Catena最新文献

{"title":"Comparison of alternative chemical density and physical methods for isolating soil organic matter fractions in high carbon soils","authors":"Hamond Motsi ,&nbsp;Catherine E. Clarke ,&nbsp;Ailsa G. Hardie ,&nbsp;Michele L. Francis ,&nbsp;Alastair J. Potts","doi":"10.1016/j.catena.2025.109788","DOIUrl":"10.1016/j.catena.2025.109788","url":null,"abstract":"<div><div>Soils beneath <em>Portulacaria afra</em> (spekboom) within the evergreen Albany Subtropical Thicket of South Africa contain unusually high soil organic carbon concentrations despite the biome's semi-arid climate. Fractionation of soil organic matter (SOM) into particulate organic matter (POM) and mineral-associated organic matter (MAOM) provide insights into the mechanisms of soil carbon (C) accumulation. Conventionally, sodium polytungstate (SPT) is used in density fractionation, but its high-cost limits sample numbers. Cheaper, reliable alternatives to SPT are therefore required to understand the C dynamics of spekboom and other high-C soils. Particle size (PS) sieving has been proposed as cheaper alternative to density fractionation, but its application in spekboom soils is unknown. This study compared SPT with PS fractionation and density fractionation using cheaper, highly soluble salts: sodium iodide (NaI), potassium iodide (KI), and calcium chloride (CaCl₂). The relationship between permanganate oxidizable carbon (POXC) and POM-C was also explored. High C (&gt;5 %) topsoils with a range of POM:MAOM ratios from the spekboom thicket, a pine forest, a wetland, and a grassland were evaluated. Results demonstrated that fractionation methods significantly affected (<em>p</em> &lt; 0.05) fractions in each soil. NaI showed similar performance to SPT in SOM fractionation followed by KI. The PS method used in this study compared poorly to SPT, underestimating POM-C (34–39 %) and overestimating MAOM-C (39–51 %) in high POM soils (spekboom and pine forest soils) with the opposite effect in lower POM soils (wetland and grassland). The CaCl₂ method was also not ideal due to salt entrainment. Some soils did not conform to the expected linear relationship between POM-C and POXC demonstrating that POXC method is not always suitable. Thus, NaI is proposed as a cheaper alternative to replace SPT in density fractionation of spekboom soils. Findings from this study have significant implications in appropriate method selection for SOM fractionation of high C soils.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109788"},"PeriodicalIF":5.7,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973835","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Development and application of a distributed hydrology and soil erosion model in a semi-arid catchment, China","authors":"Ang Lu ,&nbsp;Peng Tian ,&nbsp;Guangju Zhao ,&nbsp;Xingmin Mu ,&nbsp;Xiaojing Tian ,&nbsp;Chaojun Gu ,&nbsp;Lin Yang ,&nbsp;Junjian Fan","doi":"10.1016/j.catena.2026.109811","DOIUrl":"10.1016/j.catena.2026.109811","url":null,"abstract":"<div><div>Soil erosion models are efficient tools for quantifying regional soil erosion, estimating long-term soil erosion rates, and assessing the effects of land surface changes on soil erosion and sediment yield. These models offer significant advantages over costly and geographically limited field monitoring. These models are crucial for understanding hydrological and sediment dynamic on the Loess Plateau, which is highly vulnerable to soil erosion due to its complex topography, high erodible loess and frequent storms. To address the need of event-based simulations that can capture the impacts of widespread conservation practices, this study developed a novel distributed hydrology and soil erosion model. The model couples the Vertical Mixed Runoff Model (VMM) with the Morgan-Morgan-Finney (MMF) erosion model and integrates a specialized module to explicitly simulate the interception effects of terraces, which are a key soil and water conservation measure in the region. This integrated model simulates three key components at the flood-event scale, including runoff generation, soil erosion, and sediment transport. The model was calibrated and validated using data from nine flood events in the Xichuanhe catchment, a typical tributary of the Yanhe River on the Loess Plateau. The results demonstrate a high level of accuracy in runoff simulation, achieving Nash-Sutcliffe Efficiency (NSE) coefficients of 0.82 and 0.67 for the calibration and validation periods, respectively. Relative Peak Errors (RPE) were consistently below 23%, indicating a close match between simulated and observed hydrographs. For sediment simulation, the model effectively captured the overall dynamics with an average NSE of 0.80 and RPE between 2.3% and 18.7% during calibration periods, though with some discrepancies during validation periods. The model confirms the significant role of terraces in reducing runoff and sediment yield. On average, terraces could reduce total runoff volume by 12.1% and sediment yield by 17.2% during flood events. These findings demonstrated the model's effectiveness for hydrological and soil erosion simulation and its potential in evaluating soil and water conservation measures on the Loess Plateau. The model can offer a valuable tool for quantitatively assessing the effectiveness of soil and water conservation measures in this critical region and similar semi-arid environments.