Estimating the Acid-Base Properties and Electrical Charge of Organic Matter Using Spectrophotometry.

Spectrophotometric acid-base titration is a simple and powerful technique to evaluate the properties of proton binding sites of natural organic matter (NOM) at environmentally relevant concentrations. However, it is challenging to quantify the chemical charges (Q) carried by NOM at these concentrations. Based on a previous work, which relates the variation of Q with the specific UV-vis differential absorbance (ΔAλ,pH) at a given wavelength (λ) and pH of a dissolved NOM sample, the present work proposes a method to investigate any NOM sample. It determines specific features in the absorbance spectra attributed to proton-inert chromophores (A0,λ) and to the deprotonation processes of carboxylic (A1,λ) and phenolic groups (A2,λ). It enables to select sample-specific wavelength (λmid), where both these functional groups significantly contribute to the variation of absorbance with pH. The linear regression analysis of Aλmid,pH vs Q for various NOM reference samples evidenced that the sample-specific slope (SNOM) and intercept (INOM) were related to the intrinsic spectroscopic properties of the sample (A0,λmid, A1,λmid, and A2,λmid). This approach can thus be used to approximate the Q values of the NOM samples at environmentally relevant concentrations: a pre-requisite for predicting the fate and behavior of metal ions in natural systems.

Absolute and Relative Positioning of Natural Organic Matter Acid-Base Potentiometric Titration Curves: Implications for the Evaluation of the Density of Charged Reactive Sites.

Potentiometric acid-base titration curves collected on humic (nano)particles as a function of pH and salt concentration reflect the electrostatics of the particles and the amount of chemical charges (Q) they carry. In turn, the interpretation of titration data helps quantify their reactivity toward metals provided that both intrinsic chemical and nonspecific electrostatic contributions to proton binding are correctly unraveled. Establishing a titration curve requires several steps, i.e., blank subtraction, relative curve positioning with respect to the electrolyte concentration, and absolute curve positioning achieved by the estimation of particle charge Q0 at low pH. Failure to properly establish each step may lead to the misevaluation of nanoparticle charging behavior. Here, we report (i) a simple procedure to measure and position titration curves for humic substances (HS) versus salt concentration and (ii) an original approach for absolute curve positioning upon the exploitation of proton affinity spectra. The latter do not depend on Q0 and they thus constrain the titration data analysis using the soft Poisson-Boltzmann-based titration (SPBT) formalism for nanoparticles in the thick electric double-layer regime. We illustrate the benefits of our approach by analyzing titration measurements for a large range of humic nanoparticles and by comparing the outcome with results from the literature.

An easy spectrophotometric acid-base titration protocol for dissolved organic matter.

UV-vis spectrophotometric acid-base titration can characterize dissolved organic matter (DOM) acid-base properties. However, it requires incremental pH adjustment, which make the procedure time consuming and the results subjected to dilution effect. This study brings forth a new approach, referred as the "buffer method" for pH adjustments, by using carefully selected pH-buffers to adjust the pH. This, statistically validated method minimizes the pH adjustment time and lightens the laboratory work load. Chemical product cost associated with this novel method is slightly increased as compared to the previous approach, due to the necessity to use pH-buffers. • Buffer method: Acid-base titration by using buffer for pH adjustment • Buffer method validated by statistical means • Rapid, reliable and economical method.

Surface modifications at the oxide/water interface: Implications for Cu binding, solution chemistry and chemical stability of iron oxide nanoparticles.

