Cretaceous coastal mountain building and potential impacts on climate change in East Asia Authors: Jianhua Li, Shuwen Dong, Guochun Zhao, Peter A. Cawood, Stephen T. Johnston, Jian Zhang, Yujia Xin, Jinming Wang Published: 2024 ABSTRACT Crustal thickness and elevation variations control mountain building and climate change at convergent margins. As an archetypal Andean-type convergent margin, eastern Asia preserves voluminous magmas ideal for quantifying these processes and their impacts on climate. Here, we use Sr/Y and Ce/Y proxies to show that the crust experienced alternating thickening and thinning during the Late Mesozoic. We identify a noticeably thickened (50 to 55 kilometers) crust associated with tectonic shortening at 120 to 105 million years, corresponding to a >2500-meter-high coastal mountain range. Using climate simulation with the Community Earth System Model, we demonstrate that the mountain uplift changed Asian atmospheric circulation and precipitation patterns, increased inland aridity (~15%), and prompted the eastward desert expansion, contributing substantially to the arid zonal belt across mid- to low-latitude Asia. These findings—compatible with independent geological, geophysical, and climatic observations—have global implications for broadening our understanding of Earth-system interactions in the Cretaceous greenhouse world. INTRODUCTION Crustal thickness and elevation, intimately linked to mountain-building processes, influence the stability and evolution of continents, topography, atmospheric circulation, and precipitation (1–3). Quantifying the evolution of crustal thickness and elevation has implications for unraveling the interactions and feedback among tectonics, topography, and climate from an Earth-system perspective (4, 5). However, obtaining evidence of mountain building, the scale of topographic changes, and related impacts on climate has proved challenging. As an archetypal Mesozoic Andean-type convergent margin (6), East Asia offers opportunities to investigate the interactions and feedback between tectonics, topography, and climate. One of East Asia’s most notable and enigmatic features in the Cretaceous was the topographically high coastal mountain range along its eastern margin (7). This feature contrasts sharply with the present-day topographic pattern characterized by a broad, low-elevation, low-relief coastal plain to the east and a high-elevation plateau (i.e., the Tibetan Plateau) (8) to the west. However, quantifying how the high topography links to tectonics and affects the climate poses a major challenge. Recent recognition of Jurassic-Cretaceous arc magmas along the eastern margin of East Asia testifies to the subduction of the Paleo-Pacific plate (9, 10), which gave rise to a northeast (NE) trending, Andean-type orogenic belt (6, 10) extending >1000 km into the continental interior (Fig. 1, A and B). As part of the circum-Pacific orogenic system, this orogenic belt is essential for understanding Andean-type tectonics, paleocrustal and paleoelevation evolution, and climate patterns at convergent continental margins. Alternating contraction and extension dominated its tectonic evolution and topography during the Late Mesozoic (9). The most conspicuous topographic feature was the Cretaceous emergence of a prominent coastal mountain range with loosely constrained elevations of >2000 m along the continental margin (7, 11). Understanding the coastal mountain building and its impacts on Asian climate are crucial issues for (i) paleotectonic and paleogeographic reconstructions of East Asia and (ii) clarifying the tectonic-climate interactions in the context of the Cretaceous greenhouse world. In particular, the high elevations at convergent margins may affect atmospheric circulation, change precipitation patterns, and enhance chemical weathering and organic carbon burial, ultimately affecting climate patterns (12). Nevertheless, high-quality paleocrustal thickness and paleoelevation data (7) are sparse and incomplete. Fundamental questions, especially regarding the spatial-temporal variations in paleocrustal thickness and paleoelevation during alternating contraction and extension, remain uncertain due to the lack of robust constraints. Tectonic framework of the South China block, East Asia. (A) Topographic map with minor faults in blue and major boundary faults in white. Circles show the locations of igneous rocks that have been used in the present study after passing through the geochemical filters (green circles, mafic rocks; blue circles, intermediate rocks) for quantifying crustal thickness variation. (B) Simplified geological map showing major faults, folds, and igneous rocks in South China and adjacent regions. Abbreviations for extensional domes or metamorphic core complexes (28, 29): MFS, Mufushan; HS, Hengshan; LS, Lushan; YCL, Yuechengling. Global compilations indicate that the geochemistry of convergent margin magmas is sensitive to crustal thickness (13–15). Crustal thickness controls arc magma composition by affecting magmatic differentiation and modulating the degrees of crustal contamination and mantle melting (16). A series of chemical and isotopic parameters serve as excellent proxies for tracing the evolution of paleocrustal thickness in ancient subduction systems (16–19). In particular, Sr/Y and Ce/Y ratios are qualitative indicators of the pressures (i.e., crustal depths) at which partial melting and crystal fractionation occur (20–23). Given the close relationship between crustal thickness and surface elevation, the estimated paleocrustal