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109811"},"PeriodicalIF":5.7,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973918","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Trampling disturbance affects the stability of soil carbon pools in urban park green spaces by disrupting soil aggregates and altering the composition of organic carbon components","authors":"Wei Tang,&nbsp;Xiurong Wang,&nbsp;Rong Zou,&nbsp;Bingyang Shi,&nbsp;Jiaqi Chen","doi":"10.1016/j.catena.2025.109765","DOIUrl":"10.1016/j.catena.2025.109765","url":null,"abstract":"<div><div>While trampling disturbance affects soils, its impact on urban park soil carbon cycling remains unclear. This study elucidates the mechanisms by which varying trampling intensities influence urban park soil carbon pools, thereby providing theoretical foundations for soil restoration and park management. Four gradient sampling plots (0-3 m, 3-6 m, 6-9 m, 9-12 m) were established along park trail edges. Soil samples were collected using a five-point sampling method, with trampling intensities (heavy, moderate, light) classified based on changes in soil bulk density (compared to control) using the Jenks Natural Breaks method. Results indicated that heavy trampling reduced the &gt;2 mm aggregate proportion by 30.95 %, mean weight diameter by 11.62 %, and geometric mean diameter by 16.62 %, thereby decreasing soil stability. Heavy trampling decreased total soil organic carbon by 34.45 % while increasing the &lt;0.053 mm aggregate contribution by 40.58 %, facilitating carbon migration from &gt;2 mm to &lt;0.053 mm aggregates. Heavy trampling transiently increased labile organic carbon but decreased recalcitrant and inert fractions, while suppressing microbial biomass carbon. Structural equation modeling confirmed that trampling directly (<em>P</em> &lt; 0.05) and indirectly (via reduced aggregate stability and carbon activity, <em>P</em> &lt; 0.01) affected carbon dynamics. Trampling disturbance significantly compromises soil carbon pool stability and sequestration capacity through multiple mechanisms: disrupting aggregate stability, reducing organic carbon content, altering carbon fraction composition, and suppressing microbial activity. Compared to simply increasing carbon inputs, minimizing physical disturbance proves more effective for maintaining soil carbon sink functionality, providing a scientific basis for urban park carbon management.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109765"},"PeriodicalIF":5.7,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973834","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Response of hydrological processes to event- and annual-scale precipitation extremes in a rocky mountainous area of northern China","authors":"Yuxin Wu ,&nbsp;Xinxiao Yu ,&nbsp;Guodong Jia ,&nbsp;Zihe Liu ,&nbsp;Honghong Rao","doi":"10.1016/j.catena.2026.109803","DOIUrl":"10.1016/j.catena.2026.109803","url":null,"abstract":"<div><div>In the context of global change and human activities, the increasing frequency of heavy rainfall and drought events due to uneven precipitation distribution poses significant challenges to ecosystem hydrology. Forest and grass ecosystems can exhibit distinct hydrological responses to climate extremes, yet their short- and long-term reactions across multiple temporal scales remain inadequately quantified. Using field standard runoff plots and vertically-distributed soil moisture monitoring systems over 13 years, this study investigated the effects of extreme rainfall events on P (precipitation), R (runoff), ET (evapotranspiration), and SWC (soil water content) in forest and grass ecosystems within a rocky mountainous area of Beijing. At the event scale, NG (natural grass) showed faster and more pronounced hydrological responses in shallow soil layers compared to SF (secondary forest). Storm events induced significant water transport in deep soils across both NG and SF. Deep water movement was regulated by root systems and pre-storm SWC, with SF demonstrating greater soil water retention capacity. At the annual scale, the primary water output was ET, accounting for 75.42% ∼ 107.69% of NG and 78.68% ∼ 121.55% of SF. Multi-year R was markedly higher in NG than in SF, although mean annual R accounted for only 6.74% to 12.97% of annual P. Deep SWC was significantly greater in extremely wet years compared to extremely dry years, highlighting the role of extreme precipitation in deep soil water recharge and retention. Both ecosystems exhibited multi-year soil water deficits, although heavy rainfall at the end of the growing season occasionally resulted in annual water surplus. These findings provide new insights into the hydrological dynamics of rocky mountain ecosystems under increasing climate variability.