296 REVIEW

Arsenic in groundwater in the southern lowlands of Nepal and its mitigation options: a review Barbara Mueller

Abstract: As in several other countries of Southeast Asia (namely Bangladesh, India, Myanmar, China, Vietnam, and Cambodia) arsenic (As) concentrations in the groundwater of the lowlands of Nepal (the so called Terai) can reach concentrations that are unsafe to humans using the groundwater as drinking water. Whereas Bangladesh has received much international attention concerning the As crisis, Nepal was more or less neglected. The first report about As contamination of the groundwater above toxic levels in Nepal was published in 1999. Twenty-four percent of samples analyzed (n = 18 635) from the Terai Basin exceeded the WHO guideline of 10 ␮g/L. Since the first overall survey from 2001, only sporadic information on the situation has been published. The geological and geochemical conditions favour the release of the contaminant as As can be easily solubilized in groundwaters depending on pH, redox conditions, temperature, and solution composition. The thin alluvial aquifers of the Terai are some of the most severely As contaminated. These sediments constituting a hugh proportion of the Terai aquifers are derived from two main sources: (i) sediments deposited by large rivers that erode the upper Himalayan crystalline rocks, and (ii) weath- ered meta-sediments carried by smaller rivers originating in the Siwalik forehills. The generally low redox potential and low 2− 3− − SO4 and high DOC, PO4 , and HCO3 concentrations in groundwater signify ongoing microbial-mediated redox processes favoring As mobilization in the aquifer. Other geochemical processes, e.g., Fe-oxyhydroxide reduction and carbonate dissolu- tion, are also responsible for high As occurrence in groundwaters. Originally, gagri filters (a two-filter system with chemical powder) and later iron (Fe)-assisted biosand filters were commonly used to remove As and Fe from well water in Nepal—these two options were believed to be the best treatment option at household levels. This review focus on the description of the overall situation, including geogenic issues, occurrence of As in the sediments of the Terai, mechanisms for the release of As to the groundwater, and mitigation options.

Key words: arsenic, arsenic contamination, release of arsenic to the groundwater, removal of arsenic, mitigation. Résumé : Comme dans plusieurs autres pays de l’Asie du Sud-Est (a` savoir le Bangladesh, l’Inde, le Myanmar, la Chine, le Viêt-Nam, le Cambodge) les concentrations d’arsenic dans les eaux souterraines des plaines du Népal (connues sous le nom de Terai) peuvent atteindre des concentrations qui sont dangereuses pour les humains qui utilisent l’eau souterraine comme eau potable. Tandis que le Bangladesh a reçu beaucoup d’attention internationale concernant la crise d’arsenic, le Népal a été plus For personal use only. ou moins négligé. Le premier rapport sur la contamination des eaux souterraines par l’arsenic au-dessus des niveaux toxiques au Népal a été publié en 1999. Vingt-quatre pour cent des échantillons analysés (n = 18 635) du bassin de Terai excédait la ligne directrice de l’OMS, soit 10 ␮g/L. Depuis la première enquête globale de 2001, on a publié que des informations sporadiques sur la situation. Les conditions géologiques et géochimiques favorisent le rejet du polluant puisque l’arsenic peut être facilement solubilisé dans des eaux souterraines en fonction du pH, de la condition d’oxydo-réduction, de la température et de la compo- sition de solution. Les aquifères alluviaux minces du Terai sont parmi les plus sévèrement contaminés par l’As. Ces sédiments constituant une très grande partie des aquifères de Terai proviennent de deux sources principales, soit (i) des sédiments déposés par les grandes rivières qui érodent les roches cristallines du Haut-Himalaya, (ii) des sédiments métamorphisés météorisés portés par de plus petites rivières prenant leur source dans les contreforts du Siwalik. Le potentiel d’oxydo-réduction généralement bas, 2− 3− − les concentrations faibles de SO4 , et élevées de COD, de PO4 et d’HCO3 dans les eaux souterraines signifient qu’ilyades processus continus d’oxydo-réduction par intermédiaire microbien favorisant la mobilisation d’As dans l’aquifère. D’autres processus géochimiques, par exemple, la réduction des oxydes-hydroxydes de fer et la dissolution de carbonates sont aussi responsables de la présence élevée d’As dans les eaux souterraines. À l’origine, on utilisait généralement des filtres Gagri (système a` deux filtres avec poudre chimique) et par la suite des filtres de sable bio aidés de fer pour éliminer l’arsenic et le fer de l’eau de puits au Népal—ces deux options étaient censées être la meilleure option de traitement au niveau des ménages. Cette revue se penche sur la description de la situation globale (des questions géogéniques, la présence d’arsenic dans les sédiments du Terai, les mécanismes de rejet d’arsenic dans les eaux souterraines, les options d’atténuation). [Traduit par la Rédaction] Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 Mots-clés : arsenic, contamination a` l’arsenic, rejet d’arsenic dans les eaux souterraines, élimination de l’arsenic, atténuation.

Introduction are hazardous to human health if geological and geochemical condi- In the groundwaters of several countries of Southeast Asia tions favour the release of this contaminant. The World Health (namely Bangladesh, India, Nepal, Myanmar, China, Vietnam, and Organisation (WHO) has imposed a drinking water guideline with Cambodia), arsenic (As) can naturally reach concentrations that a value of 10 ␮g/L for As. When this value is exceeded, health risks

Received 3 August 2016. Accepted 16 December 2016. B. Mueller. Deptartment of Environmental Science, University of Basel, 4056 Basel, Switzerland. Email for correspondence: [email protected]. Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from RightsLink.

Environ. Rev. 25: 296–305 (2017) dx.doi.org/10.1139/er-2016-0068 Published at www.nrcresearchpress.com/er on 23 December 2016. Mueller 297

