中国岩溶

Formation hydro-mechanisms and emergency response of clustered karst waterlogging in southwestern Guangxi

JIANG Guanghui, GUO Fang, QIN Jianming, HUANG Yongyun, CHEN Haoyun, ZHANG Qinjun, KANG Zhiqiang, ZHANG Yong, LIU Xiaoming

2026, 45(3): 443-452. doi: 10.11932/karst20260301

Abstract FullText HTML PDF(3507KB) icon

14

5

Abstract:
Clustered karst waterlogging triggered by extreme rainfall exhibits noval characteristics of chain-type disaster propagation, underscoring the urgent need for in-depth investigations into its hydrological mechanisms and emergency response systems. This study focuses on the karst waterlogging events in Baise City, southwestern Guangxi, during the period from September to October of 2025, employing hydrogeological surveys and monitoring, remote sensing interpretation, and hydrochemical tracing to reveal the formation and disaster transfer mechanisms of karst waterlogging under extreme rainfall conditions induced by typhoons. The findings indicate that three typhoon processes accumulated 689 mm of rainfall, with a daily maximum of 169 mm. Intense rainfall induced supersaturation of water volumes within the conduit-fracture spaces of karst aquifers, resulting in a decline rate of waterlogging accumulation of merely 2 to 3 cm·d⁻¹. The flood recession half-life of subterranean rivers extended to 9.8 days, with the recession coefficient decreasing to 0.07 d⁻¹, indicating slower recession than the maximum flood events of 2023 and 2024. These observations demonstrate that under extreme hydrological conditions, the karst conduit system operates under a "full water and slow flow" state, increasing disaster risks. The waterlogging water bodies remained stagnant for up to 35 days, leading to elevated chlorophyll concentrations and CDOM (Chromophoric Dissolved Organic Matter) levels averaging higher than those in groundwater, rivers, and reservoirs, indicating a distinct eutrophication trend. This study proposes a climate-hydrology-water environment transmission mechanism for karst waterlogging disasters,identifying that the blockage of sinkholes,mining actvities,and engineering disturbances induce hydrological cascade responses in subterranean rivers and degrade drainage capacity-key factor exacerbating waterlogging. During emergency response to waterlogging, pumping and drainage played a pivotal role, with a single depression achieving staged intensive pumping of 2.6 million m³ and a total drainage volume of 5.0 million m³—marking an unprecedented scale of dewatering operations.The study proposes a waterlogging mitigation framework centered on subterranean river basin management, emphasizing an integrated solution that combines top-down dredging of the sinkholes-karst conduits-discharge outlets continuum with basin-wide collaborative regulation.

Migration simulation of heavy metal pollutants in goaf of monoclinic coal seam under rainfall leaching and seepage

YUAN Guoyu, QIN Ronggao, XIE Yiming, WU Yingying, LI Yong, WANG Gang, CHENG Fangling

2026, 45(3): 453-465, 495. doi: 10.11932/karst20260302

Abstract FullText HTML PDF(17298KB) icon

12

3

Abstract:
This study takes the typical coal mining area of Dongshan town, Qujing City, Yunnan Province as the study area. As a representative region of karst landforms in Southwest China, the study area features dense distribution of coal seams and non-ferrous metal minerals. The karst pipelines-fracture networks are intertwined, resulting in an extremely complex karst groundwater systems. The strata in the study area exhibit an overall monoclinic structure, with rock formations oriented at 27°∠9° and strata sloping toward the northeast, forming a typical monoclinic hydrogeological structure. Continuous seepage of Acidic Mine Wastewater (AMD), caused by rainfall infiltration and leaching, has servely polluted both surface water and groundwater in the region. Consequently, this area among the most heavily polluted groundwater zones in the karst region in Southwest China, and tracing the source of groundwater pollution remains challenging.The purpose of this study was to conduct an in-depth analysis of the dominant flow activities within monoclinic hydrogeological structures in coal mine goafs under the influence of rainfall, as well as to investigate the migration mechanism of heavy metal pollutants in groundwater. The research primarily focused on: (1) The effects of dominant flow and fracture development on the migration rate of heavy metals in the goaf; (2) The influence of the collapse funnels on the migration direction and distance of heavy metals. By combining laboratory seepage tests with numerical simulation method, both an experimental sandbox model and a numerical goaf model were developed. Four experimental conditions were established in the indoor sandbox experiment: T1 simulated the natural leaching of coal gangue exposed to rainfall in an open-air environment; T2 simulated the migration of heavy metals driven by dominant flow within the goaf. T3 simulated the migration of heavy metals following roof collapse in the goaf; and T4 simulated the migration of heavy metals in naturally buried coal gangue strata under rainfall. The sandbox was filled with soil and coal gangue from a goaf coal mine in Dongshan town, Qujing City, Yunnan Province. Leachate obtained through static leaching supplemented the dominant flow, and changes in the concentrations of Mn²⁺ and Zn²⁺ were reflected throughout the process. Additionally, to further simulate the long-term migration behavior of elevated concentrations of Mn²⁺ in goaf strata, the numerical model was employed to represent the migration of heavy metals in monoclinic tectonic strata by setting different working conditions and using the porous medium dilute mass transfer equation and Richards equation.The results of the sandbox experiment show that the dominant flow through the middle and lower fractures of the backfill of the close coal seam significantly enhances the migration, diffusion and accumulation of Zn²⁺ and Mn²⁺. Additionally, the presence of the overlying soil layer in the T3 experiment causes the hysteresis of Mn²⁺ release. Numerical simulation results directly show that the collapse funnel in the collapse area accelerates the migration of Mn²⁺ pollution plumes, with the collapse of the goaf increasing the migration distance of Mn²⁺ by 9.65 m. Moreover, the dominant flow within the monoclinic structure after leaching shortens the downward migration period of Mn²⁺ by 8 years and extends the horizontal migration distance by 29.85 m. Mn²⁺ released from coal seam #6 penetrated the strata horizontally and reached surface rivers within 15 years, shortening the migration period by 10 years. Furthermore, after the collapse of the goaf, the upward migration and diffusion distance of Mn²⁺ increased by 9.65 m, indicating that the collapse funnel provides an favorable channel for the upward migration of Mn²⁺.The following conclusions are drawn: (1) The presence of dominant flow in monoclinic goaf generates acidic mine wastewater faster than fracture flow, resulting in the heavy metal Mn²⁺ to more readily contribute to regional groundwater pollution through dominant diffusion in the aquifer of the karst area in Southwest China. (2) The existence of collapse funnels in the monoclinic goaf facilitates the rapid upward migration of heavy metal Mn²⁺, which further leads to surface water pollution. (3) Long-term simulation results show that acidic mine wastewater in the goaf creates a low-concentration area of Mn²⁺ near the dominant flow side of the downward rock strata of the goaf under the action of dominant flow.

Hydrochemical characteristics and controlling factors of surface water and groundwater in a typical coal mining area, southwestern Hunan

