e.g., for the Molasse Basin: Schlunegger et al., 1994, 1998; Kempf et al., 1999; Anne et al., 2017; Abdul Aziz et al., 2008; e.g., for the Po Valley and Adriatic foreland: Fontana et al., 2014). The high b value of the “sandstone” class can be also explained with the inclusion of recent alluvial sediments in the “sandstone” class. Besides the weathering acting at the soil-air interface, in fact, they are also exposed to chemical dissolution due to groundwater, which contributes to the basin streamflow. Furthermore, in alluvial sediments, which are usually located in flat and low-elevation areas, the resident time of water in the soil-air interface increases, facilitating the chemical dissolution processes. The presence of large amounts of alluvial sediments in the “sandstone” class is evident from the analysis of Table 1. “Sandstone” slope and elevation mean values are indeed much lower than those observed in the other lithologies. Therefore, results show that carbonates (in the three forms of “pure carbonate,” “mixed carbonate,” and “total carbonate”), as expected, have a strong positive correlation with water alkalinity. Surprisingly, the results also show that the correlation is even stronger for “sandstone.” This fact can be explained considering that (1) the “sandstone” class includes cemented and uncemented deposits also composed by gravel and sand-carbonate sediments, and, (2) in the analyzed basins, “sandstone” is prevalent (Table 3) in association with “pure carbonate” and “mixed carbonate” rocks.Interestingly, the “claystone” class (outcropping in ∼13% of the area) always shows a positive correlation with water alkalinity; however, depending on the type of regression, it can be low (10 lithological classes, lm; 4 classes, nnl) or high (4 classes, lm). The significance associated with the value of the coefficients is low, apart from the medium value estimated by the 4 classes lm regression. We explain such behavior with the presence in the study area of other carbonate-rich lithologies that hide the “claystone” influence on CO2 consumption. Moreover, we observe that the positive values of the b coefficient for “claystone” can be due not only to silicate weathering, but also to the chemical dissolution of carbonates that can be present in the “claystone” class rocks.“Igneous and metamorphic rocks” are scarcely represented in the area (∼11%), and their correlation with water alkalinity is always negative or equal to zero (see Tables 5 and 6). The significance of the “igneous and metamorphic rocks” coefficients is generally very low; it is also low in the case of the 4 classes lm regression, where it results in medium significance, where the corresponding P value is larger than those obtained for the other coefficients. Consequently, the present study shows a negligible contribution of volcanic rocks (acid, mafic, and intermediate) to atmospheric CO2 consumption. On the contrary, in the literature, it is shown that these lithologies, with b values ranging from 1.5 × 10−4 to 4.5 × 10−6 mol L−1, do provide a contribution (even if small) to atmospheric CO2 consumption. This different behavior is due, of course, to the fact that volcanic rocks constituted by silicate minerals occupy a small percentage (∼10%) of the whole study area. Moreover, we maintain that the negative or zero b values for the “igneous and metamorphic rocks” class (see Table 6) is due to the fact that b was estimated in basins where these lithologies are associated with more abundant carbonate minerals, which are more soluble than silicates (Table 2). This is different from what was done (1) by Bluth and Kump (1994) and Amiotte-Suchet et al. (2003), who considered monolithological basins, and (2) by Hartmann (2009), who considered multilithological basins, but who excluded basins containing more than 0.05% of carbonate sedimentary rocks.Regarding the “metamorphic rocks” class, the result (very low significance level; see Table 5) is different from the outcomes of other authors (Bluth and Kump, 1994; Amiotte-Suchet et al., 2003; Hartmann, 2009), which could be explained by the fact that we used our own classification scheme for the definition of the lithological map (made available along with the present manuscript). As an example, we believe that either the inclusion or the exclusion of some types of rocks into the metamorphic class, based on the analysis of the protoliths, can have a relevant influence on the estimation of the contribution of this class to the CO2 consumption. We conclude that, in this study area, these lithologies (“igneous and metamorphic rocks”) do not significantly contribute to the atmospheric CO2 consumption process.Considering Equation 9, the amount of atmospheric CO2 consumed by chemical weathering can be estimated starting from river water alkalinity. For this reason, the calibration parameter b of Equation 11 expresses the capability of different lithologies to consume atmospheric CO2 by chemical weathering (high value of b corresponds to high capability to consume atmospheric CO2). Consequently, the analysis shows that the lithologies more prone to consume atmospheric CO2 are, from the higher to the lower: “sandstone,” “carbonates,” and “claystone,” while the contribution of “igneous and metamorphic rocks” is negligible.Finally, we applied Equation 12 to estimate the fluxes of atmospheric CO2 consumed by chemical weathering in the “short-term,” ϕ(CO2)short, within the study area. The fluxes were calculated considering “sandstone” composed (1) mainly by silicate rocks (silicate-sandstone scenario) and (2) mainly by carbonate rocks (carbonate-sandstone scenario). In Equation 12, we considered the b values obtained from the lm regression analysis performed using 10 lithologies (see Table 5). For “gypsum evaporites,” “metamorphic rocks,” “mafic rocks,” “acid rocks,” and “intermediate rocks,” where the Significance level of the b values was very low (see Table 5), we considered b equal to 0. In the silicate-sandstone scenario, the a parameter was considered equal to 0.5 for “pure carbonate” and “mixed carbonate” categories and 1 for the remaining rock categories. In the carbonate-sandstone scenario, the a parameter was considered equal to 0.5 for “pure carbonate,” “mixed carbonate,” and “sandstone” categories and 1 for the remaining rock categories. Runoff values (RO in Eq. 12) were estimated considering the daily discharge at the time of the two sampling campaigns (in spring season and in winter season) and considering the mean annual discharge, Q(s), Q(w), and Q(my) in Table 4.As expected, in the carbonate-sandstone scenario, the ϕ(CO2)short values were systematically lower than in the silicate-sandstone scenario, since the a parameter for “sandstone” for this scenario was considered equal to 0.5, in contrast to the carbonate-sandstone scenario, where the a parameter was considered equal to 1. The percentage difference between the mean annual fluxes estimated considering the carbonate-sandstone scenario and the silicate-weathering scenario (see Eq. 10) was on average 26.99%, with the minimum value of –41.25% for Sesia and the maximum value of –13.74% for Var (see Table 7).A comparison between the contribution (1) of silicate weathering, ϕ(CO2)S(my)-sil, and (2) of carbonate weathering, ϕ(CO2)S(my)-carb, in the two scenarios (silicate-sandstone and carbonate-sandstone scenarios) is shown in Table 8. As expected, the table shows that considering the carbonate-sandstone scenario, the contribution of silicate weathering is 14.51% of the total ϕ(CO2)S(my), and it increases up to 64.72% in the silicate-sandstone scenario. It is evident that the contribution of ϕ(CO2) from silicates is really dependent on the assumptions made about the chemical composition of the sandstone rocks. In the present work, the results indicate that, in the study area, the sandstone rocks contain a relevant component of carbonate rocks. The general implication of our results is that the estimation of CO2 consumption in areas where sandstone rocks are relatively abundant cannot be made without a careful evaluation of the carbonate content of the lithotypes that were included in the sandstone class. As a consequence, attention should be paid to the choice of the coefficients adopted from the literature (for the “sandstone” class) for calculating the CO2 fluxes in any given area different from the areas used for the calibration of the parameters themselves.The approach here presented is valid in nonpolluted areas, for temperate climates, and for lithologies without pyrite (see “Weathering Estimation” section). The absence of pyrite is important because, as highlighted, for example, by Moon et al. (2007), pyrite oxidation generates sulfuric acid, which could weather surrounding carbonate and silicate minerals. Since atmospheric CO2 is not consumed in this process, not considering pyrite oxidation could lead to an overestimation of the atmospheric CO2 consumption by silicate weathering. Since both pyrite oxidation and gypsum dissolution lead to an increase of SO4− in river waters (e.g., Berner and Berner, 1996), distinguishing between the gypsum and pyrite sources of SO4− in river waters is important for a reliable estimation of the fluxes of atmospheric CO2 consumed by chemical weathering (Moon et al., 2007). We know that the presence of pyrite in the Alps is well documented (e.g., Kappler and Zeeh, 2000; Lavrič and Spangenberg, 2003; Rantitsch, 2007; Gainon et al., 2007; Grachev et al., 2008; Bernard et al., 2010; Herlec et al., 2010; Sanders et al., 2010; Sabatino et al., 2011; Pálfy and Zajzon, 2012). In addition, in the river waters sampled by Donnini et al. (2016), the samples with a relevant sulfate enrichment were located in the southwest French Alps (Roia, Var, Isere, Durance) and in the southeastern Italian Alps (Tagliamento). Since the presence of gypsum in Triassic carbonate rocks is well documented both in the southwest French Alps (e.g., Olivier et al., 2009) and in the southeastern Italian Alps (e.g., Stefanini, 1976; Longinelli and Flora, 2007), we think that it is more reasonable to consider the SO4− enrichment in these river waters as a consequence of evaporite dissolution, rather than pyrite oxidation. For this reason, we maintain that our simplification—which considers the pyrite oxidation negligible—could lead to only a slight overestimation of atmospheric CO2 consumption.A comparison of the fluxes of atmospheric CO2 fixed by chemical weathering obtained in this work with those available from the literature shows that minor differences exist at regional scale. In particular, a comparison of the ϕ(CO2)short values obtained in the present work with the ϕ(CO2)short values estimated by Donnini et al. (2016) for the same area shows that the range of atmospheric CO2 fixed by chemical weathering within the 33 main Alpine river basins is quite similar.In Donnini at al. (2016), during the spring season, ϕ(CO2)short ranges from 2.60 × 104 mol km−2 yr−2 (Durance) to 2.03 × 106 mol km−2 yr−2 (Livenza), and during the winter season, ϕ(CO2)short ranges from 2.48 × 104 mol km−2 yr−2 (Durance) to 2.04 × 106 mol km−2 yr−2 (Lech).In this work, during the spring season, in the silicate-weathering