Regional crustal thickness and precipitation in young mountain chains

Abstract

Crustal thickness is related to climate through precipitation-induced erosion. Along the Andes, the highest mountains and thickest crust (≈70 km) occur at 25° south, a region of low precipitation. Westerly winds warm passing over the Atacama Desert; precipitation is modest in the High Andes and eastward over the Altiplano. Severe aridity, hence low erosion rates, helps to account for the elevated volcanogenic contractional arc and high, internally draining plateau in its rain shadow. Weak erosion along the north-central arc provides scant amounts of sediment to the Chile–Peru Trench, starving the subduction channel. Subcrustal removal might be expected to reduce the crustal thickness, but is not a factor at 25° south. The thickness of the gravitationally compensated continental crust cannot reflect underplating and/or partial fusion of sediments, but must be caused chiefly by volcanism-plutonism and contraction. Contrasting climate typifies the terrane at 45° south where moisture-laden westerly winds encounter a cool margin, bringing abundant precipitation. The alpine landscape is of lower average elevation compared with the north-central Andes and is supported by thinner continental crust (≈35 km). Intense erosion supplies voluminous clastic debris to the offshore trench, and vast quantities are subducted. However, the southern Andean crust is only about half as thick as that at 25° south, suggesting that erosion, not subcrustal sediment accretion or anatexis, is partly responsible for the thickness of the mountain belt. The Himalayas plus Tibetan Plateau, the Sierra Nevada plus Colorado Plateau, and the Japanese Islands exhibit analogous relationships between crustal thickness and climate.

Atopic of intense geomorphic research involves quantifying the competing influences of contractional and accretionary tectonic processes plus uplift versus surficial, precipitation-induced erosion in controlling the topography of active mountain belts (1–5). Because of the orographic effect, climate and surface elevation must be intimately coupled in terms of cause and effect. Reflecting regional attendant gravitative equilibrium (isostacy) the thickness of the mountain belt crust is also partly a function of rain and snowfall. Using the Andes, Himalayas, Sierra Nevada, and Japanese Islands as examples, the modern thickness of the continental crust is shown to be related in part to present-day annual precipitation. Consequently, the interplay of igneous activity, sedimentary and tectonic accretion, compressional shortening, subcrustal removal, exhumation, and climate all play important roles in determining the regional crustal thickness of active mountain belts.

Contrasts Between Pacific- and Alpine-Type Mountain Belts

Andean-type convergent plate boundaries evolve where thousands of kilometers of oceanic lithosphere are consumed without the introduction of substantial amounts of continental crust into the subduction zone. This plate-tectonic realm produces an outboard, largely metasedimentary accretionary trench complex, a medial, longitudinal forearc basin, and an inboard calcalkaline volcanic-plutonic arc. The relatively narrow trench depositional prism consists of a low-temperature, low-heat-flow belt in which folds are overturned oceanward and thrust faults roughly parallel the inclined subduction zone, whereas the broad magmatic arc is typified by open folding and a high-temperature, high-thermal-flux regime (6–10). The former is generated exclusively on oceanic crust (e.g., Franciscan, Aleutian, and Chile–Peru trench systems); the latter is constructed on a preexisting basement consisting of continental margin or island arc plus/minus older oceanic crust (e.g., Sierran, Indonesian, and Andean arcs).

Collision-type convergent plate boundaries form where subduction rollback (11) of ocean lithosphere results in the insertion of a continental promontory, microcontinent, or island arc beneath a nonsubducted, continental crust-capped plate. A salient of the downgoing continental crust may descend to great depth imbedded in the sinking oceanic lithosphere because the overall density of the dominantly oceanic plate exceeds that of the underlying asthenosphere. During underflow, the thickness of the suprasubduction-zone continental crust is increased by underplating, contraction, and amalgamation/continental collision. Thrust sheets dipping beneath the stable plate and a paucity of calcalkaline igneous activity characterize such collisional mountain belts. Typical examples include the Urals, Alps, and Himalayas (12–16). Of course, continental lithosphere can be carried down beneath young, hot, thus less dense, oceanic lithosphere (e.g., Oman and Sulawesi; refs. 17 and 18), but such cases are uncommon because most oceanic lithospheric plates are negatively buoyant (and roll back) relative to continental plates.