XU Zhen, ZHENG Jie, DAI Liangliang, LI Teng

2026, 45(3): 466-480. doi: 10.11932/karst2026y006

Abstract FullText HTML PDF(7241KB) icon

453

304

Abstract:
The southwestern Hunan region is a significant coal resource accumulation zone in China. Within its coal mining areas, carbonate rocks and coal-bearing clastic rock strata are distributed in an interbedded manner, resulting in complex geological structures. Hydrogeochemical processes are jointly influenced by mining activities and karstification. This study focuses on a typical small watershed in southwestern Hunan, with a total area of approximately 20.51 km². The landform is predominantly characterized by dissolution-tectonic features of low mountains and wide valleys, and denudation-tectonic features of clastic rock hills and valleys. The regional strata are primarily composed of the Quaternary System of the Cenozoic Erathem and the Carboniferous System of the Upper Paleozoic Erathem. Groundwater types mainly include pore water in loose rocks, fissure-karst water in carbonate rocks, and pore-fissure water in clastic rocks. To systematically elucidate the hydrochemical properties and identify the major controlling factors of surface water and groundwater in this typical coal mining area of southwestern Hunan, 16 surface water and 13 groundwater samples (including 4 mine water samples) were collected. A multi-indicator comprehensive analysis was conducted using mathematical statistical analysis, Piper trilinear diagrams, Gibbs diagrams, and ion ratio methods to systematically investigate the hydrochemical composition, spatial distribution patterns, and formation mechanisms of various water bodies in the region.The results indicate that both surface water and groundwater in the study area are generally weakly alkaline, with pH values ranging from 7.06 to 8.33 and TDS concentrations varying between 236 and 884 mg·L⁻¹. Groundwater exhibits limited influence from mine water, with hydrochemical types predominantly being HCO₃·SO₄-Ca, followed by HCO₃-Ca. For surface water, the major hydrochemical indicators (${\rm{SO}}_4^{2-}$, Ca²⁺, Mn, TDS, and toxic metals) exhibited the spatial distribution pattern: Shiqiao Creek (north Branch) > Shiyan Creek (downstream) > Shijing Creek (south Branch). In contrast, the trends for pH and ${\rm{HCO}}_3^{-}$ concentration were the opposite. The North Branch (Shiqiao Creek), influenced by mine water input, has a SO₄-Ca hydrochemical type. The south Branch (Shijing Creek), less affected by mining activities, is primarily of the HCO₃·SO₄-Ca type. After their confluence, the main stream of Shiyan Creek generally exhibits a SO₄-Ca hydrochemical type.Gibbs diagrams show that most samples from the study area plot within the rock weathering dominance field, indicating that the chemical composition of the water bodies is primarily controlled by mineral dissolution within the aquifers, with relatively weaker influences from evaporation concentration and atmospheric precipitation. However, most mine water samples deviate from the model's distribution range, and their spatial heterogeneity may originate from geochemical disturbances caused by historical coal mining activities. The end-member diagram illustrating the relative contributions of rock weathering and dissolution shows that regional water samples are concentrated towards the carbonate rock end-member, with some samples trending towards the silicate rock end-member. Ion ratio analysis indicates that Ca²⁺, Mg²⁺, and ${\rm{HCO}}_3^{-}$ are mainly derived from the dissolution of carbonate minerals. ${\rm{SO}}_4^{2-}$ primarily originates from the dissolution of gypsum interbeds within the Zimenqiao Formation limestone and the oxidation of pyrite in the coal-bearing strata. Cl⁻ and ${\rm{NO}}_3^{-}$ are mainly attributed to inputs from human activities such as domestic sewage and agricultural fertilization. Mine water samples exhibit significantly higher concentrations of Fe and Mn and are locally acidic, indicating that historical mining disturbances have enhanced sulfide oxidation, creating an acidic environment that promotes the dissolution and migration of metallic elements. In surface water, the concentrations of ions like ${\rm{SO}}_4^{2-}$, Ca²⁺, and TDS show a positive correlation with the intensity of mine water input, reflecting the pronounced impact of mining activities on surface water chemistry.Principal Component Analysis results indicate that the groundwater hydrochemistry is primarily controlled by three factors: PC1 accounts for 39.65% of the total variance and reflects the dominance of carbonate rock dissolution in controlling the regional groundwater chemistry; PC2 explains 36.08% of the variance, characterizing the influence of anthropogenic inputs of Cl⁻ and Na⁺; PC3,representing gypsum dissolution,contribute 11.35% to the total variance. The surface water chemistry is predominantly governed by three factors: PC1, accounting for 54.28% of the total variance, characterizes the dual influence of gypsum dissolution and sulfide oxidation from coal seams on the surface water hydrochemistry; PC2, with a variance contribution of 24.69%, reflects the coupled effects of silicate weathering from the Carboniferous Ceshui Formation and anthropogenic NaCl input; PC3, with a variance contribution of 11.80%, represents ${\rm{NO}}_3^{-}$ input from agricultural fertilization.Integrated multi-indicator analysis reveals that rock weathering/dissolution and anthropogenic activities are identified as the two main factors controlling the hydrochemical evolution in the study area. Specifically, carbonate rock dissolution governs the variations of Ca²⁺, Mg²⁺, and ${\rm{HCO}}_3^{-}$, while sulfide oxidation and gypsum dissolution drive ${\rm{SO}}_4^{2-}$ enrichment. In contrast, anthropogenic activities primarily regulate the distribution of Cl⁻ and ${\rm{NO}}_3^{-}$. This study systematically reveals the hydrochemical characteristics and formation mechanisms of surface water and groundwater in a typical coal mining area of southwestern Hunan, clarifying the hydrochemical relationships and spatial distribution patterns within the "surface water-groundwater-mine water" ternary system. The results demonstrate that the coupled effects of carbonate rock weathering/dissolution and sulfide oxidation from coal-bearing strata jointly determine the chemical evolution of regional water bodies, while human activities have intensified ion migration in localized areas. The comprehensive multi-indicator analytical methodology proposed in this study is able to provide a scientific basis and methodological reference for identifying hydrochemical characteristics, preventing and controlling pollution, and protecting regional water resources in karst coal mining areas.

Hydrochemical drivers and enhanced chemical weathering mechanisms in a subtropical karst river under anthropogenic impacts: A case study of the Wengjiang River Basin

WANG Song, LIU Fan, HOU Tao, ZHOU Zhenzhao, LI Ming, WANG Zibo, CHEN Jianqing, GUO Fang

2026, 45(3): 481-495. doi: 10.11932/karst2025y031

Abstract FullText HTML PDF(2895KB) icon

444

299

Abstract:
Taking the Wengjiang River Basin, a first-order tributary of the Beijiang River in the Pearl River system, as a case study, this study unravels the hydrogeological processes and chemical weathering regimes synergistically driven by anthropogenic activities and bedrock lithology. It accurately identifies the controlling factors of solute sources across wet and dry seasons and quantifies the anthropogenic imprint. By integrating hydrochemical isotopic,and ionic ratio analyses, this study systematically elucidates the hydrogeochemical evolution patterns characteristics of subtropical karst regions under monsoon climate conditions.This study reveals that hydrogeological processes in the Wengjiang River Basin are strongly governed by the dominant subtropical monsoon climate, exhibiting pronounced seasonal variability. During the high-flow season, river discharge is primarily recharged by meteoric precipitation characterized by depleted isotopic signatures, resulting in a rapid hydrological responses dominated by quickflow pathways. Conversely, baseflow during the low-flow season is sustained by groundwater subjected to significant evaporative fractionation, as evidenced by reduced d-excess values. This clearly signals intensified evaporation, leading to a marked evaporative concentration effect on the river's hydrochemistry. Furthermore, reservior regulation and agricultural irrigation further attenuates flow velocities and weaken water cycle intensity during the dry season. Lithology serves as the primary controll on the spatial distribution of hydrochemistry; carbonate weathering dominates the solute composition, establishing a Ca-HCO₃ hydrochemical type. In areas underlain by silicate rocks, Total Dissolved Solids (TDS) are significantly lower, and chemical compositions are more susceptible to precipitation inputs. While carbonate weathering remains the principal source of solutes year-round, the contribution of anthropogenic activities escalates markedly during the low-flow season.A key finding of this study is the quantification of anthropogenic contributions to catchment-scale chemical weathering rates. Calculation indicate that the total chemical weathering rate in the Wengjiang River Basin is 78.7 t·km⁻²·a⁻¹. Notably, sulfuric acid-induced carbonate weathering accounts for an approximate 19.8% increase in the weathering rate, identifying human activities as a significant geochemical agent driving the intensification of weathering processes. Anthropogenic disturbances-exemplified by the introduction of allochthonous acids from mining activities-not only alter natural weathering trajectories but also perturb riverine ionic compositions. This effect is particularly pronounced during the low-flow season when the hydrological buffering capacity diminishes. The results further demonstrate that the hydrogeochemical characteristics of the Wengjiang River Basin emerge from the synergistic interplay of lithological background, seasonal hydrological dynamics, and high-intensity human activities. These findings underscore the pivotal regulatory role of anthropogenic disturbances in governing geochemical cycling and weathering mechanisms within karst systems.Collectively, this study systematically deciphers the hydrochemical drivers and ion sources in the Wengjiang River Basin through three interconnected perspectives: bedrock lithological control, seasonal hydrological variability, and anthropogenic disturbance. The findings not only provide a critical scientific basis for water resource protection and sustainable watershed management under intense human disturbance, but also advance the understanding of geochemical cycling in subtropical monsoon regions with complex geological settings. Furthermore, this study underscores the imperative to integrate anthropogenic impacts into future hydrogeochemical modeling and management strategies.

