Wetlands act as natural water filters, removing pollutants and improving water quality. They provide critical habitat for diverse plants and animals, supporting biodiversity and helping to reduce flooding by absorbing excess rainwater and buffering storm surges. This paper develops an analytical framework to describe the concentration and dispersion process of a bicomponent contaminant in wetlands under the influence of surface winds. The vertical velocity distribution is derived from the momentum equation to characterise the flow field under surface-wind forcing, and the resulting velocity structure is illustrated using contour plots. Based on Mei's multiscale homogenisation technique, the common environmental dispersivity is derived, while the longitudinal mean concentrations and critical lengths are obtained separately for components A' and B'. This study reveals that, under co-flow conditions, as wind intensity increases, the peak longitudinal mean concentration for both contaminants decreases. Wind direction and intensity strongly modify both the hydrodynamic structure and the longitudinal spreading of the contaminants. Co-flow enhances environmental dispersion, while counter-flow can suppress the flow and produce localized recirculation under sufficiently strong wind forcing. The long-time concentration profiles of both contaminants remain smooth, symmetric, and bell-shaped, with increasing wind intensity reducing the peak concentration and broadening the contaminant plume. The skewness and excess kurtosis remain small and decay towards their Gaussian limits with time, indicating weak but finite transient non-Gaussian behaviour. The critical lengths for contaminants A' and B' under a specified water-quality threshold were determined, which may be useful in wetland restoration assessment and environmental risk management. The proposed model offers a quantitative framework for both natural and artificial water systems to optimise wetland restoration and improve sediment management.

Agricultural non-point source pollution contributes over 45% of total nitrogen (TN) and 65% of total phosphorus (TP) loads in China's Yangtze River Economic Belt, with its severity amplified by the synergistic interaction between climate change and land use transformation. This study investigates whether Best Management Practices (BMPs) can mitigate this climate-land-use amplification in the Bailin River Basin. Using SWAT and CLUE-S models, we integrated bias-corrected CMIP6 climate projections under four Shared Socioeconomic Pathways (SSP1-2.6, SSP2-4.5, SSP3-7.0, SSP5-8.5), land-use transition scenarios, and spatially explicit BMP implementation strategies under two land-use pathways (natural development vs. territorial spatial planning). Results show that climate-land-use synergistic effects increase TN and TP loads by 118.23 t and 13.68 t relative to baseline under unmitigated conditions. Strategic BMPs implementation, specifically 20% fertilizer reduction combined with vegetative buffer strips, reduced TN and TP loads by 39.19% and 48.58%, substantially weakening the amplification effect. Under the territorial spatial planning scenario combined with BMPs, TN and TP loads were reduced by 16.5% and 22.0% relative to baseline, indicating that BMPs can reduce pollution loads under moderate climate stress (SSP2-4.5). Our findings demonstrate that BMPs function not merely as pollution mitigators but as partial buffers of climate-land-use amplification. While the quantitative reductions reported here are specific to the Bailin River Basin and the scenario framework examined, the integrated modeling approach and the mechanistic understanding of BMP-climate-land use interactions are transferable to comparable agricultural watersheds facing similar pressures.

The widespread use of neonicotinoid pesticides, characterized by high water solubility and low soil adsorption, poses a significant threat to groundwater quality due to their high leaching potential. This study investigated the efficacy of biochar as a soil amendment for immobilizing thiamethoxam (TMX) and clothianidin (CLO). Through a combination of batch adsorption-desorption experiments and dynamic soil column studies, we demonstrate that biochar amendment induces strong and nearly irreversible sorption, increasing the soil's adsorption capacity by over 20-fold at a 3% application rate. The adsorption process was governed by chemisorption and best described by the pseudo-second-order kinetic and Freundlich isotherm models, indicating multi-mechanistic synergy including pore-filling and π-π interactions. Critically, column leaching and elution experiments identified a dosage threshold of ≥1% (w/w) biochar, which achieved complete interception of freshly applied pesticides and eliminated the mobilization risk of aged residues. Mechanistic analysis revealed that adsorption promoted the formation of larger, stabilized soil-biochar-pesticide aggregates, a key colloidal-scale process for co-immobilization. These findings underscore the potential of biochar as an effective in-situ strategy to mitigate the environmental mobility of neonicotinoids and protect water resources.

