Lithologic composition of the Earth's continental surfaces derived from a new digital map emphasizing riverine material transfer

Authors


Abstract

[1] A new digital map of the lithology of the continental surfaces is proposed in vector mode (n ≈ 8300, reaggregated at 0.5° × 0.5° resolution) for 15 rock types (plus water and ice) targeted to surficial Earth system analysis (chemical weathering, land erosion, carbon cycling, sediment formation, riverine fluxes, aquifer typology, coastal erosion). These types include acid (0.98% at global scale) and basic (5.75%) volcanics, acid (7.23%) and basic (0.20%) plutonics, Precambrian basement (11.52%) and metamorphic rocks (4.07%), consolidated siliciclastic rocks (16.28%), mixed sedimentary (7.75%), carbonates (10.40%), semi- to un-consolidated sedimentary rocks (10.05%), alluvial deposits (15.48%), loess (2.62%), dunes (1.54%) and evaporites (0.12%). Where sediments, volcanics and metamorphosed rocks are too intimately mixed, a complex lithology (5.45%) class is added. Average composition is then tabulated for continents, ocean drainage basins, relief types (n = 7), 10° latitudinal bands, geological periods (n = 7), and exorheic versus endorheic domain and for formerly glaciated regions. Surficial lithology is highly heterogeneous and major differences are noted in any of these ensembles. Expected findings include the importance of alluvium and unconsolidated deposits in plains and lowlands, of Precambrian and metamorphic rocks in mid-mountain areas, the occurrence of loess, dunes and evaporites in dry regions, and of carbonates in Europe. Less expected are the large occurrences of volcanics (74% of their outcrops) in highly dissected relief and the importance of loess in South America. Prevalence of carbonate rocks between 15°N and 65°N and of Precambrian plus metamorphics in two bands (25°S–15°N and north of 55°N) is confirmed. Asia and the Atlantic Ocean drainage basin, without Mediterranean and Black Sea, are the most representative ensembles. In cratons the influence of ancient geological periods is often masked by young sediments, while active orogens have a specific composition.

1. Introduction

[2] Surficial lithology, i.e., the nature of rocks, a key descriptor of the surface of the Earth System together with temperature, precipitation, runoff, land cover, soil type, relief, is still incompletely mapped at the global scale. Current geological maps provide partial information on rock types, for example, crystalline rocks which represent about one third of the Earth's surface but fail to differentiate the other two thirds that are occupied by sedimentary rocks [Garrels and Mackenzie, 1969, 1971; Blatt and Jones, 1975; Meybeck, 1987; Amiotte-Suchet et al., 2003] into different rock types such as sandstone, carbonate rocks or evaporites. Information on chemical and physical properties of surficial rocks is needed in many fields of Earth system science. Lithology is a key control of river chemistry, and therefore of land to oceans ion fluxes [Holland, 1978; Meybeck, 1987; Berner and Berner, 1996], of river sediment transport to oceans [Jansen and Painter, 1974; Milliman and Meade, 1983; Milliman and Syvitski, 1992; Ludwig and Probst, 1998], and its associated load of particulate carbon [Ludwig et al., 1998, 1999]. Surficial lithology is also an essential control of groundwater resources [Zektser and Loaiciga, 1993; Dzhamalov et al., 1999] and understanding coastal ribbon lithology is important to locate and characterize the direct groundwater inputs to oceans through karst or alluvial aquifers [Zektser et al., 1973; Buddemeier, 1996; Burnett et al., 2003; Slomp and Van Cappellen, 2004].

[3] The demand for global-scale lithology information has greatly increased in the last decade owing to the development of GIS tools and global-scale approaches in Earth system science [Steffen et al., 2004]. In physical geography, lithology has long been recognized as a key factor in mechanical erosion control and chemical weathering intensity [Birot, 1970; Meybeck, 1987; Summerfield, 1991, 2000; Ludwig, 1996; Ludwig et al., 1996]. Geochemists need lithology data to define the chemical composition of water in pristine conditions [Drever, 1997; Bluth and Kump, 1991, 1994; Gaillardet et al., 1999; Meybeck, 1987, 2003], to study carbonate rock dissolution and the carbon cycle at a global scale [Amiotte-Suchet, 1994; Probst et al., 1994; Amiotte-Suchet et al., 2003], to interpret the chemical and mineralogical composition of river particulate solids [Potter, 1978; Gaillardet et al., 1999]. Earth system scientists need to know the lithological composition of the Earth's surface and its related riverine fluxes now and in key periods of past geological times such as the Late Glacial Maximum [Bluth and Kump, 1994; Gibbs and Kump, 1994; Gibbs et al., 1999] and at longer geologic timescales [Galy and France-Lanord, 1999], particularly to reconstruct the past CO2 evolution which is thought to be controlled by silicate rock weathering and its related transfer of dissolved inorganic carbon to oceans [Berner et al., 1983; Gaillardet et al., 1999; Millot et al., 2002]. Present-day lithology is also of interest to geologists and geochemists studying the continuous recycling of rocks, particularly sedimentary rocks [Garrels and Mackenzie, 1969, 1971; Ronov, 1972, 1976; Ronov and Yaroshevskiy, 1976; Veizer and Jansen, 1979; Probst et al., 1994; Veizer and Mackenzie, 2003].

[4] The first lithologic maps were generated at coarse resolutions by global-scale geochemists working on river chemistry and/or the carbon cycle and to estimate coastal erosion [Bluth and Kump, 1991; Gibbs and Kump, 1994; Amiotte-Suchet et al., 2003]. The transcription of existing geological maps into a lithological map highlighting riverine transfers faces numerous difficulties. (1) “Hard” rocks such as plutonic, volcanic and metamorphic rocks are marked as such on geological maps, whereas sediments and sedimentary rocks are mostly grouped according to their ages. (2) Information on sediment lithology is often available only in the legends of local geological maps or in regional descriptions of sedimentary facies; if the information exists, it is generally provided on large-scale maps and difficult to digitize. (3) Surficial rocks are often an intimate mix of layers of different lithologies (e.g., siliciclastics and carbonates) or occur as very thin layers such as loess. (4) Lateral transitions from one rock type to another (“facies change”) can occur over short distance within the same geological formation, particularly in sedimentary rocks. (5) A secondary fine-scale mix can be achieved during orogenic processes through folding, imbrication, and flattening “condensation.” (6) Finally, transformation by metamorphic processes leads to very complex rock types.

[5] Our digital map is derived from multiple sources and observations that are valid only for the present-day period: Its aim is to describe the presently outcropping rock types, i.e., at the present sea level and with the present ice caps in Greenland and Antarctica. The main differences with previous work [Amiotte-Suchet, 1994; Gibbs and Kump, 1994; Dzhamalov et al., 1999; Amiotte-Suchet et al., 2003] are: (1) the map is in vector mode (about 8 300 polygons) while others were in raster mode, (2) a finer resolution in raster mode at 30′ (and ultimately down to 5′ or 6′) versus 1° or 2° for previous maps, (3) a finer description of rocks of 15 types plus ice and water vs. 4 to 6 for previous attempts, and (4) particular attention to the sensitivity of rock types to weathering and erosion.

[6] As in some of the previous attempts we provide global averages as well as means for the continents, the areas draining into different oceans (“ocean drainage basins”), and the lithologic composition of latitudinal bands and of major relief types. Furthermore with our database sedimentary rocks and their respective ages can be compared, and we estimate the composition of surficial rocks affected by Quaternary glaciations.

[7] This work is part of a joint effort by the University of Paris 6 and the University of New Hampshire (UNH) to gradually describe and digitize at a fine resolution (30′ × 30′ minimum) the Earth's surface features such as the river networks [Vörösmarty et al., 2000a, 2000b], the global relief [Meybeck et al., 2001], the world runoff [Fekete et al., 2002], the large reservoir dams [Vörösmarty et al., 1997], and the future of water resources [Vörösmarty et al., 2000c] and it is fully compatible with these related databases.

[8] Details on data sources and data analysis can be found in the auxiliary material (Annexes A to J).

2. Analysis of Previous Global Lithology Estimates

[9] There are very few global lithology maps or data sets, even fewer in digital format, and few that fulfill our requirements of a detailed map. Some publications on global geomorphology [e.g., Snead, 1980; Summerfield, 1991, 2000] and on the natural regions of the globe [e.g., Birot, 1970] include information and maps of the lithology but are rarely available in digital format. Existing maps include the digital lithologic one by Amiotte-Suchet [1994], published by Amiotte-Suchet et al. [2003], although this map was not available at the beginning of our study, Gibbs and Kump [1994], the UNESCO map of Hydrogeological conditions and groundwater flow [Dzhamalov et al., 1999], a few lithological maps at the continental scale published with the FAO soil map [Food and Agriculture Organization/U.N. Educational, Scientific, and Cultural Organization (FAO/UNESCO), 1975, 1986]. Another source of information, widely used as reference in the literature, is provided by the studies of global proportions of exposed rocks by Ronov and coworkers [Ronov, 1972, 1976; Ronov and Yaroshevskiy, 1976], Blatt and Jones [1975], reassessed in some global river budgets [Meybeck, 1987]. The main characteristics of these attempts are presented in Table 1. The detailed discussion of these works is given in Annexe A (see auxiliary material).

Table 1. Comparison of Previous Lithology Maps at Global Scale and Background Material Used in the Lithologic Map of the World
ReferenceType of InformationFormataResolution/ScaleSubdivision/Detail
Ronov and coworkers,bBlatt and Jones [1975]proportions of exposed rockspercentvaryingvarying
 
FAO/UNESCO [1975]lithology maps for each continentPbetween 1:20,000,000 and 1:40,000,000 (A4 format)variable rock types, including: unconsolidated sediments, loess, shifting sand, consolidated clastic sediments, carbonates, metamorphic rocks, acid and basic intrusive and extrusive rocks
 
FAO/UNESCO [1975]FAO Soil Map of the WorldV1:5,000,000four layers with varying detail
 
Snead [1980]bWorld Atlas of Geomorphic FeaturesPabout 1:100,000,000six rock types: ancient metamorphic and associated intrusive igneous rocks, well-consolidated sedimentary rocks, weakly-consolidated or unconsolidated sedimentary rocks, recent alluvium, extrusive igneous rocks: fine-grained ashy/glassy, mixed rock types in areas of complex folds and faults
 
UNESCO [Choubert et al., 1980]Geological Atlas of the WorldP1:10,000,000 (22 sheets)as Geological Map of the World, sometimes more detailed depending on the chosen sheet
 
Datong et al. [1985]Map of Soluble Rocks of ChinaP1:4,000,000different soluble rock types, distinguished according to ages
 
Ford and Williams [1989]World Map of Carbonate outcropsP1:100,000,000very general map of surface and subsurface karst regions
 
UNESCO [Dottin, 1990]Geological Map of the WorldP1:25,000,000 (4 sheets)volcanic rocks of different ages, plutonic rocks, metamorphic, Precambrian, sedimentary rocks of different ages
 
Gibbs and Kump [1994]map of global lithology for 18 ka and the present dayG2° × 2° grid cellssix rock types: carbonates, shales, sandstones, extrusive igneous rocks, shields, “complex lithology”
 
Dzhamalov et al. [1999]UNESCO Map of Hydrogeological Conditions and Groundwater FlowVvarying by latitude between 1:3,000,000 and 1:12,000,000four aquifer types: sedimentogenic-pore, sedimentogenic-fracture, karst, magmatogenic-metamorphogenic
 
Amiotte-Suchet [1994], Amiotte-Suchet et al. [2003]cGlobal Lithology MapG1° × 1° grid cellssix rock types: sandstones, carbonates, shales, plutonic and metamorphic rocks, acid volcanic rocks, basalts
 
This workGlobal Lithology MapVminimal size of polygon ∼ 10 km15 rock types plus water and ice

[10] Some geomorphology and geology work offers specific, but often very coarse, lithological information at the global scale, for loess [Summerfield, 1991], and for carbonate rocks [Ford and Williams, 1989]. Other studies focus on the local scale (105–106 km2) as in New Zealand, the Japanese Islands and Tibet [Summerfield, 1991] or on the Andes [Summerfield, 2000] and distinguish various outcrops of rock types such as basalt, metamorphic rocks, uplifted basement rocks, and soluble rock types in China [Datong et al., 1985].