Climatic Patterns and Crustal Thickness Near Young Convergent Plate Junctions

General Statement. The thickness of the continental crust is a dynamic product of competing constructional and destructional processes (19–24). The former involve primary calcalkaline magmatic additions derived from the mantle, plate tectonic convergence (contraction), transform motion (strike-slip), and/or divergence (extension), terrane suturing plus/minus accretionary prism offloading/underplating. The latter include surficial erosion, subduction-induced subcrustal removal, and delamination plus foundering of the lower, mafic crust. Pacific-type paired mountain chains consist of an outboard subduction complex and a subparallel inboard Andean arc, whereas Alpine-type mountain belts chiefly display the effects of continental collision. Ancient platforms have attained thermal and architectural stasis throughout the lithosphere, reflecting tectonic quiescence provided by long-term isolation from upper mantle circulation, and the consequent establishment of an unperturbed mantle-crust geothermal gradient. In contrast, geologically youthful mountain belts may possess thinner, or more commonly, thicker, crust than do old cratons because, although regional Airy isostatic equilibrium is closely approximated, the complex interplay between constructional and destructional forces is influenced substantially and to varying degrees by tectonism and local erosion rates.

Reflecting regional gravitative equilibrium, the mean elevation of a growing mountain belt is a time-integrated reflection of the petrotectonically accumulated volume per unit area, modified by surficial degradation (25–30). As a very rough approximation, erosion rates in modern mountain ranges track with precipitation (31). This is a gross oversimplification of a much more complex set of variables (2), including chemical and physical weathering, nature and integrity of the geologic substrate (e.g., resistance of the bedrock complex to erosion), proximity to local and regional erosional base levels, etc. Other factors being equal, greater topographic gradients promote more rapid degradation for a given precipitation rate. But as rain and snowfall decrease, contraction and uplift may block external drainage systems and produce elevated, internally draining plateaus (32–34). The regional climate itself is a multivariant function of the Earth's dynamic atmosphere plus hydrosphere structure, including wind and near-surface ocean circulation patterns (functions of latitude, positions of obstructing land masses, Coriolis effect, etc.). Annual global precipitation zonations illustrated in Fig. 1 clearly transect some mountain chains at large angles, resulting in longitudinal variations in erosional efficacy.

Fig. 1.

Modern average annual precipitation in inches around the globe (modified from figure 3 in ref. 35). The cool California, Humboldt, and Tasman currents provide moisture-laden westerly winds at higher latitudes and relatively H₂O-poor winds at lower latitudes. The warm Indian Equatorial Gyre and the Kuroshio Current bring oodles of moisture via warm, H₂O-laden winds to the south-facing Himalayas and southwestern Japan. Note the marked north-south variation of rain and snowfall along the Sierra Nevada, Chilean Andes, and Himalayas-Tibetan Plateau, as well as the Japanese Islands.

Do Climatic Processes Impact Crustal Scale Features of an Active Mountain Belt? The present synthesis suggests that the answer is yes. As an example, consider an active continental margin belt such as the contractional, volcanogenic Chilean Andes, which parallels the outboard Chile–Peru convergent plate junction and surmounts the eastward descending Nazca oceanic-crust capped plate. Fig. 2 presents general topographic and climatic relationships.

Fig. 2.

Topography, average annual precipitation in meters, and relative plate velocities along the west coast of South America (simplified from figure 1 in ref. 36). (Left) Topography and plate convergence. (Right) Average annual precipitation. The mean elevation of the Andean mountain belt at 25° south exceeds 3 km, reflecting thick continental crust, whereas ≈20° latitude farther south in Bernard O'Higgins Land, the regional elevation is somewhat less than 1 km, indicating thin continental crust.

