Karst Watershed Research Articles

Hydrological processes in multi-layered aquifers of a karst watershed with coal mining activity: Insights from hydrochemistry and isotopes

Study regionThe Laochang Karst watershed (LCKW) is located in eastern Yunnan Province, southwestern China. It is the representative karst area affected by coal-mining activities in southwestern China. Study focusIdentifying hydrological processes of multi-layered aquifers in karst watersheds is challenging due to complex natural and anthropogenic processes. This study attempts to clarify the hydrological conceptual model of the LCKW using hydrochemistry and D, O, Sr, S, and C isotopes. New hydrological insights for the regionSurface water and multi-layered groundwater have the hydrochemical types of SO4-Ca·Mg, HCO3·SO4-Ca, and HCO3-Ca. Meteoric water and condensate were the major recharge sources. The main processes dominating hydrochemical compositions consist of sulfide oxidative dissolution, carbonate dissolution, positive cation exchange, and agricultural activities. Elevated SO42− concentration in the mine water, river water and shallow coalbed water mainly originated from the oxidation of pyrite in the coal-bearing strata of the Longtan Formation. whereas the deeper layers and groundwater away from the mines were hardly contaminated by SO42− due to the presence of aquiclude. HCO3− concentrations of surface water and multi-layered groundwater were mainly derived from carbonate dissolution and soil CO2, and mine water was also influenced by atmospheric CO2. Positive cation exchange contributed to increasing Na+ concentration. Agricultural activities contributed NO3−, Cl−, and K+ ions in aquifers, especially near large karst fallout caves. A hydrological model of multi-layered aquifers in the LCKW was built based on the above results. These findings will provide valuable guidance for understanding the hydrological processes of complex karst watersheds worldwide.

Complex Hydrology and Variability of Nitrogen Sources in a Karst Watershed

Complex Hydrology and Variability of Nitrogen Sources in a Karst Watershed

Quantitatively Distinguishing the Factors Driving Runoff and Sediment Yield Variations in Karst Watersheds

AbstractDue to the coupled or interconnected relationships among frequent climate extremes, unique geological conditions, discontinuous soil distribution, rugged geomorphology, and highly heterogeneous landscapes in different karst watersheds, few studies were conducted to decouple the relative magnitudes of the climate, lithology, soil, topography, and landscape on soil erosion in karst regions. The objective of this study was to quantify the relative importance of these influencing factors on runoff and sediment yield (SY) in 40 typical karst watersheds in southwest China. To address this issue, the Pearson correlation and random forest were first to select the dominant factors influencing runoff and SY. Subsequently, the partial least squares‐structural equation model (PLS‐SEM) was used to decouple the complex relationships among runoff, SY and their potential influencing factors. Results showed that climate, lithology, soil, topography and landscape could explain 79% of the runoff variation, and only climate factors have significant impact on runoff for heterogeneous karst watersheds (P < 0.01, path coefficient (β) = 0.589). The explanation of five factors to SY variability is 59%, and the landscape has the greatest impact on SY (P < 0.01, β = −0.458). Different from runoff, climatic factors have no significant influence on SY. By elucidating a complex coupled relationship framework, this study can provide a scientific basis for the formulation of soil and water loss program, and the optimization of land resources and ecological environment sustainable development in karst watersheds.

Hydrologic Model Prediction Improvement in Karst Watersheds through Available Reservoir Capacity of Karst