[11] In conclusion we consider that previous global lithologic maps and studies are not detailed enough for river basin descriptions [e.g., Snead, 1980]. Very detailed regional descriptions are mostly qualitative [e.g., Birot, 1970]. Although some regional maps are now available [e.g. European Soil Bureau, 1998], they do not exist worldwide. Our work combined various sources (e.g., the FAO lithological maps [FAO/UNESCO, 1975, 1986]), and previous attempts [e.g., Snead, 1980; Gibbs and Kump, 1994] with multiple regional and local sources.

3. Toward a Vectorized Lithologic Map of the World

3.1. Identification and Clustering of Major Rock Types

[12] The major rock types, magmatic, metamorphic, sedimentary, can actually be divided into dozens of second-order and third-order rock types, as done in petrology. In our global mapping we are forced to simplify and to reduce the number of categories, introducing artificial clusters and groupings where specialists will see major differences. Occasionally we focus on specific rock types such as loess and evaporites for their exceptional properties with regard to mechanical erosion and chemical weathering.

[13] Our analysis led us to select 17 distinct rock classes, water and ice included, not based on the conventional classification of rocks (“geo”-oriented lithology), but on the sensitivity of rocks to chemical and mechanical weathering and their eventual transport by surface water (“hydro”-oriented lithology). Magmatic rocks were subdivided into four classes according to their plutonic/volcanic origin and to their basic/acid chemical composition. Precambrian and metamorphic are differentiated in two classes, sedimentary rocks and sediments in eight classes. Large water bodies and ice are also tabulated.

3.1.1. Basic-Ultrabasic Plutonic Rocks

[14] Basic-ultrabasic rocks (Pb) include mainly peridotites and gabbros, and generally occur in ophiolite complexes, now found in collisional orogens; a few others consist of layered intrusions (lopoliths). Actually, only the biggest outcrops could be mapped. Moores and Twiss [1995] provide an overview of these formations.

3.1.2. Acid Plutonic Rocks

[15] Acid plutonic rocks (Pa) comprise mainly granites, granodiorites and quartz-diorites, locally also diorites, possibly some orthogneisses.

3.1.3. Basic and Intermediate Volcanic Rocks

[16] Basic and intermediate volcanic rocks (Vb) include basalts of various genetic origins (continental rifting and ocean spreading processes; mantle plumes generating flood basalts: the most voluminous basalts on the continents) as well as basic and intermediate volcanic rocks produced by subduction of the oceanic crust in plate collision settings (basalts, basaltic andesites and andesites), most occurrences belonging to the circum-Pacific fire belt. Basic tephra, tuff and ashes are also in this class.

3.1.4. Acid Volcanic Rocks

[17] Acid volcanic rock (Va) occurrences on our map are generally ignimbrites, since real rhyolites are too small to be mapped. These rhyolitic pyroclastics can cover huge areas, of hundreds or thousands of km2, with a thickness of tens of meters. Here acid volcanic rocks are differentiated from other volcanic rocks owing to their different water chemistry (e.g., Ca2+/Mg2+ ratio [Meybeck, 1987]).

3.1.5. Precambrian Basement

[18] Precambrian basement (Pr) includes rocks from the middle and lower crust in the Precambrian cratons constituting the old “cores” of the continents. These rocks are medium to highly metamorphic and of predominantly granodioritic-granitic character, mainly gneiss (with few exceptions). Nonmetamorphic (post-orogenic, platform-) sediments of Proterozoic age were treated separately wherever feasible, for example, in Canada. They were assigned to the corresponding groups of sedimentary rocks.

3.1.6. Metamorphic Rocks

[19] Metamorphic rocks (Mt) are mainly middle-grade regional metamorphic rocks not attributed to the craton regions as Precambrian basement rocks. The majority belong to the deeply eroded basement of the cratons, often with migmatic gneisses. Minor non-Precambrian metamorphic rocks occur in younger, Phanerozoic orogens throughout the globe. The lithologic overlap of the Precambrian basement with the metamorphic rocks class exists in many areas (e.g., Canadian Shield) and is discussed further. Marbles, i.e., highly metamorphosed pure limestones, are quite rare and their outcrops are usually too small to be shown on global-scale maps. Mappable outcrops are included in this rock type. Occasional low-grade or high-grade metamorphism was clustered in this category. Metamorphic rocks in young Alpine mountain ranges were attributed to the next category.

3.1.7. Complex Lithology

[20] “Complex lithology” (Cl) is a mixed rock class first introduced by Bluth and Kump [1991] and used by Gibbs and Kump [1994]. This class contains the intimately mixed rock associations in the inner zones of young and medium-aged orogens. These rocks are mainly sedimentary, volcano-sedimentary and volcanic, most of them tightly folded and flattened and/or involved in duplex formations by thrusting, on a meter to a few kilometers scale. High grade metamorphic and plutonic rocks can also be found in this class, for example, through thrusting and duplex formation where the strength of the orogeny is responsible for the mixing. Therefore the complex lithology occasionally includes carbonate rocks and marbles.

3.1.8. Sedimentary Rocks

[21] Sedimentary “rocks” were differentiated into eight classes. They span the whole range of nonconsolidated, semi-consolidated and consolidated sediments as well as noncarbonated to pure carbonate rocks. The limits between and within these classes were sometimes difficult to draw. Four types of consolidated sedimentary rocks were distinguished.

3.1.8.1. Siliciclastic Sedimentary Consolidated Rocks

[22] Siliciclastic rocks (Ss) include, according to the grain-size classification in pelites – psammites – psephites, clay- to mud- and siltstones, sandstones and, comparatively rare, conglomerates. No distinction was made between shales/clay- or mudstones and pure sandstones in this first version owing to the mixtures present in most sedimentary rocks, especially in orogenic settings, for example, in flysch and molasse associations. Global volume estimates of clay-mud-siltstones versus sandstones on the continents [Ronov, 1968; Füchtbauer, 1974], give, as a rough estimate, a 2:1 distribution. Because of their higher resistance to erosion, sand and sandstones have considerably greater shares of the surface of continental platforms. Siliciclastic sediments and sedimentary rocks have small carbonate proportions (normally less than 10% in volume, occasionally up to 20%). Pyritic shales are included in this class.

3.1.8.2. Mixed Sedimentary Consolidated Rocks

[23] Mixed siliciclastic–carbonate rocks (Sm) have higher carbonate contents, ranging from 30 to 70% with medians around 40%. The most typical rock types are marl and marlstone, ideally composed of equal proportions of clay and carbonate (CaCO3) ooze. Supply of both components changes rapidly with the short- and long-term climate cycles at geological scales. Many, if not most, of the widely occurring rhythmic successions of marly beds (usually cm–dm thick) with varying clay-limestone portions are nowadays thought to be due to the Milankovich cycles, high carbonate content (biological productivity) corresponding to warmer periods [Einsele, 1992; Stanley, 1994]. On our map, however, this class usually represents intimately mixed series of siliciclastic sediments, intercalations of marl- and limestones in which the carbonate proportions are highly variable.

3.1.8.3. Carbonate Rocks

[24] Carbonate rocks (Sc) are pure limestone, chalk, dolomitic limestone and dolomite; the amount of carbonate in this class is >50% and mostly around 80%. Carbonates of all ages are included in this class, from Precambrian to Cenozoic and Quaternary to Recent reef construction carbonates as for example in Yucatan–Mexico, and Florida.

3.1.8.4. Evaporites

[25] Evaporites (Ep) are too easily soluble to survive at the Earth's surface for longer, “geologic” times. Therefore only Quaternary occurrences in arid areas are considered on our map, if they are large enough, as described by Choubert et al. [1980]; sometimes, for example, in eastern Iraq and in Iran in the Persian salt deserts, they develop on outcrops of older evaporites (in Persia of Cambrian age). In addition, a special GIS layer designates a few subsurface occurrences of ancient evaporites known to considerably affect surface river water chemistry (see full reference list in auxiliary material Annexes B and J).

3.1.8.5. Semiconsolidated to Unconsolidated Sedimentary Rocks and Sediment

[26] Semi-consolidated to unconsolidated sedimentary rocks (Su) are predominantly siliciclastic sediments of Cenozoic age. The share of mixed siliciclastic-carbonate and carbonate sediments (and equivalent sedimentary rocks) should agree with the global mean of about 20% [Ronov, 1968; Füchtbauer, 1974], with very minor proportions of volcanogenic pyroclastics. However, as carbonate sediments usually undergo rapid cementation, they mostly appear as consolidated carbonate rocks (Sc) and their share of this rock class is clearly below 20%.

3.1.8.6. Alluvial Deposits

[27] Alluvial deposits (Ad) are ubiquitous unconsolidated Quaternary alluvia, predominantly of Holocene age, mostly siliciclastic sediments found in fluvial and coastal plains and/or qualified as fluvisols and gley soils in plains by FAO/UNESCO [1975, 1986]. Glacial and fluvioglacial deposits were clustered in this type.

3.1.8.7. Loess

[28] Loess (Lo) has been specially differentiated owing to its primary importance for river sediment load and chemistry as seen for example in rivers draining the Chinese loess plateaus. Loess is mostly of Pleistocene to, locally, Holocene age [Taylor et al., 1983] and is specifically shown on coarse-scale geological maps. The thickness of this semi-consolidated and easily erodible Quaternary formation ranges from less than a meter to 150–200 m, for some of the Chinese loess. It is generally characterized by a moderate carbonate content (10–15% [see Catt, 1988]). In addition to the basic work by Pécsi [1990] and Pécsi and Richter [1996], multiple regional and local references were used in the global loess mapping (see auxiliary material Annexe B).