The oxidation of magnetite into maghemite and its coating by natural organic constituents are common changes that affect the reactivity of iron oxide nanoparticles (IONP) in aqueous environments. Certain ubiquitous compounds such as humic acids (HA) and phosphatidylcholine (PC), displaying a high affinity for both copper (Cu) and IONP, could play a critical role in the interactions involved between both compounds. The adsorption of Cu onto four different IONP was studied: magnetite nanoparticles (magnNP), maghemite NP (maghNP), HA- and PC-coated magnetite NP (HA-magnNP and PC-magnNP, respectively). According to the results, the percentage of adsorbed Cu increases with increasing pH, irrespective of the IONP. Thus, protonation/deprotonation reactions are likely involved within Cu adsorption mechanism. Contrary to the other studied IONP, HA-magnNP favor Cu adsorption at most of the pH tested including acidic pH (pH = 3), suggesting that part of the active surface sites for Cu2+ were not grabbed by protons. The Freundlich adsorption isotherm of HA-magnNP provides the highest sorption constant KF (bonding energy) and n value which supports a heterogeneous sorption process. The heterogeneous adsorption between HA-magnNP and Cu2+ can be explained by both the diversity of the binding sites HA procured and the formation of multidendate complexes between Cu2+ and some of the HA functional groups. Such favorable adsorption process was neither observed on PC-coated-magnNP nor on maghNP, whose behaviors were comparable to that of magnNP. On another hand, HA and PC coatings considerably reduced iron (Fe) dissolution from magnNP as compared with magnNP. It was suggested that HA and PC coatings either provided efficient shield against Fe leaching or fostered dissolved Fe re-adsorption onto the functional groups at the coated magnNP surfaces. Thus, this study can help to better understand the complex interfacial reactions between cations-organic matter-colloidal surfaces which are relevant in environmental and agricultural contexts. This work showed that magnetite NP properties can be affected by surface modifications, which drive NP chemical stability and Cu adsorption, thereby affecting the global water chemistry.

Are nanoplastics able to bind significant amount of metals? The lead example.

The nanoscale size of plastic debris makes them potential efficient vectors of many pollutants and more especially of metals. In order to evaluate this ability, nanoplastics were produced from microplastics collected on a beach exposed to the North Atlantic Gyre. The nanoplastics were characterized using multi-dimensional methods: asymmetrical flow field flow fractionation and dynamic light scattering coupled to several detectors. Lead (II) adsorption kinetics, isotherm and pH-edge were then carried out. The sorption reached a steady state after around 200 min. The maximum sorption capacity varied between 97% and 78.5% for both tested Pb concentrations. Lead (II) adsorption kinetics is controlled by chemical reactions with the nanoplastics surface and to a lesser extent by intraparticle diffusion. Adsorption isotherm modeling using Freundlich model demonstrated that NPG are strong adsorbents equivalent to hydrous ferric oxides such as ferrihydrite (log Kadsfreundlich=8.36 against 11.76 for NPG and ferrihydrite, respectively). The adsorption is dependent upon pH, in response to the Pb(II) adsorption by the oxygenated binding sites developed on account of the surface UV oxidation under environmental conditions. They could be able to compete with Fe or humic colloids for Pb binding regards to their amount and specific areas. Nanoplastics could therefore be efficient vectors of Pb and probably of many other metals as well in the environment.

Experimental evidence of REE size fraction redistribution during redox variation in wetland soil.

The evolution of rare earth element (REE) speciation between reducing and oxidizing conditions in a riparian wetland soil was studied relative to the size fractionation of the solution. In all size fractions obtained from the reduced and oxidized soil solutions, the following analyses were carried out: organic matter (OM) characterization, transmission electron microscopy (TEM) observations as well as major and trace element analyses. Significant REE redistribution and speciation evolution between the various size fractions were observed. Under reducing conditions, the REEs were bound to colloidal and dissolved OM (<2µm size fractions). By contrast, under oxidizing conditions, they were distributed in particulate (>2µm size fraction), colloidal (<2µm size fraction), organic and Fe-enriched fractions. In the particulate size fraction, the REEs were bound to humic and bacterial OM embedding Fe nano-oxides. The resulting REE pattern showed a strong enrichment in heavy REEs (HREEs) in response to REE binding to specific bacterial OM functional groups. In the largest colloidal size fraction (0.2µm-30kDa), the REEs were bound to humic substances (HS). The lowest colloidal size fraction (<30kDa) is poorly concentrated in the REEs and the REE pattern showed an increase in the middle REEs (MREEs) and heavy REEs (HREEs) corresponding to a low REE loading on HS. A comparison of the REE patterns in the present experimental and field measurements demonstrated that, in riparian wetlands, under a high-water level, reducing conditions are insufficient to allow for the dissolution of the entire Fe nano-oxide pool formed during the oxidative period. Therefore, even under reducing conditions, Fe(III) seems to remain a potential scavenger of REEs.

Does As(III) interact with Fe(II), Fe(III) and organic matter through ternary complexes?

