River toxicity assessment using molecular biosensors: Heavy metal contamination in the Turag-Balu-Buriganga river systems, Dhaka, Bangladesh.

Pollution in rapidly urbanising cities and in delta systems is a serious problem that blights the lives and livelihoods of millions of people, damaging and restricting potable water supply and supplies to industry (Whitehead et al, 2015, 2018). Employing new technology based on luminescent molecular biosensors, the toxicity in the rivers around Dhaka in Bangladesh, namely the Turag, Tongi, Balu and Buriganga, has been assessed. Samples taken at 36 sites during medium and low flow conditions and during the Bishwa Ijtema Festival revealed high levels of cell toxicity, as well as high concentrations of metals, particularly aluminium, cadmium, chromium, iron, zinc, lithium, selenium and nickel. Chemical analysis also revealed low dissolved oxygen levels and anoxic conditions in the rivers at certain sites. The bacterial molecular biosensors were demonstrated to be fast, with results in 30 min, robust and a highly sensitive method for the assessment of water toxicity in the field. Furthermore, the biosensor toxicity analysis correlated with the metals data, and a multivariate regression relationship was developed relating toxicity to key metals, such a selenium, zinc and chromium. The resulting model has been validated against split samples and the Bishwa Ijtema Festival data. The combination of modelling and the molecular biosensor technology provides a new approach to detecting and managing pollution in urban river systems.

Optimizing land management strategies for maximum improvements in lake dissolved oxygen concentrations.

Eutrophication and anoxia are unresolved issues in many large waterbodies. Globally, management success has been inconsistent, highlighting the need to identify approaches which reliably improve water quality. We used a process-based model chain to quantify effectiveness of terrestrial nutrient control measures on in-lake nitrogen, phosphorus, chlorophyll and dissolved oxygen (DO) concentrations in Lake Simcoe, Canada. Across a baseline period of 2010-2016 hydrochemical outputs from catchment models INCA-N and INCA-P were used to drive the lake model PROTECH, which simulated water quality in the three main basins of the lake. Five terrestrial nutrient control strategies were evaluated. Effectiveness differed between catchments, and water quality responses to nutrient load reductions varied between deep and shallow lake basins. Nutrient load reductions were a significant driver of increased DO concentrations, however strategies which reduced tributary inflow had a greater impact on lake restoration, associated with changes in water temperature and chemistry. Importantly, when multiple strategies were implemented simultaneously, resultant large flow reductions induced warming throughout the water column. Negative impacts of lake warming on DO overwhelmed the positive effects of nutrient reduction, and limited the effectiveness of lake restoration strategies. This study indicates that rates of lake recovery may be accelerated through a coordinated management approach, which considers strategy interactions, and the potential for temperature change-induced physical and biological feedbacks. Identified impacts of flow and temperature on rates of lake recovery have implications for management sustainability under a changing climate.

Modelling heavy metals in the Buriganga River System, Dhaka, Bangladesh: Impacts of tannery pollution control.

Heavy metal pollution from tanneries is a global problem in many rapidly developing economies. Effluent discharges into rivers cause serious problems for water quality, damaging ecology and threatening the livelihoods of people, especially in developing urban centres which often have a high concentration of factories. The industry intensive capital area of Bangladesh is impacted with high levels of metals pollution in rivers in the Greater Dhaka Watershed. Sampling and modelling studies have been undertaken to assess pollution in the Buriganga River System in Dhaka. The process based, dynamic model INCA (Integrated Catchments) model has been used to simulate metals along the Buriganga River System in Central Dhaka. Observed and simulated metals concentrations are high, and the model shows that the proposed transfer of the tannery industry upstream helps to reduce the pollution significantly downstream. However, moving the industry upstream may be counterproductive as it is discharged into the upper reaches of the river. This will create pollution upstream unless the newly constructed effluent treatment system can operate at a high level.

Water quality modelling of the Mekong River basin: Climate change and socioeconomics drive flow and nutrient flux changes to the Mekong Delta.

