Non-point source pollution estimation in the Pingqiao River Basin,China, using a spatial hydrograph-separation approach

For efficient and targeted management, this study demonstrates a recently developed non-point source (NPS) pollution model for a year-long estimation in the Pingqiao River Basin (22.3 km²) in China. This simple but physically reasonable model estimates NPS export in terms of land use by reflecting spatial hydrological features and source runoff measurements under different land-use types. The NPS export was separately analysed by a distributed hydrological model, a spatial hydrograph-separation technique, and an empirical water quality sub-model. Simulation results suggest that 57 890 kg of total nitrogen (TN) and 1148 kg of total phosphorus (TP) were delivered. The results, validated with observed stream concentrations, show relative errors of 23.3% for TN and 47.4% for TP. Countermeasures for urban areas (5.3% of total area) were prioritized because of the high contribution rate to TN (14.1%) and TP (26.2%) which is caused by the high degree of runoff (8.5%) and pollution source.     

Contribution of baseflow nitrate export to non-point source pollution

As a common pollutant of nitrogen in groundwater, nitrate contamination has become a major concern worldwide. Baseflow, one of the dominant hydrological pathways for nitrate migration to streamflow, has been confirmed as a leading nitrate source for stream water where groundwater or subsurface flow contaminated heavily by nitrate. That is, sufficient improvements of water quality may not be attained without proper management for baseflow, even if non-point sources (NPS) pollutants discharged through surface runoff are being well managed. This article reviews the primary nitrate sources, the main factors affecting its transport, and the methodologies for baseflow nitrate estimation, to give some recommendations for future works, including: (1) giving sufficient consideration for the effects of climatological, morphological, and geological factors on baseflow recessions to obtain more reliable and accurate baseflow separation; (2) trying to solve calibration and validation problems for baseflow loads determining in storm flow period; (3) developing a simple and convenient algorithm with certain physics that can be used to separate baseflow NPS pollution from the total directly in different regions, for a reliable estimation of baseflow NPS pollution at larger scale (e.g., national scale); (4) improving groundwater quality simulation module of existing NPS pollution models to have a better simulation for biogeochemical processes in shallow aquifers; (5) taking integrated measures of “source control”, “process interception” and “end remediation” to prevent and control NPS nitrate pollution effectively, not just only the strict control of nutrients loss from surface runoff.     

Non‐point source pollution estimation using a modified approach

Non‐point source (NPS) pollution from agricultural land is increasing exponentially in many countries of the world, including India. A modified approach based on the conservation of mass and reaction kinetics has been derived to estimate the inflow of non‐point source pollutants from a river reach. Two water quality variables, namely, nitrate (NO₃) and ortho‐phosphate (o‐PO₄), which are main contributors as non‐point source pollution, were monitored at four locations of River Kali, western Uttar Pradesh, India, and used for calibration and validation of the model. Extensive water quality sampling was done with a total of 576 field data sets collected during the period from March 1999 to February 2000. Remote sensing and geographical information system (GIS) techniques were used to obtain land use/land cover of the region, digital elevation model (DEM), delineation of basin area contributing to non‐point source pollution at each sampling location and drainage map. The results obtained from a modified approach were compared with the existing mass‐balance equations and distributed modelling, and the performances of different equations were evaluated using error estimation viz. standard error, normal mean error, mean multiplicative error and correlation statistics. The developed model for the River Kali minimizes error estimates and improves correlation between observed and computed NPS loads. Copyright © 2007 John Wiley & Sons, Ltd.     