Up until now, only a small number of studies have been dedicated to the binding processes of As(III) with organic matter (OM) via ionic Fe(III) bridges; none was interested in Fe (II). Complexation isotherms were carried out with As(III), Fe(II) or Fe(III) and Leonardite humic acid (HA). Although PHREEQC/Model VI, implemented with OM thiol groups, reproduced the experimental datasets with Fe(III), the poor fit between the experimental and modeled Fe(II) data suggested another binding mechanism for As(III) to OM. PHREEQC/Model VI was modified to take various possible As(III)-Fe(II)-OM ternary complex conformations into account. The complexation of As(III) as a mononuclear bidentate complex to a bidentate Fe(II)-HA complex was evidenced. However, the model needed to be improved since the distribution of the bidentate sites appeared to be unrealistic with regards to the published XAS data. In the presence of Fe(III), As(III) was bound to thiol groups which are more competitive with regards to the low density of formed Fe(III)-HA complexes. Based on the new data and previously published results, we propose a general scheme describing the various As(III)-Fe-MO complexes that are able to form in Fe and OM-rich waters.

Thiol groups controls on arsenite binding by organic matter: new experimental and modeling evidence.

Although it has been suggested that several mechanisms can describe the direct binding of As(III) to organic matter (OM), more recently, the thiol functional group of humic acid (HA) was shown to be an important potential binding site for As(III). Isotherm experiments on As(III) sorption to HAs, that have either been grafted with thiol or not, were thus conducted to investigate the preferential As(III) binding sites. There was a low level of binding of As(III) to HA, which was strongly dependent on the abundance of the thiols. Experimental datasets were used to develop a new model (the modified PHREEQC-Model VI), which defines HA as a group of discrete carboxylic, phenolic and thiol sites. Protonation/deprotonation constants were determined for each group of sites (pKA=4.28±0.03; ΔpKA=2.13±0.10; pKB=7.11±0.26; ΔpKB=3.52±0.49; pKS=5.82±0.052; ΔpKS=6.12±0.12 for the carboxylic, phenolic and thiols sites, respectively) from HAs that were either grafted with thiol or not. The pKS value corresponds to that of single thiol-containing organic ligands. Two binding models were tested: the Mono model, which considered that As(III) is bound to the HA thiol site as monodentate complexes, and the Tri model, which considered that As(III) is bound as tridentate complexes. A simulation of the available literature datasets was used to validate the Mono model, with logKMS=2.91±0.04, i.e. the monodentate hypothesis. This study highlighted the importance of thiol groups in OM reactivity and, notably, determined the As(III) concentration bound to OM (considering that Fe is lacking or at least negligible) and was used to develop a model that is able to determine the As(III) concentrations bound to OM.

Robust Method Using Online Steric Exclusion Chromatography-Ultraviolet-Inductively Coupled Plasma Mass Spectrometry To Investigate Nanoparticle Fate and Behavior in Environmental Samples.

The foundation of nanoscience is that the properties of materials change as a function of their physical dimensions, and nanotechnology exploits this premise by applying selected property modifications for a specific benefit. However, to investigate the fate and effect of the engineered nanoparticles on toxic metal (TM) mobility, the analytical limitations in a natural environment remain a critical problem to overcome. Recently, a new generation of size exclusion chromatography (SEC) columns developed with spherical silica is available for pore sizes between 5 and 400 nm, allowing the analysis of nanoparticles. In this study, these columns were applied to the analysis of metal-based nanoparticles in environmental and artificial samples. The new method allows quantitative measurements of the interactions among nanoparticles, organic matter, and metals. Moreover, because of the new nanoscale SEC, our method allows the study of these interactions for different size ranges of nanoparticles and weights of organic molecules with a precision of 1.2 × 10(-2) kDa. The method was successfully applied to the study of nanomagnetite spiked in complex matrixes, such as sewage sludge, groundwater, tap water, and different artificial samples containing Leonardite humic acid and different toxic metals (i.e., As, Pb, Th). Finally, our results showed that different types of interactions, such as adsorption, stabilization, and/or destabilization of nanomagnetite could be observed using this new method.

Interactions between natural organic matter, sulfur, arsenic and iron oxides in re-oxidation compounds within riparian wetlands: nanoSIMS and X-ray adsorption spectroscopy evidences.