The Mekong delta is recognised as one of the world's most vulnerable mega-deltas, being subject to a range of environmental pressures including sea level rise, increasing population, and changes in flows and nutrients from its upland catchment. With changing climate and socioeconomics there is a need to assess how the Mekong catchment will be affected in terms of the delivery of water and nutrients into the delta system. Here we apply the Integrated Catchment model (INCA) to the whole Mekong River Basin to simulate flow and water quality, including nitrate, ammonia, total phosphorus and soluble reactive phosphorus. The impacts of climate change on all these variables have been assessed across 24 river reaches ranging from the Himalayas down to the delta in Vietnam. We used the UK Met Office PRECIS regionally coupled climate model to downscale precipitation and temperature to the Mekong catchment. This was accomplished using the Global Circulation Model GFDL-CM to provide the boundary conditions under two carbon control strategies, namely representative concentration pathways (RCP) 4.5 and a RCP 8.5 scenario. The RCP 4.5 scenario represents the carbon strategy required to meet the Paris Accord, which aims to limit peak global temperatures to below a 2 °C rise whilst seeking to pursue options that limit temperature rise to 1.5 °C. The RCP 8.5 scenario is associated with a larger 3-4 °C rise. In addition, we also constructed a range of socio-economic scenarios to investigate the potential impacts of changing population, atmospheric pollution, economic growth and land use change up to the 2050s. Results of INCA simulations indicate increases in mean flows of up to 24%, with flood flows in the monsoon period increasing by up to 27%, but with increasing periods of drought up to 2050. A shift in the timing of the monsoon is also simulated, with a 4 week advance in the onset of monsoon flows on average. Decreases in nitrogen and phosphorus concentrations occur primarily due to flow dilution, but fluxes of these nutrients also increase by 5%, which reflects the changing flow, land use change and population changes.

Estimating uncertainty in terrestrial critical loads and their exceedances at four sites in the UK.

Critical loads are the basis for policies controlling emissions of acidic substances in Europe and elsewhere. They are assessed by several elaborate and ingenious models, each of which requires many parameters, and have to be applied on a spatially-distributed basis. Often the values of the input parameters are poorly known, calling into question the validity of the calculated critical loads. This paper attempts to quantify the uncertainty in the critical loads due to this "parameter uncertainty", using examples from the UK. Models used for calculating critical loads for deposition of acidity and nitrogen in forest and heathland ecosystems were tested at four contrasting sites. Uncertainty was assessed by Monte Carlo methods. Each input parameter or variable was assigned a value, range and distribution in an objective a fashion as possible. Each model was run 5000 times at each site using parameters sampled from these input distributions. Output distributions of various critical load parameters were calculated. The results were surprising. Confidence limits of the calculated critical loads were typically considerably narrower than those of most of the input parameters. This may be due to a "compensation of errors" mechanism. The range of possible critical load values at a given site is however rather wide, and the tails of the distributions are typically long. The deposition reductions required for a high level of confidence that the critical load is not exceeded are thus likely to be large. The implication for pollutant regulation is that requiring a high probability of non-exceedance is likely to carry high costs. The relative contribution of the input variables to critical load uncertainty varied from site to site: any input variable could be important, and thus it was not possible to identify variables as likely targets for research into narrowing uncertainties. Sites where a number of good measurements of input parameters were available had lower uncertainties, so use of in situ measurement could be a valuable way of reducing critical load uncertainty at particularly valuable or disputed sites. From a restricted number of samples, uncertainties in heathland critical loads appear comparable to those of coniferous forest, and nutrient nitrogen critical loads to those of acidity. It was important to include correlations between input variables in the Monte Carlo analysis, but choice of statistical distribution type was of lesser importance. Overall, the analysis provided objective support for the continued use of critical loads in policy development.

Modelling metaldehyde in catchments: a River Thames case-study.

The application of metaldehyde to agricultural catchment areas to control slugs and snails has caused severe problems for drinking water supply in recent years. In the River Thames catchment, metaldehyde has been detected at levels well above the EU and UK drinking water standards of 0.1 µg l-1 at many sites across the catchment between 2008 and 2015. Metaldehyde is applied in autumn and winter, leading to its increased concentrations in surface waters. It is shown that a process-based hydro-biogeochemical transport model (INCA-contaminants) can be used to simulate metaldehyde transport in catchments from areas of application to the aquatic environment. Simulations indicate that high concentrations in the river system are a direct consequence of excessive application rates. A simple application control strategy for metaldehyde in the Thames catchment based on model results is presented.

Impacts of climate change on in-stream nitrogen in a lowland chalk stream: an appraisal of adaptation strategies.