scenario, ϕ(CO2)S(s) ranges from 3.93 × 104 mol km−2 yr−2 (Durance) to 3.71 × 106 mol km−2 yr−2 (Livenza), and in the carbonate-weathering scenario, it varies from 3.27 × 104 mol km−2 yr−2 (Durance) to 2.28 × 106 mol km−2 yr−2 (Livenza). During the winter season, in the silicate-weathering scenario, ϕ(CO2)S(w) ranges from 3.48 × 104 mol km−2 yr−2 (Durance) to 3.33 × 106 mol km−2 yr−2 (Lech), and in the carbonate-weathering scenario, it varies from 2.89 × 104 mol km−2 yr−2 (Durance) to 2.33 × 106 mol km−2 yr−2 (Lech). Considering the mean annual discharge, in the silicate-weathering scenario, ϕ(CO2)S(my) ranges from 7.73 × 104 mol km−2 yr−2 (Durance) to 5.26 × 106 mol km−2 yr−2 (Isonzo), and in the carbonate-weathering scenario, it ranges from 6.24 × 104 mol km−2 yr−2 (Durance) to 4.34 × 106 mol km−2 yr−2 (Isonzo).Overall, the difference between the mean ϕ(CO2)short estimated in the present work, where the mean values of ϕ(CO2)S(s), ϕ(CO2)S(w), and ϕ(CO2)S(my) are 9.43 × 105 mol km−2 yr−2, 9.81 × 105 mol km−2 yr−2, and 1.52 × 106 mol km−2 yr−2 in the silicate-weathering scenario, and 6.89 × 105 mol km−2 yr−2, 7.05 × 105 mol km−2 yr−2, and 1.09 × 106 mol km−2 yr−2 in the carbonate-weathering scenario, and the mean ϕ(CO2)short estimated by Donnini at al. (2016) (4.69 × 105 ± 1.03 × 105 mol km−2 yr−2 in spring season, and 5.35 × 105 ± 0.97 × 105 mol km−2 yr−2 in winter season) is less than one order of magnitude.Quite similar values were estimated by Gaillardet et al. (1999) for the Rhine, Rhone, and Po basins, respectively 5.42 × 105 mol km−2 yr−2, 8.56 × 105 mol km−2 yr−2, and 1.12 × 106 mol km−2 yr−2, showing that the mean ϕ(CO2)short of these three rivers is much higher than the world average CO2 consumed by chemical weathering (2.46 × 105 mol km−2 yr−2) estimated by the same authors.We maintain that the data-driven estimation of the CO2 consumption rates in the Alpine region presented here is more objective than the rates estimated using literature values, since the new b parameters presented here were obtained using measured data only.The results of the present study highlight the importance of discriminating rocks according to their mineralogic composition, paying close attention to the presence of minor carbonate components in rock categories usually considered dominated by silicates, like metamorphic rocks, and, as highlighted by Hartmann et al. (2009), like sandstone and shale (in the present work denominated claystone). It is well known, in fact, that these lithologies could contain a small carbonate content (e.g., Jacobson and Blum, 2003; Emberson et al., 2018). The nonnegligible contribution of carbonates to atmospheric CO2 consumption of silicate-dominated rock categories was stressed by Hartmann et al. (2009, p. 189), who stated that, at global scale, “about 12.6% of the carbonate CO2 consumption can be attributed to silicate dominated lithological classes.” The same authors highlighted that the global contribution of carbonate sedimentary rocks has been overestimated in the past, being ∼25% in Hartmann et al. (2009), in contrast to ∼40% in Gaillardet et al. (1999), Munhoven (2002), and Amiotte Suchet et al. (2003).Alpine-Geo-LiM is a high-resolution (scale 1:1,000,000) geo-lithological map of the Alps. It represents a novel map when compared with published global lithological maps (Gibbs and Kump, 1994; Amiotte-Suchet and Probst, 1995; Amiotte-Suchet at al., 2003; Dürr et al., 2005; Hartmann and Moosdorf, 2012; Moosdorf et al., 2010) for two main reasons. First of all, the lithological classes used to map the study area were chosen by considering the mineralogic composition of the outcropping rocks. Particular attention was paid in discriminating metamorphic rocks, which were classified according to the chemistry of protoliths. The class “metamorphic rocks” included only the rocks for which data on protoliths were unavailable or unclear. The role of different lithologies in atmospheric CO2 consumption by chemical weathering was estimated by means of a multilithological approach and by applying a linear multiple regression for predicting water alkalinity based on lithologies. The analyses confirmed that carbonates are lithologies highly prone to consuming atmospheric CO2, as previously stated by several authors (Bluth and Kump, 1994; Amiotte-Suchet et al., 2003; Hartmann, 2009). The present work also shows that the “sandstone” category, which includes quartzite, and also arkose, graywacke, and conglomerate, could have a nonnegligible carbonate component (Garrels and Mackenzie, 1971; Mottana et al., 2009) and play an important role in consuming atmospheric CO2. Moreover, the linear multiple regression analyses showed that the contribution of igneous rocks in atmospheric CO2 consumption is negligible.The second novel feature is that Alpine-Geo-LiM is being released in vector format together with the procedure used for the definition of the map and the original data in order to allow the replicability and reproducibility of the product (see Donnini et al., 2018).M. Donnini was supported by a grant from the Fondazione Assicurazioni Generali, and A. Zucchini was partially supported by the research projects of Paola Comodi, Francesco Frondini, and Diego Perugini of the Department of Physics and Geology of the University of Perugia.M. Donnini mainly contributed to application of the geochemical models and to verification of the accuracy of the map with respect to Alpine geology, I. Marchesini mainly contributed to the geographical and statistical operations, and A. Zucchini mainly contributed to the mineralogical-petrographic considerations useful for elaborating the geo-lithological classification of the map. M. Donnini wrote the paper, which I. Marchesini and A. Zucchini then revised internally.The Alpine geo-lithological map (Alpine-GeoLiM) represents a portion of the Geo-Lithological Map of Central Europe (Geo-LIM), released in GPKG and in PDF format in Donnini et al. (2018) - https://doi.org/10.5281/zenodo.3530257. Along with the map, we provide: (1) the original national geological maps of Germany, Italy, Slovenia, France, Switzerland, and Austria, used for creating the map; and (2) a script that can be used to replicate the classification and the union of the original maps.We are grateful to the two reviewers, who significantly contributed to the overall quality of the manuscript.