Arsenic (As) is a toxic and ubiquitous element which can be responsible for severe health problems. Recently, Nano-scale Secondary Ions Mass Spectrometry (nanoSIMS) analysis has been used to map organomineral assemblages. Here, we present a method adapted from Belzile et al. (1989) to collect freshly precipitated compounds of the re-oxidation period in a natural wetland environment using a polytetrafluoroethylene (PTFE) sheet scavenger. This method provides information on the bulk samples and on the specific interactions between metals (i.e. As) and the natural organic matter (NOM). Our method allows producing nanoSIMS imaging on natural colloid precipitates, including (75)As(-), (56)Fe(16)O(-), sulfur ((32)S(-)) and organic matter ((12)C(14)N) and to measure X-ray adsorption of sulfur (S) K-edge. A first statistical treatment on the nanoSIMS images highlights two main colocalizations: (1) (12)C(14)N(-), (32)S(-), (56)Fe(16)O(-) and (75)As(-), and (2) (12)C(14)N(-), (32)S(-) and (75)As(-). Principal component analyses (PCAs) support the importance of sulfur in the two main colocalizations firstly evidenced. The first component explains 70% of the variance in the distribution of the elements and is highly correlated with the presence of (32)S(-). The second component explains 20% of the variance and is highly correlated with the presence of (12)C(14)N(-). The X-ray adsorption near edge spectroscopy (XANES) on sulfur speciation provides a quantification of the organic (55%) and inorganic (45%) sulfur compositions. The co-existence of reduced and oxidized S forms might be attributed to a slow NOM kinetic oxidation process. Thus, a direct interaction between As and NOM through sulfur groups might be possible.

How does organic matter constrain the nature, size and availability of Fe nanoparticles for biological reduction?

Few studies have so far examined the kinetics and extent of the formation of Fe-colloids in the presence of natural organic ligands. The present study used an experimental approach to investigate the rate and amount of colloidal Fe formed in presence of humic substances, by gradually oxidizing Fe(II) at pH 6.5 with or without humic substances (HS) (in this case, humic acid--HA and fulvic acid--FA). Without HS, micronic aggregates (0.1-1 µm diameter) of nano-lepidocrocite is obtained, whereas, in a humic-rich medium (HA and FA suspensions at 60 and 55 ppm of DOC respectively), nanometer-sized Fe particles are formed trapped in an organic matrix. A proportion of iron is not found to contribute to the formation of nanoparticles since iron is complexed to HS as Fe(II) or Fe(III). Humic substances tend to (i) decrease the Fe oxidation and hydrolysis, and (ii) promote nanometer-sized Fe oxide formation by both inhibiting the development of hydroxide nuclei and reducing the aggregation of Fe nanoparticles. Bioreduction experiments demonstrate that bacteria (Shewanella putrefaciens CIP 80.40 T) are able to use Fe nanoparticles associated with organic matter about eight times faster than in the case of nano-lepidocrocite. This increase in bioreduction rate appears to be related to the presence of humic acids that (i) indirectly control the size, shape and density of oxyhydroxides and (ii) directly enhance biological reduction of nanoparticles by electron shuttling and Fe complexation. These results suggest that, in wetlands but also elsewhere where mixed organic matter-Fe colloids occur, Fe nanoparticles closely associated with organic matter represent a bioavailable Fe source much more accessible for microfauna than do crystallized Fe oxyhydroxides.

Dynamic structure of humic substances: rare earth elements as a fingerprint.

Whereas humic substances are known to play a key role in controlling metal speciation and trace element mobility within soils and waters, the understanding of their structure is still unclear and remains a matter of debate. Several models of humic substance structure have been proposed, where humic substances were composed of either: (i) macromolecular polyelectrolytes that can form molecular aggregates or (ii) supramolecular assemblies (molecular aggregates) of small molecules without macromolecular character, joined together by weak attraction forces. This experimental study was designed and dedicated: (i) to follow the size of organic molecules versus ionic strength or pH by the combined means of ultrafiltration and aromaticity data and rare earth element (REE) fingerprinting, and (ii) to investigate the pH and ionic strength effect on the distribution of associated rare earth elements in soil solution. This study supports the presence of supramolecular associations of small molecules and probably the presence of macromolecules in the bulk dissolved organic matter. By contrast to ionic strength, pH appeared to be the major parameter playing on the stability of the humic substance structure. Humic substances displayed dynamic structures, which evolved with regard to pH. Low pH led to a destabilization of the humic substance conformation. This destabilization had an impact on the trace element distribution in soil solution, as assessed by REE data, and conversely, the destabilization degree of humic substances seemed to be influenced by the metal ion charge.