The impacts of climate change on nitrogen (N) in a lowland chalk stream are investigated using a dynamic modelling approach. The INCA-N model is used to simulate transient daily hydrology and water quality in the River Kennet using temperature and precipitation scenarios downscaled from the General Circulation Model (GCM) output for the period 1961-2100. The three GCMs (CGCM2, CSIRO and HadCM3) yield very different river flow regimes with the latter projecting significant periods of drought in the second half of the 21st century. Stream-water N concentrations increase over time as higher temperatures enhance N release from the soil, and lower river flows reduce the dilution capacity of the river. Particular problems are shown to occur following severe droughts when N mineralization is high and the subsequent breaking of the drought releases high nitrate loads into the river system. Possible strategies for reducing climate-driven N loads are explored using INCA-N. The measures include land use change or fertiliser reduction, reduction in atmospheric nitrate and ammonium deposition, and the introduction of water meadows or connected wetlands adjacent to the river. The most effective strategy is to change land use or reduce fertiliser use, followed by water meadow creation, and atmospheric pollution controls. Finally, a combined approach involving all three strategies is investigated and shown to reduce in-stream nitrate concentrations to those pre-1950s even under climate change.

Fate and transport of polychlorinated biphenyls (PCBs) in the River Thames catchment - Insights from a coupled multimedia fate and hydrobiogeochemical transport model.

The fate of persistent organic pollutants (POPs) in riverine environments is strongly influenced by hydrology (including flooding) and fluxes of sediments and organic carbon. Coupling multimedia fate models (MMFMs) and hydrobiogeochemical transport models offers unique opportunities for understanding the environmental behaviour of POPs. While MMFMs are widely used for simulating the fate and transport of legacy and emerging pollutants, they use greatly simplified representations of climate, hydrology and biogeochemical processes. Using additional information about weather, river flows and water chemistry in hydrobiogeochemical transport models can lead to new insights about POP behaviour in rivers. As most riverine POPs are associated with suspended sediments (SS) or dissolved organic carbon (DOC), coupled models simulating SS and DOC can provide additional insights about POPs behaviour. Coupled simulations of river flow, DOC, SS and POP dynamics offer the possibility of improved predictions of contaminant fate and fluxes by leveraging the additional information in routine water quality time series. Here, we present an application of a daily time step dynamic coupled multimedia fate and hydrobiogeochemical transport model (The Integrated Catchment (INCA) Contaminants model) to simulate the behaviour of selected PCB congeners in the River Thames (UK). This is a follow-up to an earlier study where a Level III fugacity model was used to simulate PCB behaviour in the Thames. While coupled models are more complex to apply, we show that they can lead to much better representation of POPs dynamics. The present study shows the importance of accurate sediment and organic carbon simulations to successfully predict riverine PCB transport. Furthermore, it demonstrates the important impact of short-term weather variation on PCB movement through the environment. Specifically, it shows the consequences of the severe flooding, which occurred in early 2014 on sediment PCB concentrations in the River Thames.

An INCA model for pathogens in rivers and catchments: Model structure, sensitivity analysis and application to the River Thames catchment, UK.

Pathogens are an ongoing issue for catchment water management and quantifying their transport, loss and potential impacts at key locations, such as water abstractions for public supply and bathing sites, is an important aspect of catchment and coastal management. The Integrated Catchment Model (INCA) has been adapted to model the sources and sinks of pathogens and to capture the dominant dynamics and processes controlling pathogens in catchments. The model simulates the stores of pathogens in soils, sediments, rivers and groundwaters and can account for diffuse inputs of pathogens from agriculture, urban areas or atmospheric deposition. The model also allows for point source discharges from intensive livestock units or from sewage treatment works or any industrial input to river systems. Model equations are presented and the new pathogens model has been applied to the River Thames in order to assess total coliform (TC) responses under current and projected future land use. A Monte Carlo sensitivity analysis indicates that the input coliform estimates from agricultural sources and decay rates are the crucial parameters controlling pathogen behaviour. Whilst there are a number of uncertainties associated with the model that should be accounted for, INCA-Pathogens potentially provides a useful tool to inform policy decisions and manage pathogen loading in river systems.

Bioremediation of acid mine drainage: an introduction to the Wheal Jane wetlands project.

Acid mine drainage (AMD) is a widespread environmental problem associated with both working and abandoned mining operations. As part of an overall strategy to determine a long-term treatment option for AMD, a pilot passive treatment plant was constructed in 1994 at Wheal Jane Mine in Cornwall, UK. The plant consists of three separate systems; each containing aerobic reed beds, anaerobic cell and rock filters, and represents the largest European experimental facility of its kind. The systems only differ by the type of pre-treatment utilised to increase the pH of the influent minewater (pH<4): lime-dosed (LD), anoxic limestone drain (ALD) and lime free (LF), which receives no form of pre-treatment. The Wheal Jane pilot plant offered a unique facility and a major research project was established to evaluate the pilot plant and study in detail the biological mechanisms and the geochemical and physical processes that control passive treatment systems. The project has led to data, knowledge, models and design criteria for the future design, planning and sustainable management of passive treatment systems. A multidisciplinary team of scientists and managers from the U.K. universities, the Environment Agency and the Mining Industry has been put together to obtain the maximum advantage from the excellent facilities facility at Wheal Jane.