The highest mountains and intensely shortened, thickest continental crust (55–75 km) occur in the north-central Andes at 20–25° south (25, 37–41). This is the atmospheric realm of high-pressure descending, heating Hadley cells. Cold upwelling water and the north-flowing Humboldt Current lie directly offshore. Along this continental margin, the weak westerly winds are relatively cold, thus low in H₂O content; they warm as they pass over the Atacama Desert, and yearly precipitation is only ≈0.2–0.3 m in the High Andes (36). Air parcels continuing eastward beyond the continental divide sink over the Altiplano (3.6-km average elevation); low humidity characterizes this region, with yearly precipitation rates of ≈0.3 m. Low precipitation plus low erosion rates partially account for the elevated volcanic-plutonic arc and the internally draining high plateau on the east (34, 42, 43). Only minor amounts of sediment are transported eastward toward the Amazon Basin and westward toward the Chile–Peru Trench. Because erosion is weakly developed along the north-central cordillera, little volcanoclastic debris is carried down the Chile–Peru subduction channel (24, 44). Sediment-starved subduction zones might be anticipated to constitute sites of active subcrustal erosion, as is characteristic of some segments of the Chile–Peru continental margin (45). Accordingly, the great regional thickness of the continental crust in this sector must be mainly a consequence of calcalkaline volcanism-plutonism combined with major tectonic shortening and feeble erosion, and cannot be caused by accretionary offloading or partial fusion of subducted materials.

The climatic situation is markedly different around 45° south. Here, the temperature contrast between the Humboldt Current and the land is reversed, with the continent being slightly cooler than the nearby ocean. Low-pressure Ferrell cells in the atmosphere rise at this latitude, supplying abundant H₂O-laden winds to the cool terrestrial surface. Consequently, the rugged, fjord-rich but much lower elevation Bernard O'Higgins Land is richly endowed with ≈2–4 m of annual rain and snowfall, producing snowfields and glaciated valleys (36). The alpine landscape is supported by a continental crust ≈35 km thick (46, 47). Reflecting moderate precipitation downwind, eastward stream drainage is external and a high plateau is lacking as sediments escape to the Atlantic margin. On the western slopes, heavy precipitation, vigorous stream flow, and active erosion transport voluminous clastic debris westward to the offshore Chile–Peru Trench, and vast amounts of quartzofeldspathic material are subducted (24). Nevertheless, the southern Andean crust is only about half as thick as that at 25° south, demonstrating that intense erosional decapitation (a reflection of climatic patterns), rather than sediment underplating plus/minus anatexis is at least partly responsible for the modest crustal thickness.

Although divided into segments of differing subduction inclinations, extents of recent glaciation, and amounts of tectonic shortening, the physicochemical nature of the Nazca plate is roughly similar in both north-central and southern Andes; thus, primary calcalkaline igneous contributions to crustal thickness should be roughly comparable in both portions of the mountain belt. A substantial sediment cushion in the subduction zone at 45° south might be expected to inhibit subcrustal erosion and instead favor accretionary offloading and growth, but if such processes are operating (45), they have not generated a massively thick arc in Bernard O'Higgins Land.

Ongoing continental collision and intracrustal thrusting accounts for the great thickness (65–80 km) of the Himalayan continental crust (48–52). Fig. 3 shows annual precipitation as a function of geography for the Himalayas and Tibetan Plateau. Because precipitation zonation patterns coincide with the east-west orientation of the range, the Himalayas lack a strong axial gradient in erosion rates. The remarkable mean elevation of this mountain chain (>5 km) is chiefly a reflection of compressional tectonism. The structurally induced relief on the southern side of the belt is somewhat modified by extremely rapid erosion (4, 5, 53). H₂O-saturated air derived from the Indian Equatorial Gyre is drawn northward in response to low-pressure systems created by parcels of hot air ascending over central Asia. Juicy monsoonal winds cool as they pass over India and encounter the Himalayan Range, resulting in annual precipitation rates approaching 3–10 m (43, 53). To the north, the adjacent, internally draining Tibetan Plateau lies in the Himalayan rain shadow and thus may owe part of its great elevation and crustal thickness (54, 55) to aridity; yearly precipitation is <0.5 m, accounting for the attendant low erosion rates.

Fig. 3.

Average annual precipitation in mm, major peaks, drainage network, and gauging stations in the Himalayas and adjacent Tibetan Plateau (after figure 1 in ref. 53). Outlined study area of Findlayson et al. (53) includes both eastern and western syntaxes (maximum uplift areas). Note the pronounced rain shadow north of the Himalayas.