Double pH control on humic substance-borne trace elements distribution in soil waters as inferred from ultrafiltration.

Colloidal dissolved organic carbon (DOC) is an important carrier phase for trace elements (TE) in subsurface environments. As suggested by previously published field observations, preferential sorption of DOC onto mineral surfaces tends to enrich the solid phase in humic acids. This DOC fractionation may affect the mobility of TE. pH is known to play an important role in the stability of colloids. This study was therefore dedicated to identifying the influence of DOC fractionation on TE mobility. Sequential extraction has been used to provide information on the possible TE carriers within soil (as exchangeable, weak acid soluble, reducible, oxidizable, and nonextractible metal fractions). Batch experiments were carried out to investigate the influence of pH on the detachment of colloids and associated TE. Different groups of elements were identified according to TE behavior during pH changes. Several elements displayed increasing concentrations with decreasing pH. These concentrations can represent an important fraction of the total soil concentration. By contrast, other elements showed increasing concentrations following increasing pH, in association with an increasing amount of colloids in soil solution. Concerning this latter group, two colloidal carrier phases were identified during the pH increase: (i) the first one concerned the majority of elements, which were associated with humic substances remaining in solution, and (ii) the second one involved several TE rather associated with nanooxides. Therefore, DOC fractionation plays a key role in the TE concentration in soil solution during pH changes.

Bacillus subtilis bacteria hinder the oxidation and hydrolysis of Fe2+ ions.

Bacteria are known to associate closely with secondary iron oxides in natural environments, but it is still unclear whether they catalyze their precipitation. Here, Fe2+ ions were progressively added to various concentrations of Bacillus subtilis bacteria in permanently oxic conditions while maintaining the pH at 6.5 by adding a NaOH solution at a monitored rate. The iron/ bacteria precipitates were characterized by wet chemistry, SEM, and XRD. Abiotic syntheses produced nanolepidocrocite, and their kinetics displayed a strong autocatalytic effect. Biotic syntheses led to the formation of tiny and poorly crystallized particles at intermediate bacterial concentrations and to a complete inhibition of particle formation at high bacterial concentrations. The occurrence of the autocatalytic effect was delayed and its intensity was reduced. Both the oxidation and the hydrolysis of Fe2+ ions were hindered.

Insights into colloid-mediated trace element release at the soil/water interface.

Organic or inorganic colloids play a major role in the mobilization of trace elements in soils and waters. Environmental physicochemical parameters (pH, redox potential, temperature, pressure, ionic strength, etc.) are the controlling factors of the colloidal mobilization. This study was dedicated to follow the colloid-mediated mobilization of trace elements through time at the soil/water interface by means of an experimental approach. Soil column experiments were carried out using percolating synthetic solutions. The percolated solutions were ultrafiltrated with various decreasing cutoff thresholds to separate the different colloidal phases in which the dissolved organic carbon and trace element concentrations were measured. The major results which stem from this study are the following: (i) The data can be divided into different groups of organic compounds (microbial metabolites, fulvic acids, humic acids) with regard to their respective aromaticity and molecular weight. (ii) Three groups of elements can be distinguished based on their relationships with the colloidal phases: the first one corresponds to the so-called "truly" dissolved group (Li, B, K, Na, Rb, Si, Mg, Sr, Ca, Mn, Ba, and V). The second one can be considered as an intermediate group (Cu, Cd, Co, and Ni), while the third group gathers Al, Cr, U, Mo, Pb, Ti, Th, Fe, and rare earth elements (REE) carried by the organic colloidal pool. (iii) The data demonstrate that the fulvic acids seem to be a major organic carrier phase for trace elements such as Cu, Cd, Co, and Ni. By contrast, the trace elements belonging to the so-called colloidal pool were mostly mobilized by humic acids containing iron nanoparticles. Lead, Ti, and U were mobilized by iron nanoparticles bound to these humic acids. Thus, humic substances allowed directly or indirectly a colloidal transport of many insoluble trace elements either by binding trace elements or by stabilizing a ferric carrier phase. (iv) Finally, the results demonstrated also that REE were mostly mobilized by humic substances. The REE normalized patterns showed a middle REE downward concavity. Therefore, as previously shown elsewhere humic substances are a major control of REE speciation and REE fractionation patterns as well since the humic substance/metal ratio was the key parameter controlling the REE pattern shape.