中文翻译：

---

一个新的高山地貌岩性图（Alpine-Geo-LiM）及其对全球碳循环的影响

河流水域的化学成分提供了通过化学风化过程确定的大气中二氧化碳的量度。由于控制这些过程的主要因素是岩性和径流以及隆升和侵蚀，因此我们引入了一个新的阿尔卑斯山简化的地质岩性图（Alpine-Geo-LiM），该图采用了一种与最常用的方法相符的岩性分类。有关通过化学风化估算大气中二氧化碳消耗量的文献。该地图与33条主要高山河流的碱度数据一起使用（1）研究采样水中碳酸氢盐浓度与相应流域的岩性之间的关系，以及（2）量化化学物质消耗的大气CO2风化。这些分析证实（如文献所知），碳酸盐岩是极易消耗大气二氧化碳的岩性。此外，分析表明，砂岩（可能具有不可忽略的碳酸盐成分）在消耗大气中的CO2中起着重要作用。另一个结果是，在含岩性的多岩性盆地中，岩性更易消耗大气中的CO2，火成岩对大气CO2消耗的贡献可忽略不计。与已发布的全球岩性图相比，Alpine-Geo-LiM具有几个新颖的特征。一种新的特征是由于在区分变质岩方面的关注，这些变质岩是根据原石的化学性质进行分类的。第二个新颖的特征是，用于定义地图的过程可在Web上获得，以允许产品的可复制性和可再现性。碳是宇宙中第四大最丰富的元素（Morgan和Anders，1980年; Anders和Ebihara，1982年），它在地球环境中起着至关重要的作用。该元素在四个汇之间不断迁移：海洋，大气，生态系统和地圈（荷兰，1978；伯纳，2003；库普等人，2009）。考虑到现象的时间尺度，“短期”碳循环（短于1 my）不同于“长期”碳循环（长于1 my）。在文献中假定1 my阈值与Ca2 +在海洋系统中的停留时间一致（Donnini等人，2016）。在“短期”碳循环中，碳在海洋，生物圈，土壤和大气等表层系统中快速交换，其中还考虑了人为产生的CO2。在“长期”碳循环中，碳在地圈和海洋-大气系统之间缓慢交换。在这里，大气中CO2的浓度主要来自火山活动和变质作用所产生的CO2与硅酸盐和碳酸盐的风化所消耗的大气中CO2的平衡（Berner等，1983; Berner，1991,1994,2004,2006）。 ; Berner和Kothavala，2001; Gislason和Oelkers，2011; Li和Elderfield，2013）。由于化学风化产生的溶质丰富了河流的溶解负荷，因此河水的成分可以被视为化学风化过程的良好指标（ Mackenzie和Garrels，1966年；Garrels和Mackenzie，1971年；梅贝克（Meybeck），1987年；塔迪（Tardy），1986年；Probst，1992年；Gaillardet et al。，1999；Viers等人，2007年；Berner和Berner，2012年）。从了解河水的化学成分和流速以及流域的岩性开始，可以使用两种不同的方法来计算化学风化所消耗的大气中的二氧化碳（Hartmann，2009； Hartmann等， 2009）：（1）反向和（2）前进方法。两种方法都假定流域内发生的唯一反应是由于碳酸的存在而使硅酸盐和碳酸盐发生了变化。相反的方法是使用质量平衡方程，通过考虑特定的岩性末端元素来区分风化产物（Garrels和Mackenzie，1967； Meybeck，1987； Gaillardet等，1999）。因此，溶解在河流水中的阳离子之间的化学计量关系可以很好地估算出参与变化过程的大气中二氧化碳的摩尔数（Probst等人，1994； Amiotte-Sutt，1995； Amiotte-Sutt和Probst，1996; Boeglin和Probst，1998; Mortatti和Probst，2003; Donnini等人，2016）。根据时间尺度，为了量化化学风化消耗的大气CO2，必须考虑不同的反应（Huh，2010; Donnini et al。，2016）。正向方法假设岩性和径流（即每单位面积）是河水中碳酸氢盐浓度的主要控制因素，这是化学风化所消耗的大气中二氧化碳的量度。对于特定的岩性，径流与化学风化通过经验关系所消耗的大气中的二氧化碳有关。这样，就有可能量化化学风化所消耗的大气中的二氧化碳（Bluth和Kump，1994; Amiotte-Sutt和Probst，1993a，1993b，1995; Probst等，1994; Amiotte-Shitt等，2003 ； Hartmann，2009； Hartmann等人，2009）。对岩石的性质有很好的了解对于建立二氧化碳消耗与岩性之间的经验关系至关重要。正如Amiotte-Cheet等人所强调的。（2003）和Moosdorf等。（2010年），地质图通常很少提供有关岩石的化学和物理性质的信息，重点是岩石的年龄，其变形，地层和结构位置。缺乏信息是有问题的，尤其是对于沉积岩，它们在造山带中非常丰富（Doglioni，1994； Einsele等，1996； Clift等，2001），并且化学成分变化很大（Amiotte-Cheet等，2003）。此外，获取有关变质岩原岩的信息通常并不容易。在文献中，已在全球范围内发布了一些岩性图，并在下文进行了说明。Gibbs和Kump（1994）提出了2°×2°的全球岩性图，分为以下六种岩石类型：（1）碳酸盐岩，（2）页岩，（3）砂岩，（4）挤压性火成岩，（5）盾构区域（包括侵入性火成岩和变质岩），以及（6）“复杂岩性”（在2°×2°网格单元中难以辨别单一岩石类型）。该岩性图与7.5°×4一起使用。5°全球径流图，利用Bluth和Kump（1994）的径流与碳酸氢根通量之间的关系来计算全球河流碳酸氢盐通量.Amiotte-Sutt和Probst（1995）绘制了1°×1°的全球二氧化碳消耗量图（全球二氧化碳通量侵蚀模型[GEM-CO2]），从粮食及农业组织（FAO）和联合国教育，科学及文化组织（UNESCO）发布的简化的岩性和土壤图开始（FAO-UNESCO，1971年，1975年） ，1976，1978，1979，1981），并利用梅贝克（Meybeck，1986）估计的关系考虑了200多个法国整体岩性盆地（Amiotte-Sutt和Probst，1993a，1993b）。Amiotte-Cheet和Probst（1995）定义了岩石风化消耗的总大气/土壤CO2通量，,（CO2）短，是在给定时间内每单位面积消耗的CO2摩尔数。在该地图中，考虑了以下七个岩性：（1）岩石和变质岩，（2）砂岩和砂岩，（3）酸性火山岩，（4）蒸发岩，（5）玄武岩，（6）页岩，以及（7）碳酸盐岩。（2003年）拟订了一个1°×1°的全球岩性图，其中考虑了六种岩石类别：（1）砂岩和砂岩，（2）页岩，（3）碳酸盐岩，（4）混合侵入性火成岩和变质岩（即盾构）岩石），（5）酸性火山岩和（6）玄武岩。与Gibbs和Kump（1994）提出的地图相比，Amiotte-Chett等人的地图。（2003年）的分辨率更高（1°×1°vs. 2°×2°），并且信息量更大，因为在吉布斯和库普（1994）的地图中，约27％的总曝光量是“复杂岩性”。 ），因此，它们并没有得到精确的表征（Amiotte-Cheet等，2003）。与Amiotte-Chett和Probst（1995）相似，Amiotette-Shitt等人估计ϕ（CO2）短。Meybeck（1986）通过ϕ（CO2）短波与径流之间的关系（2003）提出了更详细的全球岩性图。（2005）：1：25,000,000的比例。与吉布斯（Gibbs）和库普（Kump）（1994）以及阿米特（Amiotte）（2003年），这是两个基于网格的栅格地图，Dürr等人的地图。（2005）为矢量格式，包含8300个多边形。该地图考虑了15种岩石类别（不包括水和冰）：（1）酸性火山岩，（2）基本火山岩，（3）酸性深成岩，（4）基本深成岩，（5）前寒武纪基底，（6） （7）固结硅质碎屑岩，（8）混合沉积岩，（9）碳酸盐岩，（10）半至非固结沉积岩，（11）冲积物，（12）黄土，（13）沙丘，（14）蒸发岩和（15）复杂岩性（沉积物，火山岩和变质岩混合在一起）。与露头岩性一起，该图还包含其他三个主题层，其中包含其他地质信息（主要的地下蒸发岩发生，地质和最大第四纪冰期作用范围的极限）。另一幅矢量格式的全球岩性图（名为GLiM）由Hartmann提出。和Moosdorf（2012）。地图包含1：1,000,000比例的1,235,400个多边形。继Moosdorf等。（2010年），地图包含三个级别的信息（图层）。第一个是强制性的，代表一般岩性。它考虑了15种岩性（不包括水和冰）：（1）蒸发岩，（2）变质岩，（3）酸性深成岩，（4）基本深成岩，（5）中深成岩，（6）火成碎屑，（7）碳酸盐沉积岩，（8）混合沉积岩， （9）硅质碎屑沉积岩，（10）未固结沉积物，（11）酸性火山岩，（12）基本火山岩，（13）中火山岩，（14）前寒武纪岩石，（15）复杂岩性。第二层和第三层可选地包含有关特定岩石属性的信息。（2016年）展示了阿尔卑斯山的岩性图。该地图与33个主要高山河流水域的主要元素浓度一起，通过应用MEGA地球化学代码估算了高山地区化学风化过程中的大气CO2消耗量（Amiotte-Sutt，1995；Amiotte-Cheet和Probst，1996），它实现了反向方法。该地图以1：1,000,000的比例绘制，并考虑了八种岩性：（1）酸性火成岩，（2）混合碳酸盐，（3）黏土和粘土岩，（4）碎片，（5）镁铁质岩，（6）变质岩岩石，（7）纯碳酸盐岩石和（8）砂岩。在本文中，我们采用了一种新的高分辨率（1：1,000,000比例）简化的阿尔卑斯山地貌图（称为Alpine-Geo-LiM）岩性分类（10个岩性类别：（1）“纯碳酸盐”，（2）“混合碳酸盐”，（3）“石膏蒸发岩”，（4）“酸性岩”，（5）“镁铁质岩”，（6 ）“中级岩石”，（7）“砂岩”，（8）“粘土石”，（9）“变质岩”和（10）“豌豆”），并符合反向和正向方法。Alpine-Geo-LiM来源于意大利，法国，德国，瑞士，奥地利和斯洛文尼亚的国家地质图，它代表了先前在Donnini等人中发表的地图的实现。（2016）。此外，它与用于构建地图的代码一起发布（Donnini等人，2018）。尽管我们使用了与Donnini等人相同的输入数据。（2016），Alpine-Geo-LiM与Donnini等人发表的地图不同。（2016）的岩性分类，即唐尼尼等人的八个岩性类别。（2016年）对比了Alpine-Geo-LiM的10个岩性分类，以及对变质岩原石的更准确分析。而且，与Donnini等人不同。（2016），Alpine-Geo-LiM以矢量格式连同用于制作地图的信息程序和原始数据一起发布（参见Donnini等人，2018）。由于我们提供了岩性图，因此我们将Alpine-Geo-LiM定义为地质岩性图，但同时也提供了用于创建该图的原始图层和过程。此外，我们在属性表中发布了原始地质信息（附录A，B和C1）.Alpine-Geo-LiM，以及2011年和2012年采样的33条主要阿尔卑斯河的碱度（Donnini等人。 （2016年）：（ 1）研究采样河水中HCO3-浓度与相应流域岩性之间的关系，并应用正演方法，（2）量化化学物质消耗的大气CO2阿尔卑斯山（欧洲中南部; 图1）是白垩纪产生的碰撞带，目前存在欧洲和非洲（也称为亚得里亚海或普利亚）大陆边缘的汇合带，这导致了地中海地区海洋的封闭（特吕皮，1960年；弗里施，1979年） ; Tricart，1984; Haas等人，1995; Stampfli等人，2001; Dal Piaz等人，2003; Schmid等人，2004; Pfiffner，2014）。