Indication of evaporation effects by hydrogen and oxygen isotopes in the water bodies of the Tanglanghe River Basin, Yiyuan county, Shandong Province

ZHU Xiaoyan, ZHANG Zhuo, HUANG Fen

2026, 45(3): 496-507. doi: 10.11932/karst20260303

Abstract FullText HTML PDF(4439KB) icon

9

1

Abstract:
This study takes the Tanglanghe River Basin in Yiyuan county, Shandong Province as the study area. To investigate the influence of evaporation on the regional water cycle, the stable isotopic compositions of groundwater and surface water samples collected along the Tanglanghe River were systematically measured in August 2024 and April 2025. The analytical results reveal distinct seasonal variations and spatial patterns in the isotopic signatures, highlighting the indicative significance of δ¹⁸O and deuterium excess (d-excess) for evaporation processes in the basin. The major findings are summarized as follows.(1) During the wet period in August 2024, which coincides with the peak of the East Asian Summer Monsoon rainfall, the δD-δ¹⁸O trend line for all water bodies in the basin lies to the Global Meteoric Water Line (GMWL). Notably, both the slope and intercept of the local water line derived from this period are higher than those of the previously established Local Meteoric Water Line (LMWL). This statistical relationship strongly indicates that atmospheric precipitation is the predominant and direct source of recharge for the basin's hydrological system. Therefore, the Tanglanghe River can be definitively classified as a rain-fed fluvial system, with its flow regime primarily dictated by the timing and magnitude of precipitation events within its catchment area. In stark contrast, during the dry period of April 2025, the isotopic compositions of all sampled water bodies exhibited a pronounced enrichment in the heavier isotopes (δ¹⁸O and δD). This systematic shift toward heavier isotopic values directly reflects the dominance of intense evaporation under arid conditions.(2) Building on this temporal contrast, a marked distinction also exists between the isotopic behavior of groundwater and surface water within the Tanglanghe River Basin. Groundwater samples exhibit a notably narrow range of variation in both δD and δ¹⁸O values, and these values demonstrate minimal seasonal fluctuation in samples between the wet and dry periods. This temporal and spatial homogeneity reflects the profound regulatory capacity, storage effect, and mixing homogenization inherent to subsurface flow systems. Conversely, surface water display pronounced seasonal isotopic variability. During the wet period, surface water isotopes are relatively depleted (more negative δ-values), whereas during the dry period, they are significantly enriched (more positive δ-values). Specifically, the isotopic enrichment ranking during the wet period is groundwater > river water > reservoir water (from less negative to more negative), whereas this order reverses dramatically in the dry period to river water > reservoir water > groundwater. This pattern indicates that the water is directly recharged by atmospheric precipitation. During the wet period, runoff is rapid and residence time is short, resulting in inadequate mixing and homogenization of seasonal signals. In the subsequent dry season, the reservoir experiences particularly intense evaporation due to its larger exposed surface area and extended hydraulic retention time, leading to heavy isotope enrichment.(3) Further analysis of environmental controls on isotopic fractionation revealed a significant negative correlation between δ¹⁸O values and water temperature. Notably, this relationship proved statistically stronger than the correlation observed between deuterium excess (d-excess) and temperature, underscoring the greater sensitivity of δ¹⁸O to thermal dynamics during evaporation.(4) Finally, by leveraging the isotopic mass balance approach and the distinct isotopic enrichment observed between the wet and dry periods, the proportion of water loss due to evaporation along the main stem of the Tanglanghe River during the dry period of April 2025 was estimated. The calculated evaporative loss fraction ranges from 4.62% to 19.40%. A distinct spatial trend is evident, with the evaporation ratio progressively increasing from the upstream reaches toward the downstream sections. This downstream increase in evaporative loss is governed by two primary factors. First, the intensity of groundwater recharge, as substantial groundwater inflow, characterized by more depleted and stable isotope signatures, can buffer the evaporation signal. Second, variations in the morphology of the river surface, including wider channel widths and slower flow velocity, increase the exposure time and the surface area available for evaporation flux, thereby accelerating isotope enrichment.

Response characteristics and accounting method optimization of karst carbon sinks to exogenous acid input: A case study of the Diaojiang River Basin

QIN Luhui, LI Jianhong, TANG Xiangling, GAN Zhijie, XIAO Qiong, SUN Ping'an

2026, 45(3): 508-524. doi: 10.11932/karst2026y018

Abstract FullText HTML PDF(8050KB) icon

13

1

Abstract:
Fully exploiting and enhancing the potential of karst carbon sinks is one of the important pathways for achieving China’s carbon peak and carbon neutrality goals. However, the input of exogenous acids and the weathering of silicate rocks significantly compromise the accuracy of assessing karst carbon sink fluxes. To reveal the patterns of dynamic influence of exogenous acid input and silicate weathering on karst carbon sinks, it is imperative to establish a revised method for carbon sink accounting. Accordingly, this study focuses on the Diaojiang River Basin in Guangxi, a typical karst area affected by both mining activities and complex geological conditions. The basin, located in northern Guangxi, experiences a subtropical monsoon climate characterized by distinct wet and dry seasons. Its lithology is complex, mainly including limestone, dolomite, and clastic interlayers. The upstream area of the basin suffers from severe sulfate pollution due to gypsum dissolution and historical mining activities, while the mid- to downstream areas are dominated by carbonate rocks and are continuously affected by human activities.This study employs a carbon isotope end-member mixing model with the stoichiometric relationships of carbonate and silicate weathering to quantify the proportions of three main sources of ${\rm{HCO}}_3^{-}$ in the river: silicate weathering, carbonate dissolution by carbonic acid, and carbonate dissolution by sulfuric acid. Based on this, the traditional hydrochemical-discharge method is systematically revised.The main results include the following five aspects: (1) At the Chehe section, the cations are predominant Ca²⁺, while anions are mainly ${\rm{SO}}_4^{2-}$ and ${\rm{HCO}}_3^{-}$; at the Jiusi and Baiwang sections, cations also are dominated by Ca²⁺, but anions are primarily ${\rm{HCO}}_3^{-}$. The hydrochemical types of the basin show clear spatial differentiation: The upstream Chehe section is classified as SO₄·HCO₃-Ca type, while the mid- to downstream Jiusi and Baiwang sections are both of the HCO₃-Ca type, reflecting the combined effects of geological background and human activities. (2) The ratios of Na⁺/(Na⁺+Ca²⁺) at all sections of the Diaojiang River Basin range from 0.02 to 0.20, and the ratios of Cl⁻/(Cl⁻+ ${\rm{HCO}}_3^{-}$) range from 0.01 to 0.07, both below 0.5, indicating that the river water chemistry is significantly influenced by rock weathering. The relationships of Ca²⁺/Na⁺ versus Mg²⁺/Na⁺ and Ca²⁺/Na⁺ versus ${\rm{HCO}}_3^{-}$/Na⁺ reflect that both silicate and carbonate rocks contribute to weathering processes at all three sections. The equivalent ratios of [Ca²⁺+Mg²⁺]/[ ${\rm{HCO}}_3^{-}$+${\rm{SO}}_4^{2-}$] at all sections are close to 1, confirming the involvement of sulfuric acid in the weathering process of carbonate rocks. (3) Analysis of the molar ratio relationships between 2${\rm{SO}}_4^{2-}$/ ${\rm{HCO}}_3^{-}$ and 2(Ca²⁺+Mg²⁺)/${\rm{HCO}}_3^{-}$ further supports the occurrence of upstream gypsum dissolution and the role of sulfuric acid in carbonate weathering. (4) The average contribution ratios of silicate weathering at the Chehe, Jiusi, and Baiwang sections are 2.9%, 1.6%, and 1.4%, respectively; the average contribution ratios of sulfuric acid weathering are 33.4%, 26.4%, and 21.3%, respectively. (5) Using the revised method, the annual carbon sink amounts at the Chehe, Jiusi, and Baiwang sections are calculated to be 2.26×10⁵ kg C, 1.96×10⁶ kg C, and 3.00×10⁷ kg C, respectively, which are significantly different from the unrevised results.In summary, the input of exogenous acids and the weathering of silicate rocks significantly alter the hydrochemical composition and carbon sink capacity of karst rivers. The revised carbon sink accounting method proposed in this study is able to effectively enhance the accuracy of watershed-scale karst carbon sink assessments under complex geological backgrounds and anthropogenic impacts, providing a reliable scientific basis for related research and policy development.

