Parameter uncertainty analysis of the non-point source pollution in the Daning River watershed of the Three Gorges Reservoir Region, China

The generation and formation of non-point source pollution involves great uncertainty, and this uncertainty makes monitoring and controlling pollution very difficult. Understanding the main parameters that affect non-point source pollution uncertainty is necessary to provide the basis for the planning and design of control measures. In this study, three methods were adopted to do the parameter uncertainty analysis with the Soil and Water Assessment Tool (SWAT). Based on the results of parameter sensitivity analysis by the Morris screening method, the ten parameters that most affect runoff, sediment, organic N, nitrate, and total phosphorous (TP) were chosen for further uncertainty analysis. First-order error analysis (FOEA) and the Monte Carlo method (MC) were used to analyze the effect of parameter uncertainty on model outputs. FOEA results showed that only a few parameters had significantly affected the uncertainty of the final simulation results, and many parameters had little or no effect. The SCS curve number was the parameter with significant uncertainty impact on runoff, sediment, organic N, nitrate and TP, and it showed that the runoff process was mainly responsible for the uncertainty of non-point source pollution load. The uncertainty of sediment was the biggest among the five model output results described above. MC results indicated that neglecting the parameter uncertainty of the model would underestimate the non-point source pollution load, and that the relationship between model input and output was non-linear. The uncertainty of non-point source pollution exhibited a temporal pattern: It was greater in summer than in winter. The uncertainty of runoff was smaller compared to that of sediment, organic N, nitrate, and TP, and the source of uncertainty was mainly affected by parameters associated with runoff.     

Parameter uncertainty analysis of non-point source pollution from different land use types

Land use type is one of the most important factors that affect the uncertainty in non-point source (NPS) pollution simulation. In this study, seventeen sensitive parameters were screened from the Soil and Water Assessment Tool (SWAT) model for parameter uncertainty analysis for different land use types in the Daning River Watershed of the Three Gorges Reservoir area, China. First-Order Error Analysis (FOEA) method was adopted to analyze the effect of parameter uncertainty on model outputs under three types of land use, namely, plantation, forest and grassland. The model outputs selected in this study consisted of runoff, sediment yield, organic nitrogen (N), and total phosphorus (TP). The results indicated that the uncertainty conferred by the parameters differed among the three land use types. In forest and grassland, the parameter uncertainty in NPS pollution was primarily associated with runoff processes, but in plantation, the main uncertain parameters were related to runoff process and soil properties. Taken together, the study suggested that adjusting the structure of land use and controlling fertilizer use are helpful methods to control the NPS pollution in the Daning River Watershed.     

Improved calibration of a rainfall‐pollutant‐runoff model using turbidity and electrical conductivity as surrogate parameters for total nitrogen

To address data scarcity for calibration of rainfall‐pollutant‐runoff (RPR) models, we evaluated the suitability of nutrient levels estimated based on surrogate parameters as a novel source of data, using runoff of total nitrogen (TN) in the Tegiru basin as a case study. A linear regression equation was developed for estimating TN based on turbidity and electrical conductivity; this expression was then used to generate TN data (n = 113) for calibration of a catchment‐specific RPR model. Using solely the estimated TN concentrations for calibration, the model accurately predicted TN concentrations (21% error based on measured TN, n = 13) and revealed runoff trends during periods in which TN measurements were lacking. Finally, we utilised this model to show that TN runoff was highest during months with frequent and high intensity rainfall. In summary, this study demonstrates the applicability of surrogate parameters to extend data on difficult‐to‐monitor nutrient loads for model calibration.     