thickness data provide a basis for reconstructing paleoelevation and mountain-building processes (3, 24, 25). Extensive subduction-related magmatism dominated the Late Mesozoic tectonic evolution of East Asia (26, 27). Resulting magmatic belts cover an area of >200,000 km2 and trend NE–southwest (SW) parallel to regional structures in South China (Fig. 1B). Magmatic activity includes two prominent flare-ups associated with extensional events at 150 to 120 million years (Ma) and 105 to 80 Ma (Fig. 2). The intervening 120- to 105-Ma magmatic decline coincides with a remarkable, regional-scale episode of tectonic shortening (28, 29). These magmas are ideal for tracking changes in paleocrustal thickness and paleoelevation. On the basis of their geochemical and isotopic data, we conduct a multidisciplinary study to examine dynamic interactions among tectonics, topography, and climate. First, we investigate spatial-temporal variations in paleocrustal thickness by combining Sr/Y ratios of intermediate-felsic rocks and Ce/Y ratios of arc basalts. Second, we use the paleocrustal thickness data to estimate the distribution and paleoelevation of the coastal mountain range. Last, we use an approximate paleogeographic reconstruction coupled with climatic modeling to assess how the high topography would affect atmospheric moisture patterns and climate variability across East Asia. Our findings provide crucial insights into the linked processes of crustal thickening, mountain building, and climate change along Andean-type convergent margins. Age distribution of Late Mesozoic igneous rocks from South China. The graph shows two magmatic flare-ups (150 to 120 Ma and 105 to 80 Ma) separated by a pulse of weak magmatism. RESULTS Paleocrustal thicknesses calculated from Sr/Y proxy Crustal thickness is the main parameter controlling the development of adakite-like signatures by affecting the stabilization of amphibole and/or garnet and the destabilization of plagioclase (3). Sr and Y show pressure-dependent differences in the partitioning behaviors into plagioclase, garnet, and amphibole (21). Sr/Y ratios of subduction-related magmas reflect the pressures (i.e., crustal depths) at which melting and magmatic differentiation occurred (22). We compiled a comprehensive database of whole-rock geochemical, geochronological, and isotopic analyses from 273 Jurassic-Cretaceous plutons and volcanoes from eastern South China (https://figshare.com/articles/dataset/Dataset_of_South_China_magmatic_rocks/25115198). Geochemical data were screened for Sr/Y outliers (see Materials and Methods) using the principles identical to those used for quantifying crustal thickness variations in the North American Cordillera (19, 21). First, we filtered data to include only intermediate-felsic rocks with 55 to 70 weight % (wt %) SiO2 and 1 to 6 wt % MgO. Second, we discarded data subsets with anomalously high Sr/Y (>40) and low Rb/Sr (<0.05) ratios to exclude the effects of plagioclase accumulation during magma ascent and emplacement (fig. S1), as plagioclase accumulation results in highly elevated Sr contents (30, 31). Third, we discarded data subsets of highly fractionated (Rb/Sr > 0.32 and Sr/Y < 10; fig. S1) magmas because plagioclase crystallization modifies the compositions of primary magmas by removing Sr from remaining melts (21). Fourth, we removed analyses of mineralized and highly altered rocks with a loss on ignition (LOI) ≥ 4 wt % because high LOI values indicate that these rocks underwent post-magmatic alteration (e.g., weathering) that could affect the abundances of fluid-mobile elements (e.g., Rb and Sr) and Sr/Y ratios (32). The remaining analytical data that passed through these geochemical filters are used to estimate paleocrustal thickness. These filtered analyses—with assigned ages, latitudes, and longitudes—are provided in table S1. We calculated the paleocrustal thickness using an empirical equation from the global geochemical database for subduction-related rocks (3, 21)z=1.11×Sr/Y+8.05,R2=0.90(1)where z is the paleocrustal thickness and Sr/Y is the whole rock average ratio. This equation has been corroborated and formulated by examining calculated crustal thicknesses with geophysically derived Moho depths (21). A number of studies have demonstrated its applicability to Paleozoic to Cenozoic subduction systems (3, 14, 19, 21, 22). Figure 3A plots the Sr/Y ratios and calculated paleocrustal thickness data against ages. Following Mantle and Collins (20), we use the maximum ratios for comparison to reduce the potential effects of crustal thickness inhomogeneities. Using OriginPro 2024b, we have conducted the polynomial regression analysis and obtained a curve fitting for estimating the variations of Sr/Y ratios and paleocrustal thickness data through time (see Materials and Methods). Our calculated results reveal a steady decrease in Sr/Y through the late Jurassic into the Early Cretaceous until 125 to 120 Ma. Then, the ratio increases rapidly until ca. 110 Ma, followed by another systematic decrease in Sr/Y from ca.110 to ca. 95 Ma (Fig. 3A). The progressive decline in Sr/Y between 161 and 125 Ma (Fig. 3A) implies a decrease in paleocrustal thickness from 50 to 60 km to 20 to 30 km over this interval. The onset of subsequent Sr/Y increase at 125 to 120 Ma implies a tectonic switch from crustal thinning to thickening. With continuous crustal thickening, the crust appears to attain a maximum thickness of 50 to 55 km during 110 to 105 Ma (Fig. 3A). Decreasing Sr/Y from 105 to 95 Ma indicates renewed crustal thinning and a decrease in thickness