Spatial and temporal distribution characteristics and control mechanisms of the karst inorganic carbon sink in the Guangxi karst region from 2002 to 2022

TIAN Huanjie, XIAO Qiong, SUN Ping'an, CHEN Fajia, ZHANG Xinyao

2026, 45(3): 525-538. doi: 10.11932/karst2026y016

Abstract FullText HTML PDF(13445KB) icon

18

1

Abstract:
Carbonate rock chemical weathering is a fundamental yet complex process within the global carbon cycle and constitutes an important long-term sink for atmospheric carbon dioxide (CO₂). During carbonate dissolution, CO₂ derived from the atmosphere and/or soil is consumed and transformed into Dissolved Inorganic Carbon (DIC), predominantly in the form of bicarbonate (${\rm{HCO}}_3^{-}$). This DIC is subsequently transported downstream through groundwater and surface runoff systems and may ultimately be stored in the ocean over long timescales. Because this process reduces atmospheric CO₂ concentrations and, to some extent, mitigates the release of soil carbon to the atmosphere, carbonate weathering is widely regarded as an important natural mechanism for buffering anthropogenic climate warming. However, the magnitude of karst inorganic carbon sinks at the regional scale remains highly uncertain. This uncertainty is particularly pronounced in extensive karst regions, where strong spatial heterogeneity in lithological conditions, aquifer connectivity, recharge processes, and land-surface environments makes it difficult to upscale point-based hydrochemical observations to the regional level. Consequently, significant discrepancies persist in regional estimates of karst carbon sinks, and the relative contributions of climatic and environmental factors to DIC production and export have yet to be fully resolved. The Guangxi Zhuang Autonomous Region, located in southern China, is one of the most extensive and representative karst regions in the country. It is characterized by widespread carbonate rock outcrops, well-developed epikarst and subsurface drainage systems, and hydrogeochemical processes that are highly sensitive to monsoonal climatic forcing. The region has a subtropical humid monsoon climate, abundant precipitation, and diverse geomorphological settings, all of which provide favorable conditions for carbonate weathering and dissolved carbon export. At the same time, Guangxi exhibits marked regional variability in climate, hydrology, topography, and karst development, making it an ideal study area for reassessing karst inorganic carbon sinks and exploring their controlling mechanisms at the regional scale. Against this background, the present study systematically re-evaluates the karst inorganic carbon sink in Guangxi from 2002 to 2022 and employs the Geodetector model to investigate its controlling factors.This study employs multidisciplinary methodology, integrating field observations, remote sensing data, machine learning modeling, and spatial attribution analysis. First, in consideration of the geological and hydrogeological setting of the karst region in Guangxi, a relatively large dataset of measured groundwater DIC concentrations was compiled. High-resolution remote sensing data were also selected to characterize environmental conditions relevant to karst carbon processes. Based on these multi-source datasets, machine learning models were trained and validated to reconstruct the spatiotemporal distribution of DIC concentrations in karst groundwater across Guangxi from 2002 to 2022. Compared with conventional point-based observational approaches, this method enhances the regional representativeness of the estimates and helps compensate for the limited spatial and temporal coverage of measured data. Additionally, the incorporation of remotely sensed variables provides feasible proxy indicators for certain environmental conditions that are challenging to monitor continuously over extended periods. Second, the reconstructed DIC concentration fields were combined with long-term discharge data, and the karst inorganic carbon sink generated by carbonate weathering was recalculated using the classical and widely applied hydrochemical flux method, thereby providing a more accurate representation of both temporal fluctuations and spatial variability in dissolved carbon export. Finally, the Geodetector model was employed to quantify the explanatory power of potential influencing factors through both factor detection and interaction detection, with the aim of identifying the dominant controls on the spatial differentiation of the karst carbon sink in Guangxi.The results show that: (1) From 2002 to 2022, the cumulative karst inorganic carbon sink in Guangxi reached 5.03×10⁷ tCO₂, with an annual average of 2.40 ×10⁶ tCO₂·a⁻¹. (2) Interannual variability in the karst inorganic carbon sink was pronounced, with a marked increasing trend from 2003 to 2006, followed by an overall fluctuating upward trend thereafter. The maximum annual sink occurred in 2008, reaching 3.56×10⁶ tCO₂·a⁻¹, whereas the minimum was recorded in 2009 at 1.44×10⁶ tCO₂·a⁻¹, the former being approximately 2.5 times the latter. (3) The mean karst inorganic carbon sink intensity during the study period was 30.24 tCO₂·km⁻²·a⁻¹, with clear spatial heterogeneity. Higher values were mainly concentrated in central and northeastern Guangxi, whereas lower values occurred in the southwestern and southeastern parts of the region. (4) Geodetector analysis indicates that precipitation is the primary factor controlling the spatiotemporal variation of the karst inorganic carbon sink in Guangxi, consistent with the strong dependence of karst process on water availability. Further interaction detection reveals that the explanatory power of the interaction between any two factors is greater than that of any single factor, indicating that the formation and differentiation of the karst carbon sink exhibit significant nonlinear coupling characteristics. Among these interactions, the interaction between temperature and precipitation shows the highest explanatory power.

Estimation of karst carbon sink fluxes from rock weathering and analysis of their spatiotemporal dynamics in Yunnan Province

HUANG Xunchao, YE Jiangxia, YIN Xiaojie, WANG Yan, LIU Zeng, TANG Jimin

2026, 45(3): 539-552. doi: 10.11932/karst20260304

Abstract FullText HTML PDF(10048KB) icon

11

1

Abstract:
Research on the global transfer of "carbon sources" and "carbon sinks" has revealed the potential existence of previously unidentified carbon sinks within the continental biosphere. Among these, the dynamic chemical process of rock weathering constitute a major carbon sink capable of sequestering atmospheric CO₂. The carbon sink from rock weathering constitutes an integral component of this "missing carbon sink". Notebly, the distinctive biogeochemical process of rock, particularly in karst landscapes, has been increasingly recognized as a critical and persistent carbon sink. Carbonate minerals dissolve through carbonic acid formed from atmospheric or soil-derived CO₂ and water, generating bicarbonate ions that are transported via river systems to the oceans, thereby sequestering carbon over geological timescales. Therefore, accurately estimating the magnitude of the carbon sink from rock weathering and understanding its spatiotemporal dynamics are essential for regional carbon assessments, climate change mitigation strategies, and policy formulation in karst-dominated regions. Yunnan Province, situated in Southwest China, is one of the country's primary karst development areas, characterized by diverse soluble and silicate rock types, making it an ideal natural laboratory for such investigations. This study aims to quantify carbon sink fluxes from rock weathering in Yunnan Province from 2001 to 2020, analyze its spatiotemporal evolution characteristics, and project future trends under ongoing environmental changes.Based on a detailed classification of soluble and insoluble rocks within Yunnan Province, this study employed the well-established Global Erosion Model for CO₂ consumption (GEM-CO₂) model.This model estimates the CO₂ consumption flux ($ {{F}}_{{{\text{CO}}_{2}}} $) resulting from the chemical weathering of different rock types by integrating surface runoff data with specific lithological coefficients ($ a $) that represent the weathering potential of each rock. The fundamental equation is ($ {{F}}_{{{\text{CO}}_{2}}}=Q\cdot a $), where ($ Q $) is the annual surface runoff. Runoff ($ Q $) was calculated using the Turc method, which estimates actual evaporation ($ E $) based on annual precipitation ($ P $) and temperature ($ T $) data, with ($ Q $) derived as ( $ P $-$ E $). High-resolution spatial data, including a 1∶500,000 geological map (resampled to a 1-km grid) and long-term meteorological datasets (precipitation and temperature) from 2001 to 2020, were utilized as primary inputs. Following flux estimation, a suite of robust statistical and geospatial methods was applied. The Theil-Sen Median trend analysis and Mann-Kendall significance test were used to detect monotonic trends in the carbon sink fluxes over the two-decade period. The coefficient of variation (CV) was calculated to assess the stability and variability of fluxes. Finally, to predict the future persistence or reversal of these trends, the Hurst exponent, derived from Rescaled Range (R/S) analysis, was calculated and combined with Sen's slope to generate a map of future trend predictions.The key findings of this study are multifaceted, (1) From 2001 to 2020, the estimated annual average CO₂ consumption flux attributable to rock weathering in Yunnan Province was 187.79 × 10³ mol·km⁻²·a⁻¹. This corresponds to a total CO₂ consumption mass of approximately 63.12×10⁶ t and a total carbon mass of 17.38×10⁶ t over the two decades. Significant disparities were observed among different types of lithology in their contribution to the carbon sink. Carbonate rocks exhibited the highest average annual CO₂ consumption flux at 582.91×10³ mol·km⁻²·a⁻¹, accounting for 53.95% of the total CO₂ consumed, thus establishing them as the predominant rock type driving the weathering carbon sink in the province. This was followed by shale (207.22×10³ mol·km⁻²·a⁻¹), acid volcanic rocks (87.55×10³ mol·km⁻²·a⁻¹), sandstone (56.29×10³ mol·km⁻²·a⁻¹), and metamorphic and plutonic rocks (51.2×10³ mol·km⁻²·a⁻¹). (2) The spatial distribution of the rock weathering carbon sink flux demonstrated a distinct pattern characterized as "high in the surrounding areas and low in the central region". High-flux zones were primarily concentrated in eastern Yunnan, with values gradually increasing from north to south, and in the border regions of western Yunnan. In contrast, low-flux zones were predominantly located in central and northern Yunnan. The coefficient of variation (CV) for the flux across the province ranged from 0 to 0.64, indicating varying degrees of temporal stability. Areas of high variability (CV ≥ 0.35) constituted 14.75% of the province's area and were mainly situated in central Yunnan (around Kunming, Chuxiong, and Dali), forming a pattern of "high variability in the center and low variability in the surrounding areas". This pattern suggests that the carbon sink in central Yunnan is more sensitive to interannual environmental fluctuations, potentially influenced by intensive human activities and specific lithological compositions (sandstone and shale). (3) Temporal trend analysis revealed a slight overall declining trend in total CO₂ consumption mass from 2001 to 2020, with a linear slope of -0.0055, although significant interannual fluctuations were observed. Theil-Sen analysis indicated that areas exhibiting a decreasing trend (approximately 76.93% of the province's area with significant change) were more extensive than those showing an increasing trend (22.98%), with the most pronounced decreases occurring in northwestern and central Yunnan and increases concentrated in northeastern Yunnan (e.g., Zhaotong). The Hurst exponent for the carbon sink flux across the province ranged from 0.30 to 0.63, indicating a complex mix of future behaviors. Approximately 57.18% of the area exhibited a Hurst index less than 0.5, suggesting an anti-persistent trend (future reversal of the past trend), while 42.82% displayed an index greater than 0.5, indicating a persistent trend. By integrating Sen's slope and the Hurst index, future predictions were derived. It is projected that areas with an increasing trend in the carbon sink flux from rock weathering will cover approximately 39.86% of the province, mainly concentrated in eastern and northwestern Yunnan. Conversely, areas predicted to experience a continuous decrease will account for approximately 34.86% of the province, predominantly distributed in central, northeastern, and parts of eastern Yunnan.This study provides a comprehensive quantitative assessment of the carbon sink from rock weathering in Yunnan Province over the past two decades. The results highlight the dominant role of carbonate rocks and reveal significant spatial heterogeneity and temporal variability in this carbon sink, primarily driven by the interplay of lithology, climate (precipitation and temperature), and potentially human activities. The predictive analysis offers valuable insight, suggesting a slight net increase in future sink areas despite recent declining trends. These findings provide crucial scientific evidences for understanding the regional carbon budget, validate the efficacy of the GEM-CO₂ model for large-scale assessments, and support the formulation of targeted ecological restoration and carbon management policies in Yunnan Province and similar karst regions worldwide.