are likely to occur. Excess uptake of As causes a range of adverse stitutes less than 20% of Nepal’s surface, it contains over half of health effects like characteristic skin lesions including pigmenta- the total arable land and is home to about 50% of the Nepalese tion changes, mainly on the upper chest, arms, and legs, and population, i.e., 30 million inhabitants. Groundwater is the main keratoses of the palms of the hands and soles of the feet, and as source of water for drinking and irrigation in the Terai area. Over the most severe effect, cancer (Smith et al. 2000; Adhikari and 90% of the Terai population draws groundwater from tube wells Ghimire 2009). for drinking, household use, and irrigation (Guillot et al. 2015). Arsenic itself is not found in high abundance in the Earth’s According to some publications, 25 058 tube wells in the Terai continental crust; it is less abundant than several of the “rare-earth” region have been tested for As, of which 5686 tube wells (22.7%) elements. Unlike the rare-earth elements, however, As is com- exceed the WHO As guideline (As = 0.01 mg/L) and 1916 tube wells monly concentrated in sulphide-bearing mineral deposits, and it (7.6%) exceed the Nepal Interim As Standard (As = 0.05 mg/L) has a strong affinity for pyrite, one of the more ubiquitous min- (Panthi et al. 2006). It is estimated that there are perhaps 200 000 erals in the Earth’s crust. Arsenic is also concentrated in hydrous tube wells in the Terai region and that 3.5 million Nepalese have iron (Fe)-oxides and clay minerals. Arsenic can be easily solubi- no access to drinking water that does not exceed the WHO As lized in groundwaters depending on pH, redox conditions, tem- guideline (Mahat and Shrestha 2008; Mahat and Kharel 2009; perature, and solution composition. Many geothermal waters Pokhrel et al. 2009). In the most recent report from the National contain high concentrations of As. Natural As in groundwater at Arsenic Steering Committee/National Red Cross Society (NASC-NRCS ␮ concentrations above the drinking water guideline of 10 g/L is 2011), the total database covers 1.1 million wells tested between the not uncommon. A small number of source materials are now recog- years 2003 and 2008. Approximately 1.73% showed values above nized as significant contributors to As in water supplies: organic-rich the Nepal drinking water standard of 50 ppb, while approxi- or black shales, Holocene alluvial sediments with slow flushing mately 5.37% of tube wells contain 11–50 ppb of As concentration. rates, mineralized and mined areas (most often gold deposits), volca- The percentage of all tube wells exceeding 50 ppb varies from nogenic sources, and thermal springs. Two other environments can 0.05% of the wells in the district of Jhapa to 11.69% in the district of lead to high As: (i) closed basins in arid-to-semi-arid climates (espe- Nawalparasi. cially in volcanogenic provinces) and (ii) strongly reducing aqui- The most severe As contamination is prevalent in several dis- fers, often composed of alluvial sediments but with low sulphate tricts of the Terai, namely Nawalparasi, Bara, Parsa, Rautahat, concentrations. Young sediments in low-lying regions of low hy- Rupandehi, and Kapalivastu (Shrestha et al. 2014). Maharjan et al. draulic gradient are characteristic of many As-rich aquifers. Ordi- (2005) reported that 29% of more than 20 000 tube wells had As nary sediments containing 1–20 mg/kg (near crustal abundance) of concentrations exceeding the WHO guideline (10 ␮g/L), that the As can give rise to high levels of dissolved As (>50 ␮g/L) if initiated prevalence of arsenicosis varied between 1.3% and 5.1% (average of by one or both of two possible “triggers”—an increase in pH above 2.6%; see NRCS–ENPHO 2002; Yadav et al 2011) among four inde- 8.5 or the onset of reductive Fe dissolution. Other important fac- pendent surveys, and that approximately 0.5 million people in tors promoting As solubility are high concentrations of phos- Terai were at risk of consuming water with an As concentra- phate, bicarbonate, silicate, and (or) organic matter in the ground tion >50 ␮g/L, the maximum permissible limit for Nepal (Shrestha waters. These solutes can decrease or prevent the adsorption of et al. 2003). It was found that overall prevalence of arsenicosis arsenate and arsenite ions onto fine-grained clays and especially among the subjects ≥15 years old was 6.9%, which was comparable Fe-oxides. Arsenite tends to adsorb less strongly than arsenate, with those found by the same examiner in As-contaminated areas often causing arsenite to be present at higher concentrations. The in Bangladesh, and that males had prevalence twice as high as

For personal use only. geologic and groundwater conditions that promote high As con- females, which could not be explained by the difference in the centrations are now quite well known and help identify high-risk exposure level Maharjan et al. (2005). These reports have alerted areas (Nordstrom 2002; Smedley and Kinniburgh 2002). The water the decision makers of the government as well as non-governmental table within the Indo-Gangetic Basin, including the Terai, alluvial agencies involved in controlling the water supply. As a conse- aquifer is typically shallow (<5 m below ground level). Abstraction quence, in 2003 the National Arsenic Steering Committee (NASC) of groundwater can also influence As flux: it can flush aqueous As was formed, involving major stakeholders from the drinking wa- from the aquifer; irrigation pumping protects deeper groundwa- ter in some instances, by creating a hydraulic barrier, but it seems ter and sanitation sectors (Shrestha et al. 2003). The NASC worked that high-capacity deep pumping may draw As down to levels in in collaboration with the Environment Public Health Organiza- the Bengal aquifer system that are otherwise of good quality tion (ENPHO) to perform testing on 18 635 tube wells in 20 Terai (MacDonald et al. 2016). districts, under a program called the State of Arsenic in Nepal Whereas Bangladesh has received much international atten- 2003. All data collected revealed that the concentration of As tion concerning the As crisis (e.g., Hug et al. 2011 and references varied both spatially and seasonally, suggesting the possibility of therein), Nepal was more or less neglected, though the population spatial variation owing to geospatial conditions such as latitude, of the southern lowlands of Nepal (the so called Terai, the Indo- longitude, and depth of tube well. The temporal distribution of As Gangetic Plain of southern Nepal) face the same As contamination showed seasonal dependence with lower concentration in winter of the groundwater (Nakano et al. 2014). The study of As concen- and higher concentration in summer (Yadav et al. 2012). trations in the groundwater in Nepal began only after the severity Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 Geological situation of the Terai region of the As contamination problem in the Bengal delta was recognized in 1998. The first report of As contamination in groundwater above Nepal is a landlocked country in South Asia, located between toxic levels in Nepal was made from the Terai Basin (Sharma 1999). Tibet to the north and India to the south, east, and west. With a Twenty-four percent of samples analyzed (n = 18 635) from the Terai total land area of 147 181 km2, the country is characterized by a Basin exceeded the WHO limit of 10 ␮g/L (Shrestha and Shrestha diverse, rugged, and undulating topography, geology, and in gen- 2004). Since the first overall survey conducted by WHO (2001), only eral by a cold climate. Nepal is predominantly mountainous, with sporadic information on the situation has been published. Avail- elevations ranging from 64 m above sea level to 8848 m at the peak of able documents later indicated that the region of As contamina- the world’s highest mountain, Sagarmatha (Everest), within a span of tion extends into 24 districts in Nepal, including all 20 Terai 200 km. Approximately 6000 rivers and rivulets, with a total districts and four hill districts (Bhattacharya et al. 2003; Neku and drainage area of about 194 471 km2, flow through Nepal, whereof Tandukar 2003; Shrestha et al. 2003; FAO 2004; Tandukar et al. 76% of this drainage area is contained within Nepal. The topo- 2005; Panthi et al. 2006; Maharjan et al. 2006; Pokhrel et al. 2009; graphic variations in Nepal are largely controlled by geology (BGS Emerman et al. 2010; Thakur et al. 2011). Although the Terai con- Report 2001; Thakur et al. 2011). The geology of Nepal marks the