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109803"},"PeriodicalIF":5.7,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973915","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Disentangling the impacts of climate, catchment, and morphological characteristics on hydrological drought propagation and recovery","authors":"Xuan Luo ,&nbsp;Xuan Ji ,&nbsp;Yi Zou ,&nbsp;Siqi Wang ,&nbsp;Xinbei Liu ,&nbsp;Xiaodong Wu ,&nbsp;Jianxing Li ,&nbsp;Yungang Li","doi":"10.1016/j.catena.2026.109799","DOIUrl":"10.1016/j.catena.2026.109799","url":null,"abstract":"<div><div>Hydrological drought (HD) propagation and recovery dynamics exhibit spatiotemporal heterogeneity, which is driven by coupled climate-catchment interactions. However, the relative contributions and role of these drivers remain poorly quantified. This study developed an attribution approach integrating process-based modeling and machine learning to disentangle multi-factor controls on HD propagation and recovery in the Lancang-Mekong River Basin. By employing a Variable Infiltration Capacity model to reconstruct natural runoff, we quantified meteorological-to-hydrological drought propagation time (PT) and recovery lag time (RT) through event-based analysis. A multi-stage predictor selection strategy, combined with Optimal Parameter-based Geodetector (OPGD), Recursive Feature Elimination (RFE), and Extreme Gradient Boosting (XGBoost) algorithms, addressed spatial heterogeneity and multicollinearity among driver factors while retaining predictive power. Finally, SHapley Additive exPlanations (SHAP) were incorporated to quantify feature contributions and support the interpretation of nonlinear driver-response relationships. Results revealed that RT (mean: 3.45 months) consistently exceeded PT (mean: 1.12 months) across all sub-basins, with drought duration and severity amplifying PT and RT. The hybrid OPGD-RFE-XGBoost model achieved higher R² (50% increase for PT and 42% increase for RT) using only 18% and 35% of the original features compared to XGBoost model. SHAP analysis identified climate as the primary controlling factor for both PT (contribution: 45.5%) and RT (51.8%). Catchment and morphological characteristics exhibited subordinate influence on PT (41.3%) and RT (32.7%). By integrating spatially structured predictors and explainable machine learning, this study establishes an analytical framework to decouple climate-catchment interactions in drought cascades, highlighting the role of underlying characteristics in drought propagation and recovery.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109799"},"PeriodicalIF":5.7,"publicationDate":"2026-01-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973919","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Natural susceptibility and human intervention in gullies sidewalls in tropical environments of southeastern Brazil","authors":"Estêvão Botura Stefanuto ,&nbsp;Cenira Maria Lupinacci ,&nbsp;Higor Lourenzoni Bonzanini ,&nbsp;Danilo Marques de Magalhães ,&nbsp;Filipe Carvalho ,&nbsp;Estela Nadal-Romero ,&nbsp;Melina Fushimi ,&nbsp;Fabiano Tomazini da Conceição ,&nbsp;Archimedes Perez Filho","doi":"10.1016/j.catena.2025.109783","DOIUrl":"10.1016/j.catena.2025.109783","url":null,"abstract":"<div><div>It is crucial to distinguish between gully erosion responses derived from the inherent fragility of the environment and those induced by human activities. This study examines gully sidewall erosion in relation to soil physical variables and land use, with particular attention to livestock farming. Sidewall data were collected and analysed during both wet and dry seasons, including grain size, soil resistance to penetration, and infiltration. Land use data were also compiled before and after the implementation of a restoration technique. These data were compared with three years of erosion monitoring on the gully sidewall. The results indicate that tropical environments with two well-defined seasons (summer and winter) also require careful consideration of transitional periods (spring and fall), as soil drying and crack formation increase erosion risk. Sandy fractions demonstrate high susceptibility to erosion; however, the role of finer fractions, especially silt, is equally critical, since their presence, together with higher soil penetration resistance during dry periods, indicates areas with high erosive potential. Infiltration measured under field conditions shows high variability, which complicates the assessment of its contribution to gully development. The adoption of simple measures, such as fencing off eroded areas, proved effective in reducing soil loss from gullies in pastures under livestock farming. Overall, the findings underline that the combined effect of environmental fragility and grazing pressure constitutes a major driver of gully erosion in tropical regions. Recognising these interactions is fundamental for designing appropriate soil conservation strategies in livestock-dominated landscapes.