Balancing the increasing demand for ecosystem services (ESs) with their limited supply is a critical global challenge for human society. However, the drivers shaping the spatial heterogeneity of the comprehensive supply-demand balance (CSDB) remain unclear. To address this, we developed an integrated framework synthesizing multi-source datasets to assess ES supply-demand balance, analyze coupling coordination and trade-offs/synergies, and identify driving factors. Focusing on the Loess Plateau (LP), from 2000 to 2020, we analyzed the spatiotemporal dynamics of four ESs: water yield (WY), carbon sequestration (CS), soil conservation (SC), and food production (FP). Furthermore, interpretable machine learning was employed to identify dominant CSDB drivers. Results show: (1) While the overall supply of all four ESs increased, demand was significantly concentrated in urban areas and their peripheries. This led to persistent and intensifying deficits in WY, CS, and FP in urban centers, whereas forest-covered regions largely maintained a balance or surplus. SC remained in surplus across most areas, with only minor localized deficits in the Plateau Wind-Sand Region (PWSR). Consequently, the CSDB exhibited a pattern of high surplus in forest zones and deteriorating deficits in urban areas. (2) The Coupling Coordination Degree (CCD) of multi-service supply-demand balances was generally higher in the central region and decreased from southeast to northwest. Approximately 51% of the area was in a state of serious imbalance. Significant differences were observed between supply-demand balance pairs: WY-CS exhibited strong synergy; WY-SC, CS-SC, and CS-FP showed weak synergy; while WY-FP and SC-FP were dominated by trade-offs. (3) The spatial heterogeneity of the CSDB was primarily explained by human activity-related variables in the model (contribution >70% in most subregions), with population density showing a substantial negative contribution. Precipitation served as a secondary key factor, promoting CSDB surplus. Additionally, five supply-demand balance clusters were identified to propose spatially differentiated governance pathways. This study provides a scientific basis and theoretical foundation for alleviating ES supply-demand imbalances, optimizing regional resource allocation strategies, and promoting sustainable ecosystem management practices on the LP.

Field–ditch–pond (FDP) systems alleviate nitrogen (N) loss from rice production through interception and purification; however, optimal strategies for N-loss regulation remain unclear. Therefore, we evaluated different water management strategies for N-reduction efficiency under FDP systems using FDP-NPS model in typical rice-growing regions of China. The ditch–pond proportions were 1.87–6.91%, and annual rainfall was 228.0–2713.2 mm. The “low‑water-level management” strategy for FDP system, consisting of water‑saving irrigation in fields and low‑water‑level regulation in ditch–ponds, synergistically intercepted over 93.72% of runoff from rainfall events. The N reduction efficiencies (IRNload) were 42.68–89.55% in Northeast Plain, 46.01–85.94% in Mid-lower Yangtze River Basin, 27.18–76.44% in Southeast Coast, and 13.85–38.82% in Upper Yangtze River Basin. Adjusting the effective water depth of ditch–pond systems increased the high-intensity regulation of IRNload by 7.74–36.23% over low- and medium-intensity regulation. The flexible regulatory strategies can further adapt to diverse rainfall events across dry, normal, and wet years while maintaining stable IRNload with slight nationwide fluctuations (0.22–14.57%). Thus, region-specific optimization of water regulation should be prioritized over blind ditch–pond expansion for N pollution control and high-standard farmland construction.

The current understanding of inorganic solute transport, hydrochemical mechanisms, and water quality status in the surface waters of the Yellow River headwaters (HWYR) remains limited. This study investigated the spatial variation, formation mechanisms, and hydrochemical ion sources while assessing surface water quality. The results indicate that HWYR surface waters are weakly alkaline (pH: 8.30–8.51) and fresh (total dissolved solids (TDS): 308–397 mg/L; mineralization: 296–378 mg/L; total hardness (TH): 200–233 mg/L). The Ngoring Lake exhibited significant spatial heterogeneity in hydrochemical parameters. The confluence of the Gyaring Lake and Duoqu River showed peak Ca2+, SO42−, total phosphorus, and TH, while the mainstem Yellow River had the highest TDS and mineralization. Riverine total nitrogen and NO3− exceeded lake concentrations, indicating nitrogen cycling significantly drives hydrochemical processes. HCO3− dominated anions (>40

