•Distinct diurnal hydrochemical variations were found in all the ponds at the Shawan Karst Test Site.•These variations were determined by coupled action of land uses and metabolism of submerged plants in the ponds.•More organic carbon was produced in pond waters with higher concentrations of DIC, indicating DIC fertilization effect.•The biological carbon pumping (BCP) effect, similar to that in the oceans, was found in each pond.•The BCP fluxes ranged from 156 to 493 t C km− 2 a− 1, depending on the land-use-controlled DIC concentrations in the ponds.High-resolution hydrochemical data from five spring-fed ponds are presented to demonstrate the effect of different land uses and aquatic biological processes on the carbon cycle at a karst-analog test site. The results show that hydrochemical parameters including pH and the concentrations of HCO3−, Ca2 +, NO3−, partial pressures of CO2 (pCO2) and dissolved O2 (DO) as well as carbon isotopic compositions (δ13C) of HCO3– in the pond water all displayed distinct diurnal variations, while those of the spring water itself were rather stable. The coupled dynamic behaviors of pCO2, DO and NO3− indicate a significant influence from the metabolism of submerged plants in the ponds. In the afternoon, when photosynthesis is the strongest, the pCO2 of the five pond waters was lower even than that of the ambient atmosphere, demonstrating the existence of a “biological carbon pumping (BCP) effect”, similar to that in the oceans. It was determined that, in October (autumn), the BCP fluxes in the five spring-fed ponds were 156 ± 51 t C km− 2 a− 1 in P1 (Pond 1 – adjoining a bare rock shore), 239 ± 83 t C km− 2 a− 1 in P2 (adjoining uncultivated soil), 414 ± 139 t C km− 2 a− 1 in P3 (adjoining land cultivated with corn), 493 ± 165 t C km− 2 a− 1 in P4 (adjoining grassland) and 399 ± 124 t C km− 2 a− 1 of P5 (adjoining brushland), indicating the potentially significant role of aquatic photosynthesis in stabilizing the carbonate weathering-related carbon sink. In addition, by comparing the DIC concentrations and fluxes of DIC transformed into autochthonous organic matter (AOC) in the five ponds, the so-called “DIC fertilization effect” was found in which more AOC is produced in pond waters with higher concentrations of DIC. This implies that the carbon cycle driven by aquatic biological processes can be regulated by changing land use and cover, the latter determining the DIC concentrations. Further, the rock weathering-related carbon sink is underestimated if one only considers the DIC component in surface waters instead of both DIC and AOC.

•Shawan Simulation Test Site was established to simulate varied land uses in karst.•Land uses determined soil CO2 and water supplies for carbonate dissolution.•A new parameter, LCIC (Land use Change Impact on CO2 flux), was proposed.•LCIC calculates the impact of land use change on CO2 flux.•Carbonate weathering-related carbon sink may be regulated by changing land uses.In the study of global climate change, a major focus of research into the carbon cycle is to determine the fate of missing carbon sinks. Carbonate weathering-related carbon sinks as a result of water‑carbonate-CO2-aquatic phototroph interactions may make a major contribution. Establishing optimal land uses, which determine soil CO2 concentrations and water supplies for carbonate dissolution, may be a feasible and effective way to increase the sink potential. Elucidating both the hydrological and the hydrochemical behavior under different land uses is critical for rational planning of land use changes to increase the carbon sink. Given the complexity within natural karst catchments, the Shawan Simulation Test Site was established at Puding, Southwest China, to simulate the influence of land uses with controlled carbonate test beds - bare rock, bare soil, crop land, grass land, shrub land. Soil CO2 concentrations, hydrochemical parameters (pH, major ion concentrations) and the ‘spring’ (artificial drain) discharge were intensively measured from Sept. 1, 2015 to Aug. 31, 2016 to investigate the carbon and water responses to different land uses in different seasons. In the vegetated land uses (crop, grass or shrub), DIC increased due to the increase of soil CO2 resulting from stronger microbial activities and root respiration in summer and autumn growing seasons. In the bare rock and soil cases, there was also an increase in DIC in summer and autumn, due to decomposition of prior organic matter within soils and/or rock pores. The average DIC concentration ranking, high to low, was grass land, shrub land, crop land, bare soil, bare rock. Soil CO2 concentration was thus the dominant of DIC concentration, which is a key multiplier of carbon sink fluxes (CSF = 0.5 ∗ [DIC] ∗ RD, where RD is depth of runoff, [DIC] is DIC concentration, and 0.5 because in carbonate dissolution, only half of the [HCO3−] is of atmospheric carbon origin). However, runoff depth ranked almost in reverse order, i.e., from high to low, bare rock, bare soil, crop land, shrub land, grass land. The CSF ranking, from high to low, was grass, crop, shrub, bare rock, bare soil. A new parameter, LCIC (Land use Change Impact on CSF) is defined to compare the impacts of land use change on [DIC] and RD, and evaluate their combined effects on CSF. Compared to bare rock, the absolute values of LCIC (| LCIC | s) were > 1, and CSFs were larger for the three tanks with vegetation cover; CSF is smaller for bare soil, | LCIC | < 1. Finally, it was found that grass land may constitute an optimal land-use for increasing the carbonate weathering-related carbon sink that is critical for carbon management to counter global warming.