A comparison of SWAT, HSPF and SHETRAN/GOPC for modelling phosphorus export from three catchments in Ireland

Recent extensive water quality surveys in Ireland revealed that diffuse phosphorus (P) pollution originating from agricultural land and transported by runoff and subsurface flows is the primary cause of the deterioration of surface water quality. P transport from land to water can be described by mathematical models that vary in modelling approach, complexity and scale (plot, field and catchment). Here, three mathematical models (soil water and analysis tools (SWAT), hydrological simulation program-FORTRAN (HSPF) and système hydrologique Européen TRANsport (SHETRAN)/grid oriented phosphorus component (GOPC)) of diffuse P pollution have been tested in three Irish catchments to explore their suitability in Irish conditions for future use in implementing the European Water Framework Directive. After calibrating the models, their daily flows and total phosphorus (TP) exports are compared and assessed. The HSPF model was the best at simulating the mean daily discharge while SWAT gave the best calibration results for daily TP loads. Annual TP exports for the three models and for two empirical models were compared with measured data. No single model is consistently better in estimating the annual TP export for all three catchments.     

Modeling effectiveness of agricultural BMPs to reduce sediment load and organophosphate pesticides in surface runoff

Quantifying effectiveness of agricultural BMPs at the watershed scale is a challenging issue, requiring robust algorithms to simulate not only the agricultural production system but also pollutant transport and fate. This research addresses the challenge to simulate performances of BMPs in reducing organophosphates (OPs) runoff at the watershed scale. The SWAT model is calibrated and validated following a sensitivity analysis combining Latin Hypercube sampling and One-factor-At-a-Time simulation. The calibrated model is then applied in the Orestimba Creek Watershed to simulate BMPs including buffer strips, sediment ponds, vegetated ditches, use reduction, and their combinations. BMP simulation suggested that sediment ponds trap 54-85% of sediment from field runoff, but less than 10% of dissolved diazinon and chlorpyrifos. Use reduction can reduce pesticide load in a close-to-linear fashion. Effectiveness of vegetated ditches and buffers depends on their physical dimension and vegetation cover. Combining individual BMPs provides enhanced mitigation effects. The combination of vegetated ditches, buffer strips and use reduction decreases diazinon and chlorpyrifos load by over 94%. This study has suggested that the SWAT model reasonably predicts BMP effectiveness at the watershed scale. Results will assist decision making in implementing BMPs to reduce pesticide loads in surface runoff.     

Soil erosion and sediment yield and their relationships with vegetation cover in upper stream of the Yellow River

Soil erosion is a significant concern when considering regional environmental protection, especially in the Yellow River Basin in China. This study evaluated the temporal-spatial interaction of land cover status with soil erosion characteristics in the Longliu Catchment of China, using the Soil and Water Assessment Tool (SWAT) model. SWAT is a physical hydrological model which uses the RUSLE equation as a sediment algorithm. Considering the spatial and temporal scale of the relationship between soil erosion and sediment yield, simulations were undertaken at monthly and annual temporal scales and basin and sub-basin spatial scales. The corresponding temporal and spatial Normalized Difference Vegetation Index (NDVI) information was summarized from MODIS data, which can integrate regional land cover and climatic features. The SWAT simulation revealed that the annual soil erosion and sediment yield showed similar spatial distribution patterns, but the monthly variation fluctuated significantly. The monthly basin soil erosion varied from almost no erosion load to 3.92 t/ha and the maximum monthly sediment yield was 47,540 tones. The inter-annual simulation focused on the spatial difference and relationship with the corresponding vegetation NDVI value for every sub-basin. It is concluded that, for this continental monsoon climate basin, the higher NDVI vegetation zones prevented sediment transport, but at the same time they also contributed considerable soil erosion. The monthly basin soil erosion and sediment yield both correlated with NDVI, and the determination coefficients of their exponential correlation model were 0.446 and 0.426, respectively. The relationships between soil erosion and sediment yield with vegetation NDVI indicated that the vegetation status has a significant impact on sediment formation and transport. The findings can be used to develop soil erosion conservation programs for the study area.     