3.1.8.8. Dunes and Shifting Sands

[29] Dunes or shifting sand (Ds) correspond to late Pleistocene and Holocene dunes and were delimited according to the global FAO soil map [FAO/UNESCO, 1975, 1986]. They were differentiated here in order to reflect arid regions where eolian accumulation and transport were the dominant surface processes with very little chemical weathering and river transport.

3.1.9. Miscellaneous Categories

[30] Where Quaternary sedimentary layers are thin, as for loess, geological maps tend to omit them but they may still be of importance for water chemistry and erosion. They were mapped where their thickness, typically exceeding 5 m, may significantly affect river material transfer.

[31] “Water bodies” (Wb) correspond to large lakes only. “Ice” (Ig) corresponds to Alpine-type glaciers and glaciers found in the Canadian Archipelago.

3.2. Step Determination of 17 Classes of Lithology

[32] The vectorized lithologic map of the world was developed in eight main steps (Figure 1) through an iterative process with multiple checking, comparisons and decisions concerning the identification, selection and clustering of various rock types, keeping in mind that the most soluble and most erodible rock types should be identified as far as possible. Detailed steps are presented in auxiliary material Annexe C.

Figure 1.

Schematic steps to identify the lithology classes from various sources (shaded: principal sources). Main sources are presented in Table 1; additional sources are given in auxiliary material Annexe B.

[33] The regional and/or specific references (see auxiliary material Annexes B and J) were used as a correction tool not only for the lithology map but also for the geology base map where major differences with the detailed regional data became apparent. The different sources are of unequal quality, most of them are themselves compilations and not primary sources. Care was taken to trust only the most reliable or primary information. The uncertainties and limitations of our map at this stage are discussed further.

[34] The logical eight-steps approach was not always possible and many checking and trial-error loops were necessary particularly at the step 5 stage. During this iteration many polygons were shape-corrected and adjusted to match finer maps and about 700 new polygons were added to the step 1 polygon set, resulting in about 8300 polygons. In some areas, substantial modifications were made, for example, in Taiwan and New Zealand. The shortest polygon length is around 7 km and the finest resolution possible with this map is about 10 × 10 km. A finer resolution would not add significant detail. The geological age information (steps 1 and 2) was kept in the final database. A specific layer was created for known and significant occurrence of subsurface evaporites.

[35] As the map published by Amiotte-Suchet et al. [2003] was not available in digital form when this work was completed [Dürr, 2003], we could only use the paper version from 1994; however, the results of Amiotte-Suchet et al. [2003] are used here for comparisons and for complementary estimates of proportions of exposed rocks.

3.3. Digitized Version

[36] The original map is in vector format and can be transformed for raster analysis at different resolutions. The resolution adopted for the databases used for global hydrological studies (such as that of Vörösmarty et al. [2000a, 2000b, 2000c]) is nowadays mainly 30′ × 30′ or 0.5° × 0.5°, about 50 km at the equator; that is, a 50,000 km2 river basin is described by at least 20 cells. When our map was transformed from vector to raster mode the initial size of the grid cells was set very fine, corresponding to grid cells with a 1 × 1 km resolution. By this procedure, no basic information is lost, and a given 30′ × 30′ grid cell will not only contain the value of the dominant lithology in the cell, but percentages of all lithologies present in the cell.

[37] Four thematic GIS layers are included in the database: (1) lithology at the surface (minus soil), (2) major subsurface evaporite occurrences (important for surface water chemistry) where identified, (3) geology, i.e., the base map amended in accordance with the lithologic map, where the ages of the different rock types can be determined, and (4) limits of maximum Quaternary glaciation extent.

[38] The authors can be contacted for use of the GIS vector map and data. High-resolution images of the map can be found at http://www.sisyphe.jussieu.fr/Internet/lithology/.

4. Main Results

4.1. Global Distribution of Terrestrial Surficial Rocks for 17 Rock Classes

[39] The present-day lithological map of the world is presented in Figure 2. The GIS processing of this vector-mode map generated the specific lithological composition of each continent (Antarctica and Greenland ice caps excluded) and of the whole globe (Table 2).

Figure 2.

Present surficial lithology of the world.

Table 2. Present Surficial Lithology for the Nonglaciated Area of Continentsa
 North America, 22.3 M km2South America, 17.9 M km2Australasia, 9.0 M km2Africa, 30.1 M km2Asia, 43.9 M km2Europe, 9.8 M km2Global, 133.0 M km2
G%C%G%C%G%C%G%C%G%C%G%C%ΣG%
  • a

    Antarctica and currently glaciated parts of Greenland are not considered. G%, proportions in percent of the global total; C%, proportions in percent of the individual continents. Σ G%: Wb + Ig = 0.57%; magmatic rocks (Pb+Pa+Vb+Va) = 14.1%; Pr+Mt = 15.6%; crystalline rocks (magmatic rocks +Pr+Mt) = 29.7%; total “hard” rocks (crystalline rocks + 1/2 Cl) = 32.4%; sedimentary rocks (Ss+Sm+Sc+Ep) = 34.6%; loose sediments (Su+Ad+Lo+Ds) = 29.7%; total “soft” rocks (1/2 Cl + sedimentary rocks + loose sediments) = 67.0%. Boldface denotes continental values above the double of the global mean, and italics denote values below half of the global mean. Three dots denote zero values. For lithology codes, see Table 3.

Wb0.261.570.010.040.0020.020.130.570.120.360.030.410.55
Ig0.010.040.020.110.0020.020.02
Pb0.010.040.010.110.020.070.130.390.050.610.20
Pa1.036.130.443.300.7010.332.179.582.437.360.466.217.23
Vb1.418.370.936.940.233.440.994.362.096.340.101.325.75
Va0.312.290.030.440.641.940.0040.050.98
Pr0.764.522.9021.550.679.894.3819.392.276.880.547.3311.52
Mt2.8316.830.070.500.355.110.170.730.651.980.010.084.07
Cl1.116.580.282.070.375.440.200.902.848.610.658.855.45
Ss2.3814.192.0415.170.9614.224.3219.124.6514.081.9326.0616.28
Sm1.7710.560.433.191.2117.880.562.493.7211.280.050.717.75
Sc2.5715.310.261.900.466.782.4810.973.169.571.4719.9510.40
Ep0.010.060.020.330.020.100.070.210.12
Su1.016.001.5311.421.2017.773.3114.632.056.200.9512.8210.05
Ad1.146.763.5326.310.568.222.9613.076.7820.560.516.9215.48
Lo0.523.090.695.150.782.370.628.452.62
Ds0.914.020.621.870.020.201.54
Σ %16.8100.013.4100.06.8100.022.6100.033.0100.07.4100.0100.0

4.1.1. Magmatic Rocks

[40] Magmatic rocks represent 14.1% of the continental surface and they are differentiated in four classes.

4.1.2. Basic-Ultrabasic Plutonic Rocks

[41] Basic-ultrabasic plutonic rocks (Pb) are rare (0.27 M km2, 0.2% of the globe). Most occurrences are ophiolite complexes, relics of uplifted ocean floor within collisional orogens. Mappable ophiolites can be found mainly in the Alpine-Himalayan orogenic system from the Alps to the Dinarides (former Yugoslavia) and Hellenides (Greece), in Turkey, on Cyprus and in Northern Syria, in the Iranian Zagros mountains and as a big mass in Oman. They are also found in the recent collisional chains of the Insulinde (Sulawesi, Halmahera etc.), Philippines, and on the southwest Pacific islands of New Guinea, New Caledonia and New Zealand. Sixty-four percent of total Pb belong to Asia, while they are nearly absent from North America (3% of total Pb). A second and very rare type of Pb occurrences are lopolithic layered intrusions in the continental crust, only the biggest examples were mapped, namely, the well-known Bushveld Massif in South Africa, extending over 65,000 km2 and up to 7 km thick, and the Great Dyke in Zimbabwe.

4.1.3. Acid Plutonic Rocks

[42] Acid plutonic rocks (Pa) (9.61 M km2, 7.23% of the globe) are much more abundant than plutonic basic rocks, relatively more so in Australia and Africa and, to a lesser extent, in South America.

4.1.4. Basic and Intermediate Volcanic Rocks

[43] Basic and intermediate volcanic rocks (Vb) (7.64 M km2, 5.75%) combine basalts and andesites because of their petrologic-geologic close relationship. Vb magmas have low viscosities and easily reach the surface; Vb rocks are therefore much more abundant than Acid volcanic rocks (Va). Vb rocks occur mainly in orogenetic active regions, for example, in the cordilleras of North and South America, as well as in the eastern parts of Asia and Africa. They are less frequent in Europe and Australasia.

[44] Basalts are mainly related to extension within the Earth's mantle and crust. Major extended basalt outcrops are found as “flood basalt” provinces, up to several km thick: Late Precambrian basalts of North America (Keweenawan, Lake Superior, now mostly covered by younger sediments); Mesozoic basalts in Siberia (Tunguska plateau), in the Karoo (South Africa), and on both sides of the South Atlantic (Paraná/Brazil and Etendeka/Namibia); Cenozoic basalts of the North Atlantic province (eastern Greenland, Iceland), northwestern India (Deccan), Ethiopia, and in northwestern North America (Columbia and Snake River plateaus).

[45] Basaltic andesites and andesites are associated here with basalts, although we recognize their potential differences in chemical weathering and erosion. It was not possible at this stage to differentiate and map these rock types as they are generally found together, as a result of the subduction of the oceanic crust, particularly in the circum-Pacific “ring of fire”: in the Andes and in the Middle Americas, the Cascades, the Alaskan and Aleutian volcanoes and on the other side in Kamchatka, on the Kuril, Japanese and Philippine islands. Subduction of the Indian Ocean is responsible for the Indonesian volcanoes and their Vb lavas.

4.1.5. Acid Volcanic Rocks

[46] Acid volcanic rocks (Va) are uncommon (1.31 M km2 global, 0.98%) as they extrude slowly and usually do not reach the surface but stay inside the continental crust and solidify into coarse crystalline rocks. They cover significant areas only in South America and Asia. Apart from a few Precambrian shield outcrops in northern South America, most acid volcanic rocks are Mesozoic and Cenozoic ignimbrites, found particularly in the Andean cordillera and in continental and island volcanic arcs around the West Pacific, especially in eastern Siberia and China and on the Japanese islands. The youngest ignimbrites characterize the active volcanic region of North New Zealand and the Lake Toba (Sumatra), the latter an eruption some 73,000 years ago of about 2500–3000 km3 of dense-rock equivalent pyroclastic material [Chesner et al., 1991]. On our map, acid volcanic rocks are absent from North America and Africa and relatively rare in Europe.

4.1.6. Precambrian Basement

[47] Precambrian basement (Pr) is one of the most common rock types (15.37 M km2, 11.52% of the globe), particularly well represented in Africa (38% of total Pr) and South America (25% of total Pr), less so in Europe (5% of total Pr). Large parts of the North American shield have been attributed to the metamorphic rocks (Mt) category in steps 1 and 2.

4.1.7. Metamorphic Rocks

[48] Metamorphic rocks (Mt) (5.36 M km2, 4.07%) according to our definition are only found in large proportions in North America; they are also quite common in Australasia. On all other continents they are rare.