to <40 km by ca. 95 Ma (Fig. 3A). Paleocrustal thickness reconstruction. (A and B) Plots of Sr/Y and Ce/Y ratios in the magmatic rocks and the corresponding calculated crustal thicknesses for the South China magmatic belt, East Asia. The solid lines are based on the polynomial regression analysis. They show the corresponding curves fitting for estimating the pattern of inferred crustal thickness variations through time. Time constraints of crustal shortening and extensional episodes are from (28, 29). Paleocrustal thicknesses calculated from Ce/Y proxy Arc basalt composition varies as a function of crustal thickness, with the maximum Ce/Y of arc basalts increasing with the Moho depth (20). Several studies have demonstrated that Ce/Y ratios in basalts can be applied to the reconstruction of crustal thickness (3, 20). In particular, quantitative estimates of changes in paleocrustal thickness evolution of the Median Tectonic Zone (Median Batholith) of New Zealand since 400 Ma have been obtained using maximum Ce/Y ratios of arc basalts (20). In contrast to widely distributed arc-related intermediate magmas in East Asia, mafic magmas are less abundant and confined to the coastal area. They are subalkaline and belong to the calc-alkaline series (27). Following the pioneering work of Mantle and Collins (20), we filtered geochemical data for mafic rocks from South China by SiO2 (45 to 53 wt %), MgO (>4 wt %), and LOI (<4 wt %) (table S2). A systematic pattern emerges (Fig. 3B) when plotting Ce/Y ratios against ages using the following equationz=18.0505×ln(Ce/Y)+21.5587,R2=0.8979(2)where z represents the paleocrustal thickness (3, 20). The 135 to 120 Ma Ce/Y data are less abundant and show some scatter (Fig. 3B). Our results reveal that the maximum Ce/Y ratios increased progressively from ~1.2 at ca. 120 Ma and reached a maximum (~3.9) at ca. 105 Ma (Fig. 3B). During the Late Cretaceous, Ce/Y ratios steadily decreased, dropping to 2.1 at ca. 80 Ma (Fig. 3B). Calculations of crustal thickness from Ce/Y ratios indicate that the paleocrustal thickness increased from ~25 km to 45 to 50 km between 120 and 105 Ma and then progressively decreased to 30 to 40 km by ca. 80 Ma (Fig. 3B). The overall Ce/Y pattern and the maximum Ce/Y correspondence are broadly consistent with those obtained from the Sr/Y proxy (Fig. 3B). The highest Ce/Y ratios at ca. 105 Ma correspond to a maximum crustal thickness of 45 to 50 km. Paleoelevation estimates Crustal thickness isostatically controls topography and elevation (12). We assume local isostatic compensation [Airy, (33)], motivated by a weak South China lithosphere (34). Because of the Cenozoic decline of the coastal mountain range (7), the current crustal and mantle parameters (e.g., density and velocity values) related to isostatic compensation may differ from those in the Cretaceous. Given the Andean margin tectonic setting of South China in the Mesozoic (9, 29), we use the empirical equations relating elevation and crustal thickness from the Andes for quantifying the Mesozoic isostatic compensation in South China. To establish an equation with appropriate parameter values applicable to more Andean-type orogenic belts, we present a dataset by compiling all available elevation and crustal thickness data from the Andes (Fig. 4). This dataset contains 198 samples with reliable elevation (0 to 5000 m) and crustal thickness (25 to 75 km) data (table S3) (33, 35–39). Despite some scatter in the dataset, it reveals a linear correlation between elevation and crustal thickness (Fig. 4). Assuming Airy isostatic equilibrium, the elevation h may be determined from the crustal thickness z from the equationh=(1−ρc/ρm)×(z−z0)(3)where ρc and ρm are the average densities of the crust and the lithospheric mantle, respectively. z0 is a reference crustal thickness. We determine a reference crustal thickness (z0) of 25.0 km for the best-fitting regression line, with a correlation coefficient of R2 = 0.83 (Fig. 4). The best-fitting regression line is broadly comparable to those determined from global arcs and subduction zones (3, 12). The regression analysis indicates an optimal ρc/ρm ratio of 0.9 (Fig. 4). To test its reliability, we calculated average crustal and mantle densities by integrating the compiled data locations (table S3) and their corresponding density information from the lithospheric structure of the Central Andes (40). The calculated results (ρc = ~2.93 g/cm3 and ρm = ~3.26 g/cm3; Fig. 4) fall within the observed crustal and mantle density values (2.85 to 2.95 g/cm3 and 3.20 to 3.35 g/cm3) in the Central Andes (41, 42) and coincide with the predicted ρc/ρm ratio (Fig. 4), validating the reliability of the regression. Using the regression equation, we convert crustal thickness to elevation, assuming that the Late Mesozoic crust of eastern South China was in isostatic equilibrium (Fig. 5A and table S4). Our results suggest that eastern South China attained a relatively high elevation of >2500 m between 120 and 105 Ma due to isostatic surface uplift associated with crustal thickening, giving rise to a coastal mountain range along the convergent margin (Fig. 5B and table S4). Elevations subsequently decreased to ~1000 m in the Late Cretaceous (Fig. 5A). Continental elevation (h) versus crustal thickness (z) in the Andes. The plots quantify the isostatic compensation in the Andean-type orogenic belts, inspired from figure 1 in (12). Data from the Western Cordillera, Altiplano-Puna Plateau, Eastern Cordillera, and adjacent regions come from (33, 35–38). Data from the South Andes come from (37, 39). The stippled purple line represents the best-fitting regression line, with a correlation coefficient of R2 = 0.83. Data show some scatter around the best-fitting regression line within a band, possibly due to (i) local flexural rigidity of the lithosphere (33) and (ii) effects of a dense eclogitic lower crust (58). Paleoelevation reconstruction of eastern South China in East Asia. (A) Time series of paleoelevation (km) between 180 and 80 Ma. On the basis of the polynomial regression analysis, the solid lines show the corresponding curves that fit to estimate the pattern of paleoelevation variations over time. Colored triangles represent the minimum paleoelevation data obtained from other studies, including estimates from carbonate-clumped isotope paleothermometry (7), carbon and oxygen isotopes (60), and atmosphere-ocean general circulation modeling (61). The blue dotted line represents the paleoelevation evolution inferred from fission-track thermochronology (62). (B) Changes in paleocrustal thickness and paleoelevation across eastern South China during the late Early Cretaceous (120 to 105 Ma). The solid lines show the fitted curves to estimate the pattern of paleoelevation and paleocrustal thickness variations. To evaluate the reliability of paleoelevation estimates, we adopt another approach to reproduce the paleoelevation variations through time. Given that elevation correlates with crustal thickness due to the isostatic equilibrium (12), this approach is primarily based on the covariations between elevation and crustal thickness. We start by investigating more easily measurable parameters in situ, including the present-day elevation, crustal thickness [fig. S2A, (10)], and average crustal and mantle densities (43). Having these parameters and integrating geographic coordinates combined with location data, we recalculate paleoelevation using the following equationsh=Δh+hpd(4)Δh=(1−ρc−pd/ρm−pd)×(z−zpd)(5)where h is the calculated paleoelevation, Δh is the variance for elevation, hpd is the present-day elevation, zpd is the present-day crustal thickness (44), z is the calculated paleocrustal thickness, ρc–pd is the present-day crustal density, and ρm–pd is the present-day mantle density. Integrating the data locations (fig. S2A and table S5) and their corresponding density information (43), we calculated the average crustal and mantle densities (ρc–pd = ~2.85 g/cm3 and ρm–pd = ~3.20 g/cm3), which fall within the observed crustal and mantle density values (2.75 to 2.95 g/cm3 and 3.15 to 3.35 g/cm3) of eastern South China (43). The results of calculations from these equations provide a reasonable set of paleoelevation estimates (fig. S2B), which are broadly consistent with those from the previous approach (Fig. 5A). Among these estimates, >95% of the elevations are reproduced within ±500 m (fig. S2C). Although this approach appears more straightforward, one disadvantage is that it assumes the present-day isostatic equilibrium to be the same as that presented during the Late Mesozoic Andean-type mountain building. However, they should be different due to the Cenozoic collapse of the mountain range (45, 46). For example, the sensitivity of various parameters (e.g., ρc, ρm, and ρc/ρm) to crustal thickness variations tends to change with the isostatic equilibrium mode. Despite different approaches, their calculation results generally agree. The integrated results lead to a better solution, constraining the paleoelevation of the coastal mountain range to >2500 m. DISCUSSION Quantifying crustal thickness variation in eastern South China The Sr/Y and Ce/Y proxy estimates yield compatible patterns of crustal thickness variations (Fig. 3, A and B), validating their reliability. Previous studies have demonstrated that eastern South China acquired its present-day lower crustal and Moho features since the Late Cretaceous (47). The Late Cretaceous extension, as the most recent and intense event affecting eastern South China (47), led to lower-crustal flow that reoriented low-viscosity material and distributed strain evenly, giving rise to the present-day smooth and flat Moho revealed by geophysical observations (10). The Ce/Y proxy-estimated crustal thickness (30 to 40 km) at 90 to 80 Ma coincides nicely with the present-day Moho depths obtained by seismological techniques (e.g., seismic reflection, refraction, and receiver function) (10). A thick Late Jurassic crust (40 to 50 km) agrees with ca. 171-Ma adakitic rocks (48), compatible with the preferential formation of ca. 170-Ma coastal porphyry copper systems in a thick arc-type crust (49). Coeval tectonic shortening is evidenced by extensive northwest (NW)-vergent thrusts, folds, and molasse deposits in the South China hinterland (50) (Fig. 1B). Further to the NW, seismic reflection profiles reveal remnants of Jurassic shortening structures, including thin- and thick-skinned fold-and-thrust belts (Fig. 6A) (10). These observations indicate regional crustal shortening in the Mid-Late Jurassic, possibly induced by advancing subduction of the Paleo-Pacific slab (9). Reconstruction of the Late Jurassic-Cretaceous tectonic evolution of South China. This reconstruction is based on seismic images across South China and previous studies on deformation and magmatism (10, 28). (A) Stage 1: Late Jurassic–Early Cretaceous crustal extension associated with slab rollback and trench retreat along the Paleo-Pacific subduction zone. MCC, metamorphic core complex. (B) Stage 2: Late Early Cretaceous coastal mountain building and crustal thickening, induced by slab shallowing