Flood-related water-quality change and ecological stress are closely coupled in deltaic rivers, yet they are seldom integrated within a single management framework. This study presents a management-oriented flood-water-quality and ecological-risk screening framework for the Ichamati River, based on 36 in-situ observations collected at 12 upstream, midstream, and downstream monitoring sites during the pre-monsoon, monsoon, and post-monsoon campaigns in 2025. Water temperature, dissolved oxygen (DO), pH, and salinity ranged from 31.55 to 35.00 °C, 7.65-12.89 mg L[-1], 8.08-9.35, and 0.010-0.050 psu, respectively. Concurrent depth, velocity, and discharge proxy (Qproxy) measurements were integrated with Landsat-8 observations, principal component analysis (PCA), locally estimated scatterplot smoothing (LOESS), Random Forest (RF), and explainable machine learning (ML). The first two principal components explained 65.86% of hydrochemical variability, while a dimensionless seasonal hydro-constituent synthesis represented pre-monsoon accumulation and rainfall-triggered mobilisation, monsoon dilution and advective flushing, and post-monsoon retention and delayed release. RF model testing yielded R[2] = 0.91, RMSE = 0.12, and MAE = 0.09, with DO stress contributing 28.6%, temperature stress 22.1%, hydraulic exposure 19.9%, pH deviation 16.7%, and salinity stress 12.7%. Very High and High classes accounted for 23.6% and 25.8% of the flood-water-quality assessment, and 17.5% and 24.3% of the ecological-risk assessment. Their strong association (r = 0.82; R[2] = 0.68) identified downstream tidal reaches as integrated intervention priorities. The framework supports seasonal source control, monsoon transport surveillance, post-monsoon recovery assessment, ecological restoration, and strategic sensor deployment, providing a robust scientific basis for climate-resilient river management.

Accurate prediction of state variables in multi-well pumping systems (MWPS) is critical for real-time management of groundwater-related operations, yet remains challenging under complex hydrogeological conditions and limited observational data. This study presents a lightweight CNN-LSTM model integrated with transfer learning for short-term but high-accuracy prediction of MWPS dynamics. The framework is evaluated using field data from an in-situ uranium leaching aquifer (China), where pumping rates from 42 wells are used to predict the uranium concentrations in the mixed solution recovered from 20 extraction wells. The proposed model contains fewer than 1000 weights and can be robustly trained using 120 pairs of historical observations. Model performance degrades with increasing prediction time when relying solely on pre-trained weights. Incorporating new observations through transfer learning significantly improves prediction accuracy. It is found that fine tuning the weights in LSTM layer adaptive to new observations yields better prediction of uranium concentration than those models updating the weights only in CNN layers. This indicates the dominant role of temporal dynamics, such as time-varying pumping rates and progressive uraninite depletion, in controlling the variation of system response features in comparison to the spatial factors including the well placement and subsurface heterogeneity. Prediction accuracy can be further improved under shorter update intervals and longer input sequences, while prediction periods ranging from 1 day to 1 week have minimal influence on the model performance. The proposed framework provides an efficient and adaptive approach for high-accuracy prediction in data-limited MWPS, with potential applications in in-situ leaching, groundwater remediation, and managed aquifer systems.

Ridge-furrow (RF) cultivation is widely applied in farmlands to improve soil water, heat, and salt conditions and increase crop yields. In cold regions, these farmlands with RF often undergo freeze-thaw processes, while previous studies mainly focused on vertical dynamics of soil water and salt based on the assumption of a flat surface. Based on a well-validated model, this study examined the water and salt dynamics in farmlands with a flat surface and RF topography under freeze-thaw processes. Results show that RF topography promotes the formation of persistent focused downward flow pathways beneath the furrows under freeze-thaw processes. Compared with a flat surface, these focused flow pathways in RF topography markedly change the subsurface flow field and drive faster and deeper downward salt transport. Further quantitative analysis showed that increased ridge height leads to greater downward salt migration, whereas the RF ratio has a negligible influence on salt migration. These findings improve the understanding of topography-controlled water flow and salt transport in seasonally frozen soils and highlight the importance of considering two-dimensional porewater flow when evaluating solute redistribution under freeze-thaw conditions.