This study aimed to enhance flood forecasting accuracy in the Liangfeng River basin, a small karst watershed in Southern China, by incorporating the Available Reservoir Capacity of Karst (ARCK) into the HEC-HMS model. This region is often threatened by floods during the rainy season, so an accurate flood forecast can help decision-makers better manage rivers. As a crucial influencing factor on karstic runoff, ARCK is often overlooked in hydrological models. The seasonal and volatile nature of ARCK makes the direct computation of its specific values challenging. In this study, a virtual reservoir for each sub-basin (total of 17) was introduced into the model to simulate the storage and release of ARCK-induced runoff phenomena. Simulations via the enhanced model for rainfall events with significant fluctuations in water levels during 2021–2022 revealed that the Nash–Sutcliffe efficiency coefficient (NSE) of the average simulation accuracy was improved by more than 34%. Normally, rainfalls (even heavy precipitations) during the dry season either do not generate runoff or cause negligible fluctuations in flow rates due to long intervals. Conversely, relatively frequent rainfall events (even light ones) during the wet season result in substantial runoff. Based on this observation, three distinct types of karstic reservoirs with different retaining/releasing capacities were defined, reflecting variations in both the frequency and volume of runoff during both seasons. As a real-time environmental variable, ARCK exhibits higher and lower values during the dry and rainy seasons, respectively, and we can better avoid the risk of flooding according to its special effects.

Interpretation of suspended sediment concentration‐runoff hysteresis loops in two small karst watersheds

AbstractQuantifying suspended sediment concentration‐runoff (SSC‐Q) hysteresis can help to understand the source and transport mechanism of sediment. Due to the unique lithology, shallow soil, high infiltration capacity, complex geomorphology and heterogeneous landscape of karst watersheds in southwestern China, SSC‐Q hysteresis patterns remain unclear in this region. In the present study, the hysteresis index (HI) was used to quantify the SSC‐Q hysteresis pattern. The effects of rainfall and runoff on the HI were also evaluated from 2014 to 2019 in the Mahuangtian (2.26 km2) and Muzhaihe (0.09 km2) karst watersheds. The results showed that the clockwise loops accounted for more than 60% of the total hysteresis patterns in the two karst watersheds. In contrast, the counter‐clockwise and figure‐eight loops occupied no more than 30% and 10% of the total hysteresis patterns, respectively. These differences in the hysteresis patterns probably resulted from the spatial distribution of sediment sources and hydrological connectivity. The HI was significantly affected by the maximum discharge in the Mahuangtian watershed (P < 0.05) and by the flashiness index in the Muzhaihe watershed. This may be attributed to their different 3D hydrogeological structure, lithology and watershed size. This study can facilitate a better understanding of suspended sediment dynamics and provide important guidance for accurately implementing soil and water conservation strategies in karst areas.

Spatial-temporal changes of landscape ecological risk in the Liuchong river basin from the perspective of production-life-ecological space

Promoting the construction of ecological civilization and sustainable development in karst mountainous areas by analyzing the spatial and temporal changes of landscape ecological risks is critical in karst mountainous watersheds. In this study, the land use transfer matrix, landscape ecological risk evaluation model, ecological contribution rate of land use change, and spatial autocorrelation analysis were combined to quantitatively analyze the land use and landscape ecological risk of a typical karst watershed, Liuchong River Basin, over the past 20 years. The results revealed that: 1) From 2000 to 2020, the functional classification of land use in the Liuchong River Basin was dominated by the woodland ecological space, and the most significant shifting characteristics were the increase in the area of watershed ecological space and industrial production space and the decrease in woodland ecological space, with shifts in the middle reaches of the Liuchong River being the most drastic; 2) Generally, the change of the regional landscape pattern was related to the transformation of the land use function type of “production-life-ecological space,” and the spatial aggregation of ecological risk level showed a gradual weakening trend. 3) The conversion of the watershed ecological space to the grassland ecological and agricultural production spaces, the conversion of urban living space to the agricultural production space, and the conversion of the rural living space to the agricultural production space were the dominant factors affecting ecological improvement, whereas the conversion of the woodland ecological space to the grassland ecological space, the woodland ecological space to the agricultural production space, and the grassland ecological space to the agricultural production space contributed to ecological degradation. The study findings can be used as a reference for the coordinated development of “production-life-ecological space” in karst watersheds and provide a scientific basis for ecological environmental protection and sustainable utilization.