and trench advance. Earlier extensional structures were inactive in this stage. (C) Stage 3: Late Cretaceous crustal extension associated with renewed slab rollback during the retreating subduction. Decreasing crustal thickness during the 160 to 120 Ma interval was coeval with a magmatic flare-up and crustal extension (Fig. 2) (10). The time consistency implies a causal link among crustal extension, thinning, and magmatism. In addition to extensive magmatism, the extension led to crustal subsidence, large-scale rift basin formation (~85,000 km2), and metamorphic core complex development (29). Most extensional basins are fault-bound; they accumulated ~1000- to 3000-m–thick syn-extensional red beds (10). Structural analysis of basins and metamorphic core complexes indicates regional NW–southeast (SE) extension (29), reflecting backarc extension induced by slab rollback and trench retreat along the Paleo-Pacific subduction zone (9, 29) (Fig. 6A). Subsequent (120 to 105 Ma) crustal thickening inferred from Sr/Y and Ce/Y proxies indicates a tectonic transition from extension to compression at ca. 120 Ma (Fig. 3, A and B). Continuous crustal thickening culminated with a 50- to 55-km–thick crust by ca. 105 Ma (Fig. 3, A and B). Geological observations that corroborate the proxy-inferred crustal thickening include the presence of a regional Early-Late Cretaceous angular unconformity, basin inversion, sinistral and top-to-the-NW shearing, crustal anataxis, and high-temperature metamorphism in the coastal Changle-Nan’ao fault zone (CNFZ) (51). U-Pb (zircon) and Ar-Ar (amphibole and mica) ages constrain the time of tectonic shortening to ca. 120 to 108 Ma (51). The transition from extension to compression could be attributed to an increase in the trenchward overriding plate velocity (52) or slab shallowing and trench advance (Fig. 6B) possibly associated with the subduction of seamounts or plateaus (53). Crustal thickening commonly generates a high-density crustal root that sinks into the upper mantle (16). The lack of thick crust across eastern South China, as demonstrated by seismic reflection profiles, implies that any possible thickened crustal root has been removed, most likely associated with Late Cretaceous crustal thinning (10). Decreasing Sr/Y and Ce/Y ratios from 105 to 80 Ma validate this thinning event and indicate that the crust thinned to 30 to 40 km by ca. 80 Ma (Fig. 3, A and B). Widespread generations of grabens and metamorphic core complexes confirm extensive extension during crustal thinning (28) (Fig. 6C). Ar-Ar ages (105 to 80 Ma) of syn-kinematic fabrics in the metamorphic core complexes (28, 29) constrain the timing of extension. The extension culminated in the generation of a vast (>800-km-wide) basin and igneous province (Fig. 1B) analogous to the American Basin and Range Province (54). The extension likely resulted from renewed slab rollback during trench retreat (29, 55) (Fig. 6C). Emergence of the coastal mountain range in the late Early Cretaceous Our results provide a unified framework for assessing the coupled patterns of crustal thickening and surface uplift in eastern South China (Fig. 5B). A coastal mountain range with elevation >2500 m emerged during the late Early Cretaceous, as a result of isostatic compensation for a thickened (~50 to 55 km) crust (Figs. 5, A and B and 6B). These elevation and crustal thickness values are broadly analogous to the Sevier hinterland in the North America Cordillera, where Late Cretaceous–thickened (50 to 60 km) crust yielded a high-elevation (Nevadaplano) plateau (>3000 m) (56). Similar to most mountain-building cases in the Andes (57, 58), the coastal mountain range in East Asia formed primarily because of surface uplift induced by crustal thickening. Owing to this mountain uplift, East Asia maintained a high topography in its eastern part during the late Early Cretaceous (11, 59). Our paleoelevation estimate of >2500 m provides quantitative evidence addressing the enigma of Cretaceous eastern high topography in East Asia. This estimate is broadly consistent with other estimates (Fig. 5A) from carbonate-clumped isotope paleothermometry (>2000 m) (7), carbon and oxygen isotopes (>2500 m in the Aptian and >3000 m in the Berremain) (60), and atmosphere-ocean general circulation models (>2000 m) (61) as well as slightly lower than the loose constraints from sediment accumulations (3500 to 4000 m) (11). The inferred low paleoelevation estimates for the Late Cretaceous are compatible with paleoelevation evolution (~1400 m at ca. 94 Ma; Fig. 5A) inferred from thermal modeling of fission-track analyses (62). Nevertheless, our paleoelevation estimates are based solely on the assumption of local isostatic equilibrium; they may represent underestimates because the Sr/Y and Ce/Y proxies provide minimum estimates of crustal thicknesses (20, 21). The existence of the coastal mountain range is supported by lines of evidence from (i) tectonic exhumation and surface uplift as proxied by thermochronological ages (10, 62), (ii) Late Early Cretaceous crustal thickening as reconstructed from europium anomalies [Eu/Eu*zircon (2)] in detrital zircons (45), and (iii) adakitic granodiorites (ca. 107 Ma) derived from the thickened lower crust (63). Extensive Late Cretaceous evaporates, eolian sandstones, and desert deposits indicate a semi-arid to arid climate in East Asia (7), compatible with rain-shadow effects of the coastal mountain range (64). This elevated mountain underwent collapse during the Late Cretaceous to Early Cenozoic (45, 46), which was associated with