Anthropogenic nitrogen loading and hydrodynamic exchange jointly regulate nitrate distributions in open coastal waters, complicating source attribution based on concentration or isotope patterns alone. This study investigated nitrate sources and processing in the coastal waters of Wenzhou, East China Sea, using hydrochemical parameters, δ[15]N-NO3[-], δ[18]O-NO3[-], δ[15]N-NH4[+], conservative water-mass mixing analysis, and MixSIAR. A pronounced estuarine-to-offshore environmental gradient was observed. Most nitrate concentration and isotope observations were consistent with conservative mixing among terrestrial, Taiwan Strait Warm Water, and Kuroshio Subsurface Water end-members, indicating first-order hydrodynamic control. Local departures from the mixing relationships were compatible with superimposed biological nitrate processing, although individual transformation pathways could not be uniquely resolved. MixSIAR identified manure and sewage as the largest individual posterior source category, contributing 31-40% in summer and 35-46% in autumn. Strong isotope overlap between soil nitrogen and fertilizer limited their independent discrimination, and their combined contribution was more robust than their individual posterior proportions. Nitrate dynamics in the Wenzhou coastal waters therefore reflect the combined effects of terrestrial nitrogen inputs, water-mass mixing, and spatially heterogeneous biogeochemical processing.

The computational cost of continuum Darcy-based models in two-phase flow simulations has led to the introduction of simpler and faster approaches such as macroscopic invasion percolation (MIP). However, neglecting viscous forces limits the applicability of MIP and can result in inaccurate results if used under unsuitable conditions. The results of this study challenge the commonly accepted belief that the existence of negligible viscous forces is both necessary and sufficient to justify the use of MIP models. This study offers concrete model selection criteria by comparing the results of a continuum (CompSim) and an MIP-based model (ET-MIP) in simulating air sparging into a two-dimensional domain over a range of capillary numbers (Ca). The effects of air injection rate (Qa), soil permeability (k), and soil heterogeneity (SDlnk) were studied. The Qa range was designed to produce Ca < 10[-4], which is common in different applications, such as air sparging, CO2 storage, and electrical resistance heating. Jaccard coefficient (J) was used to compare the predicted gas flow pathways of the two models. The accuracy of MIP declined with injection rate while its applicability extended up to Ca ≈ 1.16 × 10[-6] for a more heterogeneous soil (SDlnk = 1). The results demonstrated that a sufficient level of soil heterogeneity is essential for MIP to be reliable. Equivalent heterogeneity lengths (heq,gr and heq,visc) were introduced as surrogates for heterogeneity to allow for scaling of heterogeneity between different soils and for different pairs of fluids, and a phase diagram was designed based on heq,gr and heq,visc. The diagram was divided into three regions of Zone 1: MIP recommended, Zone 2: MIP misses some side branches, and Zone 3: MIP not recommended. This diagram can be used to quickly decide if an application should use continuum modelling or if an MIP model is sufficient.

At many historically polluted sites with chlorinated solvents, dissolved compounds have diffused into aquitards by aqueous diffusion. Remediation of the residual solvents in aquifers is possible, but after source removal, back diffusion of the dissolved compounds from aquitards can sustain groundwater contamination and requires prolonged post-remediation monitoring. The objective of this work is to evaluate the applicability of compound-specific stable isotope analysis (CSIA) as a tool for monitoring processes at sites impacted by back diffusion. Diffusion experiments in agar gels were performed with the three chlorinated solvents cis-1,2-dichloroethene (CIS), trichloroethene (TCE) and chlorobenzene (CB). No significant isotope fractionation was observed for any of the compounds during back diffusion from agar gels. The isotope fractionation during sorption to either Montmorillonite, Kaolinite and Amberlite XAD-2 yielded small inverse isotope effects for CB on Montmorillonite, and for CB and TCE on Amberlite XAD-2. Analytical reactive transport equations were developed to simulate the carbon isotope fractionation of compounds during forward and back diffusion in aquitards. The model identifies the conditions under which measurable isotope shifts in TCE, used as a model compound, are generated diffusing forward into a clay aquifer for 45 years and then diffusing back for 100 years. The model indicated that back diffusion cancels small isotope effects that occurred during forward diffusion in absence of degradation. The model also showed that aquitards can only contribute to isotope fractionation when fractionating degradation reactions are present in the aquitard (e.g., abiotic reductions). The results of this study provide an improved scientific basis for applying CSIA for post-remediation process monitoring in setting with back diffusion from contaminated aquitards.