Carbon-fixing bacteria in diverse groundwaters of karst area: Distribution patterns, ecological interactions, and driving factors

The biological carbon pump in karst areas is of great significance for maintaining the effectiveness of karst carbon sinks. However, the spatial distribution and carbon-fixing potential of microorganisms in different aquifers within karst areas remain poorly understood. In this study, the distribution patterns, ecological roles, and environmental drivers of microbiota associated with CO2 fixation were investigated in karst groundwater (KW), porous groundwater (PW), fractured groundwater (FW), and surface water (SW) within a typical karst watershed, located in Guilin, southwest China. KW, PW, and FW displayed the similar community structure and indicative carbon-fixing bacteria composition, which were dominated by chemoautotrophic bacteria compared to SW. Higher abundances of indicative carbon-fixing bacteria and carbon-fixing genes, as well as richer proportions of microbial-derived DOC, indicated the more significant microbial carbon-fixing potential in KW and PW. At the profile of KW, a carbon-fixing hotspot was discovered at the depths of 0–50 m. Correlation analysis between carbon-fixing bacteria and DOC revealed that the chemoautotrophic process driven by nitrogen and sulfur oxidation predominated the microbial carbon fixation in groundwater. Co-occurrence network analysis demonstrated that carbon-fixing bacteria exhibited cooperation with other bacterial taxa in KW, while competition was the dominant interaction in PW. Moreover, carbon-fixing bacteria was found to lead bacterial assembly more deterministic in KW. The analysis of environmental factors and microbial diversity illustrated that inorganic carbon and redox state drove community variations across groundwaters. Structural equation model (SEM) further confirmed that ORP was the primary factor influencing the carbon fixation potential. This study provides a new insight into biological carbon fixation in karst aquatic systems, which holds significance in the accurate assessment of karst carbon sinks.

Complex hydrology and variability of nitrogen sources in a karst watershed.

Streams draining karst areas with rapid groundwater transit times may respond relatively quickly to nitrogen reduction strategies, but the complex hydrologic network of interconnected sinkholes and springs is challenging for determining the placement and effectiveness of management practices. This study aims to inform nitrogen reduction strategies in a representative agricultural karst setting of the Chesapeake Bay watershed (Fishing Creek watershed, Pennsylvania) with known elevated nitrate contamination and a previous documented groundwater residence time of less than a decade. During baseflow conditions, streamflow did not increase with drainage area. Headwaters and the main stem lost substantial flow to sinkholes until eventually discharging along large springs downstream. Seasonal hydrologic conditions shift the flow and nitrogen load spatially among losing and gaining stream sections. A compilation of nitrogen source inputs with the geochemistry and the pattern of enrichment of δ15N and δ18O suggest that the nitrogen in streams and springs during baseflow represents a mixture of manure, fertilizer, and wastewater sources with low potential for denitrification. The pH and calcite saturation index increased along generalized flow paths from headwaters to springs and indicate shorter groundwater residence times in baseflow during the spring versus summer. Given the substantial investment in management practices, fixed monitoring sites could incorporate synoptic water sampling to properly monitor long-term progress and help inform management actions in karst watersheds. Although karst watersheds have the potential to respond to nitrogen reduction strategies due to shorter groundwater residence times, high nitrogen inputs, effectiveness of conservation practices, and release of legacy nutrients within the karst cavities could confound progress of water quality goals.