One of the most important questions in the science of global change is how to balance the atmospheric CO2 budget. There is a large terrestrial missing carbon sink amounting to about one billion tonnes of carbon per annum. The locations, magnitudes, variations, and mechanisms responsible for this terrestrial missing carbon sink are uncertain and the focus of much continuing debate. Although the positive feedback between global change and silicate chemical weathering is used in geochemical models of atmospheric CO2, this feedback is believed to operate over a long timescale and is therefore generally left out of the current discussion of human impact upon the carbon budget. Here, we show, by synthesizing recent findings in rock weathering research and studies into biological carbon pump effects in surface aquatic ecosystems, that the carbon sink produced by carbonate weathering based on the H2O–carbonate–CO2–aquatic phototroph interaction on land not only totals half a billion tonnes per annum, but also displays a significant increasing trend under the influence of global warming and land use change; thus, it needs to be included in the global carbon budget.全球变化科学最重要的问题之一是如何平衡大气CO2收支。至今仍存在每年十亿吨级的陆地碳汇不知去向,即所谓的“遗失碳汇”问题。这些不明碳汇的位置、量级、变化和形成机制目前仍不确定,并存在巨大争议。尽管全球变化和岩石化学风化的正相关关系已体现在大气CO2的地球化学模型中,但这一反馈被认为只在地质长时间尺度起作用,因此,对于人类活动对碳收支影响的讨论时常不考虑风化碳汇的贡献。本文基于岩石风化研究的最新进展,并综合水生生态系统碳泵效应研究的成果, 发现陆地水-碳酸盐–CO2-水生光合生物相互作用产生的碳汇不仅很重要(每年近5亿吨碳),而且在气候变暖和土地利用变化的影响下呈现显著的增加趋势,因此,必须包括在现今全球碳收支的评价中。