The Wheal Jane wetlands model for bioremediation of acid mine drainage.

Acid mine drainage (AMD) is a widespread environmental problem associated with both working and abandoned mining operations. As part of an overall strategy to determine a long-term treatment option for AMD, a pilot passive treatment plant was constructed in 1994 at Wheal Jane Mine in Cornwall, UK. The plant consists of three separate systems, each containing aerobic reed beds, anaerobic cell and rock filters, and represents the largest European experimental facility of its kind. The systems only differ by the type of pretreatment utilised to increase the pH of the influent minewater (pH <4): lime dosed (LD), anoxic limestone drain (ALD) and lime free (LF), which receives no form of pretreatment. Historical data (1994-1997) indicate median Fe reduction between 55% and 92%, sulphate removal in the range of 3-38% and removal of target metals (cadmium, copper and zinc) below detection limits, depending on pretreatment and flow rates through the system. A new model to simulate the processes and dynamics of the wetlands systems is described, as well as the application of the model to experimental data collected at the pilot plant. The model is process based, and utilises reaction kinetic approaches based on experimental microbial techniques rather than an equilibrium approach to metal precipitation. The model is dynamic and utilises numerical integration routines to solve a set of differential equations that describe the behaviour of 20 variables over the 17 pilot plant cells on a daily basis. The model outputs at each cell boundary are evaluated and compared with the measured data, and the model is demonstrated to provide a good representation of the complex behaviour of the wetland system for a wide range of variables.

Chemical behaviour of the Wheal Jane bioremediation system.

The effectiveness of remediation of the highly acidic and transition metal polluted mine water discharge from the Wheal Jane Mine by the Wheal Jane Passive Treatment Plant is described. The success of the remediation required that all the system components work as predicted. The study shows considerable success in the removal of key toxic metals and clearly demonstrates the potential for natural attenuation of acid mine drainage, particularly iron oxidation, by microbial populations. The Wheal Jane Passive Treatment Plant provides the only experimental facility of its kind.

A review of arsenic and its impacts in groundwater of the Ganges-Brahmaputra-Meghna delta, Bangladesh.

Arsenic in drinking water is the single most important environmental issue facing Bangladesh; between 35 and 77 million of its 156 million inhabitants are considered to be at risk from drinking As-contaminated water. This dominates the list of stress factors affecting health, livelihoods and the ecosystem of the delta region. There is a vast literature on the subject so this review provides a filter of the more important information available on the topic. The arsenic problem arises from the move in the 1980s and 1990s by international agencies to construct tube wells as a source of water free of pathogens, groundwater usually considered a safe source. Since arsenic was not measured during routine chemical analysis and also is difficult to measure at low concentrations it was not until the late 1990s that the widespread natural anomaly of high arsenic was discovered and confirmed. The problem was exacerbated by the fact that the medical evidence of arsenicosis only appears slowly. The problem arises in delta regions because of the young age of the sediments deposited by the GBM river system. The sediments contain minerals such as biotite which undergo slow "diagenetic" reactions as the sediments become compacted, and which, under the reducing conditions of the groundwater, release in the form of toxic As(3+). The problem is restricted to sediments of Holocene age and groundwater of a certain depth (mainly 30-150 m), coinciding with the optimum well depth. The problem is most serious in a belt across southern Bangladesh, but within 50 m of the coast the problem is only minor because of use of deep groundwater; salinity in shallow groundwater here is the main issue for drinking water. The Government of Bangladesh adopted a National Arsenic Policy and Mitigation Action Plan in 2004 for providing arsenic safe water to all the exposed population, to provide medical care for those who have visible symptoms of arsenicosis. There is as yet no national monitoring program in place. Various mitigation strategies have been tested, but generally the numerous small scale technological remedies have proved unworkable at village level. The current statistics show that use of deep groundwater (below 150 m) is the main source of arsenic mitigation over most of the arsenic affected areas as well as rainwater harvesting in certain location.