Interpreting and modelling the daily extreme sediment events in karst mountain watersheds

Increasingly frequent extreme rainfall as a result of climate change is strongly damaging the global soil and water environment. However, few studies have focused on daily extreme sediment events (DESE) in heterogeneous karst watersheds based on long-term in-situ observations. This study quantitatively assessed the time effect of DESE on rainfall response, decoupled the impact of environmental factors on DESE by using structural equation modelling, and finally explored the modelling scheme of DESE based on the hybrid model. The results showed that DESE had the highest frequency of occurrence in May–July, with dispersed distribution in the value domain. Rainfall with a time lag of 1 day and a time accumulation of 2 or 3 days was an important contribution to DESE (P < 0.01, R = 0.47–0.68). Combined effects of environmental factors explained 53.6 %–64.1 % of the variation in DESE. Runoff and vegetation exerted the strongest direct and indirect effects on DESE, respectively (β = 0.66/−0.727). Vegetation was the dominant driver of DESE in Dabanghe and Yejihe (β = −0.725/−0.758), while the dominant driver in Tongzhihe was climate (β = 0.743). In the future, the risk of extreme sediments should be prevented and resolved through the comprehensive regulation of multiple paths, such as runoff and vegetation. Hybrid models significantly improved the modelling performance of machine learning models. Generalized additive model-Extreme gradient boost had the best performance, while Partial least squares regression-Extreme gradient boost was the most valuable when considering performance and input data cost. Two methods can be used as recommended solutions for DESE modelling. This study provides new and in-depth insights into DESE in karst watersheds and helps the region develop forward-looking soil and water management models to cope with future extreme erosion hazards.

Disentangling influences of driving forces on intra-annual variability in sediment discharge in karst watersheds

The intra-annual variability in sediment discharge was considerably influenced by the climate variability and vegetation dynamics. Because of the coupled or relationships between climatic and vegetation variables, it is still challenging to decouple the direct and indirect effects of climate variability and vegetation dynamics on hydrological and sediment transport processes. The purpose of this study is to decouple influences of individual driving force on intra-annual distribution of sediment discharge during 2003–2017 using the partial least squares structural equation model (PLS-SEM) method in four typical karst watersheds of Southwest China. The coefficient of variation (Cv), Completely regulation coefficient (Cr), Lorenz asymmetry coefficient and Gini coefficient were used to represent the intra-annual sediment discharge variability. Results showed that the monthly sediment discharge (190 % < Cv < 353 %) exhibited greater variability than its potential affecting factors (18 % < Cv < 101 %). From the PLS-SEM analysis, the water discharge, climate, and vegetation together explain 57 %–75 %, 64 %–79 %, and 53 %–80 % of the total variance in Cv, Cr, and Gini coefficient, respectively. Specifically, water discharge exerts the largest influence on sediment discharge variability (0.65 ≤ direct effect ≤0.97, P < 0.05), while vegetation dynamic mainly indirectly affects sediment discharge variability (−0.88 ≤ indirect effect ≤ −0.01) through influencing water discharge. The climate factors also principally indirectly affect the sediment discharge variability (−0.47 ≤ indirect effect ≤0.19) by affecting water discharge and vegetation. The PLS-SEM can effectively reveal the driving force and influencing mechanism of intra-annual sediment discharge changes, and provide an important reference for regional soil and water resources management in karst watersheds. Future studies can decouple the influences of the extreme climate, unique lithology, discontinuous soil, heterogeneous landscape, and special geomorphology on spatial variability in sediment discharge across different karst watersheds.

Extreme Learning Machine Predicts High-Frequency Stream flow and Nitrate-N Concentrations in a Karst Agricultural Watershed