Evaluation and prediction of ecological sensitivity of the Xuebaoding region based on PLUS-InVEST

WANG Lihui, WANG Jiani, LIU Yanguo, CAI Jiangrong, JIANG Ruiyang, YIN Sijie, AN Dejun, TANG Shu, LI Qiongfang, DAI Qunwei

2026, 45(3): 553-567. doi: 10.11932/karst20260305

Abstract FullText HTML PDF(24775KB) icon

11

0

Abstract:
Ecological sensitivity refers to the capacity of environmental factors to adapt to external pressures or disturbances without degrading or compromising environmental quality. In essence, it reflects the likelihood and ease of occurrence of regional ecological and environmental problems. The Xuebaoding region, located on the eastern of the Qinghai-Xizang Plateau, features abundant karst landscapes and unique ecosystem. Highly sensitive to climate change and human activities, it serves as an ideal area for studying the evolution of ecological sensitivity. It is of great scientific value and practical significance to reveal the response mechanism of the regional ecosystem under the background of land use change for optimizing the management strategy of protected areas and promoting the harmonious coexistence between human and nature. This study focuses on the Xuebaoding region, employing the PLUS-InVEST model to simulate land use changes and analyze the dynamic variations in habitat quality. It systematically evaluates environmental quality from 2000 to 2020, predicts habitat quality for 2030, and applies the Geo Detector to quantify environmental factors. Based on these quality changes, the study investigates ecological sensitivity within the study area. The results show that:(1) From 2000 to 2020, the change of land use structure showed the characteristics of "two decreases and four increases", in which the area of grassland and cultivated land continued to decrease, with a net decrease of 294.98 km² and 32.79 km², respectively. However, the area of woodland, construction land, water area and unused land all showed an increasing trend. The net transfer of woodland was 227.78 km², and the construction land increased by 166.98%.(2) From 2000 to 2020, the cultivated land area further decreased with the acceleration of urbanization, and the construction land continued to expand at an average annual rate of 8.35%. The results of habitat quality assessment show that the area of extremely low ecological environment increased by 156.90% in 2020 compared with 2000, and the spatial distribution of expansion area and construction land was highly coincident, indicating that urban expansion was the main reason for the decline of ecological environment quality and the growth of extremely low-grade areas.(3) Despite the changes in land use structure, the overall habitat quality in the study area remained above 85.2% from 2000 to 2020, and the average habitat quality remained relatively stable after it was increased from 85.26% in 2000 to 86.11% in 2010, indicating that the overall ecosystem structure in the basin remained at a high level and the ecological carrying capacity was strong. However, the forecast results in 2030 based on PLUS model show the overall accuracy, and the average value of future habitat quality will drop to 85.9%, which is lower than that in 2020, and the area of extremely low area is expected to increase by 192.47% compared with that in 2000, indicating that the potential pressure on regional ecosystems continues to increase.(4) The results of Geo Detectors show that natural factors,such as elevation,slope,temperature,and precipitation, are still the dominant factors to determine the spatial pattern of ecological environment quality in the study area. Interactive detection shows that the factors are enhanced by two factors, which indicates that human intervention is significantly affecting the evolution direction and range of ecological environment quality.

Variations in soil organic carbon and microbial community structure under different afforestation modes in karst rocky desertification areas

YANG Hui, ZHU Tongbin, WEN Dongni, YANG Lin, QIU Cheng

2026, 45(3): 568-579. doi: 10.11932/karst20260306

Abstract FullText HTML PDF(3365KB) icon

4

0

Abstract:
Afforestation is an important way for ecological restoration of karst rocky desertification area in southwest China. Studying the variation characteristics of soil organic carbon and microbial community structure under different afforestation modes can effectively indicate soil quality, which has great significance in guiding rocky desertification restoration. This study was conducted in Changchongli, Lingchuan county, Guilin City, Guangxi, a region characterized by typical karst peak-cluster depressions and valleys. The bedrock mainly consists of limestone and the soil layer is relatively shallow. Three representative afforestation modes, including Osmanthus fragrans Lour., Pinus massoniana Lamb., and Styphnolobium japonicum (L.) Schott, were selected as research subjects, with adjacent Citrus reticulata Blanco cv. 'Shatangju' serving as the control. The four sites shared similar altitude, slope aspect, slope position, and slope gradient, and all soils were calcareous. In this study, Soil Organic Carbon (SOC) content and chemical composition, together with the abundance and community structure of soil microorganisms, were examined under different afforestation modes.The results showed that:compared to orchard soils, SOC and total nitrogen contents increased significantly under O. fragrans and S. japonicum plantations, with increases of 47.6% and 45.3% for O. fragrans and 126% and 121% for S. japonicum, respectively, whereas no significant variations occurred under P. massoniana. Water Holding Capacity (WHC) also increased significantly under O. fragrans and S. japonicum, while the C∶N ratio showed no significant difference. In contrast, WHC showed no marked variation under P. massoniana but had a significantly elevated C∶N ratio. The orchard soil was acidic (pH 4.34), whereas soils under the three afforestation modes were neutral (pH of 6.90 to 7.09). Total calcium was lowest in orchard soil (2.77 g·kg⁻¹) but increased significantly in the afforested soils, reaching 11.8 g·kg⁻¹ under S. japonicum. In contrast, total phosphorus and potassium contents were significantly lower in afforested soils. Soil texture also changed following afforestation, with a reduction in clay content and increases in proportions of silt and sand. The chemical composition of SOC also differed among afforestation types. O-alkyl C dominated in soils under O. fragrans and S. japonicum, whereas aromatic C was the predominant component in P. massoniana soils. Compared to orchard soils, aromatic C in P. massoniana soils increased by 20.7% while O-alkyl C decreased by 16.4%, with no significant variations in alkyl C or carbonyl C. Conversely, O-alkyl C in soils under O. fragrans and S. japonicum increased by 9.1% and 2.4%, respectively, while, aromatic C decreased by 21.1% and 11.8%. Compared to the orchard soil (0.65), the ratio of alkyl C to O-alkyl C rose significantly to 0.79 in P. massoniana soils,whereas no significant differences were observed under O. fragrans or S. japonicum. Afforestation significantly enhanced total soil microbial PLFA abundance under all three afforestation modes, reaching 14.4, 9.00, and 16.3 nmol·g^‒1 under O. fragrans, P. massoniana, and S. japonicum, respectively. Except for fungi and Gram-positive bacteria in P. massoniana soils, the abundances of bacteria, Gram-negative bacteria, arbuscular mycorrhizal fungi, and actinomycetes were significantly higher than those in afforested soils, with the most pronounced increases observed under O. fragrans and S. japonicum. The relative abundance of microbial groups followed the order: bacteria>Gram-positive bacteria>Gram-negative bacteria> actinomycetes>fungi>arbuscular mycorrhizal fungi across all soil types. Among all afforestation modes, the relative abundances of Gram-negative bacteria, arbuscular mycorrhizal fungi, and actinomycetes significantly increased, whereas Gram-positive bacteria declined non-significantly. Fungal abundance remained low across all plantations, with fungal/bacterial ratio (F∶B) not exceeding 0.25. Compared to Shatangju orchard soil, the F∶B ratio increased significantly under O. fragrans, slightly under S. japonicum, and showed a decling trend under P. massoniana. This findings indicate that O. fragrans and S. japonicum were more conducive to maintaining soil ecosystem stability in karst regions.In karst regions, soil under P. massoniana exhibited lower SOC content dominated by recalcitrant components, whereas O. fragrans and S. japonicum promoted SOC accumulation with a prevalence of labile fractions. Overall, these two species demonstrated significantly greater potential for improving soil quality, thereby fostering more effective soil ecosystem restoration.