Nearly 18 years after the proposal of the weathering-related carbon sink concept (Berner R A. Weathering, plants and the long-term carbon cycle. Geochim Cosmochim Acta, 1992, 56: 3225–3231), it is an appropriate timing to re-evaluate its geological context with the updated dataset. Ryskov et al. (Ryskov Ya G, Demkin V A, Oleynik S A, et al. Dynamics of pedogenic carbonate for the last 5000 years and its role as a buffer reservoir for atmospheric carbon dioxide in soils of Russia. Glob Planet Change, 2008, 61: 63–69) lately claimed that in the course of soil formation for the last 5000 years the soils of Russia fixed atmospheric carbon dioxide as pedogenic carbonate during the arid periods at a rate of 2.2 kg C/(m2 a) in chernozem, 1.13 kg C/(m2 a) in dark-chestnut soil, 0.86 kg C/(m2 a) in light-chestnut soil, on the basis of carbon isotopic data; however, their interpretations of the data do not appear straightforward nor persuading, and thus their claim is likely misleading. Their interpretations are also contrary to the conclusions drawn by Dart et al. (Dart R C, Barovich K M, Chittleborough D J, et al. Calcium in regolith carbonates of central and southern Australia: Its source and implications for the global carbon cycle. Palaeogeogr Palaeoclimatol Palaeoecol, 2007, 249: 322–334) who found that Australian regolith carbonates did not capture any additional CO2; instead the carbonate was simply being remobilized from one pool to another. Here we raise comments to these explanations on the following two issues: (1) origin of pedogenic carbonate: silicate weathering vs. carbonate weathering, and (2) problems in using carbon isotopic technique to distinguish carbonates formed by silicate weathering and carbonate weathering. It is concluded that pedogenic carbonate may not be an important atmospheric CO2 sink at all, i.e. carbonate weathering-related pedogenic carbonate does not capture any additional CO2, while the CO2 capture in silicate weathering-related pedogenic carbonate is small in short-term time scales due to the slow kinetics of silicate weathering.

The locations, magnitudes, variations and mechanisms responsible for the global CO2 sink are uncertain and under debate. Here, we show, based on theoretical calculations and evidences from field monitoring results, that there is a possible important CO2 sink (as DIC-dissolved inorganic carbon) by the global water cycle. The sink constitutes up to 0.8013 Pg C/a (or 10.1% of the total anthropogenic CO2 emission, or 28.6% of the missing CO2 sink), and is formed by the CO2 absorption of water and subsequent enhanced consumption by carbonate dissolution and aquatic plant photosynthesis. Of the sink, 0.5188 Pg C/a goes to sea via precipitation over sea (0.2748 Pg C/a) and continental rivers (0.244 Pg C/a), 0.158 Pg C/a is released to the atmosphere again, and 0.1245 Pg C/a is stored in the continental aquatic ecosystem. Therefore, the net sink could be 0.6433 Pg C/a. This sink may increase with the global-warming-intensified global water cycle, the increase in CO2 and carbonate dust in atmosphere, and reforestation/afforestation, the latter increasing soil CO2, and thus the concentration of the DIC in water.

•We determined the karst-related carbon sink flux (CSF) using MPD method.•The mean CSF in SW China was approximately 9.36 t C km− 2 a− 1•CSF decreased by about 19% from 1970s to 2010s in SW China.•Runoff decrease is the major reason for the CSF decrease.Riverine carbon fluxes of some catchments in the world have significantly changed due to contemporary climate change and human activities. As a large region with an extensive karstic area of nearly 7.5 × 105 km2, Southwest (SW) China has experienced dramatic climate changes during recent decades. Although some studies have investigated the karst-related carbon sink in some parts of this region, the importance of climate impacts have not been assessed. This research examined the impacts of recent climate change on the karst-related carbon sink in the SW China for the period 1970–2013, using a modified maximal potential dissolution (MPD) method and GIS. We first analyzed the major determinants of carbonate dissolution at a spatial scale, calculated the total karst-related carbon sink (TCS) and carbon sink fluxes (CSFs) in the SW China karst region with different types of carbonate rocks, and then compared with other methods, and analyzed the causes of CSFs variations under the changed climate conditions. The results show that the TCS in SW China experienced a dramatic change with regional climate, and there was a trend with TCS decreasing by about 19% from 1970s to 2010s. This decrease occurred mostly in Guizhou and Yunnan provinces, which experienced larger decreases in runoff depth in the past 40 years (190 mm and 90 mm, respectively) due to increased air temperature (0.33 °C and 1.04 °C, respectively) and decreased precipitation (156 mm and 106 mm, respectively). The mean value of CSFs in SW China, calculated by the modified MPD method, was approximately 9.36 t C km− 2 a− 1. In addition, there were large differences in CSFs among the provinces, attributed to differences in regional climate and to carbonate lithologies. These spatiotemporal changes depended mainly on hydrological variations (i.e., discharge or runoff depth). This work, thus, suggests that the karst-related carbon sink could respond to future climate change quickly, and needs to be considered in the modern global carbon cycle model.