4.1.8. Complex Lithology

[49] Complex lithology (Cl) (7.21 M km2, 5.45%) is found mostly in younger collisional orogenic belts. In Africa and Australasia this type is nearly absent owing to the relative scarcity of younger fold belts. This will be discussed further when lithology is linked to relief.

4.1.9. Consolidated Sedimentary Rocks

[50] Consolidated sedimentary rocks include Siliciclastics (Ss, 16.3%), mixed Siliciclastic-carbonate sedimentary rocks (Sm, 7.8%) and Carbonates (Sc, 10.4%). Together, they occupy about one third (34.5%) of the global terrestrial surface area.

4.1.10. Siliciclastic Sedimentary Consolidated Rocks

[51] Siliciclastic sedimentary consolidated rocks (Ss) (21.66 M km2, 16.28%) are the most common rocks represented on our map. They are prevalent on all continents, especially in Europe.

4.1.11. Mixed Sedimentary Consolidated Rocks

[52] Mixed sedimentary consolidated rocks (Sm) (10.33 M km2, 7.75%) have large shares of all continents, but relatively smaller ones in South America, Africa and Europe, which might be due to a more thorough lithological knowledge of these regions whereby the number of questionable polygons can be reduced and more sedimentary rocks can be assigned to the Siliciclastic (Ss) or Carbonate (Sc) rock class. In China, mixed sedimentary rocks seem to have comparatively high carbonate contents.

4.1.12. Carbonate Rocks

[53] Carbonate rocks (Sc) (13.80 M km2, 10.40%) are abundant especially in North America and Europe, and less so in South America. A few Precambrian carbonates that still remain at the Earth's surface in southern Africa, Australia and in some parts of the Upper Mackenzie and Baker river basins (Canada) are included in this class. Large areas of Cretaceous or younger carbonate platforms, built during post-break-up times of the Gondwana supercontinent, are found in this class.

4.1.13. Evaporites

[54] Evaporites are extremely difficult to map on global scale. Non-Quaternary outcrops could not be mapped by polygons: Many salt domes related to Permian evaporite deposits, or others, are actually sealed by solution residues and/or covered by alluvial deposits and not mapped. However, river chemistry can trace the location of such salt domes. They are featured on our map in a specific layer. The occurrences in question are found in the eastern Peruvian Andes, the Mackenzie river basin in northwestern Canada, the Salt Fork river basin, a tributary to the Arkansas river in North Oklahoma/Kansas, United States, in the Viluy basin (a Lena tributary) in eastern Siberia, in the Jordan basin in the Near East, and in the Khorat basin drained by the Mun river in Thailand/SE-Asia. Surface Quaternary Evaporite (Ep) deposits are rare on our map (0.16 M km2 global, 0.12%). They are associated with saline soils and are found only in South America, Australasia, Africa and Asia where they correspond to present-day salt deposition in arid and generally endorheic areas (Altiplano, Lake Eyre basin, Sahara, Gulf Coast, Kara Bogaz, etc.). Detailed references on evaporites are listed in auxiliary material Annexes B and J.

4.1.14. Unconsolidated Sedimentary Rocks and Sediments

[55] Unconsolidated sedimentary rocks and sediments (Su 10.1%; Ad 15.5%; Lo 2.6%; Ds 1.5%) represent another third (29.7%) of the continental surface area.

4.1.15. Semiconsolidated or Unconsolidated Sedimentary Rocks

[56] Semiconsolidated or unconsolidated sedimentary rocks (Su) (13.38 M km2, 10.05%) are by definition of Cenozoic age. They are found in large parts of Australasia, Africa and Europe but are less significant on the other continents.

4.1.16. Alluvial Deposits

[57] Alluvial deposits (Ad) (20.62 M km2, 15.48%), of Quaternary age, form the second most common surficial rock type. They are found on all continents, with relatively minor shares in North America, Australasia and Europe.

4.1.17. Loess

[58] Loess (Lo) (3.49 M km2, 2.62%) occurs, on our map, on most continents except in Africa and Australasia. In South America loess was mapped in the Lower Paraná region (Argentina and Paraguay), in North America mainly in the lower Mississippi region. In Europe (24% of total Lo), many of the loess deposits are too small or too thin to be taken into account on our map; for example, in northern France and south-central Germany the widespread loess cover usually has a maximum thickness of only a few meters [Moores and Fairbridge, 1997]. A small local exception (30–40 m) occurs southwest of Frankfurt/Main. Thicknesses up to several tens of meters can frequently be found in southeastern Europe, especially in Hungary, where major occurrences were identified in the Pannonian plain. Furthermore loess covers large areas east of the Carpathian mountains and north of the Black Sea. Asian loess is found east of the Ural, in the border region between Kazakhstan and Russia, extending toward the upper Ob regions. Siberian loess probably contains large quantities of sand (A. Semmel, personal communication, 2001) and has thus not been taken into account on our map. The famous Chinese loess, unusually thick, is mostly of late Pleistocene age. At present the depth of the loess deposits is about 100 to 200 m. The endorheic Tarim basin west of the loess plateau and the middle East China regions also have some loess covers which were identified on our map (see specific references for loess in auxiliary material Annexes B and J).

4.1.18. Dunes and Shifting Sand

[59] Dunes and shifting sand (Ds) (2.06 M km2, 1.54%), of Pleistocene–Recent age, occur mostly in the arid belts of Africa (Sahara) and Central Asia (e.g., Takla Makan, Ala Shan, Tarim basin). In Europe and on other continents, dunes are insignificant or absent.

4.2. Regional Lithology of Continents and Ocean Drainage Basins

[60] The regional lithology of continents and ocean basins depends on their conventional limits. We have chosen here six continents (including Australasia) and four ocean drainage basins (Arctic, Atlantic, Pacific, Indian). Their detailed limits are presented in auxiliary material Annexe D.

4.2.1. Lithological Composition Per Continent

[61] The lithological composition of the six continents is presented in Table 2. Note that Australia is combined here with New Zealand and New Guinea to form the “Australasian continent,” and that Europe is separated from Asia by the Ural mountains and the North Caucasus (both mountain ranges are therefore split between the two continents).

[62] The lithological composition of each continent may be somewhat different from the global average owing to geological history, present and Quaternary distribution of climate (arid regions, glaciations), tectonics and volcanic activity. A detailed discussion on the composition of each continent can be found in auxiliary material Annexe E.

4.2.2. Lithological Composition Per Ocean Drainage Basin and Endorheic Regions

[63] Ocean drainage basins (Arctic, Atlantic, Pacific, Indian) are based on previous work on global river drainage [Vörösmarty et al., 2000a, 2000b] (called “ocean basins” by these authors and other continental hydrologists). The basins of the Mediterranean and Black seas, which are nearly enclosed, were separated from the Atlantic Ocean proper and on each continent the endorheic regions are set apart.

[64] The most striking features of this distribution are (Table 3) as follows. Mixed sedimentary rocks are scarce in the Arctic Ocean drainage basin (17.5 M km2) while siliciclastic rocks are abundant. Recent sediments such as evaporites, dunes and loess are of course nearly absent in this basin because of its ice-cover during the Pleistocene (48% of the Arctic drainage was actually ice-covered at the Late Glacial Maximum). Other rock types are close to the global average.

Table 3. Global Proportions of Surficial Lithologies for the Different Ocean Drainage Basinsa
CodeLithologyArctic OceanAtlantic OceanbIndian OceanPacific OceanMediterranean + Black SeaEndorheic (Internal)Global
  • a

    Different ocean basins denote nonglaciated areas (in percent of total area for each basin); boldface values are ≥2× the global average, italic values are ≤0.5× the global average). Three dots denote zero values.

  • b

    Atlantic Ocean does not include Mediterranean and Black Sea drainage basins.

Wbwater bodies0.70.70.10.030.80.60.5
Igpolar ice and glaciers0.080.080.02
Pbbasic-ultrabasic plutonic rocks0.340.040.380.270.250.20.2
Paacid plutonic rocks7.17.37.611.53.04.97.2
Vbbasic volcanic rocks4.33.16.414.64.84.35.8
Vaacid volcanic rocks0.530.740.213.50.040.90.98
PrPrecambrian basement8.216.317.74.411.33.911.6
Mtmetamorphic rocks5.76.23.91.90.211.84.0
Pr+Mtbasement + metamorphic13.922.521.66.311.55.715.6
Clcomplex lithology6.93.12.69.63.59.55.4
Sssilici-clastic sedimentary rocks27.018.511.310.014.214.116.3
Smmixed sedimentary rocks2.65.913.212.71.69.77.8
Sccarbonate rocks-consolidated13.19.111.26.720.68.110.4
Eprecent evaporites0.210.120.690.12
Susemi- to un-consolidated sedimentary rocks5.311.17.27.214.015.610.1
Adalluvial deposits16.313.417.414.914.219.315.5
Loloess1.73.42.66.62.22.6
Dsdunes or shifting sand0.131.150.640.104.854.41.6
Sum 100100100100100100100
Drainage area, M km2 17.545.720.719.810.718.9133.0

[65] The Atlantic Ocean drainage basin (45.7 M km2 without Mediterranean and Black seas basins) is the most representative at the global scale. It is only depleted in basic/ultrabasic plutonic rocks, although their proportion is still very uncertain.

[66] The Indian Ocean drainage basin (20.7 M km2) has above average amounts of Precambrian basement and mixed sedimentary rocks. It is depleted in acid volcanic rocks, complex lithology and dunes.

[67] The Pacific Ocean drainage basin (19.8 M km2) is somewhat different. It is highly enriched in basic volcanic (×2.5) and acid volcanic (×3.6) rocks owing to the circum-Pacific “ring of fire,” and enriched in complex lithology and mixed sedimentary rocks. This is balanced by a marked scarcity of Precambrian basement and metamorphic rocks (total of 6.3% versus 15.6% global) and carbonate rocks (despite their abundance in parts of Asia and SE Asia), but note that in the Chinese river basins, some carbonates were mapped as mixed sedimentary rocks.

[68] The Mediterranean and Black Sea drainage basins (10.7 M km2) are rich in carbonate rocks, loess (Black Sea mostly), and dunes. However, the latter are found mainly in the South Mediterranean drainage basin, i.e., the Sahara Desert which is of course not presently drained by surficial runoff and crossed only by the allogenic Nile river.

[69] Endorheic regions (18.9 M km2) are those drained toward the interior of continents. They are found on all continents (Lake Eyre basin, Chad and Okawango basins, Altiplano and Mar Chiquita basins, Great Basin, Caspian and Aral Sea basins and other Central Asia basins). Globally these regions are poor in Precambrian basement and metamorphic rocks (5.7%), and rich in arid-zone lithological tracers (i.e., evaporites and dunes) and in semi- to un-consolidated rocks.