Published by NRC Research Press 298 Environ. Rev. Vol. 25, 2017

Fig. 1. Groundwater arsenic testing in districts of various development regions of Nepal (from Thakur et al. 2011).

transition where the Southern Gondwanaland collided with the The aquifer system is highly sensitive to precipitation (Gurung Northern Eurasianland, lifting the sediments of the then Tethys et al. 2005). Sea to form the Himalayas. As a consequence, the southern and The geology of the Terai region of Nepal itself is on the whole northern parts of Nepal differ widely in their formations. The Archean comparable to the Bengal Delta Plain and is a continuation of the crystalline formations deep beneath the Alluvium of the Terai, as well Indo-gangetic trough. The Terai Plain covered by recent and older as the marine sedimentary deposits forming the high Himalayas, alluvium comprises channel sand and gravel and outwash depos- and the Siwalik formation formed by the then east–west flowing its. These fluviatile deposits are cross-bedded, eroded, reworked, rivers can be found within this confined space (Yadav et al. 2015). and redeposited because of regular shifting of stream channels. The prominent mountain chain in Nepal—the Himalayas—is Geomorphologically, the Terai Plain is divided into two zones: the built up by four major Himalayan tectonic units: (1) the Tethys Bhabar zone in the north and the main Terai zone in the south. Himalaya, delimited at the base by the South Tibetan Detachment They have diverse hydrogeological characteristics and are sepa- system; (2) the Higher Himalayan Crystallines, delimited at the rated by a line of natural springs. The Bhabar zone is a narrow base by the Main Central Thrust I; (3) the Lesser Himalaya, divided extension of a recent alluvial and colluvial fan deposit at the into upper and lower Lesser Himalaya, delimited at the base by bottom of Siwalik Hills (Kansakar 2004). It consists of thick depos- the Main Boundary Thrust; and (4) the Siwaliks, delimitated at its its of gravel, pebble, and boulder mixed with sand and silt. Sedi- For personal use only. base by the Main Frontal Thrust and the Quaternary foreland ments in the main Terai zone were deposited by braided rivers, basin. These units span a wide range of various rocks being met- which regularly changed their course. As a result, clay, silt, sand, amorphic, sedimentary, and igneous in origin, making it possible and gravel deposits of varying thicknesses occur interlayered with for their differential erosion to account for some of the ground- each other. The Terai Plain has a multiple aquifer system (Yadav water As heterogeneity we see in the foreland and delta (i.e., et al. 2011). Gurung et al. 2005; Shah 2008; van Geen et al. 2008; Guillot et al. So far the most intensively studied Terai province concerning 2015). The Terai Plain is an active foreland basin consisting of local geology and As-contaminated groundwater is Nawalparasi. Quaternary sediments that include molasse units along with This district lies in the Terai Plain as a continuation of the Indo- gravel, sand, silt, and clay. Most of the rivers in the Terai flow Gangetic Plain (Fig. 1). It has a gentle slope towards the south from from north to south. All major rivers originate in the high Hima- an elevation of 200–300 m in the north to a low of 63 m in the layas, whereas minor rivers also emanate from the nearby Siwalik south near the Indian border from the mean sea level (Upreti Hills, and therefore deposit sediments in the form of a fan along 2001). From the Indian border, Nawalparasi district extends north- the flank of the Terai basin. Fine sediments and organic material ward across Narayani River (one of the major rivers of Nepal) are deposited in inter-fan lowlands, wetlands, and swamps (Sharma alluvium then across the low gradient fan of locally derived allu- 1995). The Siwalik lithofacies are strongly diachronous, and fur- vium and finally into the Himalayan foothills (also known as Chu- ther complicated by a variable addition of micaceous sands and ria hills) (Hagen 1969). The lithology of the Terai sedimentary arkose, which are locally derived from southward-draining tribu- basin belongs to Holocene alluvium, which includes the present Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 taries from the emerging Himalayas. The most typical Siwalik day alluvial deposits, channel sand, gravel deposits, and outwash lithologies are conglomerates, “salt and pepper” micaceous sand- deposits (Yadav et al. 2014). The district has three distinct hy- stones, blue-grey siltstones, clay-stones, red (Fe-rich) shales, and drogeological zones: the Siwalik Hills, the Bhabar recharge zone, minor lignite. Potential adsorption substrates and co-precipitation and the Terai Plain unconsolidated Holocene floodplain sedi- hosts for As are common throughout the finer-grained Siwalik ments. The northern part of the district is bounded by the steeply facies, as Fe mineralization, as sulphides, or as clays. In Nepal, the sloped Siwalik Hills, which are composed of sedimentary rocks groundwater As is of relatively local provenance, being derived such as sandstone, siltstone, mudstone, shale, and conglomerates. directly from eroded Siwaliks (Stanger 2005). Immediately south lies the Bhabar zone, which is composed of High monsoon precipitation (1800–2000 mm) and year-round unconsolidated sediments that are porous, coarse such as gravel, snow-fed river systems recharge the Terai sediments, giving them cobbles, and boulder material, thereby making the Bhabar zone a high potential for groundwater resources. Shallow aquifers highly permeable (Kansakar 2004; Shrestha 2007). A major river, (<50 m) are generally unconfined or semi-confined, whereas deep the Narayani (also known as Gandaki), which descends from the aquifers (>50 m) are mostly confined by impermeable clay layers. Higher Himalaya, flows along the eastern boundary of the Nawal-