Response of soil organic carbon mineralization to temperature and moisture changes across different karst ecosystems in Nonggang, Guangxi, China

LI Jun, DU Xinru, WANG Shaohan, LONG Jifeng, PEI Guangting, WANG Aihua, HE Tongxin, ZHANG Weidong, ZHU Tongbin, SUN Jianfei

2026, 45(3): 580-592. doi: 10.11932/karst20260307

Abstract FullText HTML PDF(3140KB) icon

4

1

Abstract:
Soil Organic Carbon(SOC) mineralization is the process of SOC decomposition driven by microorganisms, followed by the subsequent release of CO₂. This process plays a critical role in the global carbon cycle and is highly sensitive to climate warming and shifts in precipitation patterns. The karst region of southwest China is inherently ecologically fragile, and historically, inappropriate land use led to severe rocky desertification. Over recent decades, however, major national ecological restoration projects have driven substantial vegetation recovery, turning the region into an important carbon sink. While considerable researches have focused on the recovery of vegetation and soil carbon stocks, systematic understanding of SOC mineralization dynamics and its response to coupled changes in temperature and moisture across the full successional sequence from farmland to primary forest remain limited. Using a laboratory incubation approach, this study aimed to clarify how SOC mineralization responds to variations in temperature and moisture across different karst ecosystems, how temperature sensitivity (Q₁₀) responses to changes in moisture availability, and what key factors regulate these processes.The study was conducted in the Nonggang National Nature Reserve in Guangxi. The area experiences a typical tropical monsoon climate, with a mean annual temperature of 22 ℃ and annual precipitation ranging from 1,150 to 1,550 mm. The soil is calcareous soil derived from limestone. In the study area, five vegetation types are distributed: farmland, grassland, shrubland, secondary forest, and primary forest. Among them, farmland, grassland, and shrubland are located around the periphery of the protected area, the secondary forest is situated in the experimental zone of the protected area, and the primary forest is located in the buffer zone of the protected area. Surface soil (0−15 cm) were collected in July, 2022. A laboratory incubation experiment was conducted to measure SOC mineralization under three temperature levels (20 ℃, 25 ℃, and 30 ℃) and three moisture levels (30%, 50%, and 70% of water-holding capacity). SOC mineralization rates were measured using the alkali absorption method, and Q₁₀ values were calculated based on these rates. Soil physicochemical properties, including SOC, total nitrogen, microbial biomass carbon, ammonium, nitrate, available phosphorus, exchangeable calcium, and pH were also determined.The results showed, (1) Elevated temperature significantly enhanced SOC mineralization rate, cumulative mineralization, and potentially mineralizable carbon. Increased moisture also promoted mineralization, but its effect was weaker and modulated by temperature. Binary quadratic model fitting indicated that temperature was the dominant driver of SOC mineralization. (2) SOC mineralization characteristics varied distinctly among ecosystems. Cumulative mineralization was higher in secondary and primary forests and lowest in farmland. Heatmap analysis and stepwise regression revealed that soil ammonium, nitrate, pH, and exchangeable calcium were the key factors influencing the mineralization process. (3) Q₁₀ values in this study ranged from 1.16 to 2.26. Farmland and grassland exhibited significantly higher temperature sensitivity than shrubland and forests. Although the main effect of moisture on Q₁₀ was not significant, a significant interaction between ecosystem type and moisture was observed. In grassland, Q₁₀ decreased markedly with increasing moisture. Stepwise regression analysis showed that nitrate and SOC content were the primary factors regulating Q₁₀.This study systematically examined the response of SOC mineralization to coupled changes in temperature and moisture across a successional sequence from farmland to primary forest in the Guangxi karst region. The findings indicate that SOC mineralization is strongly responsive to climate warming, and the magnitude was significantly regulated by ecosystem types and soil nitrogen availability. Compared with non-karst ecosystems, the Q₁₀ values observed here were generally lower, and early-successional ecosystems (farmland and grassland) displayed higher temperature sensitivity than late-successional ecosystems (shrubland and forest). This finding has important implications for regional carbon balance projections. As climate warming continues, restored ecosystems in rocky desertification regions may face an increasing risk of soil carbon loss and a potential shift from carbon sink to carbon source. This change is constrained by ecosystem nutrient availability, particularly nitrogen. This study provides experimental evidence and mechanistic insights that are essential for accurately evaluating the carbon source/sink function of karst ecosystems under ongoing global change.

Application of machine learning algorithms in seepage evaluation of karst reservoirs

FAN Zhujun, LIU Zhiwei, WANG Jiyuan, ZHANG Yong, HUANG Qibo, LI Tengfang

2026, 45(3): 593-602. doi: 10.11932/karst2026y008

Abstract FullText HTML PDF(603KB) icon

377

289

Abstract:
To date, there is still no reliable method to quantitatively predict and evaluate the seepage volume of karst reservoirs. This is because the factors influencing the leakage in karst reservoirs are numerous and highly complex. These factors include the lithology of the strata, the degree of structural development, the type of groundwater, the development of karst, hydrodynamic conditions, and anti-seepage treatment measures, etc. Most of these factors are not completely quantitative, and even random and ambiguous, making it difficult to describe them with deterministic models. Furthermore, these factors interact through complex cross-effects and dynamic mutual influences and constraints. The nonlinear relationships among these factors are too intricate to be captured by a single or even a group of simple algebraic equations.Machine learning algorithms are characterized by high accuracy and stability in simulating complex groundwater runoff and assessing reservoir seepage, which has attracted the interest from many researchers in recent years. These algorithms can automatically analyze data to identify patterns and use them to predict unknown outcome. They can automatically select features most strongly correlated with evaluation events and solve problems such as noise, missing values, and outliers, thereby improving data quality and integrity. Additionally, machine learning significantly reduces the time required for data processing. Compared to traditional statistical methods, machine learning is more reliable in identifying nonlinear patterns of input and output. Since most factors influencing reservoir leakage are qualitative and interrelated, machine learning algorithms can be adopted to study the intrinsic connections and regularity among these qualitative influencing factors and the seepage of karst reservoirs. This approach enables the development of mathematical models that transform qualitative variables into quantitative evaluations, facilitating the quantitative evaluation of seepage volume in karst reservoirs.In this study, an innovative model for predicting leakage in karst reservoirs were developed, based on Random Forest (RF), Artificial Neural Network (ANN), and Support Vector Machine (SVM). Comparative analysis demonstrated that all three approaches achieved satisfactory prediction results. Among them, the random forest model exhibited the best prediction performance: its simulated values closely matched the measured values during both training and validation phases, accurately capturing the dynamic changes in leakage volume. This model offers high prediction accuracy and strong stability, making it the preferred choice for the evaluation of reservoir leakage in karst areas. Four indicators, namely, rock mass permeability, fault development degree, seepage channel morphology, and hydrodynamic conditions, were selected as the main influencing factors of reservoir seepage. These findings provide new technical approaches and support for risk assessment of leakage in pumped storage power stations.