Nonlinear bacterial load‐streamflow response on a marine beach

Bacterial concentration (Escherichia coli) is used as the key indicator for marine beach water quality in Hong Kong. For beaches receiving streamflow from unsewered catchments, water quality is mainly affected by local nonpoint source pollution and is highly dependent on the bacterial load contributed from the catchment. As most of these catchments are ungauged, the bacterial load is generally unknown. In this study, streamflow and the associated bacterial load contributed from an unsewered catchment to a marine beach, Big Wave Bay, are simulated using a modelling approach. The physically based distributed hydrological model, MIKE‐SHE, and the empirical watershed water quality model (Hydrological Simulation Program – Fortran) are used to simulate streamflow and daily‐averaged E. coli concentration/load, respectively. The total daily derived loads predicted by the model during calibration (June–July 2007) and validation (July–October 2008) periods agree well with empirical validation data, with a percentage difference of 3 and 2%, respectively. The simulation results show a nonlinear relationship between E. coli load and rainfall/streamflow and reveal a source limiting nature of nonpoint source pollution. The derived load is further used as an independent variable in a multiple linear regression (MLR) model to predict daily beach water quality. When compared with the MLR models based solely on hydrometeorological input variables (e.g. rainfall and salinity), the new model based on bacterial load predicts much more realistic E. coli concentrations during rainstorms. Copyright © 2014 John Wiley & Sons, Ltd.     

NUTRIENT LOAD ESTIMATION METHODS FOR RIVERS

Pollutant load estimation is essential for watershed management and water pollution control. For most watersheds, only sparse water quality measurements (e.g. monthly data) are available. The influence of input data on the accuracy of non-point source pollution load estimation is studied using the water quality and stream flow data from a small watershed in Hong Kong. Comparison and analysis of the results using 8 different methods show that the accuracy of stream-flow runoff is the single most important factor for the calculation of pollutant load. Based on the results, the stream flow correction coefficient is advanced to provide a more reliable load estimation. The improved method of pollutant load estimation can be easily applied in practice since the stream-flow runoff can be measured by hydrological station or estimated with various hydrological methods.     

Pollution Source Investigation and Water Quality Management in the Carp Lake Watershed,Taiwan

In this study, a full survey of pollutant sources and water quality was conducted, followed by the application of a water quality model (Water Quality Analysis Simulation Program, WASP) to establish strategies of water quality control in Carp Lake, Taiwan. Results of the field investigation show that both point and non‐point source (NPS) pollutants were responsible for the poor water quality. The contributions of biochemical oxygen demand (BOD) from point source and NPS pollution were 45.9 and 55.1%, respectively. About 80% of total phosphorus (TP) were contributed by NPS. Additionally, point source and NPS pollution discharged 55.5 and 44.5% of NH₃–N load, respectively. The Carlson's Trophic State Index ranged from 61.9 to 69.2 showing serious eutrophic problems in Carp Lake. The calculated BOD, NH₃–N, and TP carrying capacity were approximately 2.8, 0.42, and 0.15 kg per day, respectively. However, the current pollutant loadings are approximately 3.0–5.5 times the calculated carrying capacity. With the help of the calibrated WASP model, remedial strategies for the lake water from short‐term to long‐term were developed. The completion of the small local sewer system to remove 80% of the point source pollution can serve as a short‐term goal while 40–60% of NPS removal by natural treatment systems may serve as a mid‐term goal. Furthermore, 80% of both source point and NPS pollution removal can be considered as a long‐term strategy. Results of heavy metal analysis show that the enriched sediment would be safe for agricultural applications.     