Published by NRC Research Press Mueller 299

parasi district and has had a major influence on the underlying deposits laid down during the initial uplift of the Himalayas unconsolidated Holocene fluvial deposits that comprise the flood- (Brikowski et al. 2014). The thin alluvial aquifers of the Nawal- plain aquifer system. Unlike other regions of Terai, where finer parasi district are some of the most severely As contaminated in sediments typically increase toward the south, finer sediments the Terai region. Diwakar et al. (2015) state that the alluvial sedi- predominate in the north and sand and gravels are found near the ments comprising the Terai aquifers in this district are derived Nepal–India border (S.D. Shrestha et al. 2004). In the areas with from two main sources: (i) sediments deposited by large rivers that fine-grained sediments, elevated concentrations of As are typi- erode the upper-Himalayan crystalline rocks, and (ii) weathered cally recorded (Brikowski et al. 2004, 2014; Diwakar et al. 2015). meta-sediments carried by smaller rivers originating in the Siwalik fore- Small ephemeral rivers originating from the Siwalik frontal hills. The aquifer itself is characterized by Ca-HCO3 type water and mountains disappear upon entering into the Indo-Gangetic Plain is multi-contaminated, with the WHO guideline values exceeded and reappear again in Nawalparasi. Hence, small natural ponds for As, Mn, and F in 80%, 70%, and 40% of cases, respectively. The and river meanderings were observed as characteristic geomor- middle portion of the floodplain is heavily contaminated with As, phic features of the area. Therefore, close to the frontal mountain predominantly as As(III). The river water displayed some evidence chain, the Indo-Gangetic Plain consists of boulder- to gravel-sized of reductive processes in the hyporheic zone contributing As, Fe, sediments, while soils further south consists dominantly of fine- and Mn to baseflow and also had elevated fluoride Diwakar et al. grained sediments. Guillot et al. (2015) report the lithology of (2015). Fifty-five percent of water samples collected from streams sledge core samples from five drill holes, showing various coarse that drain the Terai, sedimentary rocks of the Siwalik Group and (millimetric) to fine-grained (micrometric) sediments in the Narayani carbonate and low-grade metamorphic rocks of the Lesser Hima- basin. They distinguished light-grey to dark-grey sands; grey, laya, had ≥0.01 ppm of As (Mukherjee et al. 2009). greenish-grey to brown–grey and yellow–brown silts; and light- Provenance of the aquifer sediments is relevant for tracing the grey to black–grey, yellow–brown, and black clay with occasional source of As. As already mentioned, there are two possible sources gravel layers. Macroscopic observations showed that on average, for the Terai sediments—the Siwalik hills and the higher Himala- the drilled sediments are composed of 33% silts; 30% grey to black yas. Sediments carried from the Siwalik hills by the minor rivers clays, 27% brown clay, 9% fine-grained silt and sand, and less than seem to release more As than those carried by major rivers from 1% calcrete. Sands, silt, and clay sediments often contained micas the higher Himalayas. Rare-earth elements and other charged cat- that were occasionally massive to laminated, bioturbated, and (or) ion elements like Th, Sc, Hf, and Zr are highly immobile in most also containing roots and plant debris. Binocular observations geological processes, and thus they can be used for provenance show that the detrital minerals in the silt fraction are dominated studies. The observed enrichment of incompatible elements is also indicative of a felsic source. Sediments hosting As-contaminated by quartz, biotite, muscovite, K-feldspar, calcite, and dolomite as aquifers are therefore probably homogeneous mixtures of differ- major phases and garnet, zircon, and monazite as heavy minerals. ent types of rocks, with a felsic source. Studies of As contamina- In the region of provenance of the Narayani basin, the Tethys tion of groundwater of the Bengal delta have demonstrated the Himalaya includes 10 km of various metasedimentary rocks (lime- geological control and found that high concentration of As is stones, calc-schists, shales, and quartzites) ranging from Cam- restricted to the Holocene sediments rich in organic matter. Av- brian to Jurassic. There is also the Manaslu leucogranite emplaced erage As content of the Terai sediments is within the range of within the Tethyan rocks. The Higher Himalayan Crystallines are normal sediments (9 ppm). Abundances are greater in finer sedi- a metamorphic stack, including from base to top: paragneisses ments such as black clay (maximum 31 ppm) than in coarser (metapelites and metapsammites), gneisses with calcsilicate min-

For personal use only. sediments (silt and fine sand, 3 ppm). The sediments represent erals (diopside and amphibole), and orthogneisses representing homogeneous mixtures of a wide range of parent rocks of felsic metamorphosed Lower Paleozoic granites. The Lesser Himalaya composition. Significant As-leaching rates indicate that the Terai consists of mostly unfossiliferous metasediments and some dolo- sediments have high potential for As release, and that pH and redox mitic meta-carbonates alternating with dominant black schists, conditions play crucial roles Gurung et al. (2005). Paudyal (2011) men- aluminium-rich schists, and quartzites. Amphibolites occur in both tion that at present there exist several possible natural sources of of these groupings. The Siwaliks represent the Cenozoic foreland As in Nepal. On the basis of chemical and mineralogical analysis of basin of the Himalayan belt with local thickness of 6 km in Nepal. collected rock, minerals, soil, and water samples from different They are divided into three units having a typical coarsening-upward parts of Nepal, several primary sources of As have been identified succession. The lower unit consists of fluvial channel sandstones (Sharma 1999; Sah et al. 2003). The sulphide minerals from the alternating with calcareous paleosols; the middle unit consists of polymetallic deposit of Ganesh Himal, iron ore of the Phulchauki very thick channel sandstones with minor paleosols; and the upper area, ferruginous concretions of Tertiary deposits, bituminous unit mainly hosts conglomerates of gravelly braided river deposits coal of the Tosh area, Kalimati clay of the Kathmandu Valley, and (after Guillot et al. 2015). sediments from hot spring water all show high values of As con- Arsenic in groundwater in the Terai centration. Ferruginous quartzite, sandstone, and mudstone also show comparatively higher values of As. The above mentioned The worst affected districts in Nepal include Nawalparasi (west- minerals, rocks, and sediments could therefore represent the pri- ern region), Rautahat and Bara (central region), and Bardia (mid- mary sources of As in Nepal (Paudyal 2011). Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 western region). These districts, together with Parsa, Rupandehi, In Nawalparasi district, clays contain particularly high amounts Kapilbastu, and Banke were a priority for testing, water-supply of Fe, in the range of 21.9–59.9 g/kg (2%–6% in sediments). To- mitigation, and health screening (BGS Report 2001). The spatial gether with the high levels of Fe, high concentrations of Al were and temporal distribution of elevated groundwater As in Nepal is also extracted from the sediments (2.75–34.1 g/kg). Fe and Al in the unique in South Asia. In the Terai districts, elevated As is found sediments were positively correlated with As, with correlation exclusively in the foreland basin south of the Main Frontal Thrust, coefficients of 0.607 and 0.444, respectively. Arsenic is retained on the undisturbed floodplain. Surficial aquifers here are formed abundantly in finer particles like clay minerals, where it forms from material eroded from the thrust wedge (immediately north several different types of phases including ion exchange phases, of the Main Frontal Thrust), which is composed of earlier flood- carbonate and sulphide phases, ferric or manganous oxide and plain and later debris fan sediments exhumed by thrusting. Arse- hydroxide phases, and soil organic matter phases depending on nic occurrences are further limited to the areas immediately pH and redox potential (Eh) (Nakano et al. 2014). Yadav et al. (2015) downslope from exposures of the fine-grained Lower Siwalik describe that As concentration varied from 0.22 to 0.64 ppm (mean Formation (Smith et al. 2004), comprised of meandering stream 0.36 ppm) in sediment samples. Comparatively, a higher concen-