Rainfall runoff modelling of the Upper Ganga and Brahmaputra basins using PERSiST.

There are ongoing discussions about the appropriate level of complexity and sources of uncertainty in rainfall runoff models. Simulations for operational hydrology, flood forecasting or nutrient transport all warrant different levels of complexity in the modelling approach. More complex model structures are appropriate for simulations of land-cover dependent nutrient transport while more parsimonious model structures may be adequate for runoff simulation. The appropriate level of complexity is also dependent on data availability. Here, we use PERSiST; a simple, semi-distributed dynamic rainfall-runoff modelling toolkit to simulate flows in the Upper Ganges and Brahmaputra rivers. We present two sets of simulations driven by single time series of daily precipitation and temperature using simple (A) and complex (B) model structures based on uniform and hydrochemically relevant land covers respectively. Models were compared based on ensembles of Bayesian Information Criterion (BIC) statistics. Equifinality was observed for parameters but not for model structures. Model performance was better for the more complex (B) structural representations than for parsimonious model structures. The results show that structural uncertainty is more important than parameter uncertainty. The ensembles of BIC statistics suggested that neither structural representation was preferable in a statistical sense. Simulations presented here confirm that relatively simple models with limited data requirements can be used to credibly simulate flows and water balance components needed for nutrient flux modelling in large, data-poor basins.

Impacts of climate change and socio-economic scenarios on flow and water quality of the Ganges, Brahmaputra and Meghna (GBM) river systems: low flow and flood statistics.

The potential impacts of climate change and socio-economic change on flow and water quality in rivers worldwide is a key area of interest. The Ganges-Brahmaputra-Meghna (GBM) is one of the largest river basins in the world serving a population of over 650 million, and is of vital concern to India and Bangladesh as it provides fresh water for people, agriculture, industry, conservation and for the delta system downstream. This paper seeks to assess future changes in flow and water quality utilising a modelling approach as a means of assessment in a very complex system. The INCA-N model has been applied to the Ganges, Brahmaputra and Meghna river systems to simulate flow and water quality along the rivers under a range of future climate conditions. Three model realisations of the Met Office Hadley Centre global and regional climate models were selected from 17 perturbed model runs to evaluate a range of potential futures in climate. In addition, the models have also been evaluated using socio-economic scenarios, comprising (1) a business as usual future, (2) a more sustainable future, and (3) a less sustainable future. Model results for the 2050s and the 2090s indicate a significant increase in monsoon flows under the future climates, with enhanced flood potential. Low flows are predicted to fall with extended drought periods, which could have impacts on water and sediment supply, irrigated agriculture and saline intrusion. In contrast, the socio-economic changes had relatively little impact on flows, except under the low flow regimes where increased irrigation could further reduce water availability. However, should large scale water transfers upstream of Bangladesh be constructed, these have the potential to reduce flows and divert water away from the delta region depending on the volume and timing of the transfers. This could have significant implications for the delta in terms of saline intrusion, water supply, agriculture and maintaining crucial ecosystems such as the mangrove forests, with serious implications for people's livelihoods in the area. The socio-economic scenarios have a significant impact on water quality, altering nutrient fluxes being transported into the delta region.

Dynamic modeling of the Ganga river system: impacts of future climate and socio-economic change on flows and nitrogen fluxes in India and Bangladesh.

This study investigates the potential impacts of future climate and socio-economic change on the flow and nitrogen fluxes of the Ganga river system. This is the first basin scale water quality study for the Ganga considering climate change at 25 km resolution together with socio-economic scenarios. The revised dynamic, process-based INCA model was used to simulate hydrology and water quality within the complex multi-branched river basins. All climate realizations utilized in the study predict increases in temperature and rainfall by the 2050s with significant increase by the 2090s. These changes generate associated increases in monsoon flows and increased availability of water for groundwater recharge and irrigation, but also more frequent flooding. Decreased concentrations of nitrate and ammonia are expected due to increased dilution. Different future socio-economic scenarios were found to have a significant impact on water quality at the downstream end of the Ganga. A less sustainable future resulted in a deterioration of water quality due to the pressures from higher population growth, land use change, increased sewage treatment discharges, enhanced atmospheric nitrogen deposition, and water abstraction. However, water quality was found to improve under a more sustainable strategy as envisaged in the Ganga clean-up plan.

Assessing the impacts of climate change and socio-economic changes on flow and phosphorus flux in the Ganga river system.