Regulation of surface water quality in a Cretaceous Chalk catchment,UK: an assessment of the relative importance of instream and wetland processes

To investigate the relative importance of instream nutrient spiralling and wetland transformation processes on surface water quality, total nitrogen (TN) and total phosphorus (TP) concentrations in a 200-m reach of the River Lambourn in the south-east of England were monitored over a 2-year period. In addition, the soil pore water nutrient dynamics in a riparian ecosystem adjacent to the river were investigated. Analysis of variance indicated that TN, TP and suspended sediment concentrations recorded upstream of the wetland were statistically significantly higher (P<0.05) than those downstream of the site. Such results suggest that the wetland was performing a nutrient retention function. Indeed, analysis of soil pore waters within the site show that up to 85% of TN and 70% of TP was removed from water flowing through the wetland during baseflow conditions, thus supporting the theory that the wetland played an important role in the regulation of surface water quality at the site. However, the small variations observed (0.034 mg TN l⁻¹ and 0.031 mg P l⁻¹) are consistent with the theory of nutrient spiralling suggesting that both instream and wetland retention processes have a causal effect on surface water quality.     

城市绿地对雨水径流污染物的削减作用

为探究城市绿地调控城市降雨地表径流污染的有效性和可行性,采用室内土柱模拟试验,研究植被覆盖、径流污染物浓度、土壤层深度、地下水、水力负荷与停留时间对城市绿地削减污染物的影响。结果表明,在低、中和高3种 雨水径流污染物浓度水平（CODCR 为68、137、550 mg/L;TN是3.01、7.51、30.06 mg/L;TP为0.29、0.69、2.73 mg/L; NH+4是0.44、1.61、2.19 mg/L）,水力负荷为3.5、3.0、2.5 cm/h,持续进水1 h条件下,城市绿地具有良好且稳定的污染削减能力,对CODCR 、TN、TP、NH+4的平均削减率分别达到41.52%、78.96%、84.68%,50.21%、70.23%、60.91%,73.18%、95.88%、94.99%,62.72%、55.16%、69.98%。受土壤复氧能力和水力停留时间的限制,绿地覆盖对污染物CODCR 与TN削减率的影响不明显。城市绿地污染削减率随着降雨地表径流污染物浓度的升高而呈现逐渐增加的趋势。城市绿地对雨水地表径流污染削减作用主要发生在深度35~65 cm土层内。城市绿地对低、中和高污染浓度水平各污染物削减率随着水力负荷的增加而降低。     

华北地区陆生园林植物去除道路径流 污染物的实证研究

利用植物来削减径流污染物浓度是植物生态修复理 论的主要应用领域之一，其研究多集中在水生植物，对华北地 区城市中大量的陆生植物的研究较少，尤其是对径流污染起重 要源头控制作用的道路绿化植物及其长期修复效果更鲜有研 究，无法满足实际应用需求。与此同时，对生物滞留设施中雨 水径流长期集中入渗所带来的污染风险的研究也不足，对植物 能否起到减缓污染物累积的作用尚不明确。通过盆栽试验分析 了矮生紫薇等11种华北地区道路绿化常用的陆生园林植物对 模拟径流中TN、TP、CODCr的去除效果，以及试验过程中土 壤里TN、TP的积累量的去除率。希望能够对景观设计、环境 工程等领域针对特定目标选择特定植物进行生态修复起到一定 参考。     

Deep convolutional neural networks for uncertainty propagation in random fields

The development of a reliable and robust surrogate model is often constrained by the dimensionality of the problem. For a system with high‐dimensional inputs/outputs (I/O), conventional approaches usually use a low‐dimensional manifold to describe the high‐dimensional system, where the I/O data are first reduced to more manageable dimensions and then the condensed representation is used for surrogate modeling. In this study, a new solution scheme for this type of problem based on a deep learning approach is presented. The proposed surrogate is based on a particular network architecture, that is, convolutional neural networks. The surrogate architecture is designed in a hierarchical style containing three different levels of model structures, advancing the efficiency and effectiveness of the model in the aspect of training. To assess the model performance, uncertainty quantification is carried out in a continuum mechanics benchmark problem. Numerical results suggest the proposed model is capable of directly inferring a wide variety of I/O mapping relationships. Uncertainty analysis results obtained via the proposed surrogate have successfully characterized the statistical properties of the output fields compared to the Monte Carlo estimates.     