Published by NRC Research Press 300 Environ. Rev. Vol. 25, 2017

tration of As was observed mostly in the fine-grained clay sedi- tion onto Fe-oxide, Mn-oxide, or clay surfaces during high-redox ments (black and yellow) than in coarse-grained sediments. A medium-pH conditions (i.e., about 5.5–6.5). Subsequently, desorp- variety of Fe minerals in the Nawalparasi aquifer system are key tive release of As occurs as groundwater becomes more reducing host phases for As. Specific examples include goethite (␣-FeOOH), and alkaline (i.e., negative Eh and pH > 6.5), principally as the

authigenic pyrite (FeS2) in deeper organic-clays, and ferrihydrite byproduct of bacterially mediated FeOOH dissolution. Since the flocs (Johnston et al. 2015). Vertical distribution of Fe followed reducing agent is buried organic material, such as peat, man- similar distribution pattern as that of As, showing its higher and groves, reed-swaps, etc., and since the predominant adsorption lower concentrations in clay and fine-sand, respectively. High As- substrate is goethite or its analogues, with clay, high-Fe and high- yielding aquifers also contained higher percentages of Ca, silica, organic sedimentary environments are evidently prerequisites for Al, and Fe. Arsenic occurs generally in oxyanionic forms in aque- the modern release of As (Stanger 2005). ous environment. The hydrogeochemical data for groundwater of Arsenic (As, atomic number = 33) is a ubiquitous element, which the Terai Alluvial Plain aquifers suggest predominantly reducing ranks 20th in the earth’s crust. Arsenic exists in four oxidation − 2− − states—+V (arsenate), +III (arsenite), 0 (arsenic), and −III (arsine). Ar- character, with high HCO3 and low SO4 and NO3 concentra- − senic is unique among the heavy metalloids and oxy-anion forming tions. Elevated HCO3 levels result primarily from the oxidation of 2− elements. Its sensitivity to mobilization largely depends on the pH organic matter, while low SO4 levels are a result of sulphate reduction (Bhattacharya et al. 2003). Yadav et al. (2012) found three values typically found in groundwater (pH 6.5–8.5) under both types of tube wells that are used as a source of drinking water in oxidizing and reducing conditions. The valency and species of the Terai region that all vary by depth. These three include shal- inorganic As highly dependent on the redox conditions (Eh) and low tube wells (<50 m deep), deep tube wells (>50 m deep), and the pH of the groundwater. Arsenite, the reduced trivalent form dug wells (up to 20 m or more). A majority of them (98%) were (As (III)), is normally present in groundwater (assuming anaerobic shallow tube wells. The depth of dug wells displayed various As conditions) while arsenate, the oxidized pentavalent form (As (V)), concentrations. The depth of deep tube wells ranged from 1 to is present in surface water (assuming aerobic conditions). In gen- 183 m. Virtually all (97%) of the tested tube wells that had As levels eral, inorganic As species are more toxic than organic forms of As exceeding WHO guidelines were of a depth less than 20 m. At this for living organisms. As already mentioned, redox potential and depth range, more than 8% of tube wells had As levels above pH basically control As speciation in natural environments. Inor- 10 mg/L, while only 2% of tube wells had levels above 50 mg/L. At a ganic As primarily occurs as arsenic acid (H3AsO4) under oxidizing depth of 21–50 m, 4.7% and 1.3% of the water in tube wells had As conditions, and predominates only at extremely high Eh values concentrations that exceeded the 10 and 50 mg/L guideline levels, and low pH (<2). Within a pH range of 2–11 it is replaced by H AsO − and HAsO 2−. At low Eh values, H AsO (arsenious acid) respectively. Similarly, at a depth greater than 50 m, tube wells 2 4 4 3 3 exists up to moderately alkaline pH but is replaced by H AsO − at having an As concentration that exceeded guideline values (10 and 2 3 pH > 9.2 (Thakur et al. 2011; Zakhaznova-Herzog and Seward 2006). 50 mg/L) were significantly fewer in number. Therefore, it seems Bhattacharya et al. (2003) report that the groundwater in the that tube wells having a depth less than 20 m had on average Terai is mostly near-neutral to alkaline within a pH range of 6.1– higher As concentrations. Most of the known wells record a high 8.1. Redox potential (Eh) levels between −0.20 and −0.11 V suggest As concentration in March, and a low value in May and Septem- a fairly reduced condition in the aquifers. The groundwater is ber. A general pattern of low As – low piezometric level, high As – − − predominantly of Ca-Mg-Na-HCO3 -type with HCO3 as the princi- high piezometric level can be observed (S.D. Shrestha et al. 2004). − 2− − pal anion and low levels of Cl and SO4 . Low NO3 coupled with According to Emerman (2005), central Nepal does not contain one + For personal use only. elevated NH concentrations in this groundwater reflects the geographically limited source of As and that nearly all rivers 4 dissimilarity nitrate reduction in the aquifers. Moreover, redox showed elevated levels of As. Nearly all rivers also showed elevated lev- levels (Eh < −0.2 V) for sulphate reduction are sufficiently low, els of Cu, Co, Fe, and Ni, while fluvial Zn was very close to the which facilitates the reduction of Fe3+ and Mn4+ in the aquifer global background level. Therefore, As mineralisation may be as- sediments. The source of As in the subsurface environment is geogenic, sociated with the mineralisation of Cu, Co, Fe, or Ni, but probably and principally mobilized through natural interaction of the not with Pb–Zn mineralisation (Pb and Zn are almost always asso- aqueous phases with the aquifer sediments under anoxic condi- ciated). Bhusal and Paudyal (2014) clearly state that the distribu- tions. The sequence of redox reactions or terminal electron ac- tion and occurence of As is controlled by geological material and cepting processes prevalent in the aquifers plays a critical role in much less by topography and not by land use, artificial fertilizers, controlling the As chemistry in groundwater. The predominant pesticides, and other organic additives. terminal electron accepting processes in the sedimentary aquifers are O reduction (aerobic respiration), NO − reduction (denitrifi- Mechanisms of arsenic release to groundwater 2 3 cation and dissimilatory nitrate reduction), Mn4+ reduction, Fe3+ Since the fundamental work by Nickson et al. (2000) some sci- 2− − 4+ 3+ reduction, and SO4 reduction with oxygen (O2), NO3 ,Mn ,Fe , entific articles about the specific situation and mechanisms of As 2− and SO4 as the prominent electron acceptors. High levels of Fe release to the groundwater in the Terai in Nepal have been pub- and Mn in the groundwater together with the predominance of lished. As outlined by Nickson et al. (2000), the As in the ground- As(III) in the groundwater suggest that As is mobilized becuae of water derives from reductive dissolution of As-rich Fe-oxyhydroxides the reductive dissolution of Fe and Mn oxides and hydroxides with Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 that exists as a dispersed phase (e.g., as a coating) on sedimentary sorbed As-oxyanions in the Terai sediments. According to Panthi grains. The reduction is driven by microbial degradation of sedi- et al. (2006), the reductive desorption theory is the most likely mentary organic matter (O2 consuming, O2 as electron acceptor) explanation in which As-rich Fe-oxides break down and get dis- and the redox process that occurs after microbial oxidation of solved into water regarding the context of strongly reducing organic matter takes place as soon as dissolved O2 and NO3 are environments (Eh −110 to −200 mV) of groundwater in Nepal. disappeared. Strong correlation between dissolved organic car- Moreover, the As is thought to be closely associated with the bon (DOC) and As in groundwater suggests that the microbial oxidation-reduction process of Fe-oxides and pyrite. Evidence ex- degradation of organic matter in the sediment results in an over- ists to support oxidizing/reducing desorption of Fe-oxides and all reducing environment and facilitates the release of As in the pyrite oxidation theories of releasing As. But negative correlation 2− groundwater (Halim et al. 2009). Whilst As release by the dissolu- between As and SO4 demonstrates the As may not be directly tion of arsenious pyrite is still recognized as a minor contributing mobilized from sulphide minerals like arsenopyrite. In flooded but widespread process, a consensus view emerges in which the soils, As is mobilized into porewater owing to reductive dissolu- dominant process is, initially, the fixation of aqueous As by sorp- tion of FeIII-(hydr)oxides and to arsenate (AsV) reduction to the less