The induced polarization (IP) method has significant potential for characterizing and monitoring groundwater contamination, particularly given recent advances in extracting rich spectral information from time domain IP (TDIP) data. However, its effectiveness depends on the quality of full-waveform data, and the impact of electrode noise remains insufficiently understood in field applications, especially at contaminated sites. Sandbox experiments were designed to evaluate electrode durability and IP data errors under field relevant combinations of contaminant type and conductivity at sampling rates of 0.1, 10, and 10,000 Hz. An error assessment method combining probability distributions, Fourier transforms, and the signal to noise ratio was proposed. It reveals the frequency ranges of amplitude and phase spectral information inherent in TD data. Pb-PbCl2 electrodes remained stable across the contaminant and conductivity range. For stainless steel electrodes, low-frequency noise decreased with increasing conductivity. Differences among contaminated media at same target conductivity are interpreted as combined responses to conductivity and ionic composition rather than as isolated contaminant effects. Organic contaminants may cause anomalous trends in electrode noise. Amplitude spectra were reliable across most of the 0.001-1000 Hz range, but for weaker signal, the SNR of stainless steel electrodes fell below 10 dB at <0.004 Hz. Reliable phase ranges were 0.01-1000 Hz (Pb-PbCl2) and 0.042-1000 Hz (stainless steel) for Sample A, and significantly narrowed for Samples B and C. These findings provide insights into electrode noise, thereby improving the reliable extraction of IP information.

Effective virus retention by natural minerals depends on surface physicochemical properties, but systematic side-by-side comparisons under realistic flow conditions and mechanistic explanations for material differences remain scarce. To address this gap, we initially screened nine natural minerals and evaluated four candidates (anthracite, lignite, coal gangue, and volcanic rock) for the removal of bacteriophages MS2 and ΦX174, using both batch equilibrium assays and saturated column transport experiments. In batch tests, the removal efficiency followed the order: anthracite > lignite > coal gangue > volcanic rock. Adsorption approached equilibrium within 3 h and was thermodynamically spontaneous, endothermic, and entropy-driven. Under dynamic flow-through conditions, anthracite exhibited consistently superior performance among the tested materials, achieving removal efficiencies of 99.75% for MS2 and 98.24% for ΦX174, and cumulative adsorption reaching up to 1.24×10[11] PFU/g for the two phages combined in the long-term column experiments. Integration of transport modeling, XDLVO theory, and surface characterization indicates that electrostatic attraction is the primary mechanism governing virus retention on anthracite, with hydrophobic interactions and surface functional groups serving as complementary contributions. These results provide a quantitative basis for understanding virus-mineral interactions and identify anthracite as a material with strong retention capacity warranting further investigation under environmentally relevant conditions.

The assessment of nitrate pollution in groundwater is crucial for ensuring the safety of groundwater quality and related human health risks. This study conducted a comprehensive hydrogeochemical survey in an area strongly influenced by mining areas and human activities. A total of 236 groundwater samples were collected. The range of nitrate concentration was 0.020 mg/L to 658.0 mg/L, with 78.4% of the samples exceeding the recommended limit of 10 mg/L by the Chinese Guidelines. As the nitrate content increases, the chemical type of groundwater shifted from HCO3-Ca type to ClCa type. The human input generated by residential land was the driving factor for nitrate enrichment and salinity increase. The calculation results of Entropy Weighted Water Quality Index (EWQI) showed a range of 4.9 to 650.8, with a mean of 41.6. Based on the current situation of land use, the causes of various excessive hydrochemical parameters were analyzed, and corresponding improvement suggestions were proposed for the local situation. The non-carcinogenic risk assessment indicated that the total risk range for infants was 0.0034 to 8.7 (mean 0.78), children ranged from 0.0021 to 5.4 (mean 0.48), adult females ranged from 0.0015 to 3.9 (mean 0.34), adult males ranged from 0.0018 to 4.6 (mean 0.41). The groundwater health risk of nitrate ranged from 0.0012 to 8.7, and the threat mainly came from oral ingestion of nitrates, with minors (infants, children) being more susceptible. The results can provide a scientific perspective for the sustainable development and utilization of groundwater resources and public health protection in adjacent mining areas.