阿尔卑斯山呈弧形，可以大致细分。分为以下图1所示的不同地质区域（Dal Piaz等，2003； Schmid等，2004； Pfiffner，2014），如图1所示：东部阿尔卑斯山，北部钙质阿尔卑斯山，东南部的Eoalpine钙质阿尔卑斯山和西部阿尔卑斯山。阿尔卑斯山的亚平宁链与西北部分相连，而迪那里底群岛则与东半部相连。潘诺尼亚盆地将阿尔卑斯山向东延伸，Molasse盆地将阿尔卑斯山脉向北限制，而Po谷和亚得里亚海前陆则将链条限制在南部。汝拉山脉定义了阿尔卑斯山的西北边界。在阿尔卑斯山的外部，在北部有欧洲前陆。图1中的多边形代表了研究区域，相当于Donnini等人将阿尔卑斯山细分为33个主要的高山流域。（2016）。阿尔卑斯山的地质学可以通过以下地质领域大致描述（Rossi and Donnini，2018）：（1）阿尔卑斯山东部的奥山高山结晶岩; （2）侏罗山，北部钙质阿尔卑斯山和东南部的伊欧高山钙质阿尔卑斯山的碳酸盐岩；（3）在阿尔卑斯山西部，钙质钙质单元与晶状质体和Penninic变质-橄榄岩相混合。在高山链之外，（1）北部的莫拉塞盆地充满了几公里厚的第三纪演替；（2）南部的波谷和亚得里亚海前陆主要由冲积物组成，潘诺尼亚盆地也是如此从地貌学的角度来看，阿尔卑斯山的特征是海拔高度（masl）在1200至1300 m之间，宽阔的低地，深切的山谷和海拔4000 masl以上的山脉（最高峰是蒙特比安科（Bianco）处于4888马氏度（Mal）;达尔·皮亚兹（Dal Piaz）等人，2003年），导致强烈的地形变化（Carraro和Giardino，2004年;戈比埃特等人，2014年）。极端温度和年降水量与阿尔卑斯山的地形有关。谷底通常比周围的山脉温暖和干燥。在冬季，1500 masl以上的几乎所有降水都是雪形式。积雪持续时间大约为11月中旬至5月底，为2000马氏度（Diem等，2019）。在融化后的接下来的几个月中，冬季储存为雪和冰的降水会释放出来（欧洲环境署，2010）。高山地区的水以湖泊，蓄水层和冰川的形式存在，为欧洲许多流域提供水源，包括莱茵河，多瑙河，波河和罗纳河（Weingartner等，2007），这是欧洲最大的河流。在流量和流域面积方面。冰川覆盖约2050平方公里（Paul等，2011），占33个主要高山盆地面积的1％（Donnini等，2016）。以下各节向读者介绍（1）控制大气CO2消耗的基本方程式和（2）新型Alpine-Geo-LiM。河水的化学成分是风化过程的指标（Mackenzie和Garrels，1966; Garrels和Mackenzie，1971； Meybeck，1987； Tardy，1986； Probst，1992； Gaillardet等，1999； Viers等，2007； Berner和Berner，2012），它们与大气输入（雨），污染，生物群和蒸发岩的溶解达到溶解负荷（例如，Gaillardet等，1999； Galy和France-Lanord，1999； Roy等，1999； Moon等，2007； Wu等，2008； Gao等人，2009； Jha等人，2009； Donnini等人，2016）。硅酸盐矿物的风化反应，水解和碳酸盐溶解会消耗大气/土壤CO2，并增加溶液的碱度。硅酸盐和碳酸盐被碳酸改变的正向反应是“短期”内流域内发生的唯一反应，并由以下等式描述（Mortatti和Probst，2003； Donnini等，2016）：在“长期”阶段，并非所有的大气中的二氧化碳都会通过风化反应而永久去除，因为一些碳会返回大气中（Huh，2010； Donnini等，2016）。特别是，释放到海洋的河流溶解负荷部分以碳酸盐（等式5中的逆反应）或自生粘土（等式1和等式2中的逆反应）的形式沉淀。特别是，公式5显示，风化的CaCO3消耗了1单位的CO2（正向反应，“短期”），并且当海洋中的CaCO3沉淀后，相同量的CO2会返回大气（反向反应，“长期”）。类似的行为可能是由于CaMg（CO3）2的风化（参见等式6），其中由于风化（正向反应，“短期”）消耗的两个单位的CO2当CaMg（CO3）2在海洋中沉淀时返回大气（反向反应，“长期-术语”）。但是，通过强大的Mg2 +水溶液溶剂化壳层以及抑制Ca-形成的结晶屏障，可以防止白云石在室温下从水溶液中直接沉淀（例如Frondini等，2014，及其中的参考文献）。镁定购白云石（Montes-Hernandez等，2016）。因此，仅在理论上考虑方程式6的逆反应。含Na和K的硅酸盐矿物（轻金属和钾长石）的风化反应会消耗大气中的CO2（方程式1和方程式2的正向反应，“短期”）。当Na +和K +离子被河流运输到海洋时，它们可能会经历逆风化作用，形成自生粘土并释放出大气中的CO2（方程1和方程2的反向反应，“长期”；呵呵） ，2010）。方程式3和方程式4是不可逆的，正向反应的产物（HCO3-，Ca2 +和Mg2 +）参与方程式5所述的反向反应。特别是大气CO2的一半单位在风化过程中消耗的钙斜长石（等式3）以CaCO3的形式在海洋中沉淀（根据等式5，逆反应）。橄榄石的正向反应产物也可能发生同样的情况（公式4），当考虑方程式6的逆反应时。如前所述，最后的推测只是理论上给出的。方程1–6是流域中唯一发生的反应的假设是有效的（a）如果碳酸（H2CO3，是由河水和大气CO2之间的相互作用产生的）是风化反应中唯一的质子来源（其他酸[HNO3，H2SO4]的贡献可忽略不计； Mortatti和Probst，2003； Donnini等人，2016），以及（b）如果黄铁矿（FeS2），石膏（CaSO4•2H2O）和冰盐（NaCl）百分比为在河流流域中可以忽略不计，因此在模型计算中未考虑其溶出度（Perrin等，2008）。总体而言，上述条件适用于无污染地区，温带气候以及不含黄铁矿的岩性。两种不同的方法被用来计算化学风化所消耗的大气中的二氧化碳（Hartmann，2009； Hartmann等，2009）：（ 1）反向方法和（2）正向方法。在下文中，将对这两种方法进行简要说明。在反向方法中，考虑到正向（“短期”）和反向，化学风化所消耗的大气/土壤二氧化碳的摩尔数是根据公式1-6计算的（“长期”）反应。实际上，在河流水中测得的溶解阳离子用于估算消耗的大气/土壤CO2的摩尔数。在等式7和等式8中，ϕ（X + Y）是两种通用化学物种X和Y的通量之和在河水中，用其摩尔浓度乘以径流得出，后缀“ sil”和“ carb”表示被认为是源自硅酸盐或碳酸盐风化的化学物种（Huh，2010； Donnini等，2016）。从所测量的河水中Ca2 +和Mg2 +的浓度开始，可以通过使用特定的离子比率来区分硅酸盐风化（Ca + Mg）硅和碳酸盐溶解（Ca + Mg）carb对总河通量的贡献（Meybeck，1986，1987）。许多作者已经采用了反向方法来估算刚果，亚马逊河和尼日尔流域以及33个主要高山流域的大气CO2消耗量（ Probst等人，1994； Amiotte-Chutt，1995； Amiotte-Chutt和Probst，1996； Boeglin和Probst，1998； Gaillardet等人，1999； Mortatti和Probst，2003； Moon等人，2007； Phil等。Donnini et al。，2016）。在正向方法中，化学风化消耗的大气/土壤二氧化碳的摩尔数是根据方程式1-6计算的，仅考虑了“短期”正向反应。类似于方程式7和方程式8，在等式9中，ϕ（HCO3）是HCO3-的摩尔通量，由其摩尔浓度乘以径流量得出。后缀“ sil”和“ carb”表示所考虑的化学物种（HCO3-）是源自硅酸盐还是碳酸盐风化作用。在文献中，一组经验关系将不同岩性联系起来，从而消耗了大气/土壤CO2的通量。短期（ϕ（CO2））化学风化到径流。Probst et al。（1993a，1993b，1995）的Amiotte-Cheet和Probst。（1994），和Amiotte-Cheet等人。（2003年）估计了溶解负荷下ϕ（CO2）短径流与200多个法国单石流域径流之间的关系（Meybeck，1986，1987）。Bluth和Kump（1994）从美国，波多黎各和冰岛的约100个单石流域的溶解负荷和径流估算出类似的关系。在Hartmann（2009）和Hartmann等人中。（2009年）首次通过多变量非线性回归分析估算了ϕ（CO2）短径流与径流之间的关系，该分析从382个日本流域排放了多个岩性开始。正向方法认为岩性和径流是主要因素控制大气中CO2消耗过程的因素，以及其他因素（例如救济或土地覆盖），在区域和全球范围内的重要性都较低（Hartmann，2009； Hartmann等，2009）。大气CO2消耗的温度依赖性仅在全球玄武岩-风化模型中实现（Dessert等，2003）。从from（CO2）短径流与径流之间的经验关系开始，仅了解露头岩性和径流在给定的区域内，可以估算大气中二氧化碳的消耗摩尔数。此估算不需要有关河水化学成分的数据。前向模型已在加龙河，刚果河和亚马逊河流域（Amiotte-Chut和Probst，1993a，1993b，1995； Probst等，1994）的流域规模和日本群岛的区域规模（Hartmann，1994）中得到应用。 2009年），并在全球范围内（Amiotte-Sutt和Probst，1995年； Amiotte-Sutt等人，2003; Hartmann et al。，2009）。为详细阐述我们基于阿尔卑斯山的地理信息系统（GIS）的简化地质岩性图（比例为1：1,000,000），我们利用了矢量格式的地质层，从（1）由意大利环境保护与研究所（ISPRA; http://www.isprambiente.gov.it）发布的比例为1：500,000的意大利地质图（Bonomo等，2006），（2）瑞士联邦地形局（Swisstopo; http://www.swisstopo.admin.ch）发布的比例为1：500,000的瑞士地质图（BundesamtfürLandestopografie，2005），（3）德国的地质图1：1,000,000比例尺（BGR，2011），（4）奥地利地质调查局（GBA; http://www.geologie.ac）发布的奥地利比例尺1：500,000比例尺的地质图（Egger等，1999）。 。在），（5）比例为1：1,000,000的法国地质图（BRGM，2003），以及（6）比例为1：250,000的斯洛文尼亚的地质图（Buser，2010）。最后两个地图是从欧洲地质数据基础设施（EGDI； http：//www.europe-geology.eu/metadata）获得的。六张地图以ESRI shapefile格式发布，具有不同的坐标参考系统以及不同的准确性和信息质量。法国，德国和斯洛文尼亚的图层包含多个拓扑错误（例如，多边形边界之间的间隙，重叠的多边形边界等），并通过删除重复的边界和分别小于1 m2、600 m2和50的区域进行了纠正m2（与相邻区域的最长边界已删除）。矢量地图的属性表包含不同的属性字段，用于存储地质信息的描述。这些字段在附录B中列出（请参见脚注1）。（2010，第2页），“分类是精确性和简单性之间的不断折衷。” 在Alpine-Geo-LiM中使用的岩性分类是Gibbs和Kump（1994），Amiotte-sutt和Probst（1995）和Amiotte-sutt等人使用的6-7岩石类别之间的折衷。（2003年），以及Dürr等人使用的15种岩石类别。（2005），Hartmann和Moosdorf（2012）和Moosdorf等。（2010年），因为我们认为第一种分类过于简化，而第二种分类过于详尽。Alpine-Geo-LiM考虑了十种岩性：（1）“纯碳酸盐”，（2）“混合碳酸盐”，（3）“石膏蒸发岩，”（4个）“酸性岩石”，（5个）“镁铁质岩石”，（6个）“中级岩石”，（7个）“砂岩”，（8个）“粘土石”，（9个）“变质岩石”和（10个） ）“豌豆”。“纯碳酸盐”类别包括主要由方解石，文石（CaCO3）和白云石[MgCa（CO3）2]组成的岩石，例如石灰石，白云石和石灰华，以及大理石，原生岩是由碳酸盐岩组成的（Pettijohn，1957； Garrels和Mackenzie，1971； Boggs和Boggs，2009）。“混合碳酸盐”类别包括由碳酸盐矿物与非碳酸盐矿物混合而成的岩石。在该类别中，有不纯净的碳酸盐岩，钙钙石和泥灰岩（Pettijohn，1957年）。在“石膏蒸发岩”类别中，我们包括石膏和硬石膏。我们知道，一般而言，术语“蒸发物”是指硬石膏，石膏和石盐（Garrels和Mackenzie，1971年）。然而，由于分析的书目来源（参见附录A和C，以及其中的参考文献[脚注1]）排除了阿尔卑斯山中存在盐岩，因此在这项工作中，“石膏蒸发物”组中仅包括石膏和硬石膏。酸性岩”，“基性岩”和“中间岩”是根据（1）总碱-硅（TAS）图（Le Bas等，1986； Middlemost，1994）完成的，该图对许多常见类型进行了分类。