rollback-induced continental extension (10). The elevation of a mountain range results from a competition between tectonics and climate (65). The former increases the elevation via crustal thickening, whereas the latter decreases the elevation via climate-dependent erosion (66). Notably, >80% of the coastal surface currently consists of Cretaceous igneous and sedimentary rocks (Fig. 1B). In contrast, surface exposures of syn-orogenic high-temperature metamorphic rocks are restricted to the CNFZ of easternmost South China (Fig. 1B), corresponding to <5% of the surface. The CNFZ appears to be an especially “hot” belt, where localized exhumation has exposed high-grade, mid-crustal metamorphic rocks (67). Given the regional high heat flow feature (10), similar high-temperature metamorphic rocks are expected to remain in most of the middle crust across the coastal region. To the west of the CNFZ, the widespread volcanic cover and the absence of mid-crustal rocks at the surface (Fig. 1B) suggest relatively shallow levels of erosion. In addition to nearby inland basins (11), extension and subsidence along the continental margin of the East and South China Seas (68) created accumulation space for the eroded deposits (45). Sedimentary provenance and detrital zircon data indicate the substantial contributions of terrigenous material from the coastal mountains to marginal sea basins [e.g., the Pearl River Mouth and Tainan basins (45, 46)]. Cretaceous-Cenozoic stratigraphic records of the basins do not show signals of detritus eroded from deep crustal metamorphic rocks. Instead, they indicate derivation from upper crustal sedimentary and volcanic rocks, consistent with shallow erosion. Given the climate control on continental erosion (69), the shallow erosion may be linked to the increased aridity at that time (70). The shallow erosion is comparable to that in the Central Andes, where the synorogenic high-grade metamorphic rocks were not exhumed to the surface but, instead, were buried below volcanic and sedimentary rocks (71). Impact of the coastal mountain topography on Asian climate The coastal mountain building occurred in the late Early Cretaceous, corresponding to the Supergreenhouse period (70), with higher global atmospheric CO2 levels and considerably warmer tropical sea surface temperatures (SSTs) (72). Evidence from stratigraphy and paleoclimate reconstruction demonstrates that a tropical to subtropical climate dominated mid- to low-latitude Asia (20°N to 40°N) during the Cretaceous (70), featuring an arid zonal belt across this latitudinal band (61). This climate pattern is closely associated with the accelerated aridification in the subtropical high-pressure belt (70) and the uplift of the coastal mountain range (64). The emergence of a mountain range may have mechanically affected regional atmospheric circulation and precipitation patterns (5). Nevertheless, the contributions of the coastal mountain range to this climate pattern remain poorly quantified, primarily because of limited climate modeling studies. Here, we perform the climatic modeling using the Community Earth System Model (Fig. 7, A and B, and see Materials and Methods) to quantify the impact of the reconstructed coastal mountain topography on the Asian climate in the context of the Cretaceous greenhouse world. Compared with previous work (61), we improved several model parameters—including the elevation of the coastal mountain range, seafloor topography, and orbital configuration—to increase the reliability of modeling results (see Materials and Methods). Impact of a 2500-m–high coastal mountain range on climate change in East Asia. (A) Cretaceous global hypsometry (61). The blue rectangle shows the location of (B). (B) East Asian topography with a >2500-m coastal mountain range from 20°N to 38°N, inspired from (61). (C) Simulated Asian land surface types without a coastal mountain range development based on climate modeling. (D) Simulated Asian land surface types with the presence of a >2500-m coastal mountain range along its eastern margin based on climate modeling. (E) Area percentage (%) of each land type and desert over the mid-to-low-latitude (20°N to 40°N) Asia. The coastal mountain uplift enhances inland aridity (~15%) and accelerates eastward desert expansion until 112°E. (F) Changes in annual mean precipitation rate and vertically integrated moisture transport between the experiments with and without a 2500-m coastal mountain range along the eastern Asian convergent margin. CM, coastal mountain. Modeling results show that the climate across mid- to low-latitude Asia is generally dry due to the paleogeography, even without a coastal mountain range (Fig. 7C). In the absence of a coastal mountain range, only 60% of the East Asian region is hyperarid and arid, most of which is covered by desert, and the extent of the desert is limited to the western and central parts (Fig. 7C), much less than the distribution of Cretaceous deserts in East Asia (70). In addition to the increased surface temperatures and evaporation (61), our results suggest that adding a 2.5-km–high coastal mountain range inhibits the advection of moist air masses from the ocean and reduces the moisture supply, resulting in less rainfall in the continental interior. Given the cumulative effects, these processes increase the fraction of hyperarid and arid areas to ~75% (Fig. 7, D and E). Consequently, the desert belt extends eastward into eastern East Asia (Fig. 7D), consistent with the distribution of Late Cretaceous intermontane deserts, eolian deposits, and evaporites in