•Biomarkers and geochemical proxies were analyzed to determine OC sources.•The calculated autochthonous OC was over 65% of the total OC in the Pearl River.•Phytoplankton biomass and DIC concentration were positively related.•The positive relationship indicates DIC fertilization effect on the photosynthesis.•DIC fertilization effect may contribute to the missing carbon sink.The photosynthetic conversion of dissolved inorganic carbon (DIC) into organic carbon (OC) by using aquatic phototrophs in rivers may serve as a potential carbon sink, especially in the carbonate rock areas, thereby offering a clue for finding the missing carbon sink. However, primary-produced autochthonous OC is erroneously considered as terrestrial-derived allochthonous OC. Thus, carbonate weathering-related carbon sink is underestimated if only DIC concentrations sampled at river mouths are considered, and the transformation of DIC to autochthonous OC is neglected. Therefore, distinguishing sources of autochthonous and allochthonous OC is vital in the assessment of carbon sink. In this study, source-specific biomarkers, in association with chemical compositions and phytoplankton proxies in water samples collected from the Pearl River, were analyzed to determine OC sources. Results showed that biomarkers in the Pearl River were quite abundant, and the calculated average autochthonous OC was approximately 65% of the total OC, indicating intense in-river primary productivity. Moreover, phytoplankton biomass and DIC concentration were positively related, indicating the DIC fertilization effect on aquatic photosynthesis. High total suspended solid (TSS) on the water surface blocked the sunlight and then reduced phytoplankton production. However, in situ photosynthesis of phytoplankton could also produce autochthonous OC, even larger than the allochthonous source at sites with high DIC, and even with higher TSS concentrations. These findings comprehensively elucidated the formation of autochthonous OC based on the coupling action of rock weathering and photosynthetic activity in the riverine system, suggesting a potential direction for finding the missing carbon sink.

The magnitudes, variations, locations and mechanisms responsible for the global atmospheric CO2 sink are uncertain and under continuing debate. Previous studies have focused mainly on the sinks in the oceans, and soil and vegetation on the continents. Here, we show, based on theoretical calculations and field monitoring evidence, that there is an important but previously underestimated sink for atmospheric CO2 as DIC-dissolved inorganic carbon that results from the combined action of carbonate dissolution, the global water cycle and the photosynthetic uptake of DIC by aquatic organisms in ocean and land. The sink constitutes up to 0.8242 Pg C/a, amounting to 29.4% of the terrestrial CO2 sink, or 10.4% of the total anthropogenic CO2 emission. 0.244 Pg C/a are transferred to the sea via continental rivers and 0.2278 Pg C/a by meteoric precipitation over the seas. 0.119 Pg C/a is released back to the atmosphere again, and 0.2334 Pg C/a is stored in the continental aquatic ecosystem. Therefore, the net sink is estimated as 0.7052 Pg C/a. This sink may increase with an intensification of the global water cycle as a consequence of global warming, rising anthropogenic emissions of CO2 and carbonate dust in atmosphere, and afforestation, which increases the soil pCO2 and thus the carbonate dissolution. Fertilization with the elements N, P, C, Fe, Zn, and Si increases the organic matter storage/burial by aquatic organisms and thus decreases the CO2 return to the atmosphere. Based on the ensemble mean projection of global warming for the year 2100 by IPCC, it is estimated that the atmospheric CO2 sink will increase by 21%, or about 0.18 Pg C/a. However, the uncertainty in the estimation of this sink needs further exploration.