As a critical ecological security barrier in China, the Loess Plateau faces severe threats to regional sustainable development owing to the degradation of ecosystem service functions and loss of biodiversity. The scientific quantification of the ecological value and identification of priority conservation areas (PCAs) are urgently required to guide regional ecological protection and high-quality development. In this study, ecosystem service value (ESV) is quantified in the Loess Plateau from 2000 to 2025, and PCAs were identified using hotspot analysis and the ordered weighted averaging (OWA) algorithm. Based on the NSGA-II-PLUS model, three land-use development patterns for 2035 were proposed, and future PCA planning schemes under these scenarios were explored. The results showed that from 2000 to 2025, the ESV of the Loess Plateau will first decrease and then increase. The high-value areas were mainly located in the Lvliang Mountain and Taihang Mountain of Shanxi and the central-southern part of Shaanxi, while low-value areas were primarily located in central Inner Mongolia and the river valley regions of Shanxi and Shaanxi. The use of hotspots to delineate PCAs is not sufficiently comprehensive and cannot be directly overlaid. Using the OWA algorithm to construct indicator weights, scenario 5 (with a risk factor of 0.4) was determined to be the optimal weight. In this scenario, the total protection efficiency of the PCA was 8.76, with a protected area of 105,311 km2, accounting for 16.6 % of the study area. Compared to directly overlaying the hotspot analysis, the comprehensive scheme reduces the protected area by 51.18 % while increasing the overall protection efficiency by 39.21 %. Under the natural development pattern, the loss of ESV caused by farmland conversion can be compensated for by returning farmland to forests; however, urban expansion will result in a 12.3 % decrease in ESV in regions such as the Guanzhong Plain in Shaanxi. Under the ecological protection pattern, Shanxi had the largest protection area (45,233 km2), which was 56.5 % larger than that under the economic development pattern. The center of gravity of the overall PCAs is in Shaanxi Province. The center of the PCAs in the entire region is in Shaanxi Province, and the identified priority areas are mainly soil conservation areas. Future reserves may be expanded based on existing reserves in Shaanxi and central Shanxi. The identification of priority conservation areas and future development predictions in this study provides scientific support for the formulation of rational ecological protection policies for the Loess Plateau.

Accurate prediction of freshwater infiltration into saline soils is hindered by neglect of solute-potential feedback in conventional water-salt transport models. To address this, we developed SWSK1, a soil water-solute kinetic model extension that explicitly incorporates solute potential through an osmotic efficiency coefficient (f0) derived from semipermeable membrane theory. SWSK1 was validated in laboratory column tests on loam, loamy sand, and sandy loam at NaCl salinities of 0-5 g kg- 1. Compared to the baseline model (SWSK0), SWSK1 markedly improved cumulative infiltration predictions (nRMSE reduced from 0.048 to 0.163-0.013-0.077; Pbias constrained within +/- 8%) and reproduced salinity-induced oscillatory peaks and troughs in transient volumetric water content (mean nRMSE reduced from 0.201 to 0.127; Pbias within +/- 5%). Application to four USDA-textured soils (silt loam, loam, sandy loam, clay loam) showed that inclusion of f0 shortened infiltration times by 19%-35%, increased infiltration depths by 2-5 cm and extended solute diffusion depths by up to 7 cm, with osmotic effects plateauing above 3 g kg- 1 salinity. These findings underscore the critical role of solute-potential feedback in saline-soil infiltration modeling and provide a robust basis for enhancing water-use efficiency and salt leaching in agricultural practice.