鄱阳湖湖泊流域系统水文水动力联合模拟

本文以鄱阳湖湖泊流域系统为研究对象,鉴于该湖泊流域系统尺度较大,下垫面自然属性呈现高度空间异质性且具有流域-平原区-湖泊不同机制的水文水动力过程,为了真实描述湖泊流域间的水文水动力联系及反映不同过程间的作用机制,构建了鄱阳湖湖泊流域联合模拟模型.该模型基于自主研发的流域分布式水文模型WATLAC和湖滨平原区产流模型以及水动力模型MIKE 21 3个不同功能子模型的连接来实现该复杂系统的模拟.模型的联合采用输入-输出驱动及子模型的顺序执行进程,即将五大子流域与平原区入湖径流量作为输入条件来驱动湖泊水动力模型,模拟湖泊水位对流域入湖径流量的响应.以2000-2005年鄱阳湖流域6个水文站点的河道径流量、流域基流指数以及湖泊4个站点的水位资料来率定模型,其中各站点日径流量拟合的纳希效率系数E_ns为0.71~0.84,确定性系数R²介于0.70~0.88之间,而湖泊各站点水位拟合的纳希效率系数E_ns变化为0.88~0.98,确定性系数R²为0.96~0.98,均取得令人满意的率定结果.本文提出的鄱阳湖湖泊流域系统水文水动力联合模拟模型能较为理想再现湖泊水位对流域降雨-径流过程的响应.水位模拟结果进一步表明,该联合模型能用来获取重要的水动力空间变化特征.该模型可作为有效工具定量揭示湖泊流域系统水文水动力过程对气候变化和流域人类活动的响应.     

Distributed hydrological modelling using weather radar in gauged and ungauged basins

Distributed hydrological modelling using space–time estimates of rainfall from weather radar provides a natural approach to area-wide flood forecasting and warning at any location, whether gauged or ungauged. However, radar estimates of rainfall may lack consistent, quantitative accuracy. Also, the formulation of hydrological models in distributed form may be problematic due to process complexity and scaling issues. Here, the aim is to first explore ways of improving radar rainfall accuracy through combination with raingauge network data via integrated multiquadric methods. When the resulting gridded rainfall estimates are employed as input to hydrological models, the simulated river flows show marked improvements when compared to using radar data alone. Secondly, simple forms of physical–conceptual distributed hydrological model are considered, capable of exploiting spatial datasets on topography and, where necessary, land-cover, soil and geology properties. The simplest Grid-to-Grid model uses only digital terrain data to delineate flow pathways and to control runoff production, the latter by invoking a probability-distributed relation linking terrain slope to soil absorption capacity. Model performance is assessed over nested river basins in northwest England, employing a lumped model as a reference. When the distributed model is used with the gridded radar-based rainfall estimators, it shows particular benefits for forecasting at ungauged locations.     

The non-point source pollution in livestock-breeding areas of the Heihe River basin in Yellow River

A GIS-based distributed soil and water assessment tools (SWAT) model was used to simulate the runoff, sediment yield and the load of the non-point source pollution in the Heihe River basin, which is a tributary and main water supply source of the Yellow River. It is a typical stockbreeding area, and its industry and agriculture are not well developed. The main pollution source of the Heihe River was livestock related non-point source pollution. With GIS and remote sensing techniques, a database of non-point source pollution in the Heihe River basin was established. The SWAT model was parameterized for this area. The pollution load and transportation rules such as nitrogen were illustrated. After several years of hard work, the situations of point source pollution were more and more accurate. This paper provided an effective way to assess and calculate the pollution load in the wide agriculture area in China. With the help of historical data formulated parameters, the non-point source load and the theory of pollution load distribution were illustrated about the Heihe River basin. In 2000, the soluble N load in this area was 1.06 × 106 kg. By the simulation, the main pollution sources were in the south east of the basin, where the pasturing areas located in the south-east of Ruoergai County and in north of the Hongyuan County.     

Using a topographic index to distribute variable source area runoff predicted with the SCS curve‐number equation