4.2.3. Latitudinal Distributions of Lithologies

[70] Latitudinal variation of lithologies was first described by Amiotte-Suchet [1994] and Amiotte-Suchet et al. [2003]. We used (see Annexe F in auxiliary material) the same division into 12 latitudinal bands: 55°S–35°S, 35°S–25°S, 25°S–15°S, 15°S–05°S, 05°S–05°N, 05°N–15°N, 15°N–25°N, 25°N–35°N, 35°N–45°N, 45°N–55°N, 55°N–65°N and >65°N. For simplicity, closely related rock types that may be very difficult to differentiate were clustered into 10 types, namely: water and ice, Precambrian basement and metamorphic, plutonic acid and basic, volcanic acid and basic, dunes/evaporites and loess, complex lithology, siliciclastic sedimentary, carbonate rocks, mixed sedimentary, alluvial deposits and sedimentary semi- to un-consolidated. The distribution of re-clustered rock types is presented in Annexe F (Annexe F Table in auxiliary material) and illustrates these variations for four latitudinal bands: north of 55°N, between 15°N and 55°N, between 25°S and 15°N and south of 25°S.

[71] When the percentage of occurrence of a given rock type is normalized to the global average (part B of the auxiliary material Annexe F Figure), the latitudinal distribution of lithologies shows some remarkable features, explained by (1) the plate tectonic movements of continents leading to their present position, the current distribution of land and sea, and (2) the climate belts of the Earth, i.e., the climate-driven, principally Quaternary to Present, formation of sediments.

[72] Young and Quaternary sediments such as loess, dunes and evaporites are strongly related to the climate and only occur in large proportions in two dry belts from 15°N to 55°N and south of 25°S. Carbonates (part A of the auxiliary material Annexe F Figure) are most abundant between 15°N and 55°N as pointed out by Amiotte-Suchet et al. [2003], quite rare south of 15°N and absent south of 35°S.

[73] Complex lithology is prevalent between 35°N and 55°N where numerous mountain chains are situated. Precambrian basement and metamorphic rocks are highly developed particularly north of 55°N (Canadian and Scandinavian shields) and between 25°S and 15°N (Brazilian and Guyana shields, African shield). The proportions of total plutonics and total volcanics do not vary much with the latitude (part B of the auxiliary material Annexe F Figure). Detrital consolidated sedimentary rocks (Ss) are not very frequent between 25°N and 45°N: At these latitudes, Quaternary alluvium is dominant, possibly an indication of the active degradation of recent mountain ranges.

4.3. Rock Ages for Lithology Classes

[74] The ages of all volcanic and sedimentary formations are provided by the UNESCO Geological Map of the World [Dottin, 1990]. The share of surficial continental rocks and sediments in general is inversely proportional to their ages, since (1) they disappear in course of geologic time by erosion, feeding the exogenic recycling, or (2) are covered by younger rocks, and (3) later possibly involved in collisional orogenies: gradually immersed, tectonized and subjected to various grades of metamorphism, or eventually molten, partially reappearing by endogenic recycling. The relative global distribution of our lithology classes, in particular of volcanic and sedimentary rocks, per age is tentatively provided in Table 4 and Figure 3. The following remarks should be made.

Figure 3.

Volcanic and sedimentary rock shares for different ages. For lithology codes see Table 3.

Table 4. Distribution of Lithologies Per Geological Agesa
LithologyRecent + QuaternarybCenozoicMesozoicPaleozoicPrecambrianUndifferentiatedTotal
Cretaceous JurassicTriassicUpper (Per., Carb., Dev.)Lower (Sil., Ord., Cam.)
  • a

    Values are percent and area (M km2). Important values are in boldface. For lithology codes, see Table 3.

  • b

    Quaternary rock classes include polygons of “Recent” age from the base map [Dottin et al., 1990].

Wb
   %       100100
   Area       0.680.68
Ig
   %       100100
   Area       0.030.03
Pb
   %       100100
   Area       0.270.27
Pa
   %       100100
   Area       9.619.61
Vb
   %27.643.216.48.23.41.2  100
   Area2.113.301.250.630.260.09  7.64
Va
   %3.014.539.219.617.85.9  100
   Area0.040.190.510.260.230.08  1.31
Pr
   %      100 100
   Area      15.37 15.37
Mt
   %       100100
   Area       5.365.36
Cl
   %  6.53.332.139.50.717.8100
   Area  0.470.242.322.840.051.287.21
Ss
   % 7.348.310.423.19.81.0 100
   Area 1.5910.472.255.012.120.22 21.66
Sm
   % 19.530.09.229.99.12.3 100
   Area 2.013.100.953.090.940.24 10.33
Sc
   %1.215.439.13.510.219.511.2 100
   Area0.162.125.400.481.412.691.54 13.80
Ep
   %100       100
   Area0.16       0.16
Su
   % 100      100
   Area 13.38      13.38
Ad
   %100       100
   Area20.62       20.62
Lo
   %100       100
   Area3.49       3.49
Ds
   %100       100
   Area2.06       2.06
Total
   %21.517.015.93.69.26.613.113.0100
   Area28.6422.5921.204.8112.328.7617.4217.23133.0

[75] 1. Plutonic and metamorphic rocks have no precise age attributed to them in the base map. In total, Precambrian basement and nonmetamorphic Precambrian sedimentary rocks, representing a very long period (nearly 3 billion years) of Earth history sediment deposition and subsequent erosion, together make up 13.1%, i.e., only a few percent less than the sum of all Mesozoic (19.5%, about 185 million years) or Paleozoic rocks (15.8%, about 300 million years).

[76] 2. Most basic volcanic rocks (Vb) are Recent/Quaternary (27% of all rocks in this type) or Cenozoic (43%), the rest originate from the Mesozoic (25%), with minor portions from the Paleozoic era. The distribution of acid volcanic rocks (Va) is different: They are generally older, 59% from the Mesozoic and 24% from the Paleozoic. This distribution may be due to the fact that fresh ignimbrites are easily erodible and often covered by other volcanic products, but once diagenetically hardened or slightly metamorphosed they become very resistant.

[77] 3. Complex lithology rocks (Cl), i.e., tectonized and partially metamorphosed rocks, mixed by orogenic processes and found in mountain ranges, are mostly of Paleozoic age (71.6%). However, this may be due to bias in our construction since in younger orogens, with generally less metamorphosed storeys at the surface, rock types are more easily differentiated and mapped.

[78] The geologic base map is biased with a marked depletion for Triassic sedimentary rocks. Since, for volcanics, Dottin [1990] distinguish only Mesozoic from Paleozoic rocks, we were forced to introduce a differentiation per geological era in the tabulated version of the digitized base map such as the following.

[79] According to their general decay with increasing age, two thirds of the volcanic rocks of undifferentiated Mesozoic age were attributed to Cretaceous and Jurassic rocks and one third to Triassic rocks. Of the volcanic rocks and complex lithology of undifferentiated Paleozoic age three quarters were attributed to the Upper Paleozoic and one quarter to Lower Paleozoic rocks.

[80] The general distribution of all sedimentary rocks (Ss + Sm + Sc + Ep + 1/2 Cl due to the composition of this rock class) is inversely proportional to their ages [Garrels and Mackenzie, 1969; Ronov et al., 1980; Veizer and Mackenzie, 2003]. Quaternary (<1.81 Ma) sediments (Ad + Lo + Ds) and sedimentary rocks are spread over 26.5 M km2, Cenozoic (1.81–65.5 Ma) over 19.1 M km2, Mesozoic (65.5–250 Ma) over 23 M km2 and all Paleozoic sedimentary rocks (250–540 Ma) represent only 17.8 M km2. Finally, about 2 M km2 of Precambrian sedimentary rocks can still be differentiated from metamorphic rocks.

[81] Siliciclastic sedimentary consolidated rocks (Ss) are mostly from the Cretaceous and Jurassic period (48.3% of total), more rarely from the Upper Paleozoic (23.1%) and the Lower Paleozoic (9.8%) periods. Cenozoic Ss-type rocks are quite rare (7.3%) as most siliciclastic sediments of Tertiary and Quaternary ages are only partly (Su) or not at all consolidated (Ad). Mixed sedimentary consolidated rocks (Sm) are also found in the Cretaceous and Jurassic and Upper Paleozoic but are less abundant in the Cenozoic where they may be assigned to other sediment classes.

[82] Nearly half of the Carbonate rocks (Sc) are from the Cretaceous and Jurassic (39.1%), followed by the Lower Paleozoic (19.5%). Significant portions are from the Cenozoic (15.4%), the Upper Paleozoic (10.2%) and the Precambrian (11.2%), smaller shares from the Triassic (3.5%). Globally Quaternary shares are minor. The relative proportion of carbonate rocks among all sedimentary rocks (complex lithology included) and sediments is quite low for the Quaternary (0.6%), a cold period, and increases gradually with age: 11.1% for Cenozoic carbonates, 25.6% for Mesozoic, 23% for Paleozoic. For the Precambrian this proportion is here 77%, but this could be biased by our specific attention to Precambrian carbonate rocks. These formations, often resistant dolomites, are found for example in the Canadian and African shields.

[83] Unconsolidated rocks (alluvial deposits, Ad; loess, Lo; dunes and shifting sand, Ds), are only found in the Quaternary era, but this proportion can also be biased by the construction of the map. Their total is 26.2 M km2. Semi- to un-consolidated rocks (Su, 13,4 M km2) were attributed to the Cenozoic by construction: They represent more than one third of all nonconsolidated rocks.

4.4. Influence of Glaciations on Surficial Lithology of Continents Exposed to Riverine Erosion

[84] The continental lithology must be reconstituted to assess the riverine inputs to the oceans during past geological periods, particularly at the Last Glacial Maximum (LGM), 18,000 BP [Gibbs and Kump, 1994]. Moreover, glacial cover has a certain erosive effect by removing the nonconsolidated sedimentary cover from consolidated rocks. Furthermore the scraping action of glaciers eventually exposes fresh rocks at the surface with limited evidence of chemical weathering. Numerous problems arise in the reconstitution.

[85] 1. The sea level was lower at the LGM by about 120 m, thus large tracts of present continental shelf were exposed to weathering and river transport. In the Arctic Ocean drainage basin only parts of these continental platforms were covered by ice, the Laptev Sea shelf was not covered by ice [Kleiber and Niessen, 1999].

[86] 2. The distribution of the unconsolidated sediments as alluvial deposits and dunes during these periods is certainly shifting with climate variations. Particularly the distribution of loess deposited as eolian sediment during glacial periods depends on the ice cover distribution.

[87] 3. Volcanic activity may have changed slightly in ice-covered regions, although these changes would probably be marginal at the spatial resolution used here.

[88] The ice-cover mask used here (20.87 M km2, see auxiliary material Annexe D for details) has the size of a whole continent and corresponds to the maximum Quaternary glacial cover. The Southern Hemisphere ice cover, mostly in Patagonia, was not tabulated here. Table 5 presents the area of crystalline and consolidated sedimentary rocks, affected by glaciations, in the geometry of the present sea level; that is, no attempt was made to reconstruct the lithology of the exposed continental shelf at the LGM.