The Sierra Nevada Range in eastern California lies east of the southeast-flowing California Current. The latter, coupled with westerly winds, brings abundant moisture to the cool Pacific North-west and northern California. Farther south, the land is considerably warmer than the ocean, resulting in progressively greater aridity. Eastward from central-southern California, the Sierra achieves its maximum regional elevation in the Mount Whitney area. The crustal thickness of the southern Sierra Nevada (42–55? km) appears to exceed that of the northern Sierra Nevada (≈35 km), according to the limited amount of geological and geophysical data available (56–59). Contrasts in Sierran crustal thickness are not as marked as in the Andes and Himalayas, presumably because the constructional phase of mountain building largely ceased in the Late Cretaceous. And what about the Colorado Plateau (2.4-km average elevation), located downwind from the southern Sierra? It lies well to the leeward of the high Sierra, but the intervening Basin and Range Province may represent Neogene crustal extension of an initially somewhat larger plateau. As illustrated in Fig. 4, the average annual precipitation for the American Southwest appears to be more or less inversely correlated with elevation, and thus the thickness of the crust. Episodes of extensional tectonics clearly have severely stretched the continental crust in the intervening Basin and Range Province, but it seems possible that, in addition to volcanic-plutonic constructional processes, erosion linked to presentday climate patterns may have influenced the current regional crustal thickness in both the Sierra Nevada and Colorado Plateau.

Fig. 4.

Average annual precipitation throughout the American West in inches (modified from regional map, Oregon Climate Service at www.ocs.oregonstate.edu/prism). The Colorado Plateau lies far downwind from the high-elevation southern Sierra Nevada but is separated from it by the Neogene extensional crust of the Basin and Range Province.

Another well studied mountain belt, the Japanese island arc, exhibits strongly zoned rainfall patterns (Fig. 1). Annual precipitation in western Chugoku, Shikoku, and Kyushu exceeds that in northern Honshu, reflecting moist winds that blow off the warm, northward flowing Kuroshio Current and over southwestern Japan. However, the plate tectonic environment of the Japanese arc is quite complex, with the old Pacific lithosphere descending beneath northern Honshu and the young Philippine Sea plate sinking under western Chugoku (60). The Japan Trench is receiving only modest volumes of sedimentary debris, compared to the clastic load entering the Nankai Trough. Nevertheless, northeast Japan appears to have a continental crustal thickness of ≈40 km (61–63), in contrast to southwestern Japan where the crust is on the order of 30–35 km thick (64–66, †).

Summary

Silicic/intermediate crust is generated by calcalkaline magmatism in island arcs and continental margins situated over convergent plate junctions; some of it represents net additions to the continental crust. Paired Pacific-type mountain belts develop above zones of long-lived subduction of oceanic crust-capped lithosphere. They consist of an outboard, low-heat-flow accretionary prism deposited in and landward from the trench, and an inboard, high-heat-flow volcanic-plutonic arc. The trench assemblage consists dominantly of clastic debris derived from the nearby, contemporaneous arc, but includes tectonic fragments of the oceanic plate. A massive magmatic arc dominates the landward belt where new continental crust is added. Alpine-type chains form during underflow of an oceanic plate that transports continental crust into the subduction zone. Growth of the continental margin crust is largely a regional function of the operation of petrotectonic processes, such as calcalkaline magmatism, terrane accretion, and tectonic contraction. Mountain building results in regional crustal thicknesses that are comparable to, or exceed, those of old, cold, stable cratons (67).

But reflecting the intensity of erosional removal, the thickness of the continental crust in a geologically youthful contractional mountain belt is also to some extent a function of climatic patterns. For a given convergent plate tectonic environment, the mean elevation of a mountain chain and therefore the regional thickness of the isostatically compensated continental crust are inversely correlated with precipitation and consequent erosion. In addition, tectonic shortening (crustal thickening) and attendant rapid uplift in rain-shadow realms can defeat rapid stream erosion and promote the formation of internally draining, high-elevation plateaus typified by only modest sedimentary deposition and/or erosion.