East Asia (64). Paleogeographic reconstructions suggest that the coastal mountain range extended >1000-km north-northeast (NNE)–south-southwest (SSW), forming a prominent, longitudinally parallel, linear topographic feature (7). This remarkable mountain range likely affected atmospheric circulation and precipitation patterns by creating an orographic barrier along the East Asian continental margin, blocking humid oceanic air from moving inland, and impeding moisture transport into the continental interior (61). Our simulation results show that summer monsoons dominated the seasonal patterns of winds, precipitation, and run-off in East Asia and that the coastal mountain range emergence might have changed the wind direction by inducing an anomalous cyclonic wind circulation of moisture transport (Fig. 7F). By deflecting hot air upward, the mountain range deprived the moisture off the continental interior and the ocean, leading to increased precipitation over the coastal mountain range and nearby regions (Figs. 7F and 8). In the continental interior, the loss of moisture leads to enhanced aridity and decreased precipitation (Fig. 8). Model illustrating the impact of the coastal mountain topography on Asian climate. The mountain range might have affected atmospheric circulation and precipitation patterns by creating an orographic barrier along the continental margin, blocking humid oceanic air from moving inland, and impeding moisture transport into the continental interior. These processes reduced moisture supply, changed wind directions, decreased inland precipitation, and enhanced inland aridity, contributing substantially to the arid zonal belt of mid- to low-latitude Asia. In summary, the emergence of a coastal mountain range reduced moisture supply, changed wind directions, decreased inland precipitation, enhanced inland aridity (Fig. 8), and prompted eastward desert expansion, contributing substantially to the arid zonal belt of mid- to low-latitude Asia. These variations reflect the forcing of tectonically induced surface uplift on climate change, complementing and enhancing the mid-latitude dry-belt climate of East Asia in the Cretaceous greenhouse world. In the Early Cenozoic, the high topography in the eastern part of East Asia was reversed with the collapse of the coastal mountains, the uplift of the Tibetan Plateau, and the opening of marginal seas, forming a new landscape of low topography in the eastern part and high topography in the western part of East Asia (59). In contrast to the Cretaceous arid climate, a humid climate has prevailed in South China since the Early Cenozoic (73), when the elevation of the coastal mountain range was relatively low and insufficient to block moisture transport from the ocean. The uplift of the Tibetan plateau and retreat of the Paratethys Sea, associated with the India-Asia collision (74), resulted in monsoonal intensification in Asia (74). The final collapse of the coastal mountains, in conjunction with the enhanced monsoonal climate and the South China Sea expansion (75), might have substantially increased precipitation in South China in the Cenozoic. MATERIALS AND METHODS Sr/Y and Ce/Y proxies Crustal thickness controls the Sr/Y variability of arc magmas by affecting the stabilization of mineral phases, which fractionates Sr (plagioclase) and Y (garnet) (14, 21). Sr is compatible and partitions into plagioclase at low pressures (<10 kbar); at high pressures (>12 kbar), Sr becomes incompatible and enters the liquid phase, and plagioclase is less abundant (76). By contrast, Y is incompatible at low pressures and tends to partition into garnet and amphibole at high pressures (19). Consequently, Sr/Y is a qualitative indicator of the pressures (i.e., crustal depths) at which partial melting and crystal fractionation occur (22, 23). Chapman et al. (21) suggest that the Sr/Y ratios of intermediate magmas vary with crustal thickness in the North American Cordillera, with larger Sr/Y ratios signifying greater pressures and, hence, crustal thicknesses. On the basis of large datasets of global arc lavas, Profeta et al. (19) propose Sr/Y and La/Yb empirical correlations that are feasible in monitoring the crustal thickness evolution of ancient convergent margins. In addition to intermediate magmas, crustal thickness controls the composition of arc basalts by modulating the degree of mantle melting (19). Mantle and Collins (20) quantitatively estimate the changes in the crustal thickness of the Median Tectonic Zone (Median Batholith) of New Zealand since 400 Ma using maximum Ce/Y ratios of arc basalts. The Sr/Y and Ce/Y empirical correlations have been corroborated and formulated by examining the calculated crustal thicknesses with the geophysically derived Moho depths (20, 21, 24). Polynomial regression analysis Using OriginPro 2024b, the polynomial regression analysis was performed to model the expected maximum mean value of a dependent variable y in terms of a given x value. The fitting analysis was carried out using a general polynomial regression equationy=β0+β1x+β2x2+β3x3+…+βkxk+ε(6)where β0 means the y intercept and β1, β2,…, βk represents the partial coefficients. The polynomial order, k, is a positive number and can be adjusted manually for an optimal fitting. The random error term, ε, is independent and normally distributed with N(0,σ2). The residuals are defined asresi=yi−(β0+β1x+β2x2+β3x3+…+βkxk)(7) The residuals are normally distributed with the mean equal to 0 and the variance equal to σi2. The maximum likelihood estimates for the parameters βi can be obtained