从组合的碱（Na2O + K2O）和二氧化硅（SiO2）含量之间的关系开始，对火山岩进行分解；（2）对相同的岩体岩石图进行了改编（Le Bas等，1986； Middlemost，1994）。因此，我们将SiO2含量低于50％-52％的“基性岩”，SiO2含量在50％-52％和60％-62％之间的“中级岩”视为 和“酸性岩石”是指那些SiO2含量超过60％–62％的岩石。根据Mottana等人的分类，未包括在两个TAS图中的变质岩。（2009）。根据原生石的成分，将原片麻岩视为“酸性岩”（假设是花岗岩的原生岩成分），将蛇纹岩视为“镁铁质岩”（Mottana等，2009）。 ，我们包括了arkose，graywacke（Garrels和Mackenzie，1971）和砾岩，最后一个在来源和沉积机制方面与砂岩相似（Boggs和Boggs，2009）。此外，变质岩石英岩作为其原生质体属于“砂岩”类（Mottana等，2009）。在“粘土岩”类中，我们包括页岩，泥锌矿，粉砂岩和泥岩（Garrels和Mackenzie，1971； Gardrels和Mackenzie，1971； Mat。博格斯和博格斯，2009年），并再次考虑原石（Mottana等人，2009年），变质的千枚岩，片岩和paragneiss。仅当关于原石的信息不可用或不清楚（例如，最后，由于阿尔卑斯山中存在这些类型的沉积物，我们引入了进一步的岩性“岩浆”。由一种以上岩性组成的岩石对其分类提出了一些问题。Goldich（1938）引入了一系列风化的硅酸盐，Railsback（2006）对其进行了进一步的改性，添加了一些非硅酸盐矿物。根据该风化系列，碳酸盐的溶解度明显高于硅酸盐的溶解度，而石膏/硬石膏的溶解度高于碳酸盐的溶解度。为此原因，（1）在“石膏蒸发岩”类别中，我们还包括了由碳酸盐和石膏/硬石膏混合物组成的岩石，（2）在“碳酸盐混合岩”类别中，我们还包括了由一种以上岩性组成的岩石，其中至少有一个是由“混合碳酸盐岩”组成的（例如，由砂岩，graywacke和泥灰岩组成的岩性被认为是“混合碳酸盐岩”）。对于由多个岩性组成的硅酸盐岩，我们采用“流行原则”。我们根据参考地图属性表的不同字段中列出的最丰富的岩性对这些岩石类型进行了分类。例如，“基本岩石”类别中包括露头（矢量地图的多边形），其中不同的字段属性报告了玄武岩，曲折玄武岩和安山岩的存在，因为，按照我们的分类，玄武岩和曲折玄武岩可以被认为是“基本岩石”，而安山岩则被认为是“中间岩石”。在极少数情况下，露头是由“中间岩石”和“酸性岩石”（或“镁铁矿”）组成的岩石”）以同样的比例被认为是“酸性岩石”（或“镁铁质岩石”）。例如，由长石（“中间岩石”）和花岗岩（“酸性岩石”）组成的露头就是这种情况；它被认为是“酸性岩石”。附录C（请参见脚注1）中对高山地质进行了深入研究，以对特定的地质单元进行分类。新的Alpine-Geo-LiM是“中欧地质-岩性图”的一部分（Geo-LiM; Donnini et al。，2018），它以矢量格式发布，可以在网址（https://doi.org/10.com）上免费下载。5281 / zenodo.3530257，Donnini等，2018）。地图由12,001个多边形组成。地图中存在一些非常小的多边形（由于沿着研究区域的边界切割了地图）。最大的多边形面积约为11,197.5 km2，平均多边形大小约为16.5 km2。使用GRASS GIS（图2）进行构建地图的预处理（清除拓扑错误）和处理（联合，相交和分类）步骤（ Neteler和Mitasova，2008； Neteler等人，2012）（一种开源GIS软件）和PostgreSQL（PostgreSQL-http://www.postgresql.org），一种开源关系数据库管理系统（RDBMS），具有它的PostGIS空间扩展（PostGIS; http://www.postgis.org）。结果地图的属性表由以下10个字段组成：litho_irpi，rsil_mm，orig1，orig2，orig3，orig4，orig5，orig6，orig7和国家/地区。litho_irpi油田是由上述10种岩性类别之一组成的（“酸性岩”，“基性岩”，“中间岩”，“变质岩”，“砂岩”，“粘土岩”，“纯碳酸盐岩”，“混合岩性”碳酸盐岩”，“石膏蒸发岩”和“豌豆”）。orig1，orig2，orig3，orig4，orig5，orig6和orig7字段包含源自六个原始地质图（意大利，瑞士，德国，奥地利，法国，斯洛文尼亚；请参阅附录B [脚注1]）的原始地质信息，在国家/地区字段中包含国家/地区的名称。基于原始数据构建地图的命令行和查询与此处的地质岩性图一起提供。Alpine-Geo-LiM如图2所示。用于区分Alpine-Geo-LiM中不同岩性的颜色来自于美国地质调查局（USGS）为美国地质图所采用的岩性图例。图例和红绿蓝（RGB）代码可通过USGS在Web上获得（https://mrdata.usgs.gov/catalog/lithclass-color.php）。高山地区如表1所示，估计面积为197,773平方公里，与Donnini等人所定义的主要高山河流域（图3）相对应。（2016）。该表显示，碳酸盐岩是阿尔卑斯地区最丰富的类型，其中“混合碳酸盐”占23.75％，“纯碳酸盐”占20.82％，总计为44.57％。其次是“砂岩”（26.99％），“粘土石”（12.87％）和火山岩（占7。38％的“酸性岩石”，2.69％的“基性岩石”和0.43％的“中间岩石”，总计为10.50％）。“变质岩”占研究区域的1.81％，而“豌豆”和“石膏蒸发岩”占研究区域的不到1％（分别为0.48％和0.08％）。一小部分（2.69％）被湖泊和冰川形式的“水”覆盖。有关每种高露岩类型的丰度数据（表1中的面积百分比），与Donnini等人计算的百分比非常相似。（2016年），然而，它低估了粘土岩的百分比，并没有完全区分变质岩。查看表1中报告的标高和坡度值，并基于分辨率为25 m的欧洲数字标高模型（EU-DEM； Bashfield和Keim，2011； https：//www.eea.europa。eu / data-and-maps / data / copernicus-land-monitoring-service-eu-dem），我们发现“粘土岩”，“酸性岩”，“基性岩”，“变质岩”和“中间岩”具有最高的平均仰角（从1664 m到1993 m）和具有相对较低的标准偏差值的坡度值（从23.95°到28.32°）。这是由于以下事实：这些岩性构成了阿尔卑斯山最高山脉的晶体地块（例如Monte Bianco，4808 masl; Monte Rosa，4634 masl; Dent Blanche，4357 masl）。高平均海拔值（1662 m ）也与“水”（湖泊和冰川）相关，其标准偏差较高（1301 m）；此外，“水”与中等斜率值（12.55°）和相对较高的标准偏差（14.72°）相关。“这种较大的变化与以下事实有关：“水”类包括湖泊（通常位于山谷中）和冰川（位于高海拔地区）。“纯碳酸盐岩”的平均海拔和平均坡度分别等于1279 m和22.55°。阿尔卑斯山最高的石灰岩山海拔很高（例如，奥尔塔尔钙质阿尔卑斯山东南部的奥尔特斯，高3905马尔；格兰泽布鲁，高3857马尔；马拉莫拉达达，高3343马尔），这一事实证实了这一点。在北部钙质阿尔卑斯山的水位为3036马斯尔；在汝拉山脉的山脚的水位为1720马斯尔。表1显示，主要由不纯碳酸盐岩，钙钙石和泥灰岩组成的“混合碳酸盐岩”的平均仰角和坡度值分别为1157 m和18.38°。相对于其他岩性而言，归类为“砂岩”类别的岩石的高度和坡度相对较低（分别为698 m和8.50°，相关标准偏差分别为519 m。和10.27°）。这是由于以下事实：我们在岩性中包括了由于地块侵蚀而产生的砾岩和非胶结沉积物。“石膏蒸发岩”的平均仰角（603 m）和坡度（11.50°）较低，而“豌豆” ”（555 m； 2.14°）。最后的观察结果不足为奇，因为在植被平坦的湿地上，“豌豆”是由于植被退化而形成的（Bracco等人，2004年）。估算每个流域露头岩石的比例（表2）。我们考虑了Donnini等人定义的所有33个流域（图3）。（2016年），包括四个莱茵次流域（林斯，罗伊斯，阿尔卑斯山莱茵河和阿勒河）。冰川和湖泊在通过化学风化估算大气中的二氧化碳消耗量方面存在一些不确定性。安德森等。（1997年）和安德森（2005年）的研究表明，冰川增加了盆地内的径流，但并未增强硅酸盐的风化过程。关于湖泊，科尔等。（2007年）和Tranvik等。（Huttunen等人，2009年）（2009年）表明淡水（湖泊，河流和水库）在大气碳（CO2和CH4）的运输和生产中均起作用，证实了湖泊中大气CO2和CH4的产生。（2003），Del Sontro等。（2010），Diem等。（2012年）和Pighini等人。（2018）。在这项工作中 我们决定简化方法，就像考虑水文网络的一部分一样考虑冰川和湖泊。在高山-地质-LiM中，冲积山谷和阶地的沉积物被归类为“砂岩”类。在这里，由于构成分水岭上部的岩石的侵蚀，运输和沉积，砾岩和未胶结的沉积物就位了。结果，可以认为这些材料的岩性应反映上游地块的岩性。在33个流域中检查砂岩是否主要与碳酸盐岩或硅酸盐岩有关是否有意义。为了更好地研究“砂岩”类与其他岩性的联系，我们使用了著名的无监督k均值算法（Hartigan和Wong，1979年）在R软件中实现（R Core Team，2016年），并强加了四个集群。表3显示了代表这四个团簇中心的岩性百分比。集群1显示的“砂岩”百分比等于19.72％，与以下各项有关：（1）73.80％的碳酸盐岩（“混合碳酸盐”和“纯碳酸盐”碳酸盐”；（2）6.24％的“粘土岩”和火成岩和变质岩（“酸岩”，“基性岩”，“变质岩”和“中间岩”），以及（3）其他岩性的0.26％ （“豌豆”和“石膏蒸发岩”。）第2组显示出类似比例的“砂岩”（15.02％），而碳酸盐岩的比例相对较低（28.66％），而“粘土石”和火成岩的比例较高。变质岩（56.32％）。其他岩性（“豌豆”和“石膏蒸发岩”）可以忽略不计（0.01％）。在第3组中，“砂岩”百分比等于20.74％，碳酸盐岩为64.46％，“粘土岩”和火成岩和变质岩为14.75％，其他岩石为0.06％。第4组显示“砂岩”的浓度相对较高（40.79％）紧随其后的是碳酸盐岩（37.86％）和“粘土岩”以及火成岩和变质岩（20.29％）。至于其他类群，其他岩性所占的比例很低（1.06％）。分析表明，仅在类群2中，“砂岩”类与“粘土岩”以及火成岩和变质岩的比例较高。团簇（包括特征为“砂岩”浓度高的团簇4），“砂岩”与高比例的碳酸盐岩有关。由于“砂岩”岩性是由地块的侵蚀产生的，因此我们认为在研究区域中，（2016年）和2011-2012年水文年春季和冬季在流域出口附近采样的河水碱度（Donnini等人，2016年）。排放量数据来自不同来源，包括国际，国家，和地方当局（请参阅表4中的参考资料）。对于大多数集水区，数据以流量（m3 / s）的形式提供；对于只有测量站的阶段测量值（m）可用的集水区，通过使用流域当局提供的等级曲线或使用经验数据将阶段测量值转换为估计流量（m3 / s）。表4显示了两次采样活动中的流量显着不同，平均PD [Q]（w / s）为3.06％，变异系数等于23.79％。此外，表4显示，有22条河流（几乎占考虑河流的70％）在冬季流量最低，而在春季流量最高（PD [Q]（w / s）> 0），这是冰川和河流流量最大的河流。融雪为主的政权。Q（s）和Q（w）与Q（my）的比较突出了这样一个事实，即在两次采样活动时测量的几乎所有流量值都低于考虑到每日排放量而估算的年平均排放量在一年的水文中。相反，表4显示，在季节性测量中，碱度值相对于流速变化较小，平均PD [HCO3]（s / w）为17.38％，变异系数为1.40％。更一般地，似乎流速和碱度之间没有相关性。表5中显示了考虑了10个岩性类别而进行的线性多元回归分析得出的系数，其中b表示每种岩性的估计系数；标准误差测量b估计中的标准误差；P值表示b值偶然等于0的概率；显着性水平是表示分析可靠性的P值的文字分类。高b值与相关的HCO3−浓度相关，因此确定了更容易因化学风化而消耗大气CO2的岩性。相反，b值低表示来自相应岩性的化学风化作用降低了大气CO2消耗。小P值表示预测变量（[HCO3]）和响应（SR）变量之间的相关性较弱。高P值对应于分析的低显着性（表5中的显着性水平）。图6显示了两次采样活动中测得的高山河流的碱度值（观测到的碱度）与通过应用预测的同一高山河流的碱度值公式11（预测碱度）。测量值和拟合值之间的线性拟合（穿过零）的确定系数（R2）接近1（0.95），而残差的中位数，均值和标准偏差为–3.742×10−图5分别为–1.92×10−9和5.335×10−4，突出显示了该模型再现观测数据的能力。在表5中，岩性从最高b值到最低b顺序排列。“酸性岩石”和“中间岩石”具有很低的显着性和P值。这些结果显示出“皮亚特”的令人惊讶的行为，其标定参数b为正且非常显着。