Evaluation of the EUROSEM model for predicting the effects of erosion-control blankets on runoff and interrill soil erosion by water

The European Soil Erosion Model (EUROSEM, Morgan et al., 1998) is an event-based soil erosion model which predicts runoff and sediment discharge for different environmental conditions. Applying geotextiles or erosion-control blankets (ECB’s) on the soil surface significantly affects surface seal formation and topsoil properties and therefore controls runoff and soil erosion rates during a rainfall event. Since these within-storm changes of soil surface characteristics and hydrological conditions are not incorporated in EUROSEM, errors in runoff and soil erosion predictions may occur for soil surfaces covered with ECB’s.Therefore, the objective of this paper is to evaluate and improve the performance of a research version of the physically-based erosion model EUROSEM (EUROSEM-2010; Borselli and Torri, 2010) for simulating the effects of ECB’s on runoff and interrill soil erosion by water during intense simulated rainfall events. Results of model simulations are compared with experimental results of interrill erosion using biological (i.e. natural) ECB’s and simulated rainfall. Because ECB’s applied on the soil surface retard seal formation, the differences between observed and predicted runoff rates and sediment discharges are rather high during the first 20-30 min of the simulated rainstorm. Therefore, a simple approach is proposed to cope with the dynamic evolution of some soil characteristics, i.e. saturated hydraulic conductivity, soil erodibility and soil cohesion, during an intense rainfall event. This time-dependent approach improves the predictions of runoff rate and sediment discharge during the first 20-30 min of a rainfall event and increases the model efficiency (i.e. a measure for the goodness of fit) from 0.84 to 0.98 and from 0.48 to 0.68 for the total runoff volume and soil loss, respectively. For most conditions, the predicted final sediment discharge is still considerably larger than the observed values, which can be partly attributed to the deposition of sediment in the bare soil patches (i.e. inter-weave open areas) of the ECB’s, which is not simulated by EUROSEM in this study. This model approach increases our understanding of the effects of ECB’s on within-storm changes in hydrological conditions and soil surface characteristics.     

武汉墨水湖底泥中总氮、总磷污染特征分析

为了研究武汉市墨水湖底泥中总氮、总磷的污染特征,在湖区选取4个取样断面采集底泥柱状样,测定其总氮、总磷含量及变化。结果表明：在平面上,墨水湖底泥中总氮、总磷的含量在离岸50 m和100 m处的变化不大;在垂直分布上,总氮、总磷的含量在底泥表层含量较小,在10～40 cm深度处含量达到最大,在40 cm以下的深度,总氮和总磷的含量逐渐变小并趋于稳定,因此底泥清淤深度不宜小于40 cm。     

Spatial analysis of water quality trends in the Han River basin, South Korea

Spatial patterns of water quality trends for 118 sites in the Han River basin of South Korea were examined for eight parameters-temperature, pH, dissolved oxygen (DO), biochemical oxygen demand (BOD), chemical oxygen demand (COD), suspended sediment (SS), total phosphorus (TP), and total nitrogen (TN). A non-parametric seasonal Mann-Kendall's test determined the significance of trends for each parameter for each site between 1993 and 2002. There are no significant trends in temperature, but TN concentrations increased for the majority of the monitoring stations. DO, BOD, COD, pH, SS, and TP show increasing or decreasing trends with approximately half of the stations exhibiting no trends. Urban land cover is positively associated with increases in water pollution and included as an important explanatory variable for the variations in all water quality parameters except pH. Topography and soil factors further explain the spatial variations in pH, COD, BOD, and SS. BOD, COD, SS, and TP variations are consistently better explained by 100m buffer scale analysis, but DO are better explained by the whole basin scale analysis. Local water quality management or geology could further explain some variations of water quality. Non-point-source pollution exhibits strong positive spatial autocorrelation as measured by Moran's I, indicating that the incorporation of spatial dimensions into water quality assessment enhances our understanding of spatial patterns of water quality. The spatial regression models, compared to ordinary least square (OLS) models, always better explain the variations in water quality. This study suggests that spatial analysis of watershed data at different scales should be a vital part of identifying the fundamental spatio-temporal distribution of water quality.     