Published by NRC Research Press Mueller 301

competitively sorbing arsenite (AsIII). By contrast, As concentra- and geomorphological settings of the aquifers. Redox-sensitive tions in porewater are markedly lower under oxic conditions and parameters indicate generally reducing, post-oxic, metal-reducing generally dominated by AsV (Roberts et al. 2011). Furthermore, the conditions. However, redox conditions are highly spatially vari- equilibrium of groundwater with respect to carbonate minerals able (oxic to methanic), with no systematic depth variation within and their precipitation/dissolution seems to be controlling the sampled depth of the aquifers. Nakano et al. (2014) assumed that 2− overall groundwater chemistry. The low SO4 and high DOC, the brown color of the sediments in the Terai arises from the 3− − PO4 , and HCO3 concentrations in groundwater signify ongoing presence of Fe(III) and the gray color can be induced by the reduc- microbial-mediated redox processes favoring As mobilization in tion of Fe(III). The dissolution of FeOOH seems to be mainly de- the aquifer. Multiple geochemical processes, e.g., Fe-oxyhydroxides re- rived from microbial fermentation under redox condition. They duction and carbonate dissolution (pH!), are responsible for high found that microbial degradation accompanying Fe reduction re- As occurrence in groundwaters Bhowmick et al. (2013). leased As attached on the surface of Fe-bearing solids; however, The generally sub-oxic conditions, dominance of As(III) and Fe2+ the released As coupled with dissolution of Fe can be continuously

species, and positive correlation between As and both NH3 and resorbed on the surface of solid phases like aluminosilicates (clay UV-absorbance at 254 nm suggests that oxidation of organic mat- minerals) and silty sediments. Another possibility of resorption ter coupled with microbial-mediated reductive processes are are crystalized Fe-bearing minerals, which might be reproduced important for mobilizing As in the aquifers in the Terai. The gen- along with As during the sediment–water interactions controlled erally low redox potential of tube well waters combined with the by microbial activity and redox condition. Microbial activity will abundance of reduced species of various redox sensitive elements be strongly affected by redox and pH changes. Upon saturation of 2+ (i.e., Fe , As(III), and NH3) clearly indicates that reductive pro- adsorption sites, the As remains in the groundwater. The dissolu- cesses are important controls on aquifer geochemistry Diwakar et al. tion of calcium-related minerals may also play an important role (2015). For example, McArthur et al. (2011) proposed that the ab- in the process of releasing As as this dissolution raises the pH sence or presence of a palaeo-weathering surface was a key con- locally, making the environment more alkaline. Alkaline condi- trol on As heterogeneity at their study site in West Bengal, India. tions favor the desorption of As from As-bearing oxides as well as − 2− They suggested that a palaeo-weathering surface formed during from organic matter. Low concentrations of NO3 and SO4 to- the last glacial maximum protects the underlying Pleistocene gether with high Fe, as found in the geochemical analysis, also aquifer from contamination with DOC- and As-enriched water. indicates reducing conditions being prevalent in Terai groundwa- According to Brikowski et al. (2014), mitigation efforts concerning ter. In sequential extraction techniques, chemical leaching by po- elevated As in groundwater in Southeast Asia are hindered by tassium chlorate and HCl releases As from sulphide and silicate persistent uncertainty about the proximal source of As and mech- phases. As exhibited in regression analysis, weak interrelation- 2− anisms for its mobilization. At the core of this uncertainty seems ship between As, Fe, and SO4 suggests the absence of a pyrite/ to be the relative roles of surficial organic clays versus deeper arsenopyrite oxidation mechanism in the present site. Further, if aquifer matrix Fe-oxyhydroxides. Temporal variations in ground- pyrite would have been oxidized, then As would have been sorbed water chemistry can serve to distinguish the contributions of these onto the resulting Fe-oxyhydroxide rather than getting released two sources, and such variation is especially pronounced in headwa- in the groundwater. The leachable As content was high in organic ter areas of the Ganges floodplain immediately adjacent to the matter phase next to sulphide/silicate phase as observed in se- Himalayan foothills (e.g., the Terai of Nepal). Monsoon recharge quential leaching analysis. This is an indication of the role of refreshes these aquifers, temporarily minimizing As concentra- microbial population and organic matter in mobility of As under