Because the traditional Soil Conservation Service curve‐number (SCS‐CN) approach continues to be used ubiquitously in water quality models, new application methods are needed that are consistent with variable source area (VSA) hydrological processes in the landscape. We developed and tested a distributed approach for applying the traditional SCS‐CN equation to watersheds where VSA hydrology is a dominant process. Predicting the location of source areas is important for watershed planning because restricting potentially polluting activities from runoff source areas is fundamental to controlling non‐point‐source pollution. The method presented here used the traditional SCS‐CN approach to predict runoff volume and spatial extent of saturated areas and a topographic index, like that used in TOPMODEL, to distribute runoff source areas through watersheds. The resulting distributed CN–VSA method was applied to two subwatersheds of the Delaware basin in the Catskill Mountains region of New York State and one watershed in south‐eastern Australia to produce runoff‐probability maps. Observed saturated area locations in the watersheds agreed with the distributed CN–VSA method. Results showed good agreement with those obtained from the previously validated soil moisture routing (SMR) model. When compared with the traditional SCS‐CN method, the distributed CN–VSA method predicted a similar total volume of runoff, but vastly different locations of runoff generation. Thus, the distributed CN–VSA approach provides a physically based method that is simple enough to be incorporated into water quality models, and other tools that currently use the traditional SCS–CN method, while still adhering to the principles of VSA hydrology. Copyright © 2004 John Wiley & Sons, Ltd.     

Assessing effects of model complexity and structure on predictions of hydrological responses using serial and parallel model design

By utilizing functional relationships based on observations at plot or field scales, water quality models first compute surface runoff and then use it as the primary governing variable to estimate sediment and nutrient transport. When these models are applied at watershed scales, this serial model structure, coupling a surface runoff sub-model with a water quality sub-model, may be inappropriate because dominant hydrological processes differ among scales. A parallel modeling approach is proposed to evaluate how best to combine dominant hydrological processes for predicting water quality at watershed scales. In the parallel scheme, dominant variables of water quality models are identified based entirely on their statistical significance using time series analysis. Four surface runoff models of different model complexity were assessed using both the serial and parallel approaches to quantify the uncertainty on forcing variables used to predict water quality. The eight alternative model structures were tested against a 25-year high-resolution data set of streamflow, suspended sediment discharge, and phosphorous discharge at weekly time steps. Models using the parallel approach consistently performed better than serial-based models, by having less error in predictions of watershed scale streamflow, sediment and phosphorus, which suggests model structures of water quantity and quality models at watershed scales should be reformulated by incorporating the dominant variables. The implication is that hydrological models should be constructed in a way that avoids stacking one sub-model with one set of scale assumptions onto the front end of another sub-model with a different set of scale assumptions.     

The composition of nutrient fluxes from contrasting UK river basins

Accurate estimates of N and P loads were obtained for four contrasting UK river basins over a complete annual cycle. The fractionation of these loads into dissolved and particulate, and inorganic and organic components allowed a detailed examination of the nutrient load composition and of the factors influencing both the relative and absolute magnitude of these components. The particulate phosphorus (TPP) loads account for 26–75% of the annual total phosphorus (TP) transport and are predominantly inorganic. The inorganic (PIP) and organic (POP) fractions of the TPP loads represent 20–47% and 6–28% of the annual TP transport, respectively. In contrast, the particulate nitrogen loads (TPN) represent 8% or less of the annual total nitrogen (TN) loads and are predominately organic. For dissolved P transport, the dissolved inorganic fraction (DIP) is more important, representing 15–70% of the TP loads, whereas the dissolved organic fraction (DOP) represents only 3–9% of the TP loads. The TN loads are dominated by the dissolved component and more particularly the total oxidized fraction (TON), which is composed of nitrate and nitrite and represents 76–82% of the annual TN transport. The remaining dissolved N species, ammonium (NH₄-N) and organic N (DON) account for 0·3–1·2% and 13–16% of the annual TN transport, respectively. The TPN and TPP fluxes closely reflect the suspended sediment dynamics of the study basins, which are in turn controlled by basin size and morphology. The dissolved inorganic nutrient fluxes are influenced by point source inputs to the study basins, especially for P, although the TON flux is primarily influenced by diffuse source contributions and the hydrological connectivity between the river and its catchment area. The dissolved organic fractions are closely related to the dissolved organic carbon (DOC) dynamics, which are in turn influenced by land use and basin size. The magnitude of the NH₄-N fraction was dependent on the proximity of the monitoring station to point source discharges, because of rapid nitrification within the water column. However, during storm events, desorption from suspended sediment may be temporarily important. Both the magnitude and relative contribution of the different nutrient fractions exhibit significant seasonal variability in response to the hydrological regime, sediment mobilization, the degree of dilution of point source inputs and biological processes. © 1998 John Wiley & Sons, Ltd.     