•Negative Δ14C value of river POC does not mean the POC is necessarily old.•It may indicate carbonate weathering coupled with present aquatic photosynthesis.•This has implications for the interpretation of organic carbon age in surface waters.Generally, negative Δ14C values of riverine particulate organic carbon (POC) are interpreted as old carbon derived from the erosion of deep soils and sedimentary rocks. Here we present natural 14C and 13C data from the carbonate-rich Pearl River Basin that discharges into the South China Sea. We found that the Δ14C values of POC and DIC (dissolved inorganic carbon) transported by the carbonate-rich river are all negative. This, however, does not mean the POC is necessarily old but indicates control of carbonate weathering (producing “old” DIC with negative Δ14C values) coupled with contemporary aquatic photosynthesis (producing new autochthonous POC but with negative Δ14C values) through the “dead carbon” effect of carbonate rocks, which was further evidenced by particular seasonal change in Δ14C values of DIC and POC (both higher in the rainy season and lower in dry season), spatial variation (both getting higher downstream), and negative correlation between δ13C and “age” of POC. This finding indicates that previous studies suggesting that riverine POC depleted in 14C is old may be problematic in carbonate-dominated river basins. The finding that river basins rich in carbonates can release “old” POC may have important implications for the interpretation of organic carbon age in rivers and coastal oceans affected by the runoff from this basin type. It also indicates that it is necessary to examine the concentrations of both DIC and autochthonous organic carbon in rivers to correctly assess the carbon sink produced by rock weathering.

It is widely accepted that chemical weathering of Ca–silicate rocks could potentially control long-term climate change by providing feedback interaction with atmospheric CO2 drawdown by means of precipitation of carbonate, and that in contrast weathering of carbonate rocks has not an equivalent impact because all of the CO2 consumed in the weathering process is returned to the atmosphere by the comparatively rapid precipitation of carbonates in the oceans. Here, it is shown that the rapid kinetics of carbonate dissolution and the importance of small amounts of carbonate minerals in controlling the dissolved inorganic C (DIC) of silicate watersheds, coupled with aquatic photosynthetic uptake of the weathering-related DIC and burial of some of the resulting organic C, suggest that the atmospheric CO2 sink from carbonate weathering may previously have been underestimated by a factor of about 3, amounting to 0.477 Pg C/a. This indicates that the contribution of silicate weathering to the atmospheric CO2 sink may be only 6%, while the other 94% is by carbonate weathering. Therefore, the atmospheric CO2 sink by carbonate weathering might be significant in controlling both the short-term and long-term climate changes. This questions the traditional point of view that only chemical weathering of Ca–silicate rocks potentially controls long-term climate change.

Huanglong, well known for its unique natural travertine landscape, was listed by UNESCO as an entry in the World’s Nature Heritage in 1992, and attracts more than one million of tourists from all over the world each year. However, the landscape has undergone significant degradation (notably, serious decay of travertine) during the past two decades as the tourist numbers have increased remarkably. To understand the variations of travertine deposition rates and their controlling factors, especially the impact of tourism activities, paired water and modern travertine samples deposited on plexiglass substrates were taken along the Huanglong stream at regular intervals from early May to early November in 2010 (i.e., in the wet season). The travertine deposition rates have declined significantly compared to those in early 90s in all four subsystems in the Huanglong Ravine. The largest decrease (89.5%) occurred at the lowest sampling site. The reduction in travertine deposition most likely resulted from the phosphate pollution caused by the tourism activities. In spite of an increase in concentrations of Ca, calcite saturation, and water temperature, which facilitate calcite precipitation, deposition rates decreased because of inhibition by PO43- ions. Seasonally, three control patterns of travertine deposition rates were distinguished along the Ravine. They are control by water-temperature, control by dilution of rainwater and snow-melting water and control by PO4-inhibition of calcite precipitation.Highlights► Travertine deposition at Huanglong decreased greatly compared to early 1990s. ► The decrease in deposition may result from the phosphate inhibition of calcite. ► It was the tourism activities that introduced more phosphate into water. ► Water-temperature and dilution showed also controls on travertine deposition.