Highlights A novel high-frequency dataset of streamflow, nitrate-N concentration, and soil moisture at depths throughout and below the root zone highlighted tight connectivity between soil moisture dynamics and nitrate-N concentrations spanning event to seasonal timescales. A TELM model successfully predicted both flow rate and nitrate exports from these complex systems because the time lag associated with the soil conditions was represented in model training. The inclusion of soil moisture and temperature data below the effective root zone was important to accurately capture the storm event hysteresis dynamics observed in the exported nitrate signals. Abstract. Efforts to reduce nitrogen contributions from karst agroecosystems have had variable success, in part due to an incomplete understanding of nitrogen source, fate, and transport dynamics in karst watersheds. Recent advancements in environmental sensors and data-driven artificial intelligence models may be useful in improving our understanding of system behavior and the linkages between soil hydrologic processes and karst nitrate loading dynamics. We collected 35 months of high-resolution streamflow, nitrate-N concentration, soil moisture and temperature (from 10-100 cm depths), and meteorological data in a karst agricultural watershed in the Inner-Bluegrass region of Central Kentucky. Two-layer extreme learning machine (TELM) models were developed to predict nitrate-N concentrations and flow rates as a function of meteorological and soil parameter inputs. Results suggest tight linkages between soil moisture gradients at different depths and nitrate-N concentrations at the watershed outlet. TELM modeling results supported visual observations from the high-frequency data and suggest that inclusion of both soil moisture and temperature parameters at all soil depths improved predictions of both flow rate and nitrate-N concentration (with optimal NSE values of 0.93 and 0.94, respectively, when all inputs were considered). Hysteresis analysis suggested that inclusion of the deepest soil layer (100 cm) was necessary to predict hysteresis observed during storm events. The findings of the study highlight the importance of variable activation of matrix waters in preferential flows throughout events and seasons and its subsequent impacts on nitrate-N concentrations. Results suggest that management models should incorporate vertical variability in soil hydrology to accurately characterize nitrate source and transport dynamics. Further, the results of hysteresis analysis underscore the importance of inclusion of hysteresis indices, in addition to typical model evaluation statistics, to ensure accurate representation of nutrient flow pathways. Keywords: Extreme learning machine, Karst agroecosystem, Nitrate, Water resources.

River Sediments Downstream of Villages in a Karstic Watershed Exhibited Increased Numbers and Higher Diversity of Nontuberculous Mycobacteria

The impact of residential villages on the nontuberculous mycobacteria (NTM) in streams flowing through them has not been studied in detail. Water and sediments of streams are highly susceptible to anthropogenic inputs such as surface water flows. This study investigated the impact of seven residential villages in a karst watershed on the prevalence and species spectrum of NTM in water and sediments. Higher NTM species diversity (i.e., 19 out of 28 detected) was recorded downstream of the villages and wastewater treatment plants (WWTPs) compared to sampling sites upstream (i.e., 5). Significantly, higher Zn and lower silicon concentrations were detected in sediments inside the village and downstream of the WWTP’s effluents. Higher phosphorus concentration in sediment was downstream of WWTPs compared to other sampling sites. The effluent from the WWTPs had a substantial impact on water quality parameters with significant increases in total phosphorus, anions (Cl–and N-NH3–), and cations (Na+ and K+). The results provide insights into NTM numbers and species diversity distribution in a karst watershed and the impact of urban areas. Although in this report the focus is on the NTM, it is likely that other water and sediment microbes will be influenced as well.Graphical

Comparative evaluation of daily streamflow prediction by ANN and SWAT models in two karst watersheds in central south Texas

Abstract This work compares the accuracy of streamflow estimated by a data-driven artificial neural network (ANN) and the physically based soil and water assessment tool (SWAT). The models were applied in two small watersheds, one highly urbanized and the other primarily covered with evergreen forest and shrubs, in the San Antonio region of central south Texas, where karst geologic features are prevalent. Both models predicted daily streamflow in the urbanized watershed very well, with the ANN and SWAT having the Nash–Sutcliffe coefficient of efficiency (NSE) values of 0.76 and 0.72 in the validation period, respectively. However, both models predicted streamflow poorly in the nonurban watershed. The NSE values of the ANNs significantly improved when a time series autoregressive model structure using historical streamflow data was implemented in the nonurban watershed. The SWAT model only achieved trivial performance improvement after using the SWAT-CUP SUFI-2 calibration procedure. This result suggests that an ANN model may be more suitable for short-term streamflow forecasting in watersheds heavily affected by karst features where the complex processes of rapid groundwater recharge and discharge strongly influence surface water flow.