Identifying key sources of uncertainty in the modelling of greenhouse gas emissions from wastewater treatment

This study investigates sources of uncertainty in the modelling of greenhouse gas emissions from wastewater treatment, through the use of local and global sensitivity analysis tools, and contributes to an in-depth understanding of wastewater treatment modelling by revealing critical parameters and parameter interactions. One-factor-at-a-time sensitivity analysis is used to screen model parameters and identify those with significant individual effects on three performance indicators: total greenhouse gas emissions, effluent quality and operational cost. Sobol's method enables identification of parameters with significant higher order effects and of particular parameter pairs to which model outputs are sensitive. Use of a variance-based global sensitivity analysis tool to investigate parameter interactions enables identification of important parameters not revealed in one-factor-at-a-time sensitivity analysis. These interaction effects have not been considered in previous studies and thus provide a better understanding wastewater treatment plant model characterisation. It was found that uncertainty in modelled nitrous oxide emissions is the primary contributor to uncertainty in total greenhouse gas emissions, due largely to the interaction effects of three nitrogen conversion modelling parameters. The higher order effects of these parameters are also shown to be a key source of uncertainty in effluent quality.     

河北承德中部伊逊河红旗地区土壤生源要素空间分布格局及其影响因素

土壤生源要素的含量和空间分布特征对指导土壤肥力调控,优化土地利用具有重要意义。以承德市伊逊河红旗周边地区为例,采集不同深度和典型剖面土壤样品,综合运用描述性统计、地统计半方差函数模型和冗余分析等方法,论述了研究区土壤生源要素碳、氮、磷和钾的空间分布格局与影响因素。结果表明,研究区表层土壤TN含量平均为848.74 mg/kg,总体较为缺乏;土壤Corg平均含量为1.35%,总体属中度水平;TP和TK含量相对丰富。TN、Corg、TP和TK土地质量地球化学等级属较丰富以上水平的样品占比分别为7.71%、26.18%、35.08%和77.85%。深层土壤TN和TP含量显著低于表层土壤,垂向分布上TN和TP含量总体呈随深度增加而降低的趋势,TK含量呈先下降后上升的趋势。土壤TP和TK空间分异以结构性变异为主,具有强烈的空间自相关性;TN和Corg分布具有中度的自相关性,空间分布受土地利用类型控制,也受农业活动等人为因素影响;土壤TP空间分布受成土母质和矿业活动因素共同影响,土壤TK空间分布则主要受成土母质自然因素控制。     

潜流湿地控制暴雨径流中氮磷冲击负荷中试研究

进行了3种人工潜流湿地（无植物空白床、芦苇床、茭草床）控制暴雨径流中氮、磷浓度的冲击负荷试验，考察了潜流湿地对氮、磷的去除效果及其耐冲击负荷能力。结果表明，潜流湿地能够有效控制暴雨径流中的氮、磷浓度，且具有较强的耐冲击负荷能力，其中以芦苇床潜流湿地对暴雨径流中氮、磷的去除效果为最好。处理滇池流域重现期为5a的单场降雨，试验床体面积占汇水面积的比例〉3．7％时，芦苇床湿地对总磷的去除率高达92％，对总氮的去除率为63％；当冲击负荷流量增大后，处理系统面积占汇水面积的比例减小到2％时，芦苇床对总氮和总磷的去除率分别为38．6％和52．6％。在相同水力停留时间下，潜流湿地的冲击试验与前期的连续流试验结果相比，冲击负荷会降低系统的脱氮除磷率，降低的脱氮率〈10％，降低的除磷率在15％左右。     

Mesocosm study of enhanced bioretention media in treating nutrient rich stormwater for mixed development area