•Significant diel CO2 exchange cycles in karst spring-fed ponds are found.•The diel cycles are driven by biological carbon pump (BCP) in the ponds.•The BCP caused carbon sequestration and decreased CO2 emission.Whether carbonate weathering could produce a stable carbon sink depends primarily on the utilization of dissolved inorganic carbon (DIC) by aquatic phototrophs (the so-called Biological Carbon Pump-BCP effect). On this basis, water temperature (T), pH, electrical conductivity (EC) and dissolved oxygen (DO) were synchronously monitored at 15-min resolution for one and two days respectively in January and October 2013 in Maolan Spring and the spring-fed midstream and downstream ponds in Maolan Nature Reserve, China. A thermodynamic model was used to link the continuous data to allow calculation of CO2 partial pressures (pCO2) and calcite saturation indexes (SIC). A floating static chamber was placed on the water surface successively at all sites to quantify CO2 exchange flux between atmosphere and water so as to evaluate the BCP effect. Results show that, in both winter and autumn, remarkable diel variations of hydrochemical parameters were present in the midstream pond where DO, pH, and SIC increased in the day and decreased during the night while EC, [HCO3−], [Ca2+] and pCO2 showed inverse changes mainly due to the metabolic processes of the flourishing submerged plants, with photosynthesis dominating in the day and respiration dominating at night. However, hydrochemical parameters in the spring and downstream pond show less change since few submerged plants developed there. It was determined that the BCP effect in the midstream pond was 285 ± 193 t C km−2 a−1 in winter and 892 ± 300 t C km−2 a−1 in autumn, indicating a potential significant role of terrestrial aquatic photosynthesis in stabilizing the carbonate weathering-related carbon sink.

Karst processes-related carbon cycle, as a result of the water–carbonate rock–CO2 gas–aquatic organism interaction, significantly affects global carbon budget. In karst areas, soil CO2 is a major chemical driving force for the karst processes and has significant impact on the geochemical processes of the water–rock–gas–organism system. Currently, there have been many studies mainly focusing on the hydrochemical responses of the epikarst water system to weather conditions. However, few studies examine the direct correlation between the hydrochemical parameters in epikarst systems and soil CO2. We chose an epikarst spring system at Chenqi, Puding, SW China to monitor both the concentration of soil CO2 and hydrochemical parameters at high-resolution (every 15 min) during July 2010–December 2011 covering a complete hydrologic year, and to investigate the response of hydrochemical changes to soil CO2 and weather conditions. It was found that both soil CO2 and rainfall are the major driving forces for the epikarst hydrochemical variations. The soil CO2 effect on hydrochemical variations was reflected in all seasonal, diurnal and storm-scales. There was an increase in pCO2 and electrical conductivity (EC) but a decrease in pH caused by the increase in soil CO2 in spring-summer growing season, while a decrease in pCO2 and EC but an increase in pH caused by the decrease in soil CO2 happened in autumn–winter dormant season. Similar variations were also found on diurnal scales but with a time lag of a few hours between hydrochemical changes and soil CO2 change during dry season, showing effect of the groundwater recharge mode as well as the complexity of the supply path (quick flow by conduit or slow flow by fracture). During rainy seasons, however, hydrochemical changes in epikarst groundwater were regulated by both dilution and soil CO2 effects. Under high-intensity rainfall, the dilution effect was dominant, indicated by a quick decrease in EC, pH and calcite saturation (SIc) but a quick increase in pCO2. In contrast, under low-intensity rainfall, the soil CO2 effect was dominant, indicated by an increase in EC and pCO2 but a decrease in pH and SIc. To sum up, this study has shown the high sensitivity and variability of epikarst processes to the environmental change, implying that the role of karst processes in the global carbon cycle needs to be reappraised based on high-resolution monitoring strategy.Highlights► Soil CO2 is the major driving force for the karst hydrochemical variations. ► The soil CO2 effect is reflected in all seasonal, diurnal and storm-scales. ► The role of karst processes in the global carbon cycle needs to be reappraised.