Natural processes, combined with human activities, determine the inherent quality of regional water supply and demand. However, the interaction between artificial vegetation restoration and water supply-demand dynamics remains insufficiently understood, particularly in arid and semi-arid regions. This study focuses on the Jinghe River Basin (JRB) in the central Loess Plateau, aiming to investigate the changes in supply and demand of ecosystem water yield services and analyze factors affecting the water supply-demand relationship during the vegetation restoration, using the InVEST model, scenario analysis, and the Geodetector. The key findings are as follows: (1) Water shortages in the basin are concentrated in the southern and northern areas, while the southwestern and central areas exhibit water surpluses. Between 2000 and 2020, the water supply showed a first decline and then increase trend, whereas water demand exhibited an opposite trend, with 2009 and 2013 as turning points, respectively. These changes mitigated the basin's water scarcity. (2) Assuming constant meteorological factors and fixed water use indicators, water supply in the vegetation restoration areas decreased by 47.4 million m3, but water demand decreased by 89 million m3, indicating that vegetation restoration did not threaten human water availability. (3) Water supply is primarily influenced by meteorological factors, while water demand and the water supply-demand ratio (WSDR) are mainly driven by socio-economic factors. The influence of precipitation on water yield (q = 0.69) outweighs that of land use (q = 0.14), indicating that the reduction in water yield due to vegetation restoration is offset by the increased precipitation. This study provides insights into the spatiotemporal dynamics of water yield services and the relationship between vegetation restoration and water supply-demand in the arid and semi-arid regions.

The spatiotemporal patterns and driving factors of drought-flood abrupt alternations (DFAA) have been investigated across several regional and watershed scales; however, comprehensive examination at the global scale is lacking. Here, we employed the long period drought-flood abrupt change index (LDFAI), derived from an ensemble of 40 output datasets from eight Coupled Model Intercomparison Project phase 6 (CMIP6) models, to assess the spatiotemporal patterns, drivers, and future projections of global DFAA. The results indicate that DFAA are influenced by various anthropogenic forcings, and greenhouse gas emissions exert the most significant impact. The changes in the intensity of global DFAA (1950–2014), attributed to natural forcing (NAT), anthropogenic aerosols (AER), and greenhouse gas (GHG) forcing, accounted for 5.65%, 14.57%, and 33.55%, respectively. The rates of change of the DFAA intensity under shared socioeconomic pathways (SSPs) from 2014 to 2100 were estimated to be 21.73% (SSP1-2.6), 45.37% (SSP2-4.5), 63.1% (SSP3-7.0), and 69.51% (SSP5-8.5). This means that under high radiative forcing, the regional rivalry and fossil-fuel development models will lead to a significant increase in DFAA. These findings can aid in the development of global adaptive policies related to DFAA.

The export coefficient model (ECM) remains widely applied in estimating agricultural Non-point Source Pollution (NPSP) due to its simplicity, minimal parameter requirements, and relatively high accuracy. However, its reliance on empirical export coefficient (EC) limits its ability to accurately quantify pollutant loads in large and complex watersheds. This necessitates the development of more advanced approaches for improved pollutant load estimation. To overcome these challenges, the EC-ICM integrates environmental factors with optimized EC values for various land use types, enhancing its adaptability for watershed management. Empirical EC values were derived using genetic algorithm (GA) and Latin hypercube sampling, then improved EC and corrected EC were employed to estimate pollutant discharge and water inflow across land uses. The model incorporates multiple factors-such as surface runoff, topographic influence, landscape interception, soil erosion, pollutant production, water leaching, and cost-distance-allowing for more accurate NPSP load assessments. Pollution factors were classified using the natural breaks method, with Entropy Weight method determining weights for a comprehensive multi-factor evaluation and risk-level assignment. Compared to ECM optimized solely with GA, the EC-ICM demonstrates improved accuracy, reducing the relative error of total nitrogen and total phosphorus by 9.66% and 6.68%, respectively. Land use contributes the highest share of TN loads, particularly from cropland and grassland, followed by livestock and population sources. TP loads are primarily attributed to livestock and poultry farming, followed by land use and population sources. The Longdong Loess Plateau, responsible for approximately 12% of total NPSP loss, is identified as a high-risk area. Targeted zoning management strategies based on risk analysis prioritize these high-risk regions, providing practical recommendations for pollution control and comprehensive watershed environmental management. Future research can further explore the impact of improving temporal resolution, future climate change and combining hydrodynamic models on the ability to simulate the amount of pollutants entering the river.