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109783"},"PeriodicalIF":5.7,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973833","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Methodological advancements in soil erosion: a meta-analysis of organic matter content and erosion in the Mediterranean region","authors":"Tom Avikasis Cohen ,&nbsp;Gabriel Cotlier ,&nbsp;Ioannis Daliakopoulos ,&nbsp;Pandi Zdruli ,&nbsp;Shahar Baram ,&nbsp;Anna Brook","doi":"10.1016/j.catena.2026.109793","DOIUrl":"10.1016/j.catena.2026.109793","url":null,"abstract":"<div><div>Topsoil organic matter content (OMC), particularly within the upper 30 cm, is critical for soil fertility, ecosystem functioning, and resilience. In Mediterranean environments, OMC is highly vulnerable to soil water erosion (SWE), a process intensified by increasingly frequent and intense rainfall. While SWE research has largely focused on mineral soil loss, its effects on OMC, especially the balance between depletion in source areas and accumulation in depositional zones, remain poorly quantified at regional scales. This study applies advanced analytical approaches, including multilevel meta-analytic models, to examine the spatial dynamics of OMC redistribution under SWE and to evaluate how scale and methodology influence reported outcomes. A meta-analysis was conducted using data from 71 studies across Mediterranean landscapes. Due to the limited availability of land-use-specific studies, the analysis adopts a broad regional perspective rather than stratification by management type. Results reveal extremely high heterogeneity, driven primarily by methodological differences, spatial scale, and timing. Mean effects varied by study type: field experiments showed a weakly negative association between SWE and OMC (β = −0.0421, k = 68), rainfall-based studies showed a positive relationship (β = 0.7209, k = 66), and simulation studies indicated a mild positive trend (β = 0.0835, k = 14), all with high heterogeneity (I² &gt; 90 %). Overall, the findings demonstrate that SWE does not uniformly result in OMC loss but often reflects spatial translocation. These results highlight the need for improved methodological consistency and ecologically grounded frameworks to interpret OMC dynamics in Mediterranean erosion-prone landscapes.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109793"},"PeriodicalIF":5.7,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973832","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Slope orientation regulates the joint influence of ephemeral plants and microorganisms on the soil multifunctionality of the mobile sand dunes","authors":"Xinmeng You ,&nbsp;Xiaodong Yang ,&nbsp;Lu Gong ,&nbsp;Yihu Niu ,&nbsp;Xue Wu ,&nbsp;Xiaochen Li ,&nbsp;Qian Guo","doi":"10.1016/j.catena.2026.109818","DOIUrl":"10.1016/j.catena.2026.109818","url":null,"abstract":"<div><div>In desert dune ecosystems, climate change-and human disturbance-induced ecological processes have exacerbated the expansion risk of mobile dunes, substantially impairing soil functions maintenance and performance. As key biological functional groups in the dune ecosystem, ephemeral plants and soil microorganisms play crucial roles in regulating soil multifunctionality (SMF). However, the ecological mechanisms underlying this regulation remain unclear. To address this gap, we conducted a study on typical mobile dunes at the southern margin of the Gurbantunggut Desert. We established plots across four dune slope orientations to investigate the structure and diversity of ephemeral plant communities, employed high-throughput sequencing to analyze soil microbial community composition, and comprehensively evaluated SMF and its driving mechanisms. Results showed that SMF was significantly higher at the dune bottom than on the middle positions. Biodiversity and single functional indicators varied among different slope orientations. Piecewise structural equation modeling (SEM) analysis revealed that microbial diversity (bacterial ACE and fungal ACE) exerted a direct and significant positive effect on SMF. In contrast, plant diversity (Shannon and MNTD) imposed significant negative impacts through the dominant-species effect and their interactions with microorganisms. Our findings indicated that slope orientation, as a key environmental filtering factor, regulated soil physicochemical factors such as soil moisture, electrical conductivity, and pH. This regulation indirectly affected the structure and interaction of plant and microbial communities, which in turn modulated SMF. Collectively, slope orientation-driven environmental heterogeneity, microbial functional complementarity, and the dominant-species effect govern the spatial differentiation of SMF in arid desert dune ecosystems.