Rainfall, spring stage, water temperature, pH and conductivity in the paired karst spring catchments of Chenqi and Dengzhanhe, which shared the same climatic condition but different land use/land cover (LULC) at Puding, Guizhou Province, SW China, were monitored by two high-resolution multi-parameter auto-recordable instrument of CTDP300 during the hydrological year of September 2007–September 2008. Other monthly hydrogeochemical and carbon isotopic (δ13C) variations in the paired karst catchments during the same hydrological year were also investigated. A thermodynamic model was used to link the continuous data to monthly hydrogeochemical data allowing the calculation of CO2 partial pressure (pCO2) and calcite saturation index (SIc) on a continuous basis. The primary study objective was to understand how the karst processes and karst hydrogeochemistry respond to different LULC, which is essential to assessing the karst-related carbon cycle. Marked seasonal and storm-scale variations were found for pH, conductivity, pCO2, SIc and δ13C of the two springs, indicating that both springs were dynamic and variable systems. However, there were differences in the magnitude and direction of the variations of these features between the two springs. The higher pCO2 and HCO3- concentration and lower pH, SIc and δ13C in Chenqi spring than those in Dengzhanhe spring tend to be related to the difference in LULC between Chenqi and Dengzhanhe spring catchments: in the Chenqi spring catchment, there was larger soil cover and the paddy land was located in the discharge area, both of which produced and kept more CO2 (a major driving agent for the karst processes) and lower δ13C in the soil-aquifer system, while in the Dengzhanhe spring catchment area, there was larger bare carbonate rock occurrence and the paddy land was located mainly in the recharge area. Moreover, the pH increased and pCO2 decreased generally in Chenqi spring after rainfall, possibly due to more carbonate dissolution in the larger soil cover rich in limestone fragments in the spring catchment, while the pH decreased and pCO2 increased generally in Dengzhanhe spring after rainfall. All these differences show that soil cover and land use pattern played important roles in the karst processes. In other words, the karst hydrogeochemistry and the karst-related carbon cycle could be regulated effectively by different LULC. In addition, the higher concentrations of Ca2+, SO42-, Mg2+ and conductivity of Dengzhanhe spring were due to the dissolution of more gypsum and dolomite minerals in the strata of Dengzhanhe spring catchment. Therefore, the karst hydrogeochemical parameters, including pH, conductivity, HCO3-, Ca2+, Mg2+, SO42-, pCO2, SIc, and δ13CDIC, could serve as good indicators of different LULC and the other environmental changes.

Water samples and modern endogenic (thermogene) travertine calcite deposited on plexiglass substrates in travertine pools and a ramp stream were collected along the Huanglong Ravine, Sichuan, SW China at regular ∼10 day intervals from early May to early November in 2010, including both wet and dry conditions. Temporal and spatial variations in the δ13C and δ18O values of the modern travertine were examined to understand their potential for paleoclimatic and paleoenvironmental interpretations. It was found that δ13C and δ18O of travertine formed in the ramp stream were low in the warm rainy season and high in the cold dry season. Their positive correlation was mainly due to dilution and rainfall seasonal effects on δ13C and δ18O values, respectively, i.e., low δ13C values were caused by dilution by overland flow with depleted δ13C values and reduced CO2-degassing in the warm rainy season while low δ18O values of travertine were because of low δ18O values of water induced by seasonal variation in oxygen isotopic ratios of rainwater. Meanwhile, kinetic effect on oxygen isotopic fractionation during ramp travertine deposition existed and reduced this positive correlation. In contrast, the δ13C and δ18O values of the pool travertines displayed a converse behavior which was caused mainly by the temperature effect. Low δ18O values and high δ13C values in the warm rainy season were correlated chiefly with the higher water temperatures. Therefore, the δ13C and δ18O values of the travertine may be used for paleo-rainfall or paleotemperature reconstruction respectively. This study demonstrates that endogenic travertine, like epigenic (meteogene) tufa, may be a suitable candidate for high-resolution paleoclimatic and paleoenvironmental reconstructions. However, since travertines deposited under differing hydrodynamic conditions (e.g., pools with still water contrasted to fast flow streams) have different climatic responses, it is necessary to check the depositional facies of fossil travertine samples before they can be used for palaeoclimate (temperature and/or rainfall) reconstruction.