Study on vertical bearing behavior of pile foundations overlying karst cavities

YANG Bai, XIAO Kang, PAN Zongyuan, QIN Chao, HUANG Xiang

2026, 45(3): 603-614. doi: 10.11932/karst2025y027

Abstract FullText HTML PDF(2576KB) icon

3

0

Abstract:
Karst regions are extensively distributed acorss China, where rock-socketed pile foundations are widely adopted in the construction of bridges, buildings, and other infrastructure owing to their high bearing capacity, minimal settlement, and favorable deformation control performance. However, when a karst cave exists beneath the pile tip, its load transfer mechanism, failure mode, and ultimate bearing capacity may deviate markedly from those of conventional pile foundations. Consequently, the stability of the cave roof and the interactions among the pile, overlying soil, bedrock, and underlying cavity constitute critical determinants of foundation safety. To investigate the vertical bearing behavior of pile foundations with underlying karst caves, this study integrates physical model tests with three-dimensional numerical simulations to examine the effects of cave roof thickness and rock-socketed ratio on failure characteristics, load–displacement response, axial force transfer, lateral resistance distribution, and ultimate bearing capacity.The physical model tests were designed based on the geological conditions of a karst site in Jiangnan District, Nanning, Guangxi, where the overlying layer is mainly red clay and the bedrock is moderately weathered dolomite. With a geometric similarity ratio of 30 and a volumetric weight similarity ratio of 1.2, nine model piles were fabricated in total. In the cave-roof-thickness test series, five piles shared a constant rock-socketed ratio of 0.5 while the cave roof thicknesses varied from 1D to 5D. In the rock-socketed-ratio test series, five piles maintained a uniform cave roof thickness of 3D with rock-socketed ratios ranging from 0.2 to 0.6; one specimen was shared between the two test series. The slow maintained load procedure was adopted to obtain the load–displacement curves at the pile top and the internal force responses along the pile shaft. Concurrently, a corresponding full-scale (1∶1) numerical model was developed using PLAXIS 3D. The Mohr–Coulomb constitutive model was assigned to both soil and rock masses, whereas the pile body was simulated as a linear elastic material. Interface elements were embedded at the pile shaft and pile tip to characterize the pile–soil and pile–rock contact behaviors. The interface strength reduction factors were determined via direct shear tests, yielding values of 0.5 for the pile–soil interface and 0.8 for the pile-rock interface.The results indicate that pile foundation with underlying karst caves exhibit a composite failure mode: columnar shear failure occurs along the pile shaft within the surrounding soil and rock mass, accompanied by trumpet-shaped punching failure of the cave roof beneath the pile tip. The displacement contour plots derived from numerical simulations are generally consistent with the failure morphology observed in the physical model tests, demonstrating that the established numerical model can effectively reproduce the failure mechanism of the pile-rock–cavity system. As the cave roof thickness increases, both the height and the scope of influence of the failure zone expand; conversely, these dimensions diminish with increasing rock-socketed ratio. These findings reveal that the roof thickness and the socketing condition collectively govern the diffusion of pile-tip loads within the rock mass and the subsequent punching failure process of the cave roof.The load-displacement curves of pile foundations with underlying karst caves exhibit distinct abrupt-failure characteristics. During the initial loading stage, the pile-head displacement increases approximately linearly with the applied load.Upon reaching the ultimate bearing capacity, any further load increment triggers a sharp surge in displacement, indicating sudden punching failure or instability of the cave roof. In contrast, piles without underlying cavities display a gradual load-displacement response without any pronounced inflection point. This discrepancy reflects the competitive relationship between the shaft resistance mobilized along the pile-soil/rock interface and the punching resistance offered by the cave roof. When the cave roof is relatively thin, the stress-affected zone beneath the pile tip readily penetrates the roof, rendering punching failure the dominant failure mechanism. Conversely, when the roof thickness is sufficient, the pile-tip load diffuses more effectively within the rock mass, leading to stable bearing behavior of the pile foundation.The axial force and shaft resistance profiles further elucidate the load transfer mechanism of pile foundation with underlying karst caves.At various load levels, the axial force along the pile shaft decreases progressively with depth, exhibiting a notably higher attenuation rate within the rock mass than within the overlying soil layer. Piles with underlying cavities predominantly behave as end-bearing piles. The distribution of shaft resistance displays a characteristic L-shaped pattern. Owing to the relatively low cohesion and internal friction angle of the red clay, the shaft resistance in the soil layer remains modest and mobilizes rapidly with increasing vertical load. In contrast, the rock mass possesses superior strength parameters; once relative shear displacement occurs at the pile-rock interface, the rock shaft resistance increases substantially, forming a distinct inflection point near the soil-rock interface. Both the rock shaft resistance and the pile tip resistance escalate continuously with rising vertical load, indicating that the bearing capacity of such piles is collectively governed by the rock shaft resistance, the pile tip resistance, and the stability of the underlying cave roof.The numerical results show good agreement with the physical model tests, with relative errors in ultimate bearing capacity ranging from −22.2% to 15.8%. Upon validating the reliability of the numerical model, parametric analyses were further extended to cover cave roof thicknesses from 1D to 10D and rock-socketed ratios from 0.2 to 1.0. The results indicate that the ultimate bearing capacity increases with cave roof thickness, exhibiting an approximately linear growth trend within the 1D–6D range; beyond 6D, the increase gradually plateaus. Conversely, the ultimate bearing capacity decreases with increasing rock-socketed ratio, showing an approximately linear decline as the ratio rises from 0.2 to 0.6; when the ratio exceeds 0.6, the rate of decrease diminishes markedly. These findings provide valuable references for evaluating the bearing behavior of rock-socketed piles in karst regions, determining rational design parameters, and implementing effective safety control measures.

Applied research on integrated "air-ground-cave" surveying of the world's supercave, the Miao Chamber Cavern：A case study of the Ziyun Miao Chamber

ZHOU Wenlong, SONG Xiaoqing, LUO Ji, LI Huaibing, YANG Jiafang, ZHAO Shiqi, MO Guifen

2026, 45(3): 615-627, 642. doi: 10.11932/karst2026y001

Abstract FullText HTML PDF(12798KB) icon

323

219

Abstract:
As a world-class rare super cavern, the Ziyun Miao Chamber presents extraordinary challenges for traditional surveying techniques due to its vast spatial dimensions, intricate multi-level cave structures, and complex geological formations. To address the limitations of early exploration efforts, this study develops and implements an innovative integrated "air-ground-cave" surveying technology system. By synergizing advanced methodologies, including drone-based aerial photogrammetry, terrestrial laser scanning (TLS), and simultaneous localization and mapping (SLAM) technology, a comprehensive spatial data acquisition and multi-source data fusion framework was established. This system enabled, for the first time, high-precision, full-coverage 3D surveying and seamless registration of both surface karst landforms and multi-tiered underground cave networks within the Miao Chamber.The research outcomes, derived from technical summarization and extensive data analysis, yield the following key findings: (1) Data correction and spatial referencing: building upon the foundational 2014 Sino-British joint scientific survey, this study employed SLAM technology to rectify point cloud deviations from prior measurements of the Miao Chamber. By integrating real-time kinematic (RTK) positioning with the Qianxun Continuously Operating Reference Station (CORS) service, an absolute spatial reference framework was established for the entire cave system. This advancement not only enhances measurement accuracy but also provides reliable geo-spatial data for re-examining the developmental history and evolutionary mechanisms of the Miao Chamber.(2) Spatial visualization and geomorphological analysis: Utilizing the fused and registered 3D spatial datasets, point cloud slicing analysis tools were applied to generate intuitive visualizations of the positional relationships between surface karst features and multi-level subterranean caves. These analyses reveal critical spatial correlations and composite geomorphological patterns, offering new insights into the formation processes of this world-class super cavern. The findings highlight the co-evolutionary dynamics between the Miao Chamber and the overlying karst landscape, driven by long-term hydrological and tectonic interactions.(3) Technological advancements in cave surveying: the integrated "air-ground-cave" surveying approach effectively overcomes the constraints of traditional techniques in exploring giant cave systems. By combining aerial, terrestrial, and underground data acquisition methods, this system achieves highly accurate 3D reconstructions of surface topography and subterranean structures. The methodology not only enhances surveying accuarcy but also sets a benchmark for future explorations of similar large-scale karst environments.(4) Structural controls and hydrological pathways: geological investigations demonstrate that the development of the Miao Chamber is predominantly governed by four sets of hierarchically organized joints with varying orientations. Current structural assessments indicate that the cave system remains in a relatively stable state. Paleo-hydrological evidence suggests that ancient surface flows were divided into at least two distinct pathways (Zhongdong and Xiaochuandong). Modern surface runoff continues to recharge the underground system through karst conduits and fractures, sustaining the dynamic interplay between surface and subsurface processes.(5) Paradigm for future cave exploration: the successful application of the integrated "air-ground-cave" surveying technology system in the Ziyun Miao Chamber establishes a pioneering framework for 3D spatial measurements of giant subterranean spaces and complex geological structures. This methodology not only enhances scientific understanding of karst systems but also provides a replicable model for global cave exploration, particularly in regions with similar geomorphological and hydrological conditions.In conclusion, this study advances the field of karst geomorphology by delivering a high-precision, multi-dimensional surveying solution for super caves. The findings contribute to a deeper comprehension of cave formation mechanisms, structural stability, and hydrological connectivity, while the developed techniques offer transformative potential for future research and exploration in subterranean environments worldwide.