Table 5. Distribution of Present Consolidated (Crystalline and Sedimentary Combined) Rocks Affected by Glaciations (Maximum Ice Extent Coverage Through One of the Last Ice Ages) for the Northern Hemisphere in the Geometry of Present Sea Levela
LithologyPast Ice Cover (Area in M km2)Global Present, M km2Covered/Present, %
North AmericaAsiaEuropeGlobal
  • a

    For lithology classes, see Table 3. Three dots denote zero values.

Pb0.0080.010.050.070.2725.9
Pa1.060.890.342.439.6125.3
Vb0.380.690.071.347.6417.5
Va0.210.241.3118.3
Pr + Mt4.360.900.715.9120.7328.5
Cl1.000.660.462.217.2130.7
Ss2.081.501.144.7621.6622.0
Sm0.720.760.031.5310.3314.8
Sc1.660.340.402.3713.8017.2
Total11.285.963.1920.8792.5622.5

[89] About 22% of present crystalline rocks and consolidated sedimentary rocks were covered by ice at the maximum ice cover. This estimate is 13% higher than previous figures used by Gibbs and Kump [1994] (20.9 M km2 versus 18.4 M km2). If Precambrian basement rocks (Pr) and metamorphic rocks (Mt) are reclustered, there is little change in the overall proportions of rock types exposed to weathering: Pr + Mt are slightly over-covered by ice (28.5%) while consolidated sedimentary rocks (Ss + Sm + Sc) are slightly under-covered (18.9% altogether). Complex lithology (Cl) is also over-covered with ice and both acid and basic volcanics slightly under-covered. However, these changes must be considered with caution since we did not consider the exposed continental shelf which was estimated at 15.9 M km2 by Gibbs and Kump [1994]. Such mapping of continental shelf was not possible at this stage as only few maps depicting shelf geology exist.

4.5. Lithology, Relief Types, and Tectonic Domains

[90] Another way to look at the distribution of rock types is to consider their proportions in the 15 global relief classes as defined by Meybeck et al. [2001]. We further simplified here this classification to seven clusters only from “plains” cluster to “high mountains” cluster. The plains cluster corresponds here to the sum of plains (category 1 of Meybeck et al. [2001]) and mid-altitude plains (category 2), the lowlands cluster to the sum of lowlands (category 4) and very low plateaus (category 6), the hills cluster to rugged lowlands (category 5) and hills (category 11), the mid-plateaus cluster to high-altitude plains (category 3), low plateaus (category 7) and mid-altitude plateaus (category 8), the mid-mountains cluster to low mountains (category 12) and mid-altitude mountains (category 13), the high plateaus cluster to high plateaus (category 9) and very high plateaus (category 10) and the high mountains cluster to high mountains (category 14) and very high mountains (category 15) (see details on relief definition and clusters in auxiliary material Annexe G).

[91] The lithology distribution is presented here as per cent of rock types in each relief class (Figure 4; high plateaus essentially correspond to the Bolivian Altiplano and the Tibetan plateau; high mountains correspond to the main Alpine orogens, i.e., the Alps, Caucasus, Hindu Kush, Pamir, Karakorum, Himalaya, etc., and parts of the North-American cordilleras and the Andes; see Meybeck et al. [2001] for a complete list of major geomorphic regions of the world and their relief classes attribution).

Figure 4.

Lithological composition of major relief classes at the global scale: percent of each rock type in a given relief class and total area (M km2) for each relief class. Cratons: sum of Plains, lowlands, mid plateaus + 1/2 hills + 1/2 mid-mountains; Active orogens s.str.: sum of 1/2 hills + 1/2 mid-mountains + high plateaus + high mountains. Bold values are ≥2× the global average; values in italics are ≤0.5× the global average; Wb and Ig are not considered. For lithology codes see Table 3. Relief classes are reclustered from Meybeck et al. [2001] (see auxiliary material Annexe F).

[92] The relief typology and the lithology map were established from completely different databases. Their comparison through GIS analysis can therefore be made. However, combining different databases may result in the loss of some areas, not defined in any of the original databases. This concerns here some parts of coastal regions, defined slightly differently in the lithology database and the relief database, as well as some missing grid cells in the relief typology database. Owing to these effects and the limitation, here, to areas south of 78°N, the summation of relief classes area does not result in a global continental landmass area of about 133 M km2, but of 127.9 M km2 or 96% of the global continental landmass being defined here. This relative error could not be fixed and is believed to be equally distributed. It was assumed that it does not affect a relief type or a lithology in particular and only the relative proportions of lithology classes are presented in Figure 4. Some clusters as those concerning the mild relief revealed very marked relationships, although most of them were expected.

[93] The plains cluster corresponds by definition to the flattest regions of the globe; they are therefore occupied mainly by Quaternary alluvial deposits (32.5% of plains and 47.5% of all Ad), by Tertiary unconsolidated sediments (15.8% of plains and 35.7% of all Su), by loess (6.0% of plains, 50.7% of all Lo) and dunes (3.4% of plains, 48.8% of all Ds). Consolidated detrital sedimentary rocks (Ss), together with carbonate rocks (Sc), also make up large portions of the plains (13.2% and 9.2%, respectively). All “hard” rocks are clearly under-represented, particularly the basic volcanics (0.5% of the plains compared to 7.2% in all other relief types).

[94] The lowlands cluster is constituted by sedimentary rocks and sediments such as Alluvial deposits (13.7% of lowlands and 22.5% of all Ad), siliciclastic rocks (Ss = 19.9% of lowlands and 30.6% of all Ss) and carbonates (Sc = 12.6% of lowlands and 31.1% of all Sc), but a considerable part of the lowlands, 27,4%, is actually formed by the cratonic crystalline basement (Pa, Pr, Mt) without any sedimentary cover. As a result Precambrian and Metamorphic rocks (Pr and Mt) are over-represented (21.2% of lowlands versus 15.5% global). Volcanic rocks occur fairly rarely in lowlands.

[95] The mid-plateaus cluster is similar to the lowlands cluster in most of their lithologic characteristics with a dominance of Precambrian basement rocks (Pr 15.7%), consolidated detrital sedimentary rocks and sediments (Ss 22.1%, Su 14.8%). Dunes (Ds) are still over-represented, for example, in Central Africa, the Near East and Central Asia. Volcanic rocks (mainly Vb) are under-represented (4.6% versus 6.7% globally) with respect to global values.

[96] Loess blanketed the world during Pleistocene glaciation periods. As this deposit is highly erodible (with maximum yields up to 10,000 t km−2 yr−1 in China), it could not be preserved in dissected relief and is mostly found now (73.6%) in very flat regions, i.e., plains and lowlands.

[97] All dissected relief clusters, hills, mid-mountains, high plateaus and high mountains, show similar characteristics.

[98] 1. Nonconsolidated sediments (Su and Alluvial deposits: Ad) are markedly under-represented while consolidated sedimentary rocks and crystalline “hard” rocks are mostly over-represented, except for metamorphic rocks (Mt) which are less abundant in mid-mountains, and Precambrian basement (Pr), seldom encountered in high mountains.

[99] 2. Basement (Precambrian) rocks (Pr) are mostly found in mid-mountains; these areas are presumably the uplifted parts of cratons as for example in East Africa.

[100] 3. Shales and sandstones (Ss) are scarce in high mountains and high plateaus, which have however a great deal of Mixed sedimentary rocks (Sm). Sm rocks are generally found on our map in south-central Asia, in and around Tibet, for which an appropriate rock type differentiation was not possible, as our main sources (in particular, Datong et al. [1985]) show carbonate occurrences intimately mixed with Siliciclastic rocks.

[101] 4. Complex lithologies are found essentially in the dissected relief types (76.0%), especially in the mid-mountains (45.9% of total Cl). As this rock class was originally defined in association with younger orogenic regions, this distribution was to be expected.

[102] 5. Volcanic rocks, either basic or acid, are over-represented in the most dissected relief classes (74.3% of all volcanic rocks are found in these regions). The occurrence of flood basalts, 40% of basic volcanics, with relief actually is dependent on their ages: The youngest examples, the Cenozoic Ethiopia Trap and the basaltic Columbia River plateau of late Tertiary age, respectively, correspond to the high plateaus and mid-mountain relief category, the Cretaceous-Tertiary Indian Deccan Traps to mid-plateaus to lowlands, the slightly older Iceland formation to hills, the Jurassic-Cretaceous Paraná basalts to lowlands. The Permo-Triassic central Siberian basalts are the only flood basalt of that age still presently found in mid-mountains. The remaining basic volcanics are generated mostly by subduction of the oceanic lithosphere in orogenic (in a strict sense) settings, and therefore found in the active “ring of fire” around the Pacific with relief classes from hills to high mountains.

[103] A different, comprehensive geodynamic view of global relief and lithology distributions, can be adopted when the continents and their crust are split into two main tectonic domains: (1) the young active “collisional orogens” (orogens sensu stricto, s.str.), usually mountain chains with strong relief, (2) the cratons, i.e., the older consolidated parts of continents usually peneplained and often covered by tabular sediments, dissected only under particular circumstances, for example along rifted margins. Parts of cratons may be affected by distensive movements, crustal rifting, formation of deep grabens with high shoulders accompanied by basic volcanism (e.g., East Africa), creating “orogens” in a broader sense (sensu lato, s.l.).

[104] Both types of endogene orogenic (s.l.) activities, i.e., collisional as well as distensive tectonic processes and associated volcanic phenomena, create the first-order relief (“mountains”), which is then dissected and eventually leveled by running water. Glacial modeling plays a role only locally and during certain geological periods.

[105] All dissected relief cluster types (“hills,” “mountains” (mid and high), and “high plateaus”) represent 37.4% of the global continental areas actually affected by recent orogenic (s.l.) activity. The remaining 62.6% of the continents with very moderate relief (“plains,” “lowlands,” “mid plateaus”) represent the presently nonactive, an-orogenic and geologically relatively quiet parts of the continents.

[106] As an initial attempt to differentiate the lithology of active collisional orogens (s.str.) from cratons, it is assumed that half of the “hills” and “mid-mountains” relief types can be attributed to cratons, the other half to active collisional orogens. Consequently, about 78% of the continental surface area correspond to cratons, and the remaining 22% to young, active collisional orogens (s.str.). The average lithological composition of these two ensembles was recalculated and is tentatively proposed in Figure 4.

[107] As defined the typical craton lithology is different from the continental average. There are some marked discrepancies between cratons and active orogens (s.str.).

[108] 1. The proportion of basic volcanics (Vb) is 2.8 times higher in active orogens. This is particularly significant in North and South American mountain ranges.

[109] 2. The proportion of “complex lithology” (Cl) is 3 times higher in active orogens; this finding is expected as consequence of the definition originally used for this class.

[110] 3. Unconsolidated sedimentary rocks (Su), Alluvial deposits (Ad), loess (Lo), dunes (Ds) and evaporites (Ep) are half as abundant in active orogens (15.4% altogether) while they cover about 33.7% of the flat cratons area.

[111] 4. The sum of plutonic rocks + basement + metamorphic rocks is nearly identical in active orogens (25.2%) and in cratons (22.6%). This similarity is also found for consolidated rock types such as carbonates.