Seasonal and spatial variations in the δ13C and δ18O values of the modern endogenic (thermogene) travertine deposited in a calcite-depositing canal at Baishuitai, Yunnan, SW China were examined to understand their potential for paleoclimatic and paleoenvironmental implications. The sampling sites were set in the upstream, middle reach and downstream of the canal, and the modern endogenic travertine samples were collected semimonthly to measure their δ13C and δ18O values. It was found that both δ13C and δ18O values of the endogenic travertine were low in the warm rainy season and high in the cold dry season, and correlated with each other. The low δ18O values in warm rainy season were mainly related to the higher water temperature and the lower δ18O values of rainwater, and the low δ13C values are caused by the dilution effect of overland flow with low δ13C values in the warm rainy season and the reduced CO2-degassing of canal-water caused by the dilution effect of the overland flow. The linear negative correlation between the travertine δ18O (or δ13C) values and rainfall amount may be used for paleo-rainfall reconstruction if one knows the δ18O (or δ13C) values of the fossil endogenic travertine at Baishuitai though the reconstruction was not straightforward. It was also found that there was a progressive downstream increase of the δ18O and δ13C values of the travertine along the canal, the former being mainly due to the preferential evaporation of H216O to the atmosphere and the latter to the preferential release of 12CO2 to the atmosphere during CO2-degassing. However, the downstream increase of the travertine δ18O and δ13C values was less intensive in rainy season because of the reduced evaporation and CO2-degassing during the rainy season. To conclude, the downstream travertine sites could be more favorable for the paleo-rainfall reconstruction while the upstream travertine sites are more favorable for the paleo-temperature reconstruction. So, this study demonstrates that endogenic travertine, like epigenic (meteogene) tufa, could also be a good candidate for high-resolution paleoclimatic and paleoenvironmental reconstruction.

Travertines are potential archives of continental paleoclimate. Records of stable carbon and oxygen isotopic composition (δ13C and δ18O) in laminated travertine deposits from endogene spring waters show regular cyclic patterns which may be due to seasonal change in climate determinants such as temperature and rainfall. In this study, δ13C and δ18O measurements of three travertine specimens that grew naturally over the eight years, 2004–2011, at upstream, middle and downstream sites in a canal at Baishuitai, SW China, are presented. They exhibit clear seasonal variations that generally correlate with biannual laminations. Specifically, δ13C and δ18O values show significant positive correlation with each other for the three travertine specimens, with the correlation coefficients increasing downstream along the canal. To reveal the factors governing the seasonal and spatial variations in δ13C and δ18O values, newly formed travertines precipitated on Plexiglas substrates are also examined. Both δ13C and δ18O of the substrate travertines are low in the summer/rainy season and high in the winter/dry season, showing a great consistency with the patterns in the natural travertines. Spatially, isotope values increase downstream in both seasons, with higher increase rates in winter that are related to removal of larger fractions of dissolved inorganic carbon (DIC) from the solution and stronger kinetic isotopic fractionation in winter. Due to in-stream physicochemical processes, including CaCO3 precipitation and the associated degassing of CO2, seasonal changes in δ13C and δ18O in the travertines are amplified by two times between the upstream and downstream sites: this is opposite to trends for epigene (meteogene) tufas whose seasonal changes in stable isotope compositions are reduced downstream. We suggest in-stream physicochemical processes are a potential reason for underestimation of annual temperature ranges that are inferred from epigene tufa δ18O data.