Evaluating climate change impacts in a heavily irrigated karst watershed using a coupled surface and groundwater model

Study regionLower Apalachicola-Chattahoochee-Flint (ACF) River Basin of southeastern United States Study focusThe threats of climate change on the surface- and groundwater resources of the lower ACF River Basin of southeastern U.S. is an important concern for the long-term ecological as well as agricultural sustainability. This study developed a coupled SWAT-MODFLOW for the study region and evaluated the impacts of climate change projected under RCP4.5 and RCP8.5 emissions scenarios. New hydrological insights into the regionEvaluation of simulated streamflow and groundwater levels showed that SWAT-MODFLOW can adequately replicate the hydrology of a karstic watershed such as that present in the study domain even without the incorporation of conduit flows/karst features. Comparison to baseline conditions indicates a shift in the monthly streamflow pattern in the region with a reduction from April to June and increases in the rest of the year under future climate. The region will also likely see an increase in low-flow as well as high-flow events, thus increasing streamflow variability in the region. More frequent low flow conditions in the future can lead to increased drying of ephemeral streams threatening the ecological sustainability of the region due to habitat loss. This, along with the projected reduction in groundwater levels can lead to increased stress on water resources of the region for irrigation, thus threatening the agriculture sustainability and increasing water conflict between the neighboring states.

Spatial and Temporal Evolutionary Characteristics of Landscape Pattern of a Typical Karst Watershed Based on GEE Platform

Spatial and Temporal Evolutionary Characteristics of Landscape Pattern of a Typical Karst Watershed Based on GEE Platform

Quantifying the Effect of Driving Factors on Spring Discharge in an Industrialized Karst Watershed

Quantifying the Effect of Driving Factors on Spring Discharge in an Industrialized Karst Watershed

Simulation of the impact of land use change on surface run-off in Karst Leang Lonrong Sub-Watershed

Estimation of surface run-off in karst watersheds has not been widely carried out and the estimation method is generally developed for non-karst watersheds. This study aims to analyze the accuracy of estimation of river discharge at the outlet of the Karst Biringere Sub-watershed using the modified Soil Conservation Service Curve Number (SCS-CN) method and analyzed the impact of land use change on river discharge at the outlet of the Karst Biringere Sub-watershed. Modification of the SCS-CN method in estimating direct flow at the outlet of the Beringere Sub-watershed is influenced by the similarity of rainfall and direct flow fluctuation of river discharge. The modified SCS-CN method provides satisfactory direct flow estimation when rainfall with observed direct flow of river discharge forms a linear relationship with a strong correlation. Changes in land use with reduced forest cover into mining areas result in increased run-off and reduced storage, the larger the forest land becomes mining areas, the greater the increase in run-off and reduction in storage. Reclamation of the former mining area to be a forest reduces run-off and increases water storage.

Evaluating the effects of water exchange between surface rivers and karst aquifers on surface flood simulations at different watershed scales

The complex subsurface conditions and spatial anisotropy of water-bearing media in karst basins result in much more variable runoff generation and confluence processes and water circulation than those in nonkarst basins. Therefore, accurate flood forecasting is more difficult in karst areas, and it is challenging to accurately quantify the water exchange relationship between surface rivers and karst aquifers. To provide key technical support for accurate flood forecasts in karst basins, this study improved a distributed physical hydrologic model, the Karst-Liuxihe model, by improving the runoff generation and confluence algorithms and the parameter optimization algorithm to simulate surface floods in three different karst watersheds in southern China. Additionally, in the model calculations, this study focused on the influence of water exchange on surface flood simulations in changing watersheds. The results showed that for the largest basin, Xijiang Basin (353,100 km2), considering or not considering the water exchange between the surface river and karst aquifer had little effect on the surface flood simulation results. However, this water exchange had a great effect on the surface simulations for the smallest basin, the Beijiang Basin (1,700 km2). For instance, the simulated average peak flow showed a 29 % difference, and the total amount of flooding showed a 25 % difference between dry years and normal years. Therefore, when forecasting surface floods in small karst areas, the water exchange relationship between surface rivers and karst aquifers must be considered. The results showed that the impact of groundwater on surface rivers can be disregarded when forecasting surface floods in the large Xijiang karst basin and can be treated directly as surface flood forecasting in nonkarst areas. The application of this strategy will greatly simplify the cost of hydrologic modelling in the Xijiang basin. By improving the Karst-Liuxihe model, the accuracy of its flood simulation and general applicability in karst basins were greatly enhanced. The improvement in the karst hydrological model in this study provides a good tool for flood simulation and forecasting in the karst regions of Southwest China.