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109818"},"PeriodicalIF":5.7,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145974283","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Multi-annual evolution of coastal dunes: Transition from fixed to transgressive dunes state","authors":"Alexandre Nicolae Lerma ,&nbsp;Olivier Burvingt ,&nbsp;Bruno Castelle ,&nbsp;Bruce Ayache ,&nbsp;Nicolas Robin ,&nbsp;David Rosebery ,&nbsp;Julie Billy","doi":"10.1016/j.catena.2025.109787","DOIUrl":"10.1016/j.catena.2025.109787","url":null,"abstract":"<div><div>Most of coastal dunes located in temperate latitudes, especially in the Northern Hemisphere, are relatively stable. However, along the Gironde coast (SW France) substantial dune remobilization has been observed over the last decade following major marine erosion events during the 2013–2014 winter. This study is based on the analysis of a robust dataset including (i) 10 high-resolution Digital Terrain Models (DTMs) derived from airborne LiDAR surveys conducted over a 12-year period (2011−2023) and (ii) 7 Satellite-derived Digital maps of dune vegetation cover derived from Sentinel-2 satellite images acquired between 2017 and 2023. These morphological and biological parameters are linked to forcing parameters derived from observed wind data, to provide a comprehensive analysis of coastal dune changes related to the transition from vegetation-fixed dunes to the development of transgressive dunes. For the first time, morphological and vegetation dynamics are explored over a large spatial scale (tens of km), covering a range of initial dune morphology and sediment supply.</div><div>Dunes have transitioned from stable to transgressive states primarily driven by sediment stoss slope recycling process (cannibalism) across a gradient of alongshore variable dune sediment budget, ranging from slightly negative to notably positive (+10 to 15 m³/m/yr), Along this coast, transgressive dunes defined as dune migrating via similar stoss and lee slope migration rates, have tripled in number over the last 10 years (reaching ≈ 15 km or 17.3 % of the studied coast). At the center of the Gironde coast where dunes are heavily remobilised, the lee slope of the dune translates landward at a rate of several meters to more than 10 m/year. In the following years, dunes will probably continue to migrate and remobilise across a broader scale if no re-stabilization management plan is implemented.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109787"},"PeriodicalIF":5.7,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145923165","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deciphering sediment connectivity drivers in the loess watershed of northwestern Hebei Province using machine learning and structural equation modeling","authors":"Yibo Wang ,&nbsp;Zhen Li ,&nbsp;Yancang Li ,&nbsp;Ke Wu ,&nbsp;Zhiguo Qi ,&nbsp;Zhe Li","doi":"10.1016/j.catena.2025.109752","DOIUrl":"10.1016/j.catena.2025.109752","url":null,"abstract":"<div><div>Sediment connectivity is a crucial metric for characterizing erosion and sediment transport within watersheds, influenced by the complex interplay of multiple factors. However, current research has inadequately explored the cascading effects of these driving pathways. This study investigates a typical watershed in the loess region of northwestern Hebei, China, where 13 potential driving factors, such as topography, vegetation, and gully morphological features, are systematically extracted by combining high-precision DEMs with high-resolution GF-7 and Sentinel-2 imagery. A unique integrated analytical framework combining machine learning and structural equation modeling was developed. To assess factor relevance, Random Forest (RF) and Extreme Gradient Boosting (XGBoost) models were initially employed, followed by the selection of the optimal feature subset using Shapley values from SHapley Additive exPlanations (SHAP). Subsequently, a Partial Least Squares Structural Equation Model (PLS-SEM) was constructed to quantitatively evaluate the driving pathways influencing sediment connectivity. The results indicate that vegetation factors (<em>NDVI</em>, <em>LAI</em>), topographical factors (<em>LS</em>, <em>K</em>_p), and gully characteristics (<em>A</em>, <em>L</em>) are the primary variables identified by RF and XGBoost in combination with SHAP. With a high path value of 0.66, topography strongly increases sediment connectivity, while vegetation considerably reduces connectivity with a path coefficient of −0.45. Gully erosion contributes positively to sediment connectivity with a coefficient of 0.17 and is jointly controlled by vegetation (path coefficient = −0.10) and topography (path coefficient = 0.54). This primarily reflects topography's dominant control, which outweighs vegetation's comparatively little suppressive effect. Furthermore, regarding goodness-of-fit, the PLS-SEM model with optimal components outperformed the conventional SEM by 19.3 %. This study presents a novel multi-model integration framework that delivers innovative insights for driving factor selection, effect analysis, and cascading pathway assessment. It provides a robust analytical approach for investigating sediment connectivity mechanisms and offers theoretical guidance for optimizing soil and water conservation strategies in loess watersheds.</div></div>","PeriodicalId":9801,"journal":{"name":"Catena","volume":"264 ","pages":"Article 109752"},"PeriodicalIF":5.7,"publicationDate":"2026-01-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145973836","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"农林科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}