由于泥炭至少包含30％的有机物（Joosten和Clarke，2002），因此泥煤的溶解导致有机碳释放到大气中（例如Chow等人，2003； Bengtsson和Törneman，2004； Schwalm和Zeitz，2015; Selvam et al。，2017）; 因此，我们期望b为负值。这种差异可以通过以下事实来解释：“豌豆”的存在主要集中在伊萨尔盆地（盆地的6.84％），其中与“砂岩”（41.6％），“纯碳酸盐”（24.12％）相关，和“混合碳酸盐”（16.76％），即在二氧化碳消耗方面特别有效的岩性。表6显示了通过使用lm和nnl回归模型对岩性进行建模获得的b系数的值（R Core Team，2016）。我们观察到，在lm模型中，所有系数都是显着的（“砂岩”和“总碳酸盐”的显着性很高，“黏土”和“火成岩和变质岩”的显着性中等）。此外，我们注意到“火成岩和变质岩”的b值仍为负。该值与以下假设不符：酸，镁铁质，中间和变质岩都参与了CO2的消耗过程（方程1-4）。因此，在表6中，我们还显示了使用非负线性（nnl）回归分析获得的b系数的值。我们观察到，正如预期的那样，“火成岩和变质岩”的b值变为0，而“砂岩”和“总碳酸盐”的系数相对于通过lm回归获得的值没有显着变化。相反，使用nnl模型获得的“粘土”的b系数（6.30×10-4 mol L-1）相对于使用lm回归获得的值（2.00×10-3 mol L-1）较低。测得的浓度与lm回归的拟合值之间的拟合确定系数（R2）再次等于0.95，而残差的中位数，均值和标准偏差为–5.035×10−5，–分别为1.65×10-9和5.935×10-4。获得了nnl模型的模拟值（R2：0.94，残差中位数：5.943×10-5，残差平均值：1.453×10-5，残差标准偏差：6.187×10-4）。针对使用文献值，尤其是Amiotte Suchet等人估计的b值建立的模型，测试了使用四种岩性类别得出的线性模型。（2003）。特别地，我们假设b等于（1）“砂岩”等于1.52×10−4，（2）对于“粘土岩”等于6.27×10−4，（3）对于“总碳酸盐”等于3.17×10−3，而（4 ）“火成岩和变质岩” 9.50×10−5。使用Amiotte Suchet等人建立的线性模型之间的残差。（2003年）系数和测得的碱度的中位数，均值和标准偏差分别等于–6.74×10-4，–7.86×10-4和9.06×10-4。如预期的那样，使用Amiotte Suchet等人建立的模型。（2003）系数相对于使用四个岩性类别推导的拟议模型预测原始碱度值时精度较低（均值和中位数的绝对值较大），而精度（标准偏差较大）.RO为径流，SRi为岩性i覆盖的表面积的比例（从0到1），bi是岩性i的校准参数，a是硅酸盐岩情况下值为1且碳酸盐岩情况下值为0.5（请参见Eq 。9）。表7显示了化学风化消耗的大气CO2的通量ϕ（CO2）short，并通过应用公式12在盆地规模进行了估算。b系数的值是通过使用10种岩性进行的lm回归分析得出的（参见表5）。b值的显着性水平非常低的地方（即，对于“石膏蒸发物”，“变质岩”，“基性岩”，“酸岩”和“中间岩”（如表5所示），我们认为b等于0。 1）如果b的显着性水平很低，则表示b在统计上与0相同，并且（2）“镁铁矿岩”，“酸性岩”和“中间岩”的负b值不一致假设这些岩性参与了CO2的消耗过程（参见方程1-4）。然后考虑“砂岩”主要由硅酸盐岩（硅酸盐-砂岩情景）或碳酸盐岩（碳酸盐-砂岩情景）组成，计算CO2通量。在硅酸盐-砂岩场景中，a参数被认为等于0。仅纯碳​​酸盐和混合碳酸盐类别为5，其余岩石类别为1。在碳酸盐-砂岩情景中，对于纯碳酸盐，混合碳酸盐和砂岩类别，a参数被认为等于0.5，对于其余岩石类别，a参数被认为等于1。最后，根据Q（s），Q（w）和Q（my）计算RO值（请参见表4），从而得出每种情况的三组ϕ（CO2）S短路：ϕ（CO2）S（s） ，ϕ（CO2）S（w）和ϕ（CO2）S（my）。考虑到硅酸盐-砂岩情景，表7显示在春季，化学风化消耗的大气CO2通量为ϕ（CO2）。 S（s）的范围为3.93×104 mol km-2 yr-2（Durance）至3.71×106 mol km-2 y-2（Livenza）。通过使用Q（w）和Q（my）获得相似的值，因为：（1）ϕ（CO2）S（w）的范围为3.48×104 mol km-2 yr-2（Durance）至3。33×106 mol km-2 yr-2（Lech）和（2）ϕ（CO2）S（my）范围从7.73×104 mol km-2 yr-2（Durance）至5.26×106 mol km-2 yr。 -2（Isonzo）。同样，ϕ（CO2）S（s），ϕ（CO2）S（w）和ϕ（CO2）S（my）的平均值非常相似，分别为9.43×105 mol km-2 yr-2、9.81 ×105 mol km-2 yr-2和1.52×106 mol km-2 yr-2。考虑到碳酸盐-砂岩情景，系统地获得了最低的ϕ（CO2）短值，因为“砂岩”的参数与碳酸盐-砂岩情景相比，该参数被认为等于0.5，而碳酸盐-砂岩情景中的a参数被认为等于1。考虑到春季，ϕ（CO2）S（s）的范围为3.27×104 mol km-2 yr-2（Durance）至2.28×106 mol km-2 yr-2（Livenza）; 考虑到冬季，ϕ（CO2）S（w）从2.89×104 mol km-2 yr-2（Durance）变化到2。33×106 mol km-2 yr-2（Lech）; 考虑Q（my），）（CO2）S（my）范围为6.24×104 mol km-2 yr-2（Durance）至4.34×106 mol km-2 yr-2（Isonzo）。考虑到ϕ（CO2）S（s），ϕ（CO2）S（w）和ϕ（CO2）S（my）的平均值，获得了非常相似的值，分别为6.89×105 mol km-2 yr-2、7.05×105 mol km-2 yr-2和1.09×106 mol km-2 yr-2。在表7中，PD [ϕ（CO2）S（my）]（carb / sil）表示百分比分别根据方程10计算的，在考虑碳酸盐-砂岩情景和硅酸盐-砂岩情景的情况下计算的ϕ（CO2）S（my）的两个值之间的差。所获得的百分比差异表明，在碳酸盐-砂岩场景中，ϕ（CO2）S（my）值相对于在硅酸盐-砂岩场景中估计的通量平均为–26.94％，最小值为–41。Sesia的25％，Var的最大值–13.74％。化学风化消耗的大气CO2通量来自硅酸盐，from（CO2）S（my）-硅和碳酸盐、,（CO2）S（my ）-碳水化合物，是在两种情况下估计得出的，并在表8中进行了报告。在硅酸盐-砂岩场景中，ϕ（CO2）S（my）-sil值是在考虑“砂岩”和“粘土”风化的情况下估算的，而CO（CO2）S（my）-carb的值是考虑到“纯碳酸盐”和“混合碳酸盐”的风化作用而估算的。在碳酸盐-砂岩情景中，仅考虑“粘土岩”的风化作用估算ϕ（CO2）S（my）-sil值，而考虑“纯碳酸盐”的风化作用估算estimated（CO2）S（my）-carb值，“混合碳酸盐”和“砂岩”。相对于总ϕ（CO2）S（my），估算了两种情况下ϕ（CO2）S（my）-sil和ϕ（CO2）S（my）-carb的相对百分比。表8显示，考虑到硅酸盐-砂岩情景，硅酸盐风化的贡献为：（1）在八个流域（瓦尔，布伦塔，伊松佐，杜兰斯，明西奥，塔格里亚门托，皮亚韦和萨瓦）的25％至50％之间，（2）14个流域（伊萨尔，罗亚，伊泽尔，莱希，罗伊斯，阿尔卑斯莱茵河，伊勒，阿勒，阿迪杰，林思，恩斯，莱茵河，德拉和旅馆）的50％至75％之间；以及（3） 11个流域（罗纳，利文萨，多拉·巴尔蒂亚，塔纳罗，梅拉，阿达，奥格里奥，穆尔，波，假设盆地中发生的化学反应是方程1-6中所报告的，则通过考虑露头岩石的矿物学组成来选择Alpine-Geo-LiM的岩性类别。因此，根据原石的化学性质对变质岩进行了分类，所有无法获得或不清楚原石数据的岩石（例如，在蒙脱石，镍铁矿和变质沉积物的情况下）都包括在“变质岩”类中。 ”（占整个研究区域的1.81％）。与其他全球岩性图相比，该标准代表了一个新颖的特征（Gibbs和Kump，1994； Amiotte-Chut和Probst，1995； Amiotte-Chut等人，2003；Dürr等人，2005； Hartmann和Moosdorf，2012； Daniel等，2003）。 Moosdorf等，2010），其中在大气CO2消耗过程中行为截然不同的岩性包括在一般的“变质”类中。例如，大理石就是这种情况，大理石是由碳酸盐矿物组成的变质岩，极易消耗大气中的二氧化碳。大理石相对于其他变质岩具有非常不同的行为，例如正长石，它是一种从花岗岩/流纹岩原石衍生而来的变质岩，相对于大理石，该变质岩不易消耗大气中的二氧化碳。对于高山链条，根据原岩的化学性质对变质岩进行分类尤其令人关注，考虑到Hartmann和Moosdorf（2012）绘制的全球岩性图GLiM，变质岩非常丰富，占25.84％。整个区域。这项工作的另一个新颖特征是发布了带有用于生成地图的过程（GIS命令和数据库查询）的地图。我们决定共享这些信息，以实现研究的可重复性和可复制性，并遵循开放科学的概念（Nüst等人，2018）。根据Alpine-Geo-LiM，它表明碳酸盐岩是该地区最丰富的类型。高山地区（44.57％），其次是“砂岩”（26.99％），“粘土石”（12.87％），“火山岩”（10.50％），“变质岩”（1.81％），“豌豆”（0.48％）和“石膏蒸发物”（0.08％）。一小部分（2.69％）被湖泊和冰川形式的“水”覆盖。根据研究区中几乎所有的变质岩露头这一事实，证明了根据Alpine-Geo-LiM中原生质化学区分变质岩的努力（25。根据Hartmann和Moosdorf（2012）的研究，有84％被分配到一个特定的岩石类别，只有1.81％的研究区域仍属于一般的“变质岩石”类别，仅在无法获得或不清楚原生石的信息时才使用。该图强调了存在一个主要由结晶硅酸盐岩组成的内核，并在北部和南部被主要由碳酸盐组成的岩石束缚，最后在高山链外部的盆地中存在由砂岩组成的岩石，与Donnini等。（2016）和Rossi and Donnini（2018）。为研究流域的岩性组成与其水碱度之间的关系，遵循Hartmann等人的方法进行了三项线性多元回归分析。（2009）；参见此处的公式11。第一次分析使用线性多重（lm）回归分析工具（R Core Team，2016），并考虑了Alpine-Geo-LiM的原始10个岩性类别（表2）。由于某些岩性的稀缺性，使用线性多重（lm）和多重非负线性（nnl）回归分析（R Core Team，2016年）并考虑了四个岩性类别（“砂岩”，“ “粘土”，“总碳酸盐”和“火成岩和变质岩”）。将nnl回归分析（表6）获得的b值与考虑了整体岩性盆地的文献中的b值（Bluth和Kump，1994； Aviotte- Suchet等人，2003年），并考虑了多岩性盆地（Hartmann，2009年）。比较表明，在目前的工作中，文献值范围（1×10-3至8×10-4 mol L-1）中包括“总碳酸盐”（2.45×10-3 mol l-1）的校准参数b “粘土石”，其b值估计等于6.30×10-4 mol L-1，即文献值的数量级从2×10-4到9×10-4 mol L-1。相反，该比较还显示，“砂岩”的估计b值（4.50×10-3 mol L-1）明显高于文献值（6×10-4至6×10-5 mol L-1）。此外，从“砂岩”类别获得的b值始终大于为“纯碳酸盐”，“混合碳酸盐”和“总碳酸盐”计算的b值（请参见表5和6）。前面提到的聚类分析结果（表3）解释了这种较大的差异，强调指出，在研究领域中，“砂岩”类可能由相关的碳酸盐成分组成。附录A中解释了高山前陆中有关碳酸盐组分的存在，并且在文献中对此也有充分的说明（例如，参见莫拉西盆地：Schlunegger等，1994，1998； Kempf等，1999）。 ； Anne等人，2017； Abdul Aziz等人，2008；例如，对于Po谷和亚得里亚海前陆：Fontana等人，2014）。“砂岩”类别的高b值也可以通过在“砂岩”类别中包括最近的冲积沉积物来解释。实际上，除风化作用作用于土壤-空气界面外，它们还因地下水而暴露于化学溶解下，这有助于盆地的水流。此外，在冲积沉积物中，通常位于平坦和低海拔地区，水在土壤-空气界面中的停留时间增加，从而促进了化学溶解过程。从表1的分析中可以明显看出“砂岩”类中有大量冲积沉积物。“砂岩”的坡度和海拔平均值确实比其他岩性观察到的要低得多。因此，结果表明，碳酸盐（以“纯碳酸盐”，“混合碳酸盐”和“总碳酸盐”的三种形式）与水的碱度具有很强的正相关性。令人惊讶的是，结果还表明，“砂岩”的相关性甚至更高。可以考虑以下事实来解释这一事实：（1）“砂岩”类别包括也由砾石和碳酸盐沉积物组成的胶结和非胶结沉积物；（2）在分析盆地中，“砂岩”与“纯碳酸盐”和“混合碳酸盐”岩石相关（表3）。（有趣的是，“黏土”类（露头面积约13％）与水的碱度总是呈正相关；但是，根据回归类型的不同，它可以较低（10个岩性分类，lm； 4个分类，nnl）或较高（4个分类，lm）。除了通过4类lm回归估计的中值之外，与系数值相关的显着性较低。