In recent decades, soil erosion on the Loess Plateau has markedly declined, leading to a significant reduction in sediment entering river systems. However, the relationships among sediment connectivity, soil erosion, and critical source areas (CSAs) of sediment yield remain inadequately understood. This study addresses this gap by developing a coupled erosion and sediment yield model that integrates the Revised Universal Soil Loss Equation (RUSLE), the modified Index of Sediment Connectivity (ICR), and sediment retention estimates from check dams. The Geodetector model is employed to identify the dominant factors influencing soil erosion, while statistical analyses based on land use and land cover (LULC) data are used to identify CSAs. The results demonstrate that the coupled erosion-sediment yield model is both applicable and robust for the Yanhe watershed over the study period from 2000 to 2020. Vegetation cover, slope, and slope length are identified as the primary driving factors of soil erosion. The contribution of sediment reduction by check dams significantly declined from 60.7 % during 2000-2003 to 14.7 % during 2016-2020. The remaining 39.3 %-85.3 % of sediment reduction is attributed to changes in sediment connectivity driven by vegetation restoration and rainfall variability. The contribution from vegetation-induced sediment reduction is expected to increase over time. Areas characterized by both high sediment connectivity and high erosion intensity are identified as key zones for erosion control. Spatial analysis reveals that 76.82 % of CSAs are located in regions with high connectivity. Receiver Operating Characteristic (ROC) curve analysis identified CSAs thresholds of 10.74 % for forest cover and 27.84 % for source land types (cropland and bare ground). Based on these thresholds, CSAs occupy only 36.54 % of the watershed area but contribute 52.16 % of the total sediment yield, indicating their disproportionate impact on sediment production.

Water yield in a watershed is mainly affected by both climatic and physiographical factors, but the interaction between different factors still cannot be completely understood. The soil and water assessment tool (SWAT), geographic detector, and natural breaks were integrated to address the impact of different factors on water yield and the identification of important hydrometeorological and underlying surface variables. (1) The significance of precipitation accounted for more than 70% of the total statistics, and its q statistic reached 0.261, which was the main driving factor and contributed an important explanatory power to the spatial variation of water yield. Moreover, the two-factor combination of slope and precipitation has the strongest interactive explanatory power on spatial variation of water yield. (2) The interaction between different meteorological factors has bivariate enhancement and nonlinear enhancement effects on the spatial variation of water yield. However, there is only nonlinear enhancement between land use factors and meteorological factors because the explanatory power of different land use types on water yield is not very prominent, while their explanatory power can be enhanced when interacting with other meteorological factors, especially evaporation. (3) Adding factors does not improve the explanatory power of original single or double factors, indicating that using the SWAT model to delineate subbasins and simulate water yield may homogenize factor attributes, especially peak rainfall, and weaken the explanatory power, and the size of subbasin and distribution of rain gauges make the spatial correspondence between different factor attributes and water yield not entirely consistent. This study can provide a scientific reference for the protection of water resources, the optimization of land use strategy, and the tradeoffs of ecological hydrologic services in semiarid and arid regions.

Abstract. Distributed models require parameter sensitivity analyses that capture both spatial heterogeneity and temporal variability, yet most existing approaches collapse one of these dimensions. We present a two-step, deep learning-assisted, time-varying spatial sensitivity analysis (SSA) that identifies dominant parameters across space and time. Using SWAT for runoff simulation of the Jinghe River Basin, we first apply the Morris method with a spatially lumped strategy to screen influential parameters and then perform SSA using a deep learning-assisted Sobol' method for quantitative evaluation. A key innovation lies in the systematic sensitivity evaluation with parameters represented and analysed at both subbasin and hydrologic response unit (HRU) scales, enabling explicit treatment of distributed parameters at their native spatial resolutions. To reduce computational burden, two multilayer perceptron surrogates were trained for 195 subbasin and 2,559 HRU parameters, respectively, allowing efficient time-varying SSA of NSE-based Sobol' indices over 3- and 24-month rolling windows during 1971–1986. Results reveal structured, scale-dependent controls: spatially, sensitivity hotspots are coherent between scales but become more localized at the HRU level, reflecting heterogeneity in land use, soils, and topography; temporally, sensitivities fluctuate with runoff in the 3-month window, while event-scale variations are smoothed in the 24-month window, yielding more persistent patterns governed by storage and routing processes. The proposed framework provides a computationally efficient and unified approach for identifying scale-dependent sensitivity hotspots and hot moments, thereby supporting targeted calibration and enhancing the interpretability and predictive robustness of distributed models under nonstationary conditions.