For personal use only. tions. Post-monsoon, average groundwater compositions exhibit reducing condition. Moreover, the microbial oxidation of organic increasing trends in water–rock interaction (higher total dis- matter consumes dissolved oxygen present in the groundwater − solved solids, with cation exchange to form increasingly Na–HCO3 resulting in the formation of HCO3 . The distribution of grain size waters), as well as in As and Fe concentrations. This cycle can be of the sediments in groundwater may also play a vital role in the repeated during dry-season precipitation events as well, revealing mobility of As. It is evident from XRF analysis that high As con- a direct correlation between trends in degree of clay interaction centration was mostly associated with fine-grained clay minerals. (sodium fraction of major cations) and As concentrations. These ob- As the fine grain-size fraction has larger surface area it adsorbs the servations strongly support a model of reductive mobilization of As major part of As on their surface. Since, Fe, Mn, and Al oxides and from adjacent clays into aquifers, tempered by repeated flushing hydroxides are the major components of fine-grained particles during periods of appreciable rainfall. Surficial sediments in the and thought to retain high As under specific pH conditions, their Terai exhibit extreme heterogeneity. Highly organic clays pre- abundant percentage in Terai groundwater also suggests a reduc- dominate in the shallow hydrologic system (the upper 50–100 m tive dissolution mechanism for As release Yadav et al. (2015). Find- of surficial sediments contain >70% clay), and aquifer hydraulic ings presented by Johnston et al. (2015) provide direct XAS-based conductivities are two orders of magnitude lower than in the quantification of solid-phase As and Fe speciation in the alluvial delta. Low hydraulic conductivity of surficial fines limits infiltra- aquifer sediments of the Terai region and help to shed light on key tion, which likely enhances reducing conditions and mobilization processes controlling spatial patterns of solid-phase As/Fe specia- of As. In the Terai these factors combine to yield highly heteroge- tion. Their dataset is broadly consistent with the widely invoked neous groundwater As concentrations both in space and time, hypotheses that reductive dissolution of (near surface) Fe-oxides Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 providing a valuable setting for exploring the As mobilization and (or) reductive desorption of As(III) coupled with downward process. transport are largely responsible for As mobilization in Gangetic Mukherjee et al. (2012) state that information on groundwater floodplain aquifers (e.g., Fendorf et al. 2010). The findings also chemistry in the central Ganges basin could provide insights into strongly affirm the critical role that various Fe minerals can play recharge, provenance, and fate of solutes in As-affected areas up- as host phases for As as it undergoes redox cycling throughout the stream of the more intensively studied Bengal basin. The area floodplain landscape. Most tube wells on the Nawalparasi flood- they studied extends from the northern edge of the Indian craton plain are screened more than 15–20 m below ground level outcrops to the foothills of the Himalayas. Arsenic is probably (Gurung et al. 2005) to tap permanently saturated thin sandy lay- mobilized by reductive dissolution of Fe–Mn (oxyhydr)oxides in ers. Data presented in this article for the various floodplain sites the alluvium, with possibility of competitive anionic mobiliza- indicate that at these depths, solid phase As(III) and lower valency tion. Hence, relative to the Bengal basin, in addition to lower As-sulphide species are the dominant species, while poorly crys- groundwater abstraction influence, groundwater chemistry in their talline Fe(III)- and Fe-oxides are largely absent. Consistent with the study area reflects a greater variety of differences in the geological findings of Polizzotto et al. (2005), the paucity of Fe-oxides at the

Published by NRC Research Press 302 Environ. Rev. Vol. 25, 2017

depth of tube well screens suggests that current mobilization of of iron nails (3 cm depth), 2 kg of coarse sand (4 cm depth), and As(III) within these sedimentary facies is more likely owing to raw water. The middle filter contains the following from bottom downward transport or desorption of As(III) rather than contem- to top: a layer of polyester cloth, about 50 kg of brickbats, 2 kg of porary in situ reductive dissolution of As-bearing Fe-oxides occur- fine sand (3.5 cm depth), 1 kg of charcoal (6 cm depth), 2 kg of ring at the depth of well screens. Johnston et al. (2015) state that brickbats (3 cm depth), and filtered water from the top filter. This downstream transport is likely to be followed by some degree of filter could remove 95%–99% of the As, but there were problems floc reburial on the floodplain and therefore result in exposure of with high Fe in treated water and filter clogging because of bac- Fe(III) floc to seasonally fluctuating redox conditions. The material terial growth. This filter was quickly replaced by an As biosand is freshly precipitated, very poorly-crystalline—hence susceptible filter (Neku and Tandukar 2003; Thakur et al. 2011). to reductive dissolution—and contains readily exchangeable As at concentrations well above those of bulk sediments. Arsenic- Kanchan filters bearing authigenic pyrite occurs within 12 m of the ground surface These Fe-assisted biosand filters were constructed on the basis at various floodplain sites and close (ϳ5 m) to the current range of of As removal from water using zero-valent Fe (ZVI) media. Under seasonal water table fluctuations. While stable under reducing condi- conditions applicable to drinking water treatment, arsenate re- tions, if there is some regional lowering of water tables through moval by ZVI media involves surface complexation only and does prolonged drought, climate-induced shifts in monsoonal precipi- not involve reduction to metallic As. Under the pH and redox tation, or excess groundwater abstraction, then these materials conditions of most groundwaters and surface waters, dissolved As − 2− may be at risk of exposure and oxidation. Although oxidation of exists as As(V) (arsenate) species, H2AsO4 and HAsO4 , and As(III) 0 − pyrite may simply cause As to shift host phases and become se- (arsenite) species, H3AsO3 and H2AsO3 . Removal of As occurs questered in the resulting Fe(III)-oxides (Polizzotto et al. 2006), it is through adsorption and coprecipitation during the formation conceivable there could be consequences for mobilizing addi- of Fe(III)-hydroxides. However, acceptable levels of removal are tional As in the aquifer, especially in the short-term. achieved only when there is a filtration step to remove colloidal As (Farrell et al. 2001). Greater attention is required for the removal Mitigation options: types of arsenic removal filters of As(III) from groundwater owing to its higher toxicity and mobility, which mainly arise from its neutral state (H AsO 0) used in the Terai of Nepal 3 3 in groundwater as compared with the charged As(V) species Following the study of Sharma (1999), several organizations and − 2− (H2AsO4 and HAsO4 ), which predominate near pH 6–9. The agencies have conducted surveys into As contamination of well As(III) removal mechanism is mainly due to spontaneous adsorp- water in Nepal. In 2003, NRCS/ENPHO provided the following six tion and coprecipitation of As(III) with Fe(II) and Fe(III) oxides and types of mitigation options for arsenicosis patients in all VDCs of hydroxides, which form in situ during ZVI oxidation (corrosion). Rautahat district: (i) two-gagri (water vessel) filter, (ii) innovated Heterogeneous reactions at the corroding ZVI surface are complex dug well, (iii) As–Fe removal plant (AIRP), (iv) tube wells from As and result in a variety of potential adsorption surfaces for As(III) free aquifer, (v) modified biosand filter, and (vi) awareness pro- and As(V). Evidence has been presented showing that As(III) can be gram on nutrition. Of these, the option of two-gagri filter and removed by adsorption on nanoscale ZVI (NZVI) in a very short awareness program has been provided in Bagahi (Pradhan 2006). time (minute scale) and is strongly adsorbed on NZVI over a wide According to Nakano et al. (2014), gagri filters and Fe-assisted range of pH and anion environments (Kanel et al. 2005). Investi- biosand filters were later commonly used to remove As and Fe gations by Neumann et al. (2013) regarding the SONO household