Key influence of hydrogeological, geochemical, and geological structure factors on runoff characteristics in karst catchments

Lithology, dissolution degree, and other geological factors are essential factors affecting karst environments. However, the complex effects of multiple geologic factors on runoff characteristics have not been well quantified in karst catchments. Thus, 30 catchments and four runoff characteristics (runoff coefficient, runoff modulus, annual runoff extremes ratio, and annual runoff variation coefficient) were selected in this study to evaluate the influence of geological factors on runoff characteristics in karst catchments. The linear mixed effect model was used to quantitatively evaluate the contribution rate of each factor to the runoff characteristics, and the structural equation model was used to analyze the quantitative relationship between each factor and the runoff characteristics. The results show that among the three aspects of geological factors, hydrogeological factors have the greatest influence on runoff characteristic values. In turn, the density of sinkhole is the most influential factor among the hydrogeological factors. Geological structure factors had both direct and indirect effects on runoff characteristics (path coefficient is 0.27). Among them, the density of faults is the most influential factor. Geochemical factors affected runoff characteristics more by influencing hydrogeological factors (path coefficient −0.51) or geological structure factors (path coefficient −0.43). In conclusion, data from several watersheds directly confirm that in karst watersheds, these geological factors such as hydrogeology, geochemistry, and geological structure have a very fundamental influence on karst watershed runoff characteristics. Our results can help us further understand karst hydrological processes and improve the accuracy of karst hydrological modeling.

Temporal variations in rainfall, runoff and sediment yield in small karst watersheds

As one of the largest and most typical continuous karst distribution belts in the world, Southwest China is vulnerable to soil erosion and rocky desertification because of its unique geological conditions and heterogeneous landscapes. Understanding the rainfall-runoff-sediment transport relationships is critical to effectively controlling soil erosion in karst watersheds. This study aims to evaluate the sediment yield variability at different temporal scales (daily, monthly/seasonal, and annual) from 2014 to 2019 and to further identify the rainfall and runoff variables most associated with sediment yield in the selected Mahuangtian (2.26 km2) and Muzhaihe (0.09 km2) karst watersheds. Since rainfall and runoff variables were highly codependent, the partial least squares regression (PLSR) model was used to quantify the relative importance of rainfall and runoff variables to sediment yield. The variable importance of the projection (VIP) scores>1 was considered to be the most influential predictor of the PLSR model. Results showed that sediment yield was greatly influenced by rainfall and runoff, which can explain 70% and 85% of the variation of sediment yield in the Mahuangtian and Muzhaihe watersheds, respectively. For rainfall variables, three-day antecedent precipitation (VIP = 1.41) and five-day antecedent precipitation (VIP = 1.28) were the main first-order factors controlling sediment yield in the Mahuangtian watershed. In contrast, the rainfall amount (VIP = 1.35) was the dominant factor influencing sediment yield in the Muzhaihe watershed. For runoff variables, runoff amount, runoff coefficient, and maximum discharge were the primary variables controlling the sediment yield in these two karst watersheds. By partially eliminating the co-dependency of the variables, the PLSR approach can conveniently determine the dominant factors controlling sediment yield in heterogeneous karst watersheds.