Non‐point source pollution modelling using Soil and Water Assessment Tool and its parameter sensitivity analysis in Xin'anjiang catchment,China

Non‐point source pollution is a key issue in integrated river basin management around the world and has resulted in water contamination, aquatic ecology deterioration and eutrophication. Xin'anjiang catchment is the key drinking water source area for Hangzhou City, China. A promising model (Soil and Water Assessment Tool) was applied to assess the non‐point source pollution and its effect on drinking water. Sensitivity analysis of model parameters was carried out using the Sequential Uncertainty Domain Parameter Fitting 2 sensitivity technique. Water discharge, sediment, total nitrogen and total phosphorus load processes from 2000 to 2010 were simulated, and the spatial distributions of non‐point source pollutants were evaluated at the catchment and administrative country levels. The results show that the hydrological parameters of the Soil and Water Assessment Tool were dominantly sensitive for non‐point source pollution simulation, including CN2, RCHRG_DP, ALPHA_BF, SOL_AWC, ESCO and SOL_K and the characteristic parameters of sub‐basins (viz. HRU_SLP and SLSUBBSN). Also, water quality parameters (viz. CH_EROD, NPERCO, RSDCO and PPERCO, PHOSKD, etc.) have a significant effect on nutrients. The model performance was very satisfactory, especially for runoff, sediment and total phosphorus simulation. The non‐point source pollutant load increased from 2001 to 2010 in the whole catchment. Total nitrogen load increased from 3428 tons (0.59 ton km^?2) to 7315 tons (1.25 ton km^?2), and total phosphorus load increased from 299 tons (0.05 ton km^?2) to 867 tons (0.15 ton km^?2). The contribution of rice land was the largest, accounting for nearly 95%, followed by tea garden (3.56%), winter wheat (1.37%), forest (0.07%) and grassland (0.02%). Moreover, She County and Xiuning County contributed more than half of the non‐point source pollutants. This study was expected to provide a method and reference for non‐point source pollution quantification and to support water quality management implementation in China. Copyright © 2012 John Wiley & Sons, Ltd.     

Probabilistic calibration of a distributed hydrological model for flood forecasting

Abstract

The complexity of distributed hydrological models has led to improvements in calibration methodologies in recent years. There are various manual, automatic and hybrid methods of calibration. Most use a single objective function to calculate estimation errors. The use of multi-objective calibration improves results, since different aspects of the hydrograph may be considered simultaneously. However, the uncertainty of estimates from a hydrological model can only be taken into account by using a probabilistic approach. This paper presents a calibration method of probabilistic nature, based on the determination of probability functions that best characterize different parameters of the model. The method was applied to the Real-time Interactive Basin Simulator (RIBS) distributed hydrological model using the Manzanares River basin in Spain as a case study. The proposed method allows us to consider the uncertainty in the model estimates by obtaining the probability distributions of flows in the flood hydrograph.

Citation Mediero, L., Garrote, L. & Martín-Carrasco, F. J. (2011) Probabilistic calibration of a distributed hydrological model for flood forecasting. Hydrol. Sci. J. 56(7), 1129–1149.     