Natural mobile colloids strongly facilitate the subsurface transport of radioactive Americium(III) (Am(III)), considerably extending the transport distance of Am(III) in porous media. Assessing the long-term transport behavior of colloidal Am(III) and its potential subsurface risk is critical for the safety evaluation and emergency management of radioactive disposal sites. In this study, laboratory column experiments combined with numerical simulations were conducted to investigate colloid-facilitated Am(III) transport in unsaturated porous media. A one-dimensional unsaturated column model was constructed using Hydrus-1D to inversely determine key retention and transport parameters, followed by a parameter sensitivity analysis to quantitatively evaluate the influencing intensity of hydrodynamic and geochemical factors. On this basis, an unsaturated disposal site model was established to predict the environmental transport characteristics of colloidal Am(III) under varied leakage and rainfall scenarios. Sensitivity analysis identified highly sensitive parameters such as the medium dispersion (0.141 ≤ S ≤ 0.258) and the saturated volumetric water content (-0.165 ≤ S ≤ -0.139), as well as low sensitivity parameters such as the saturated permeability coefficient (-0.044 ≤ S ≤ -0.012) and medium density (-0.0030 ≤ S ≤ -0.0026). The scenario simulation results indicated the vertical transport depth of colloidal Am(III) in the vadose zone increased with increasing time, rainfall intensity, and leakage amount. Short term heavy rainfall (100-300 mm/d) remarkably accelerated downward transport of colloidal Am(III), leading to a maximum breakthrough depth of 0.60-1.26 m without reaching the groundwater table. Nevertheless, long-term leakage under Case S1 (> 224 years) or large volume accidental leakage under Case S3 (> 7.66 m[3]) may eventually lead to groundwater contamination, posing severe threats to subsurface environmental safety.

During the groundwater recharge to reclaimed water, dissolved organic matter (DOM) might be affected by Ca[2+]-induced aggregation and coprecipitation with CaCO3. In this study, humic acid (HA) was taken as the representative of DOM. Batch experimental methods combined with three-dimensional fluorescence and FT-ICR MS were used to investigate the changes in DOM components during the aggregation and coprecipitation in water at the molecular level. The results showed that Ca[2+] tended to bind with high aromatic and carboxyl-rich components in HA, leading to aggregation and precipitation. During this process, the contribution of number of carboxyl groups in alicyclic molecules was more significant than that of the hydrophobicity of aromatic molecules. Among alicyclic molecules, those containing CHONS exhibit a stronger tendency to aggregate and precipitate than those containing CHO, CHON, or CHOS. As the initial HA concentration decreased from 24 to 6 mg C/L, the CaCO3-normalized partition coefficient of residual HA in the supernatant decreased after aggregation and precipitation, due to the removal of highly unsaturated oxygen-containing components prone to coprecipitation with CaCO3. In this process, the effects of pH decrease and Ca-induced further aggregation were negligible. These findings are helpful to understand the effects of typical hydrogeochemical processes on HA components during reclaimed water recharge to groundwater.

This study investigated the relationship between temporal variability of microbial contamination at a karst spring and the different flow components of a karst groundwater system in western Ireland. A high frequency tryptophan like fluorescence (TLF) concentration time series was collected over 12 months as a surrogate indicator of microbial contamination. The study evaluated whether temporal variations in TLF were associated with the dominant spring flow components and assessed the relative influence of internal aquifer characteristics and rainfall intensity. Flow duration curves, master recession curves, Fast Fourier Transform spectral analysis, and continuous and discrete wavelet analyses were applied. Results characterize the spring as a triple-porosity karst system comprising fast conduit, intermediate fracture, and slow diffuse matrix flow domains. TLF variability was weakly associated with slow/baseflow but more strongly associated with intermediate and fast flow components. TLF was also strongly correlated with E. coli abundance (Pearson r = 0.703; Spearman r = 0.806; Kendall r = 0.644; p < 0.05), supporting its use as a surrogate indicator of microbial contamination. TLF temporal patterns were associated with the aquifer's internal flow architecture, although rainfall intensity or recharge alone did not fully explain the observed variability. Collectively, these findings suggest that TLF variability, and by extension microbial contamination signatures, is more closely associated with shorter timescale groundwater flow responses and internal flow partitioning than with rainfall forcing alone, highlighting the value of flow component-based monitoring for managing contamination risk in karst groundwater systems.