The spatiotemporal patterns and driving factors of drought-flood abrupt alternations (DFAA) have been investigated across several regional and watershed scales; however, comprehensive examination at the global scale is lacking. Here, we employed the long period drought-flood abrupt change index (LDFAI), derived from an ensemble of 40 output datasets from eight Coupled Model Intercomparison Project phase 6 (CMIP6) models, to assess the spatiotemporal patterns, drivers, and future projections of global DFAA. The results indicate that DFAA are influenced by various anthropogenic forcings, and greenhouse gas emissions exert the most significant impact. The changes in the intensity of global DFAA (1950–2014), attributed to natural forcing (NAT), anthropogenic aerosols (AER), and greenhouse gas (GHG) forcing, accounted for 5.65%, 14.57%, and 33.55%, respectively. The rates of change of the DFAA intensity under shared socioeconomic pathways (SSPs) from 2014 to 2100 were estimated to be 21.73% (SSP1-2.6), 45.37% (SSP2-4.5), 63.1% (SSP3-7.0), and 69.51% (SSP5-8.5). This means that under high radiative forcing, the regional rivalry and fossil-fuel development models will lead to a significant increase in DFAA. These findings can aid in the development of global adaptive policies related to DFAA.

Appropriately selected best management practices (BMPs) may be highly effective to control soil erosion. However, the environmental efficacy and economic benefit of one or a combination of BMPs in reducing sediment loads remains an open question in arid and semiarid watersheds. This study employs the widely used Soil and Water Assessment Tool (SWAT) model, the multiattribute decision-making method, and the cost-effectiveness analysis method to simulate runoff and sediment and compares the environmental and cost benefits of 44 types of single and combined BMPs at both subbasin and basin scales. The main results are as follows: (1) the SWAT model performs well in the simulation of monthly runoff and sediment in the Yanhe River watershed, meeting the accuracy requirements for model calibration, validation, and application [runoff R2>0.66 and Nash-Sutcliffe efficiency (NSE)>0.64; sediment R2>0.6 and NSE>0.51]; (2) the sediment reduction rate of returning farmland to forestland + residue cover tillage + grass waterways was 41.73%, which was the best combination in terms of environmental efficacy; the cost-effectiveness (CE) value of grass waterways was far higher than that of other measures, which was 10,864.08 kg/& YEN;; and (3) the comprehensive attribute value Z of combined BMPs was better than that of single BMPs, and the grassed waterways + strip tillage (GW+ST) showed the highest Z value of 0.92 among all BMPs. These results help provide policymakers with authentic, effective, and tailored decision-making plans of soil conservation practices. DOI:10.1061/JHYEFF.HEENG-6221.(c) 2025American Society of Civil Engineers.

Blue landscapes (BLs) are comprehensive ecosystems that are dominated by aquatic environments. They provide ecosystem services such as carbon sequestration, water purification, and soil fertility. As a plateau BL, the degree of fragmentation of the Caohai ecosystem in China has changed significantly under the combined influence of natural and human factors. Years of environmental changes in this ecosystem have resulted in transfers from croplands to wetlands, forestlands, and grasslands of 223, 165, and 852 hm2 over the past 32 years (1989-2021). The wetland area was observed to decrease and then increase. In contrast, forestlands followed the opposite trend, while grasslands increased, subsequently decreased, and increased again. The migration directions of the center of gravity of wetlands, croplands, forestlands, grasslands, and settlements were determined as northwest, west, northwest, northwest, and southwest, respectively. The landscape pattern index revealed that the large patches in the region were fragmented and the number of small patches increased. Moreover, the distribution discretization and diversity increased. We assessed ecosystem health by analyzing ecosystem changes based on ecological and landscape indicators using the vigor-organization-resilience (VOR) model. The ecosystem health assessment value ranged from 0.549 to 0.679 and increased yearly, indicating a rise in ecosystem stability. Based on our results, we propose that the government promotes sustainable ecosystem development by adjusting urban planning initiatives, developing and implementing policies that prioritize wetland protection, and integrating and maintaining blue infrastructure.