Anthropogenic climate change has impacted and will continue to impact the natural environment and people around the world. Increasing temperatures and altered rainfall patterns combined with socio-economic factors such as population changes, land use changes and water transfers will affect flows and nutrient fluxes in river systems. The Ganga river, one of the largest river systems in the world, supports approximately 10% global population and more than 700 cities. Changes in the Ganga river system are likely to have a significant impact on water availability, water quality, aquatic habitats and people. In order to investigate these potential changes on the flow and water quality of the Ganga river, a multi-branch version of INCA Phosphorus (INCA-P) model has been applied to the entire river system. The model is used to quantify the impacts from a changing climate, population growth, additional agricultural land, pollution control and water transfers for 2041-2060 and 2080-2099. The results provide valuable information about potential effects of different management strategies on catchment water quality.

Steady state and dynamic modelling of nitrogen in the River Kennet: impacts of land use change since the 1930s.

Steady state and dynamic models have been developed and applied to the River Kennet system. Annual nitrogen exports from the land surface to the river have been estimated based on land use from the 1930s and the 1990s. Long term modelled trends indicate that there has been a large increase in nitrogen transport into the river system driven by increased fertiliser application associated with increased cereal production, increased population and increased livestock levels. The dynamic model INCA (Integrated Nitrogen in Catchments) has been applied to simulate the day-to-day transport of N from the terrestrial ecosystem to the riverine environment. This process-based model generates spatial and temporal data and reproduces the observed instream concentrations. Applying the model to current land use and 1930s land use indicates that there has been a major shift in the short term dynamics since the 1930s, with increased river and groundwater concentrations caused by both non-point source pollution from agriculture and point source discharges.

Modelling acidification at Beacon Hill-a low rainfall, high pollutant deposition site in Central England.

The model MAGIC (Model of Acidification of Groundwater In Catchments) has been applied to the Beacon Hill site, near Loughborough in Central England. This site is heavily impacted by wet and dry deposition of oxides of sulphur and nitrogen. The high acid inputs have caused soil acidification and acid stream waters. Long term simulations suggest that there has been a major decline in alkalinity and pH over the past 50 years. Despite recent reductions in deposition levels, soils and streams are predicted to continue to acidify in the future. For this heavily impacted site, deposition must be reduced by 80-90% to reverse the acidification trend and allow recovery of soil and stream waters.

Recovery from acidification in the Tillingbourne catchment, southern England: catchment description and preliminary results.

Measurements of acid deposition and streamwater chemistry made in 1979-1982 and 1999-2000 are compared for a small, acid-sensitive catchment in Southeast England. The location, geology, soils, vegetation and hydrology of the catchment are described. The catchment is located on an acidic cretaceous sandstone with a low permeability clay sub-stratum. Soils are predominantly podzol and gley, with some mesotrophic peat. The catchment is forested. Mean volume-weighted concentrations in precipitation have changed approximately in proportion to emission changes. SO4(2-) has declined by 61%, H+ by 75%, both NO- and NH4+ by 37% and Cl- by 26%. Changes in wet deposition are greater, sulfate deposition declined by 69%, non-marine SO4(2-) by 73%, H+ deposition by 75%, NO3- and NH4+ by 50% and Cl- by 41%. Sulfate deposition in throughfall, a surrogate for total deposition measurement, has declined by 82% and non-marine SO4(2-) by 86%. Some of these changes are due to alterations in the tree cover and location of the collectors. In 1979-1982, the flux of NO3- and NH4+ in throughfall was less than in rainfall, 7.5 compared with 11.3 kg N ha(-1) year(-1), showing that N uptake by the canopy was greater than dry deposition of these species. However, in 1999-2000, the throughfall flux of N was greater than rainfall, 19.6 compared to 5.7 kg N ha year(-1), indicating that canopy uptake is not occurring to the same extent. Surface water was sampled at the same locations in the catchment during the two periods. At the catchment exit, mean pH increased, from 3.93 to 4.21 mg l(-1), and SO4(2-) declined from 20.2 to 16.7 mg l(-1) (18%). The decrease in SO4(2-) is much less than the reduction in deposition, suggesting that the predicted recovery is being delayed by release of sulfur from the soil. In contrast, NO3- concentrations in the catchment waters increased from 0.22 to 0.52 mg N l(-1) (133%) despite the reduction in N deposition. NH4+ concentrations were low during both study periods. It is concluded that recovery from acidification is probably occurring, but is possibly being delayed by desorption of soil S. The catchment is also showing signs of increasing N saturation, despite a reduction in N inputs.