Statistical analysis of karst spatial distribution characteristics in Guangzhou

WANG Shunsheng

2026, 45(3): 628-642. doi: 10.11932/karst2026y017

Abstract FullText HTML PDF(6961KB) icon

3

0

Abstract:
The Guanghua Basin, situated in the central-northern part of Guangzhou,China, serves as the primary study area due to its characteristic and highly concentrated karst geomorphology. The region features a complex topography that includes alluvial plains, gentle hilly terraces, and karst peak-cluster depressions, with elevations generally ranging from 5 to 150 m. Geologically, the area is characterized by a multi-layered stratigraphy. The foundational karst-prone strata primarily consist of Carboniferous and Permian carbonate rocks, notably the Hutian Formation (composed of thick-layered limestone and dolomitic limestone) and the Qixia Formation. These rock masses are extensively covered by Quaternary alluvial-diluvial deposits and Paleogene red beds. Subject to a subtropical monsoon climate with abundant, yet unevenly distributed annual rainfall ranging from 1,421 to 2,245 mm, the region maintains a shallow groundwater table. The hydrogeological network, strongly influenced by regional fault systems, creates a highly active hydrodynamic environment that continuously drives intense karstification processes along structural axes.Driven by rapid urbanization and land resource constraints in the Guangdong-Hong Kong-Macao Greater Bay Area, Guangzhou’s urban development is increasingly focused on the utilization of deep underground space. This involves the extensive construction of subway networks, comprehensive utility tunnels, and underground expressways. However, the complex karst geological conditions pose significant challenges to engineering safety. During underground tunnel excavations, active karst systems frequently trigger sudden water and mud inrushes or severe collapses that compromise structural integrity. For surface structures, foundations penetrating karst zones often encounter insufficient bearing capacity, uneven settlement, or pile suspension. Although various theoretical risk assessment models exist, their practical application is often severely limited by the lack of precise, localized spatial and statistical parameters regarding cave distribution and hydrogeological factors. Therefore, this study aims to systematically investigate the spatial distribution, morphological characteristics, filling status, and hydrogeological controlling factors of karst in this basin. The primary objective is to bridge this data gap and provide reliable, quantitative technical references for hazard identification, engineering site selection, and foundation treatment optimization in karst-prone urban environments.This study is based on a comprehensive and robust dataset compiled from 17 detailed geotechnical engineering investigation reports, covering 15 survey lines across the Baiyun, Huadu, and Tianhe districts. A total of 3,047 sets of borehole data were extracted and systematically categorized. The dataset includes 2,706 records of bedrock burial depth, 2,468 records of stable groundwater depth, and 2,291 records detailing karst cave roof depth, bottom depth, and vertical height. Statistical analyses were conducted to evaluate the spatial variability of both limestone and sandstone stratigraphy. Furthermore, the regional karst development intensity was quantitatively evaluated using the line karst rate and the borehole cave encounter rate. Finally, Pearson correlation analysis was employed to reveal quantitative relationships between spatial cave metrics (e.g., cave height, roof thickness) and various hydrogeological indicators, including groundwater depth, pH, aggressiveness of free CO₂, and total mineralization.Statistical analysis reveals that the region is predominantly characterized by shallow-covered karst. The burial depth of limestone bedrock ranges from 9.3 to 69.3 m, with an average depth of 29.83 m, and approximately 56% of the samples have a burial depth less than 30 m. Similarly, sandstone burial depths average 28.75 m. Morphologically, the karst caves predominantly exhibit a “high-cave–thin-roof” structure. In limestone strata, cave heights reach a maximum of 42.4 m with an average height of 3.28 m, presenting a high coefficient of variation of 109%. Concurrently, the average roof thickness is remarkably thin at 2.76 m, with over 80% of roof thicknesses measuring less than 6 m. Regarding internal conditions, the caves demonstrate poor structural stability; 54.1% are semi-filled, and 32.2% are completely unfilled. The predominant filling material is soft, plastic clay, accounting for 59.0%, which is highly susceptible to erosion under engineering disturbances. The overall karst development intensity is evaluated as high, featuring an average line karst rate of 24.58%, with 48% of caves classified as strongly developed according to National Standards, and a high average borehole cave encounter rate of 54.52%. The groundwater environment exhibits an average stable depth of 4.13 m and an average pH of 7.21. Correlation analysis indicates a significant positive correlation between cave height and groundwater depth (correlation coefficient 0.1335, p=0.0001), alongside a weak negative correlation with groundwater pH and total mineralization, suggesting that deeper water tables and slightly acidic conditions significantly exacerbate vertical karstification.The concealed karst in the Guanghua Basin exhibits a distinctive pattern characterized by shallow coverage, high caves with dangerously thin roofs, semi-filled or unfilled states primarily composed of soft clay, reflecting a moderate to strong degree of development. The widespread “high-cave-thin-roof” structure poses severe risks of roof collapse and structural instability for both shallow foundations and deep underground excavations. Active groundwater hydrodynamics further exacerbate these risks by promoting continuous rock dissolution. Therefore, it is strongly recommended to conduct differential engineering site investigations and implement highly targeted foundation treatments. For high-risk caves with roof thicknesses under 3 m or heights exceeding 10 m, priority should be given to grouting reinforcement or the avoidance of pile foundation. Additionally, dynamic groundwater monitoring and advanced drainage designs must be integrated to ensure engineering safety and mitigate geological disasters in water-sensitive karst zones.

Disaster-causing process of a steep columnar karst mountain collapse induced by mining activities: A case study of the Yiziyan collapse in Guizhou

DENG Zhinan, YANG Liang, JIANG Xingyuan, WU Di, YANG Yi, WANG Changhui

2026, 45(3): 643-654. doi: 10.11932/karst2026y020

Abstract FullText HTML PDF(5077KB) icon

194

23

Abstract:
To reveal the collapse disaster mechanism of steep columnar karst mountains under mining disturbance, this study taking a typical"hard-top, soft-bottom; steep-top, gentle-bottom" collapse body at Yiziyan, Jinsha County, Guizhou Province as the research object, systematically analyzes the entire deformation and failure process of the slope under multiple mining activities. It is based on field geological surveys and laboratory rock mechanics tests, conducting bottom friction physical simulation experiments, combined with PhotoInfor digital imaging system for quantitative monitoring of slope deformation and displacement. The results indicate that slope deformation caused by mining disturbance presents a four evolution stages pattern. During the initial unloading and joint development phase, unloading rebound occurs in the roof above the goaf, forming joint zones and collapse bands. The displacement areas show a crescent-shaped distribution and has a relatively minimal impact on the slope surface. At the overburden overall subsidence and slope surface response stage, the overall collapse of roof causes significant subsidence of the surface slope, the deformation of overburden continues to intensify and extends to the slope crest area. In the fissure expansion and stress redistribution stage, mining induces secondary roof unloading rebound, causing the joint zones and collapse bands to develop further upward, promoting furthrer upward development of joint zones and collapse bands, which reactivates and widens existing dissolution fissures at the slope crest. During shear failure and overturning collapse stage, the compressive shear stresses at the bottom of the rock mass are highly concentrated to form a continuous shear failure plane. The slope surface experiences intense compressive shear failure, with rear-edge tensile fractures progressively worsening. Ultimately, the steep inclined columnar unstable rock masses collapses toward the open face. The evolution of the shear strain field indicates that the maximum shear strain during mining process gradually increases from an initial value of 1.2 to 1.7. The stress concentration zone expands from within the collapse zone to the vicinity of the slope top fractures and the vertically overlapping regions of multiple mining voids. This study has confirmed that progressive deformation and stress redistribution of the overburden caused by multiple mining activities are the key mechanisms controlling the instability of columnar rock masses. The bottom friction combined with digital image measurement technology can effectively reveal the four-stage evolution pattern: "initial unloading-overburden overall subsidence-fissure expansion-shear overturning" under mining influence, clarifying the deformation characteristics and mechanical response mechanisms at each stage.

Current Issue
2026, Volume 45, Issue 3

Archive

Submission Guide

Current Issue 2026, Volume 45, Issue 3

Archive

Submission Guide

Current Issue
2026, Volume 45, Issue 3