by minimizing the chi-square, which is defined asx2=∑i=1N(yi−y^i)2σi2(8) The fitting automatically handles parameter optimization using an internal iterative optimization algorithm, yielding the corresponding fitted curve. Because of the lack of age constraints during several age periods (e.g., 125 to 116 Ma in Fig. 3A and 120 ot 116 Ma in Fig. 5A), we use the piece-wise polynomial fitting to provide an estimate of y at x based on neighboring data. The second-order, fourth-order, and third-order polynomials were used in Figs. 3 (A and B) and 5 (A and B), respectively. Climate modeling Here, we use a fully coupled atmosphere-ocean general circulation model, i.e., the Community Earth System Model version 1.2.2 (CESM1.2.2) developed by the National Center for Atmospheric Research. CESM1.2.2 includes dynamic atmosphere, ocean, land, sea ice, and runoff modules (Fig. 7, A and B), which interact with each other through a coupler (77). To explore the influence of the East Asian coastal mountains on the Asian climate in the mid-Cretaceous, two experiments, i.e., no coastal mountain (noCM) (Fig. 7C) and CM (Fig. 7D), are conducted using the global topography compiled by Sewall et al. (78). In experiment noCM, the East Asian coastal mountain range does not exist (Fig. 7C). In comparison, a north-south oriented mountain at ~22°N to 38°N along the East Asian margin is included in experiment CM (Fig. 7D). Accordingly, the differences in the experiment directly reveal the effects of the coastal mountains on the Asian climate. Given the zonally limited East Asian coastal mountains, we use a relatively high resolution of 0.9° × 1.25° for the atmosphere (CAM4.0) and land (CLM4.0) modules in the experiments. Notably, fully coupled running with a dynamic ocean at such a high resolution is not feasible due to limited computing resources. The two experiments are driven by fixed SSTs and sea ice with annual cycling, which are the monthly climatology of the third experiment (CPL). This experiment is a fully coupled run of the CESM1.2.2, in which the CAM4.0 and CLM4.0 are modeled at a resolution of 3.75° × 3.75°. The ocean (POP2) and sea ice (CICE4) modules use a gx3v7 grid having 116 and 100 grid points in the meridional and zonal directions, respectively. The CAM4.0 and POP2 have 26 and 60 vertical levels, respectively. The runoff module runs at a resolution of 0.5° × 0.5°. The abovementioned methods are commonly used to model the topographical impacts on the regional climate (61, 79). In this study, the experiment CPL runs for 3000 model years, which is long enough for Earth’s surface to reach a quasi-equilibrium state. The monthly climatology of the last 100-year SSTs and sea ice are used to force the experiments noCM and CM. These two uncoupled experiments run for 40 model years, and the data of the past 30-model years are analyzed here. The other settings in the three experiments are concordant. The used solar constant is 1350.87 W/m2, 0.75% lower than the present value of ~1361 W/m2 (80). We set the chlorofluorocarbon concentrations to 0 because they come from human release. The pCO2 concentrations are set to 1120 parts per million by volume (ppmv), consistent with the palaeo-CO2 reconstruction (81). The pCH4 and pN2O concentrations are set to 760 and 270 ppmv, consistent with the preindustrial levels. The influence of orbital forcing is not considered. We use the orbital parameters of 1990 CE in the modeling experiments. The default values of the CESM1.2.2 were used for all other atmospheric components (e.g., O3 and aerosols), all at preindustrial levels. Our modeling results are consistent with the first-order climatic conditions in the Cretaceous greenhouse world, including the average surface temperatures [20° to 25°C; (82)], aridity, and precipitation patterns (70). In comparison with previous work (61), we improved several model parameters (i.e., the elevation of the coastal mountain range, seafloor topography, and orbital configuration) to increase the reliability of modeling results. 1) Elevation of the coastal mountain range. The elevation of the coastal mountain range is one of the most sensitive parameters used to determine the modeling results. Our paleoelevation estimate of >2500 m provides the latest quantitative evidence not modeled in previous work. This estimate is compatible with independent geological, isotopic, and climatic observations (7, 45, 64). It sheds light on addressing the enigma of the Cretaceous topography-climate interaction in East Asia. 2) Seafloor topography. Seafloor topography influences ocean circulation by steering ocean flows and contributing to ocean mixing (83). The ocean circulation field transports heat and energy, leading to changes in sea level and surface temperature; it forms a critical element for modulating Earth’s climate system (84). The seafloor topography, which was less considered in previous work (61), has been added to this modeling (Fig. 7A). 3) Orbital configuration. Earth’s orbital cycling substantially influenced the East Asian Monsoon climate system during the Cretaceous (85). The orbital parameters—including Earth’s obliquity, precession, and eccentricity—varied within respective ranges during the Cretaceous. The eccentricity was set to zero in the previous work (61). The influence of precession on the Asian climate is prominent when the eccentricity is nonzero. In this study, we use the present-day orbital configuration, similar to the configuration of Earth’s orbit at around 115,000 years before present, which is used in many climate modeling studies (86).