Competition between humic acid and carbonates for rare earth elements complexation.

The competitive binding of rare earth elements (REE) to humic acid (HA) and carbonates was studied experimentally at various pH and alkalinity values by combining ultrafiltration and inductively coupled plasma mass spectrometry techniques. The results show that the REE species occur as binary humate or carbonate complexes but not as ternary REE-carbonate-humate as previously proposed. The results also reveal the strong pH and alkalinity dependence of the competition as well as the existence of a systematic fractionation across the REE series. Specifically, carbonate complexation is at a maximum at pH 10 and increase with increasing alkalinity and with the atomic number of the REE (LuCO(3)>>LaCO(3)). Modeling of the data using Model VI and recently published stability constants for complexation of REE by humic acid well reproduced the experimental data, confirming the ability of Model VI to accurately determine REE speciation in natural waters. This modeling also confirms the reliability of recently published stability constants. This work shed more light not only on the competition between carbonates and HA for REE complexation but also on the reliability of WHAM 6 and Model VI for calculating the speciation of REE with organic matter in alkaline organic-rich water.

Impact of humate complexation on the adsorption of REE onto Fe oxyhydroxide.

Adsorption experiments of rare-earth elements (REE) onto hydrous ferric oxide (HFO) were performed to evaluate the impact of organic complexation on both REE(III) adsorption and the Ce(III) oxidation rate. Scavenging experiments were performed at pH 5.2 with NaCl and NaNO3 solutions containing either free REE (III) or REE(III)-humate complexes. The log K(d)(REE) patterns obtained from HFO suspensions exhibit a slight positive Ce anomaly and an M-type lanthanide tetrad effect, in contrast with the partitioning between REE(III)-humate complexes and HFO, which yields completely flat distribution patterns. The "organic" partitioning runs yield log K(d)(REEorganic)/log K(d)(DOC) ratios (DOC = dissolved organic carbon) close to 1.0, implying that the REE(III) and humate remain bound to each other during the adsorption experiment. The lack of any positive Ce anomaly or M-type lanthanide tetrad effect in the organic experiments seems to reflect an anionic adsorption of the REE-humate complex. Adsorption onto HFO takes place via the humate side of the REE(III)-humate complexes. The oxidation of Ce(III) by Fe(III) and the proportion of surface hydroxyl groups coordinated to REE(III) at the HFO surface are the two most commonly invoked processes for explaining the development of positive Ce anomalies and the M-type tetrad lanthanide effect. However, such processes cannot proceed since the REE are not in direct contact with the HFO suspensions, the latter being shielded by PHA. The present results further complicate the use of Ce anomalies as reliable paleoredox proxies in natural precipitates. They are also further demonstration that organic matter may inhibit the lanthanide tetrad effect in geological samples.

Controls on the distribution of rare earth elements in shallow groundwaters.

Shallow groundwater samples (<0.22 microm) collected from a small catchment in Western France (Petit Hermitage catchment) were analyzed for their rare earth elements (REE), dissolved organic carbon (DOC) and trace-element (Fe, Mn, Th and U) contents, with the aim to investigate the controlling factors of REE signatures. Two spatially distinct water types are recognized in this catchment based on changes of REE concentrations and variations of Ce anomalies. These include (i) DOC-poor groundwater flowing below the hillslope domains; this type has low REE contents and records conspicuous negative Ce anomalies; (ii) DOC-rich groundwater from the wetland domains, close to the river network; this type displays much higher REE concentrations, and typically lack negative Ce anomalies. Temporal REE concentration variations were assessed in wetland waters. Results show a marked increase of the REE content at the winter-spring transition, along with variations of DOC, Fe and Mn contents and redox potential changes. Using the above data set in conjunction with previously published results on comparable shallow groundwaters, we demonstrate that factors such as aquifer-rock composition or anthropogenic input probably play a minor role in determining the REE signatures of shallow groundwaters. Most likely, the two dominant factors involved are the organic matter content and the redox status of the waters. We suggest that topography might be the ultimate key parameter by its ability to control the DOC content of groundwater at a catchment scale.