5. Discussion

5.1. Uncertainties and Limitations of Lithological Mapping

[112] A comparison with previous lithological mapping at the global and river-basin scale sheds some light on the difficulties of such attempts. Two previous global mappings [Gibbs and Kump, 1994; Amiotte-Suchet et al., 2003] and two global distributions of rock types [Blatt and Jones, 1975; Meybeck, 1987] are used for comparison (Table 6). These comparisons are not easy since the rock type differentiation is different.

Table 6. Global Proportions of Surficial Lithologies From the Lithologic Map of the World Compared to Previous Estimatesa
Rock TypeThis WorkAmiotte-Suchet et al. [2003], %Blatt and Jones [1975], %Meybeck [1987], %Gibbs and Kump [1994], %
CodeArea, M km2Percent
  • a

    Values are in percent, and also M km2 for our map. NC, not considered.

  • b

    Estimate (70 to 30% share of shales to sandstones as explained in the text).

  • c

    Includes semi- to non-consolidated rocks.

  • d

    N.B.: some carbonates contained in Mixed sedimentary (Sm) rocks class.

Water bodies, ice and glaciersWb, Ig0.80.6NCNCNCNC
 
Sandstones, sandsSs6.5b4.9b26.2cNC15.8c23.9c
 
ShalesSs15.1b11.4b25.4cNC34.4c12.6c
 
Sedimentary rocks without carbonates and evaporitesSs, Sm, Su, Ad, Lo, Ds71.453.751.6NC50.236.5
 
Carbonate rocksSc13.810.4d13.4NC15.99.3
 
Total sedimentary rocksSs, Sm, Sc, Ep, Su, Ad, Lo, Ds85.564.365.066.066.145.8
 
Intrusive igneous rocksPa, Pb9.87.4NC9.011.0NC
 
Metamorphic rocksMt5.44.1NC17.015.0NC
 
Total shield rocks (intrusive igneous + metamorphic)Pa, Pb, Mt, Pr30.623,027.526.026.020.0
 
Acid volcanic rocksVa1.31.02.3NC3.8NC
 
Basic volcanic rocks, basaltsVb7.75.85.2NC4.1NC
 
Total volcanic rocksVa, Vb9.06.87.58.07.96.8
 
Total crystalline rocksPa, Pb, Va, Vb, Mt, Pr39.629.835.034.033.926.8
 
Complex lithology/fold beltsCl7.35.5NCNCNC27.5
 
Total 133.0100.0100.0100.0100.0100.1

[113] Sedimentary rock outcrops are here estimated to 64.3%, a proportion similar to those of Amiotte-Suchet et al. [2003], Blatt and Jones [1975] and Meybeck [1987]. The notable difference with Gibbs and Kump's [1994] figure (45.8%) is due to their use of a new rock class, the “complex lithology” (27.5%), composed mainly of folded sedimentary rocks.

[114] Total crystalline rocks (29.8%) are here intermediate between Gibbs and Kump [1994] (26.8%) and all others (34 to 35%). This is probably due to the fact that some Precambrian sediments were identified here and distinguished from other basement rocks considered as crystalline.

[115] Metamorphic rocks were not distinguished by Gibbs and Kump [1994] nor by Amiotte-Suchet et al. [2003]. Our figure (4.1%) is much smaller than that previously found by Blatt and Jones [1975] (17%), a proportion adopted by Meybeck [1987] (15%). We believe that distinguishing between metamorphic rocks and Precambrian basement is very difficult: On some continents, metamorphic occurrences on shields were minimized in our basic sources while on others it was the reverse. Therefore it may be advisable in future attempts to cluster these two types as already done here in some applications.

[116] True carbonate rocks (10.4%) are here intermediate between Amiotte-Suchet et al. [2003] (13.4%) and Gibbs and Kump [1994] (9.3%). According to Ford and Williams [1989], carbonate rocks occupy about 12% of the Earth's dry and ice-free land. The extent of carbonate terrains with distinctive karst landforms and/or karst groundwater circulation is estimated to be less, around 7–10% of the area [Ford and Williams, 1989]. As for other rock categories, a precise definition of carbonate rocks cannot be proposed until the carbonate contents in sedimentary rocks have been globally mapped.

[117] Detrital sedimentary rocks, i.e., all sedimentary rocks except evaporites and carbonates, are here estimated at 53.7% compared to 51.6% by Amiotte-Suchet et al. [2003], 50.2% by Meybeck [1987] but only 36.5% by Gibbs and Kump [1994]. Again, this divergence between Gibbs and Kump and other authors is attributed to their large proportion of “complex lithology.” Our estimates of pure sandstones, conglomerates and shales, presented in Table 6, result from the split of consolidated siliciclastic rocks in proportions of 70% for shales and 30% for sandstones as proposed in section 5.3. Our estimate is very different from previous ones, probably because additional occurrences of shales and sandstones can be found in our class of “mixed sedimentary” rocks, where they are interbedded with other types of sediments.

[118] The analysis of previous work at the global scale (see section 2) and these comparisons enlighten the difficulties of lithological mapping and allow us to approximate the degree of uncertainty of global estimates including this one as follows: (1) excellent estimates (uncertainties of less than 10%): ice (Ig), major water bodies (Wb) and basic volcanics (Vb) in most regions; (2) good estimates (10–20% uncertainties) for very common (>15% global occurrence) and common rock types (5–15% occurrence): basic volcanics in a few regions (Vb), acid and basic plutonic rocks (Pa, Pb), acid volcanics (Va), undifferentiated Precambrian basement + metamorphic (Pr + Mt), siliciclastic sedimentary (Ss), carbonates (Sc), loess (Lo), semi- to un-consolidated rocks (Su), alluvial deposits (Ad) and dunes (Ds); (3) fair estimates (20 to 50% uncertainty) for intermediate rock classes: complex lithology (Cl), mixed sedimentary (Sm) and for Precambrian basement (Pr) and metamorphics (Mt) differentiated from one another; and (4) poor (>50% uncertainty): evaporites (Ep). The lack of standard rock type groupings is actually limiting the comparison with previous works at the regional or basin scales (see a full discussion for the Mackenzie River basin based on Reeder et al. [1972] results, in auxiliary material Annexe H).

5.2. Scales and Spatial Resolution

[119] The uncertainty of lithology analysis depends greatly on the adopted scales. The previous estimates of uncertainties, given for global figures, are generally not valid at the regional scale (106–107 km2); that is, very common rock occurrences are probably not known at such scale with better than a 15–20% uncertainty. This means that at the medium and large river basin level (105–106 km2), uncertainty may reach 30 to 50% for some basins. Below a basin area of 105 km2 the errors become greater. At such a scale it is therefore not advisable to use our database to assess the distribution of uncommon to very rare rock types. Also, a coarser resolution tends to underscore the extent of rare lithologies, which are generally present as multiple small outcrops at the continent surface (e.g., ultrabasic rocks). An example of resolution effect on local lithology is provided in auxiliary material Annexe I for New Zealand, comparing estimates of Gibbs and Kump [1994], Amiotte-Suchet et al. [2003] and ours.

5.3. Map Criticism and Improvements

[120] Now that a vector mode is possible and as the spatial resolution gets finer, lithology maps will attract the attention of regional geologists. All of them will probably offer well-founded criticism based on their field experience and specific knowledge and we are fully aware of having omitted some carbonate outcrops, misinterpreted some maps notably concerning sedimentary facies. The following are some of the expected criticisms.

[121] 1. Some simplifications and generalizations may appear rather “crude.”

[122] 2. Mapping lithologies that can be used for erosion, weathering or groundwater resources (“hydrologically important”) in favor of other features that can be used for tectonic or petrologic purposes is a choice that can be discussed.

[123] 3. Substantial errors probably remain in regions with complex lithology and/or where scientific knowledge is still incomplete (e.g., carbonate occurrence on the Himalayan and Tibetan plateaus). In some regions of the world, the lithology may not be as well defined as in others owing to their lack of accessibility (e.g., tropical regions of Central Africa and South America), although this type of error may have little influence on riverine transport applications when it occurs in arid regions, for example, in the Sahara and Arabia.

[124] 4. Although this map is originally targeted to water-related applications, the groundwater chemistry cannot be adequately addressed since outcrops such as limestones underneath alluvium, hidden salt layers or salt domes are not yet mapped except for the very few occurrences of major hydrologically important subsurface evaporites.

[125] 5. Some minor rock types important in riverine transport such as marble or volcanic tephra cannot yet be mapped at the global scale.

[126] 6. Although dunes were well marked on the FAO/UNESCO [1975, 1986] soil map, used here as a basis, some errors may remain; the well-known Namibian dunes are still identified as alluvial deposits here. Other errors discovered a posteriori concern missing carbonate outcrops in southern Chile or in northern Borneo.

[127] Improvements to this map should probably be made along the following lines.

[128] 1. A more extensive and intensive study and use of regional literature should complement the unavoidably eclectic and incomplete choice of this work (see auxiliary material Annexes B and J).

[129] 2. The basement of the Precambrian cratons (Pr) should systematically be differentiated from younger orogenic basements (Mt). At this stage these petrologically nearly identical types have still not been separated everywhere. Also, Precambrian greenstone belts (mainly Vb) should be recognized, where large enough.

[130] 3. In consolidated siliciclastic sedimentary rocks (Ss), shales/clay- or mudstones should now be separated from pure sandstones, as attempted on prior lithological maps [Amiotte-Suchet, 1994; Gibbs and Kump, 1994; Amiotte-Suchet et al., 2003] because of the considerable differences with regard to mechanical weathering and the important role of shales in supplying fossil organic carbon to rivers [Kao and Liu, 2002]. The estimated distribution between silt + clay + shale (87%) and conglomerate + sandstone (13%) as given by Press and Siever [1994] from sediment volumes is somewhat biased when considering outcrops. Conglomerates/sandstones are more frequently represented than shales owing to better resistance especially to mechanical weathering. Also big volumes of shales are especially formed in/on continental rises or in deep sea trenches where they are rapidly buried with other sediments (turbidity currents–flysch sinks, fast disappearing in accretionary wedges), but also in molasse basins and in lakes. Surface occurrences could be corrected to 70% for mudstone-clay-shale-silt and 30% for sand-sandstone-conglomerate (see Table 6), but these proportions should be verified.

[131] 4. Marine evaporites have been identified as dating from many geological periods [Warren, 1999; Lagny et al., 2001]: the Cambrian (Near East Asia, India, Australia, Siberia), Ordovician (Siberia), Middle and Upper Paleozoic (Sverdrup basin in the Arctic, Devonian basins in western Siberia and near Moscow, Permo-Carboniferous basins in North and South America and west of the Urals), Permo-Triassic (Europe), Jurassic (Mexico), Upper Mesozoic (Khorat basin, Mekong) and Upper Miocene/Messinian (Mediterranean) (see local references in auxiliary material Annexes B and J). These evaporite deposits have been buried by hundreds or thousands of meters of younger sediments, often initially by “impermeable” clay, but some of them rise to the Earth's surface such as salt diapirs. Their complete mapping at the global scale, where they affect local water chemistry, remains to be done. In sufficiently humid climates outcrops of evaporites do not survive as they are dissolved too rapidly, for example, in the Huallaga river basin, Upper Amazon [Stallard and Edmond, 1983; Stallard, 1985]. Evaporites in the underground in contact with groundwater also begin to dissolve, but simultaneously they cover themselves with a seal of insoluble residues. Evaporite outcrops are of course more likely to resist chemical weathering in arid regions; the diapirs of Cambrian salt in the central Iranian Lut desert are well-known as well as those in the outer Zagros belt northeast of the Persian Gulf where even some “salt glaciers” occur.