Climate and land use exert profound influences on runoff-sediment dynamics, but the interaction influence of factors such as rainfall and vegetation restoration on the multi-scale spatiotemporal distribution of connectivity levels has not been yet fully understood, especially in arid and semi-arid regions. In this study, the connectivity index, Pearson's correlation analysis and Geographical Detector Model (GDM) are combined to assess the connectivity level at different spatial and temporal scales, and to elucidate the explanation level of connectivity by each factor and its interaction. The results reveal a consistent decline in multi-year average connectivity, with high connectivity values predominantly concentrated near riverbanks and lower values typically found in the forested regions to the east and west of the basin. Moreover, the GDM reveals that in 2035, the interactive explanatory power of rainfall erosivity (R) and the aggregated weighting factor (AWC) is superior (0.343) than the interaction between rainfall erosivity (R) and AWC in 2020 (0.339). The spatial distribution pattern of the connectivity is significantly correlated with static topographic elements (especially slope factors) and is co-regulated by climate change and dynamic succession of vegetation cover. Monthly scale analysis further validates this finding: when the normalised R-value (R-n) > 0.18, high rainfall erosivity factor in July and August significantly increase the connectivity of the basin by breaking through the vegetation constraints. Spatial and temporal integration of climate, land use, and connectivity can contribute valuable insights into the dynamic interplay between surface processes and soil-water resource management.

Changing the soil and underlying surface conditions is a key practice for realizing irrigation on-site storage and infiltration. However, biochar addition and grass planting effects on soil infiltration and water retention capacity remain unclear. The effects of 0% biochar (C1), 1% biochar (C3), 2% biochar (C4), 3% biochar (C5) under ryegrass and 0% biochar (C2), 1% biochar (C6), 2% biochar (C7) and 3% biochar (C8) under Festuca arundinacea on infiltration behaviours were modelled by using sandy loessial soil columns with 'bare soil + 0% biochar' as the control (CK). (i) There is a linear relationship between cumulative infiltration and CK-C8 treatment wetting fronts (R2 >= 0.982), which showed an initial rising trend and then tended to gradual, and the influence of different treatments was primarily reflected in the middle and late infiltration stages. (ii) Both biochar and grass planting decreased the soil infiltration capacity compared with that of the CK treatment. A high biochar addition rate was beneficial for inhibiting soil water infiltration and improving water retention ability in sandy loessial soil, however, ryegrass soil infiltrabilities under 1%, 2% and 3% biochar were all stronger than that of F. arundinacea. (iii) The cumulative infiltration fitting effects in different treatments with the Kostiakov, Kostiakov-Lewis, Philip, USDA-NRCS, Horton and Green-Ampt equations were all good, although there were some differences in the infiltration rate curves under the six different fitting equations. This study is helpful in understanding effective sandy loessial soil storage ability for irrigation and efficient water resource usage. Effects of 0% biochar (C1), 1% biochar (C3), 2% biochar (C4), 3% biochar (C5) under ryegrass and 0% biochar (C2), 1% biochar (C6), 2% biochar (C7) and 3% biochar (C8) under Festuca arundinacea on sandy loam soil infiltrations were modelled with 'bare soil + 0% biochar' as the control (CK).image