A well-designed engineered soil for bioretention is important as it ensures that pollutant removal requirements are met. This laboratory study investigated the nutrient removal efficiency of bioretention media enhanced with 10% (by volume) additives from various waste materials (cockle shell, newspaper, printed paper, coconut husk and tyre crumb) and planted with Red Hot Chinese Hibiscus (Hibiscus rosa-sinensis), a common landscape shrub in tropical countries. The results showed that media enhanced with shredded newspaper demonstrated a significant improvement in total nitrogen (TN) removal (80.4%), compared to standard bioretention media (57.5%) without compromising total suspended solids (TSS) and total phosphorus (TP) removal, when dosed with actual runoff. The thick root system and rapid growth rate of the plant was proven to contribute to TN removal. This study concluded that shredded newspaper can be a potential addition to enhance bioretention media performance in treating stormwater, especially nutrient rich runoff from mixed development areas.     

Environmental factors influencing microcystin distribution and concentration in the Midwestern United States

During May–September 2000–2001, physicochemical data were collected from 241 lakes in Missouri, Iowa, northeastern Kansas, and southern Minnesota U.S.A., to determine the environmental variables associated with high concentrations of the cyanobacterial hepatotoxin microcystin (MC). The study region represents a south–north latitudinal gradient in increasing trophic status, with total phosphorus (TP) and total nitrogen (TN) values ranging between 2–995 and 90–15870 μg/L, respectively. Particulate MC values, measured by ELISA, ranged from undetectable to 4500 ng/L and increased with increasing latitude. Despite latitudinal trends, environmental variables explained <50% of the variation in MC values. Inspection of MC–TN and MC–Secchi bivariate plots revealed distinctly nonlinear trends, suggesting optima for maximum MC values. Nonlinear interval maxima regression indicated that MC–TN maxima were characterized by a unimodal curve, with maximal (>2000 ng/L) MC values occurring between 1500 and 4000 μg/L TN. Above 8000 μg/L TN all MC values were <150 ng/L. MC–Secchi maxima were characterized by exponential decline, with maximal MC values occurring at Secchi depths <2.5 m. The development of empirical relationships between environmental variables and MC values is critical to effective lake management and minimization of human health risks associated with the toxin. This study indicates MC values are linked to the physicochemical environment; however, the relationships are not traditional linear models.     

The 20th century whole-basin trophic history of an inter-drumlin lake in an agricultural catchment

Eight 1-m sediment cores were extracted from across the basin of Friary Lough, a 5.4-ha eutrophic lake in a wholly grassland agricultural catchment in Co. Tyrone, Northern Ireland. Sedimentary TP, diatom inferred TP, Ca, Na, Fe, Mn, loss-on-ignition (LOI), dry weight and density were determined in the core profiles. Core dating and correlation gave a 210Pb, 137Cs and 241Am chronology from 1906 to 1995 and enabled a whole-basin estimate of chemical and sediment accumulation rate over the 20th Century. The major changes for all parameters occurred after c. 1946. Sediment accumulation rate was most influenced by organic matter accumulations, probably of planktonic origin, and increasing after c. 1946. Inorganic sediment accumulation rate was found to be largely unchanging through the century at 10 t km(-2) yr(-1) when expressed as catchment exports. All chemical accumulation rate changes occurred after c. 1946. Total phosphorus accumulation rate, however, was found to be the only chemical to be increasing throughout the epilimnion and hypolimnion areas of the sedimentary basin at an average of 22.5 mg m(-2) yr(-1) between 1946 and 1995. The other chemical parameters showed increasing accumulation rates after c. 1946 in the epilimnion part of the basin only. Interpreted in terms of whole-basin sedimentation and catchment export processes over time, it is suggested that diffuse TP inputs are independent of sediment inputs. This corresponds to hydrochemical models that suggest soluble P as the primary fraction that is lost from grassland catchments. The increase in sedimentary TP accumulation rate, and DI-TP concentration, are also explained with regard to current models that suggest increases in runoff P concentrations from elevated soil P concentrations. Increases in eplimnion chemical and sediment accumulation rate after c. 1946 may be due to local erosion that has limited impact on lake basin sedimentation.