[132] 5. Pyritic and carbonaceous shales have a specific water chemistry and are important sources of sulfate ions [Berner and Berner, 1996; Meybeck, 1987, 2003]. They should now be distinguished from other types of shales.

[133] 6. Greenstones and ophiolites should be more carefully mapped as their water chemistry is also very specific, with marked Mg2+ dominance over Ca2+ [Cleaves et al., 1974; Meybeck, 1987]. Greenstone belts are often found in the Archean core cratons in northern North America, Southern Africa and Australia. The same remark applies to peridotites.

[134] 7. Glacial till is still imbedded here with alluvial deposits and semi- to un-consolidated sedimentary rocks and should now be treated separately.

6. Conclusions and Perspectives

[135] Although our new global lithology map has some major advantages compared to previous attempts i.e., (1) a vector mode, (2) a finer resolution with about 8300 polygons of homogeneous lithology, (3) 15 rock types instead of 4 to 6 classes, and (4) clusters of rock types specially designed for chemical weathering and mechanical erosion by surface waters, it should be regarded as a preliminary version to be questioned and improved by regional hydrologists.

[136] The most important results of this work are the following.

[137] 1. The relative proportions of plutonic (7.4% altogether) and volcanic rocks (6.7%) are not very different from those found by previous authors. We tentatively split them into acidic and basic rocks accounting for their marked differences concerning weathering and erosion.

[138] 2. The sum of unconsolidated sedimentary rocks is estimated at 29.7% compared to 34.6% for consolidated sedimentary rocks, of which nearly one third is pure carbonate rocks (10.4% global outcrop). However, complex lithologies (Cl = 5.5%), mixed sedimentary rocks (Sm = 7.8%) and loess (Lo = 2.6%), as well as semi- to un-consolidated sedimentary rocks (Su = 10.1%) and alluvial deposits (Ad = 15.5%), contain variable proportions of carbonates. Therefore the proportion of global continental area where carbonate minerals are present probably reaches 40% or more. Considering the modes of occurrences of carbonates and their sensitivity to chemical weathering, their influence on chemistry and material transport is actually found in a majority of river basins, both in solute and particulate form [Meybeck, 1982, 1993, 2003]. This should now be clarified by a systematic characterization of river basin lithological compositions at a finer resolution.

[139] 3. The concept of a “world average lithology” used as a reference is actually difficult to promote given the large variability observed between continents or ocean drainage basins: (1) In South America, Precambrian basement and metamorphic rocks are nearly twice the global average, as well as the acid volcanic rocks and loess; (2) in Australasia mixed sediments (Sm) of Paleozoic and Mesozoic ages are twice the global values; (3) Precambrian basement and metamorphic rocks are nearly half the global values in Asia, of which the general lithological composition is closest to the world's average; (4) Europe is markedly poor in volcanic rocks, has twice as much carbonates, and its composition is the most different from the world's average. Unexpected results include the loess in South America and the relative scarcity of volcanics in Africa despite the East African Rift.

[140] 4. The composition of ocean drainage basins is not equally distributed either, except for the Atlantic Ocean basin (particularly if the Mediterranean plus Black Sea drainage basins are excluded), which is close to the global proportions. The Pacific Ocean basin for instance is markedly rich in basic volcanics (×2.5) and acid volcanics (×3.6) while being poor in Precambrian basement plus metamorphic rocks. The Mediterranean plus Black Sea drainage basin is markedly richer in carbonate rocks, loess and dunes and has no acid volcanics and few acid plutonics. An important application of our database would be to link the oceanic sediment composition with ocean basin lithology for some tracers of continental inputs.

[141] 5. The latitudinal distribution of rock types confirmed the over-representation of carbonate rocks between 15°N and 65°N already described by Amiotte-Suchet et al. [2003] and the prevalence of Precambrian basement plus metamorphic rocks in two bands from 25°S to 15°N and north of 55°N. Rock types related to dry climates such as dunes, loess and evaporites, are found mainly between 15°N and 55°N.

[142] 6. As expected, relief types and distinct lithology classes are related. Unconsolidated rocks (Su of Cenozoic age) and alluvial deposits (of Quaternary age by definition) are found mainly on plains and lowlands where detrital sediments accumulate. Precambrian basement and metamorphic rocks are more abundant in mid-mountainous areas. Volcanics are found essentially, 74%, in the dissected relief types. If within the continents, Greenland and Antarctica excluded, a coarse separation is made between active collisional orogens (21.6% of the globe) and cratons (78.4%) with their sedimentary cover, the active collisional orogens are characterized by much more volcanics (×2.8 for basic volcanics, ×3 for acid ones) and by much less unconsolidated sedimentary rocks (×0.5 for alluvium, loess and dunes). However, they cannot be differentiated by their proportions of Precambrian basement plus metamorphic rocks; large shares of these rock types are incorporated into orogenic processes, mainly metamorphic rocks in young orogens.

[143] The perspectives of this work and ways of improving it are numerous following the path demonstrated by our predecessors [Gibbs and Kump, 1994; Ludwig and Probst, 1998; Ludwig et al., 1999; Amiotte-Suchet et al., 2003]. Figure 5 represents the position of surficial rocks and river systems within the Earth System as conceived by Garrels and Mackenzie [1971] and now widely promoted by the global-cycle scientific community, for example the International Geosphere Biosphere Programme [Steffen et al., 2004]. Surficial rock composition is a major control factor of chemical weathering (fluxes and interactions 2A and 2B, Figure 5), i.e., of CO2 uptake and transfer from the atmosphere to plants, of ions and nutrients (Si, P), and trace element fluxes from river to oceans (fluxes 1, 2B) and of mechanical erosion of surficial rocks and soils (fluxes 2A, 2B). Furthermore, the distribution of present-day lithology is an indicator of the relative influences of tectonics, metamorphism, volcanism and sedimentary processes during geological times (flux 3) [Ronov et al., 1980; Mackenzie and Mackenzie, 1995; Veizer and Mackenzie, 2003]. The lithology also controls the present formation and nature of continental surficial rocks accumulated in different types of Earth system filters [Meybeck and Vörösmarty, 2005] (1) in continental sinks such as internal seas and closed lake basins (flux 4B), and in floodplains and lakes found in the external drainage sinks (flux 4A) and (2) on the continental shelf (flux 5), addressed by sedimentologists, physical geographers and ocean scientists [Milliman and Syvitski, 1992; Mackenzie and Mackenzie, 1995; Summerfield, 1991, 2000].

Figure 5.

Schematic position of surficial lithology and river systems fluxes within the major Earth system components of the Anthropocene era (modified from Meybeck [2003]). Main natural fluxes: 1, biological uptake of atmospheric N2 and CO2; 2A, 2B, mechanical erosion and chemical weathering; 3, rock uplift and transformation (tectonics, metamorphism, volcanism); 4A, 4B, river transfer to oceans and internal drainage; 5, retention of river material within river systems; 6, net input to coast; 7, 8, shelf sediments and oceanic sediments. Main anthropogenic impacts: A, atmospheric pollution and climate change, N2 industrial fixation, CO2 fixation in irrigated fields; B, cropping, agrochemicals use; C, irrigation, water use; D, mining, construction; E, F, river and coast engineering; G, reservoirs, irrigation.

[144] Human activities such as deforestation and agriculture, mining, N-fixation by fertilizer, construction, industrial and urban wastes, and artificialization of river networks (impacts A to G, Figure 5), here referred to as the Anthroposphere, have gradually modified the surface of the Earth and many of its fluxes at the global scale, within a complex combination of acceleration and retention [Steffen et al., 2004; Vörösmarty and Meybeck, 2004; Messerli et al., 2000; Meybeck and Vörösmarty, 2005]. The related changes in fluxes now match or exceed the natural fluxes particularly those to the atmosphere and the river systems. This new era has been described as the Anthropocene [Crutzen and Stoermer, 2000; Meybeck, 2002, 2003; Meybeck and Vörösmarty, 2005]. In the Anthropocene, many sources of riverine material are activated by earth removal for mining or public works (impacts D) [Hooke, 2000], enhanced erosion and pollution, but new filters are also created within river systems [Meybeck and Vörösmarty, 2005] for example, reservoirs [Vörösmarty et al., 2003] and irrigated land (impacts C and G). The resulting river fluxes may either increase or decrease: In some basins the material brought by land erosion and/or human activities is actually stored in these new filters and does not reach the ocean [Meybeck and Vörösmarty, 2005]. Globally it has been estimated that at least 30% of riverine particulates are stored in reservoirs [Vörösmarty et al., 2003], and more than 90% of river systems are affected by both reservoir storage and irrigation, for example, those of the Colorado, the Aral Sea Basin, the Nile [Meybeck and Vörösmarty, 2005].

[145] The products of continental erosion and weathering are no longer regulated solely by natural drivers as lithology, climate and tectonics. Therefore, together with the improvement of the lithology database, a multiple set of databases on the Anthroposphere should now be established on land use characteristics, reservoirs, population and urbanization, mining, industries, agrochemical practices. Such bases and previous lithology distribution data have already been combined for the global modeling of carbon [Ludwig et al., 1996], phosphorus [Caraco, 1995], nitrogen [Caraco and Cole, 1999; Seitzinger et al., 2002; Green et al., 2004] and their interactions [Ver et al., 1999], and are now planned for silica modeling [Conley, 2002].

[146] It is hoped that our map will now be used by interested parties who will help to improve it. Probably the best way to improve this map would be to work at the regional scale (e.g., regional sea basins, major mountain ranges, very large river basins, major biomes of the world). We hope that such international cooperation may be organized, in the same way as within the UNESCO “International Geological Correlation Programme” (IGCP) and the Commission for the Geological Map of the World (http://ccgm.free.fr/).

Acknowledgments

[147] We are particularly indebted to Charles J. Vörösmarty, head of the Water Systems Analysis Group (University of New Hampshire), and to Stanley Glidden for the communication of the digitized version of the UNESCO Geological Map of the World which served as a basis for this work. We also greatly acknowledge the continuous support of Pamela Green and the communication of LGM limits by Richard Lammers, both at UNH. Hans Dürr's work has been funded by a Ph.D. grant from the French “Ministère de la Recherche” and by the European Union Si-WEBS programme (EU contract HPRN-CT-2002-000218). Discussions with numerous colleagues from geology departments have been appreciated, as with A. Semmel for the loess distribution in the Americas. We are also grateful to M. Gibbs for the transfer of their lithology map and his advice at the initial stages of this project.

Ancillary