Integration of gauge and radar rainfall to enable best simulation of hydrological parameters

This study is about use of spatially distributed rain in physically based hydrological models. In recent years, spatially distributed radar rainfall data have become available. The distributed radar rain is used to precisely model hydrologic processes and it is more realistic than the past practice of distribution methods like Thiessen polygons. Radar provides a highly accurate spatial distribution of rainfall and greatly improves the basin average rainfall estimates. However, quantification of the exact amount of rainfall from radar observation is relatively difficult. Thus, the fundamental idea of this study is to apply hourly gauge and radar rainfall data in a distributed hydrological model to simulate hydrological parameters. Hence the comparison is made between the outcomes of the WetSpa model from radar rainfall distribution and gauge rainfall distributed by the Thiessen polygon technique. The comparative plots of the hydrograph and the results of hydrological components such as evapotranspiration, surface runoff, soil moisture, recharge and interflow, reflect the spatially distributed radar input performing well for model outflow simulation.

EDITOR D. Koutsoyiannis

ASSOCIATE EDITOR F. Pappenberger     

Estimating the discharge of contaminants to coastal waters by rivers: Some cautionary comments

Some of the problems associated with the estimation of contaminant loads transported by rivers using infrequent samples are described. The reliability of many load estimates, particularly those for substances where particulate-associated transport predominates, is questioned. A case study of the accuracy and precision of estimates of the suspended sediment load of the River Exe at Thorverton for the period 1978–1980 is used to demonstrate the potential reliability of loads calculated using a variety of estimation procedures. Underestimation and lack of precision are shown to be important problems.     

Hydraulic characteristics and habitat suitability in presence of woody debris: A flume experiment

Data from a flume experiment were used to explore the modified hydraulic conditions and habitat suitability in streams where feeding of large woody debris (LWD) is present. Feeding of LWD was simulated by insertion of wood dowels with varying diameter and length. Two processes were mimicked, namely (i) lumped LWD load, and (ii) distributed LWD load. Lumped load may occur for wood coming either from upstream or from a tributary, and entering the stream of interest in one only section. Distributed load occurs for wood entering along the considered stream, in several sections. Distributed wood income resulted in homogeneously increased bed roughness, leading to increased flow depth and decreased velocity, whereas lumped input of wood from upstream resulted in larger local clustering and change of the flow properties, but with less influence on the distributed hydraulic properties. A method is proposed to predict bulk flow properties in presence of LWD. Then, a simple approach is used based upon the concept of wetter usable area WUA to investigate modified habitat conditions for fish species in presence of woody debris. An application to a real world case study from the literature is then shown, where increasing density of wood increases habitat availability for colonization by fish guilds.     

Modeling of non‐point source nitrogen pollution from 1979 to 2008 in Jiaodong Peninsula,China

Efforts to reduce land‐based non‐point source (NPS) pollutions from watersheds to coastal waters are ongoing all around the world. In this study, annual yield of NPS nitrogen (NPS‐N) pollution in Jiaodong Peninsula, China from 1979 to 2008 was estimated. The results showed that: from 1979 to 2008, NPS‐N yields exhibited significant inter‐annual variations and an increasing trend on decadal scale. High NPS‐N yield was mainly found in east and south parts, as well as the urbanized coastal regions in Jiaodong Peninsula. Among the 32 river basins, the three largest basins yielded more than 41.16% of the NPS‐N. However, some small coastal watersheds along the South Yellow Sea and Jiaozhou Bay had higher per unit area yield. Most of the small watersheds characterized by seasonal runoff had coastal waters pertain to mild and moderate pollution levels. The ratio of watershed area to shoreline length and the up‐stream land use had significant impacts on NPS‐N flux through the shoreline. Among the four adjacent coastal areas of Jiaodong Peninsula, Jiaozhou Bay was the most noteworthy one not only because of high levels of land‐based NPS‐N pollution but also because of its nearly enclosed structure. The combination between integrated coastal zone management and integrated river basin management, land use planning and landscape designing in Jiaodong Peninsula is recommended. Copyright © 2013 John Wiley & Sons, Ltd.     