我们通过研究区域中存在的其他富含碳酸盐岩性来解释这种行为，这些岩性掩盖了“粘土岩”对CO2消耗的影响。此外，我们观察到“粘土”的b系数的正值不仅可能是由于硅酸盐风化引起的，而且还涉及碳酸盐的化学溶解，这种碳酸盐可能存在于“粘土”类岩石中。该地区几乎没有“火成岩和变质岩”（约11％），它们与水碱度的关系始终为负或等于零（见表5和6）。“火成岩和变质岩”系数的显着性通常很低。在4类lm回归的情况下，它也很低，导致中等显着性，相应的P值大于从其他系数获得的P值。因此，本研究表明火山岩（酸，镁铁质和中间岩）对大气CO2消耗的贡献可忽略不计。相反，在文献中显示，这些岩性的b值范围为1.5×10-4至4.5×10-6 mol L-1，确实为大气中的二氧化碳消耗做出了贡献（即使很小）。这种不同的行为当然是由于以下事实：由硅酸盐矿物构成的火山岩在整个研究区域中所占的比例很小（约10％）。此外，我们认为“火成岩和变质岩”类别的b值为负值或为零（参见表6）是由于以下事实：b是在这些岩性与更丰富的碳酸盐矿物相关的盆地中估算的，而碳酸盐矿物含量更高。比硅酸盐可溶（表2）。这与Bluth和Kump（1994）和Amiotte-Cheet等人（1）所做的不同。（2003年）考虑了整体岩性盆地，（2）Hartmann（2009年）考虑了多岩性盆地，但排除了含碳酸盐沉积岩超过0.05％的盆地。关于“变质岩”类，其结果（极低的显着性水平；见表5）不同于其他作者的结果（Bluth和Kump，1994； Amiotte-Cheet等，2003； Hartmann，2009）。可以用以下事实来解释：我们使用自己的分类方案来定义岩性图（随本手稿一起提供）。例如，我们认为，基于原石的分析，将某些类型的岩石包括在或排除在变质类中，可能会对估计此类对CO2消耗的贡献产生相关影响。我们得出的结论是，在该研究区中，这些岩性（“火成岩和变质岩”）对大气CO2消耗过程没有显着贡献。考虑方程式9，可以从河水的碱度开始估算化学风化所消耗的大气CO2量。因此，公式11的校准参数b表示通过化学风化不同岩性消耗大气CO2的能力（b的高值表示消耗大气CO2的能力高）。因此，分析表明，更容易消耗大气CO2的岩性从高到低依次为：“砂岩”，“碳酸盐”和“粘土岩”，而“火成岩和变质岩”的贡献可忽略不计。 ，我们应用公式12来估算研究区域内“短期” ϕ（CO2）短时化学风化消耗的大气CO2通量。考虑“砂岩”来计算通量，其中“砂岩”主要由硅酸盐岩（硅酸盐-砂岩情景）和（2）碳酸盐岩（碳酸盐-砂岩情景）组成。在公式12中，我们考虑了使用10种岩性从lm回归分析获得的b值（请参见表5）。对于“石膏蒸发物”，“变质岩”，“基性岩”，“酸性岩”和“中间岩”，b的显着性水平很低（参见表5），我们认为b等于0在硅酸盐-砂岩情景中，“纯碳酸盐”和“混合碳酸盐”类别的a参数被认为等于0.5，而其余岩石类别则被视为1。在碳酸盐-砂岩情景中，“纯碳酸盐”，“混合碳酸盐”的a参数被认为等于0.5，”和“砂岩”类别，其余岩石类别则为1。考虑两个采样活动（春季和冬季）时的日排放量，并考虑年均排放量Q（s），Q（w）和表4中的Q（my）如预期的那样，在碳酸盐-砂岩场景中，ϕ（CO2）短值系统地低于在硅酸盐-砂岩场景中，因为此场景中“砂岩”的参数被认为是相等的到0.5，与碳酸盐-砂岩情景相反，其中a参数被认为等于1。考虑到碳酸盐-砂岩情景和硅酸盐-风化情景（参见等式10），估计的年平均通量之间的百分比差为平均26.99％，最小值为–41。Sesia为25％，Var为最大值–13.74％（请参见表7）。比较硅酸盐风化，ing（CO2）S（my）-sil和碳酸盐风化的贡献（1）表8显示了两种情况（硅酸盐-砂岩和碳酸盐-砂岩方案）中的ϕ（CO2）S（my）-carb。如预期的那样，该表显示，考虑到碳酸盐-砂岩方案，硅酸盐的贡献风化占ϕ（CO2）S（my）总量的14.51％，而在硅酸盐-砂岩情景中，风化增加至64.72％。显然，硅酸盐对ϕ（CO2）的贡献实际上取决于对砂岩岩石化学成分的假设。在目前的工作中，结果表明，在研究区域中，砂岩含有碳酸盐岩的相关成分。我们研究结果的普遍含义是，如果不仔细评估砂岩类所含岩石类型的碳酸盐含量，就无法估算砂岩岩石相对丰富的地区的CO2消耗量。因此，应注意选择文献（对于“砂岩”类）采用的系数，以计算与给定参数本身所使用的区域不同的任何给定区域中的二氧化碳通量。此处给出的结果适用于无污染地区，温带气候和不含黄铁矿的岩性（请参见“风化估算”部分）。黄铁矿的缺乏很重要，因为正如Moon等人所强调的那样。（2007年），黄铁矿氧化生成硫酸，可能会侵蚀周围的碳酸盐和硅酸盐矿物。由于在此过程中不消耗大气中的CO2，因此不考虑黄铁矿的氧化会导致硅酸盐风化对大气中CO2消耗的高估。由于黄铁矿的氧化和石膏的溶解都会导致河水中SO4-的增加（例如，Berner和Berner，1996年），因此区分河水中的SO4-的石膏和黄铁矿源对于可靠地估算水的通量很重要。化学风化消耗的大气中二氧化碳（Moon等，2007）。我们知道阿尔卑斯山中黄铁矿的存在已被充分记录（例如，Kappler和Zeeh，2000;Lavrič和Spangenberg，2003; Rantitsch，2007; Gainon等，2007; Grachev等，2008; Bernard等。 ； 2010； Herlec等人，2010； Sanders等人，2010； Sabatino等人，2011; Pálfy和Zajzon，2012年）。此外，在Donnini等人采样的河流水域中。（2016年），具有相关硫酸盐富集的样品位于法国西南部的阿尔卑斯山（Roia，Var，Isere，杜兰斯）和意大利东南部的阿尔卑斯山（Tagliamento）。由于三叠纪碳酸盐岩中石膏的存在在法国西南部的阿尔卑斯山（例如，Olivier等人，2009年）和意大利东南部的阿尔卑斯山（例如，Stefanini，1976年； Longinelli和Flora，2007年）都有充分的文献记载，我们认为考虑到这些蒸发水中溶解而不是黄铁矿氧化的结果，认为这些河流水中的SO4-富集更为合理。因此，我们认为简化（认为黄铁矿氧化可以忽略不计）只会导致对大气中CO2消耗量的高估。通过这项工作获得的由化学风化作用固定的大气CO2通量与文献中获得的通量的比较表明，区域范围内存在细微的差异。特别是，本工作中获得的ϕ（CO2）短值与Donnini等人估计的ϕ（CO2）短值的比较。（2016）对同一地区的研究表明，在33个主要的高山流域内，化学风化作用固定的大气CO2的范围非常相似。（2016），在春季，ϕ（CO2）短范围从2.60×104 mol km-2 yr-2（Durance）到2.03×106 mol km-2 yr-2（Livenza），而在冬季， ϕ（CO2）短范围从2.48×104 mol km-2 yr-2（Durance）到2.04×106 mol km-2 yr-2（Lech）。在这项工作中，在春季，硅酸盐风化情景，ϕ（CO2）S（s）的范围是3。93×104 mol km-2 yr-2（Durance）至3.71×106 mol km-2 yr-2（Livenza），在碳酸盐岩风化场景中，其变化范围为3.27×104 mol km-2 yr-2（杜兰斯）至2.28×106 mol km-2 yr-2（Livenza）。在冬季，在硅酸盐风化的情况下，ϕ（CO2）S（w）的范围从3.48×104 mol km-2 yr-2（Durance）到3.33×106 mol km-2 yr-2（Lech），在碳酸盐岩风化情景中，其变化范围为2.89×104 mol km-2 yr-2（Durance）至2.33×106 mol km-2 yr-2（Lech）。考虑到年均排放量，在硅酸盐风化的情况下，ϕ（CO2）S（my）的范围从7.73×104 mol km-2 yr-2（Durance）到5.26×106 mol km-2 yr-2（Isonzo）在碳酸盐岩风化的情况下，范围从6.24×104 mol km-2 yr-2（Durance）到4.34×106 mol km-2 yr-2（Isonzo）。work（CO2）S（s），ϕ（CO2）S（w）和ϕ（CO2）S（my）的平均值在当前工作中估算的平均值ϕ（CO2）之间的差在硅酸盐风化情景中，×105 mol km-2 yr-2、9.81×105 mol km-2 yr-2和1.52×106 mol km-2 yr-2，以及6.89×105 mol km-2 yr-2 ，在碳酸盐岩风化的情况下为7.05×105 mol km-2 yr-2和1.09×106 mol km-2 yr-2，Donnini等人估算的平均ϕ（CO2）短。（2016）（春季为4.69×105±1.03×105 mol km-2 yr-2，冬季为5.35×105±0.97×105 mol km-2·yr-2）小于一个数量级。 Gaillardet等人估计了相似的值。（1999）对莱茵河，罗纳河和波河盆地分别为5.42×105 mol km-2 yr-2、8.56×105 mol km-2 yr-2和1.12×106 mol km-2 yr-2，表明这三条河流的平均short（CO2）短缺量远高于同一作者估算的化学风化所消耗的世界平均CO2（2.46×105 mol km-2 yr-2）。由于此处给出的新b参数仅使用实测数据获得，因此此处给出的高山地区CO2消耗速率的估算比使用文献值估算的速率更为客观。本研究的结果强调了根据岩石来区分岩石的重要性。在其矿物学组成方面，要密切注意通常被认为是硅酸盐（如变质岩）为主的岩石类别中次要碳酸盐成分的存在，正如Hartmann等人所强调的那样。（2009年），就像砂岩和页岩（在本工作中被称为黏土）。事实上，众所周知，这些岩性可能含有少量碳酸盐（例如Jacobson和Blum，2003年; Emerson等人，2018年）。Hartmann等人强调了碳酸盐对以硅酸盐为主的岩石类别的大气二氧化碳消耗的不可忽视的贡献。（2009年，第189页），他指出，在全球范围内，“碳酸盐CO2消耗量的大约12.6％可以归因于以硅酸盐为主的岩性类别。” 同一作者强调指出，碳酸盐沉积岩的全球贡献在过去被高估了，在Hartmann等人的研究中约为25％。（2009年），相比之下，Gaillardet等人约为40％。（1999），Munhoven（2002）和Amiotte Suchet等。（2003）.Alpine-Geo-LiM是阿尔卑斯山的高分辨率（比例为1：1,000,000）的地质岩性图。与已发布的全球岩性图相比，它代表了一种新颖的图（Gibbs和Kump，1994; Amiotte-Sutt和Probst，1995; Amiotte-Sutt等，2003;Dürr等，2005; Hartmann和Moosdorf，2012; Moosdorf等（2010）。主要有两个原因。首先，通过考虑露头岩石的矿物学组成来选择用于绘制研究区域的岩性类别。特别注意区分变质岩，这些变质岩是根据原石的化学性质分类的。“变质岩”类仅包括原始石块数据不可用或不清楚的岩石。通过多岩性方法并通过基于岩性的线性多元回归预测水碱度，估算了不同岩性在化学风化作用下对大气CO2消耗的作用。这些分析证实，碳酸盐岩是极易消耗大气二氧化碳的岩性，如几位作者先前所言（Bluth和Kump，1994； Amiotte-Sutt等人，2003； Hartmann，2009）。目前的工作还表明，“砂岩”类别包括石英岩，还有黑糖，灰瓦克和砾岩，可能具有不可忽略的碳酸盐成分（Garrels和Mackenzie，1971； Mottana等，2009），并发挥了重要作用。消耗大气中的二氧化碳。此外，线性多元回归分析表明，火成岩在大气CO2消耗中的贡献可忽略不计。第二个新特点是Alpine-Geo-LiM以矢量格式以及用于定义地图的程序和原始程序被释放。数据以允许产品的可复制性和再现性（见Donnini等，2018）。Donnini得到了Generali Fondazione Assicurazioni基金会的资助，A。Zucchini得到了佩鲁贾大学物理与地质系Paola Comodi，Francesco Frondini和Diego Perugini的研究项目的部分支持。Donnini主要为应用地球化学模型和验证地图相对于高山地质学I的准确性做出了贡献。马尔凯西尼（Marchesini）主要为地理和统计业务做出了贡献，而西葫芦（A. Zucchini）主要为矿物岩石学方面的考虑做出了贡献，这些考虑对详细说明地图的岩性分类非常有用。M. Donnini撰写了该论文，I。Marchesini和A. Zucchini随后进行了内部修订。Alpine地质岩性图（Alpine-GeoLiM）代表了中欧地理岩性图（Geo-LIM）的一部分，于2003年发布。 GPKG和Donnini等人的PDF格式。（2018）-https://doi.org/10.5281/zenodo.3530257。连同地图一起，我们提供：（1）用于创建地图的德国，意大利，斯洛文尼亚，法国，瑞士和奥地利的原始国家地质图；（2）一个可用于复制原始地图的分类和并集的脚本。我们感谢两位审稿人，