Investigating eco-hydrology in desert grasslands is pivotal to comprehend the dynamic evolution patterns of vegetation. Nonetheless, a research void persists in understanding the eco-hydrological mutual feedback mechanisms associated with hydrological connectivity and the corresponding health index evaluation of a small watershed. This study is centered on the Shangdong River watershed in Inner Mongolia and uses SWAT (Soil and Water Assessment Tool) to simulate hydrological processes. The hydrological connectivity index (IC) was employed as a link to conduct Pearson correlation analysis and Granger causality tests on ecological and meteorological-hydrological factors. Additionally, the PSR model was utilized to assess the ecological health status of the watershed. Key findings reveal the following: (1) The NDVI in the Shangdong River watershed showed an overall upward trend from 2007 to 2018, while IC exhibited an overall downward trend. Temporally and spatially, there was a significant negative correlation between IC and NDVI. (2) During the vegetation growth season, IC serves as a pivotal link in the feedback loop of eco-hydrological processes. Temperature drives vegetation growth, which in turn affects IC. IC regulates soil moisture content and evaporation, further influencing vegetation growth, thus forming a feedback mechanism. (3) Over the study period, the Grassland Health Composite Index (GHI) demonstrated a consistent rise, averaging 0.44, signaling a suboptimal state for the grassland ecosystem. Furthermore, a negative correlation was observed between GHI and IC. Consequently, regulating IC could play a crucial role in safeguarding and rejuvenating the grassland ecosystem. This study offers theoretical and data support for understanding eco-hydrological processes and effective pasture management of the desert grassland watershed.

The exogenous hydroxylamine dosing has been proven to enhance nitrite supply for anammox bacteria. In this study, exogenous hydroxylamine was fed into a sequencing batch reactor to investigate its long-term effect on anammox granular sludge. The results showed that hydroxylamine enhanced the reactor's performance with an increase in total nitrogen removal rate from 0.23 to 0.52 kg N/m3/d and an increase in bacterial activity from 11.65 to 78.24 mg N/g VSS/h. Meanwhile, hydroxylamine promoted granulation by eluting flocs. And higher anammox activity and granulation were supported by extracellular polymeric substances (EPS) characteristics. Moreover, Candidatus Brocadia's abundance increased from 1.10 % to 3.03 %, and its symbiosis with heterotrophic bacteria was intensified. Additionally, molecular docking detailed the mechanism of the hydroxylamine effect. Overall, this study would provide new insights into the hydroxylamine dosing strategy application.

Low resolution of input data and equifinality in model calibration can lead to inaccuracy and insufficient reflection of spatial differences, thereby increasing model errors. However, the impact of input data accuracy, catchment threshold area, and calibration algorithm on model uncertainty reduction has not yet been well understood. The sequential uncertainty fitting version 2 (SUFI-2) that is linked with the Soil and Water Assessment Tool (SWAT) in the package called SWAT Calibration Uncertainty Programs (SWAT-CUP) was introduced to quantify the effects of different input data resolutions on parameter sensitivity and model uncertainty in the Jinghe River watershed, and the effects of different sub-basin delineations and other two calibration algorithms on model uncertainty were also comparatively analysed. (i) USLE_C, EPCO, ALPHA_BNK, and CN2 are the most sensitive parameters among all SWAT projects. When the change of digital elevation model (DEM) resolution is small, the sensitivity of parameters does not change obviously. When the DEM resolution changes significantly, BIOMIX, LAT_SED, USLE_K, and CH_N1 become highly sensitive parameters by replacing OV_N, SMTMP, SURLAG, and USLE_P. However, the change in land use resolution has little impact on parameter sensitivity, with only a slight change in the sensitivity ranking of specific parameters. (ii) Model uncertainty responded to changes in the resolution of DEM more than land use. Most of the runoff simulations had smaller uncertainties (P factor, R factor, percentage of bias [PBIAS]) than sediment. High resolution DEM data reduced model uncertainty, but the models with 2000 m DEM resolution also achieved small uncertainty. Small catchment threshold area leads to high uncertainty of the model, and large catchment threshold areas decrease the model uncertainty. The model has relatively good simulation effects in runoff and sediment when the catchment threshold area was 2000 km(2). (iii) The SWAT model has different simulation deviations and uncertainties in different calibration algorithms, the SUFI-2 and generalized likelihood uncertainty estimation (GLUE) algorithms show better applicability than particle swarm optimization (PSO). The NSE indicators of the three algorithms are in the following order: SUFI-2 > GLUE > PSO for runoff, and GLUE > SUFI-2 > PSO for sediment. This study helps us understand the cause, knowledge of which moves from the particular to the general by the comprehension of essence, power, and nature in reducing model uncertainty.