基于改进SWAT模型的灌溉-施肥-耕作对乌梁素海流域营养物负荷及作物产量的影响

不合理的灌溉、施肥和耕作是导致乌梁素海流域农业面源污染的主要根源,乌梁素海作为我国北方地区重要的生态安全屏障,多年来面临着湖泊水环境污染、水生态退化等问题,科学开展湖泊水环境综合治理首先要解决流域内农业面源污染问题. 研究通过修改土壤水平衡、溶质平衡、地下水平衡和作物生长等模块对SWAT 2012原始版本进行改进,并采用改进的SWAT模型构建了乌梁素海流域分布式水文模型,利用实测径流、硝态氮与总磷排放量、地下水埋深以及作物产量校正和验证模型. 基于现状情景,以玉米、葵花和小麦3种主要作物为研究对象,设置了削减灌水量、施肥量及调整耕作方式3种农田管理情景. 基于改进SWAT模型不同情景的模拟结果,计算分析各管理情景下的硝态氮与总磷负荷及对各作物产量的影响. 结果表明,改进SWAT模型具有良好的模拟效果. 不同作物削减5%夏灌水量增产最多达8.41%~10.32%,削减10%秋浇水量均明显减少硝态氮和总磷负荷. 不同作物营养物负荷均随着氮磷施肥削减比例的增大呈现逐渐降低的趋势,但下降曲线逐渐趋于平缓; 各作物产量随氮磷施肥削减比例的增加呈先增加后减少的趋势,其中玉米、小麦氮磷施肥削减比例达20%时产量开始下降,葵花氮磷施肥削减比例达25%时产量开始下降. 不同作物营养物负荷与小麦产量均随耕作方式混合深度与混合效率参数的增大逐渐减小,而玉米和葵花产量则随耕作参数增大逐渐增加. 综合分析,削减5%夏灌水量+削减20%氮磷施肥比例+模板犁耕作组合玉米产量增幅最大达36.5%;削减10%秋浇水量+削减25%氮磷施肥比例+模板犁耕作组合葵花硝态氮负荷降幅最大达42.1%;削减5%夏灌水量+削减20%氮磷施肥比例+免耕组合小麦产量增幅最大达29.1%;而削减5%秋浇水量+削减20%氮磷施肥比例+常规春耕组合小麦硝态氮负荷减少最大达27.2%,总磷负荷减少最大达18.5%. 本研究可为降低流域内面源污染、提高作物产量及减少乌梁素海营养物入湖负荷农业管理措施的实施提供理论依据.     

Hydrological impact of roadside ditches in an agricultural watershed in Central New York: implications for non‐point source pollutant transport

Nonpoint source pollution and hydromodification are the leading causes of impairment to our nation's rivers and streams. Roadside ditch networks, ubiquitous in both rural and urban landscapes, intercept and shunt substantial quantities of overland runoff and shallow groundwater to stream systems. By altering natural flowpaths, road ditches contribute not only to hydromodification but also potentially to nonpoint‐source (NPS) pollution by acting as hydrological links between agricultural fields and natural streams. Unfortunately, the impacts of these alterations on watershed hydrology and water quality are not well understood. Through a series of field measurements, including field surveys and discharge monitoring, this study examined the effect of road ditch networks on basin morphometry, field‐ and watershed‐scale hydrology, and pollutant transport in a 38 km² agricultural watershed in south‐central NY. Salient findings include the following: (i) 94% of road ditches discharged to natural streams, effectively doubling the drainage density; (ii) on average, road ditches increased peak and total event flows in their receiving streams by 78% and 57%, respectively, but displayed significant variation across ditches; and (iii) ditches intercepted large quantities of surface and subsurface runoff from agricultural fields and therefore represent efficient conduits for the transport of agricultural NPS pollutants to sensitive receiving waterbodies. Our results provide useful information for hydrologists who wish to further understand how artificial drainage may be affecting watershed hydrology and for managers and engineers tasked with designing appropriate flood and NPS pollution control measures. Copyright © 2012 John Wiley & Sons, Ltd.