For personal use only. from well water in Nepal, which are believed to be the best treat- filters used in Bangladesh (an other version of an Fe-assisted bio- ment option at household levels (Yadav et al. 2011). The remainder sand filter) showed that over 95% of the As passing the top sand of this review will focus on the description of these two household layer was removed in the composite Fe matrix (CIM) by sorption, filter types. coprecipitation, and incorporation into solids formed during the corrosion of ZVI. The continued presence of dissolved Fe(II) in Gagri filters the CIM appears to be important for the long-term operation of One of the first filters employed was the two-gagri filter system the filters. While young CIM contained large fractions of As in with chemical powder. The system, consisting of two earthen pots amorphous or poorly crystalline phases, magnetite was dominant (Nepali language: gagri), uses chemical powder (a mixture of in older CIM, consistent with an invariably deep black color. The FeCl3, NaOCl, and charcoal). Ferric chloride is the compound that transformation of As-rich Fe(III)-(hydr)oxides into magnetite is im- removes As present in affected water. The candle filter aids in portant for the following two reasons: (i) the much denser mag- filtration of the coagulants formed in the upper pot. The second netite does not lead to clogging of the filter and (ii) magnetite is pot underneath the first one receives water free from As, Fe, bac- more stable toward dissolution than freshly formed amorphous teria, and odour. This system is 90% efficient in removing As and is phases and leaches less As during milder extraction steps. Leach- below the Nepal interim standard. Further development led to the ing tests with spent CIM in a previous study have shown very low three-gagri filter system. This filter replicates the three-kulsi sys- remobilization of As, rendering used CIM nonhazardous. Because tem of Bangladesh and solves the problem of chemical powder As is removed predominantly in the CIM, the other filter compo- use. Oxidation, adsorption, precipitation, and filtration are the nents such as sand, brick chips, and the plastic components can Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 processes for removal of As and Fe in this filter. This filter system be disposed of without special care. can remove up to 95% of the As, even when the water is highly The biosand filter (the modified model now used in Nepal contaminated. Retardation of the filtration process because of is known as a Kanchan filter) as a point-of-use drinking water clogging and presence of microbes in the treated water limits the treatment option was initially designed by David Manz of the filter’s performance. Therefore, techniques for improvement of University of Calgary, Canada, in the late 1990s with support of microbiological quality should also have been used while provid- numerous organizations and individuals. The biosand filters were ing this option (B.R. Shrestha et al. 2004). The three-gagri filter is a modified to remove As and tested in Nepal jointly by Massachu- water container made of copper, brass, steel, tin, and or clay pot. setts Institute of Technology (MIT) researchers; ENPHO, Nepal; The three-gagri filter unit consists of three clay pots staggered Rural Water Supply and Sanitation Support Programme (RWSSSP), vertically witha1cmdiameter hole in the bottom of the middle Nepal; and CAWST, Canada, based on slow sand filtration and and top filters. The top and middle filters work as a reactor, and Fe-hydroxide adsorption principles (Thakur et al. 2011). Now such the bottom filter stores the treated water. The top filter contains filters can be used for removal of As, Fe, bacteria, and turbidity. the following, from bottom to top: a layer of polyester cloth, 3 kg This filter uses the process of aeration, adsorption, and filtration.

Published by NRC Research Press Mueller 303

Fig. 2. Diagram of the Kanchan Arsenic Filter, showing the location Fig. 3. Operating filter in Nawalparasi district, Nepal, October 2015. and arrangement of its components. Source: Ngai et al. (2005). Photo: B. Mueller.

As this system has a high flow rate of 30 L/h, the biosand filter has become in high demand in communities, not only for As removal but also for its higher flow rate. Field test showed that this filter removes more than 95% of the As on average and up to 99% in some cases (NRCS–ENPHO 2003; Ngai and Walewijk 2003). The filter also removes high levels of Fe—up to 99%, with an average of 95%. The microbiological quality of this treated water has been deemed satisfactory (B.R. Shrestha et al. 2004). The Kanchan Arsenic Filter™ (KAF), an award-winning house-

For personal use only. hold water filter, was constructed for simultaneous As and patho- gen removal. The KAF is constructed using locally available labour and materials and is optimized based on the local socio-economic conditions. The KAF combines the concept of a slow sand filter for intermittent use (i.e., a biosand filter base) with the innovation of gravel. There is a standing water height of 5 cm above the sand a diffuser basin containing (rusty) iron nails for As removal. Op- layer. The diffuser basin is filled with 5–6 kg of non-galvanized erating under the water quality conditions encountered in the iron nails for As removal. In addition, pathogens, Fe, and sus- Terai region of Nepal (total As < 500 mg/L, phosphate < 2 mg/L, pended material are removed from the water through a combina- pH < 8) the iron nails can last 3 years before replacement is nec- tion of biological and physical processes—mechanical trapping, essary (Ngai et al. 2006). A two-year technical and social evaluation adsorption/attraction, predation, and natural death. This filter of over 1000 KAFs deployed in rural villages of Nepal determined can treat approximately 10–15 L/h of As-contaminated water. The that the KAF typically removes 85%–90% As, 90%–95% Fe, 80%–95% filters are locally available at a cost of about 1400–1800 NRs (about turbidity, and 85%–99% total coliforms. In total, 83% of the house- US$20) per filter (Thakur et al. 2011). Figure 3 exhibits one of these holds continued to use the filter after 1 year, mainly motivated by filters operating in the district of Nawalparasi in October 2015. the clean appearance, improved taste, and reduced odour of the This ZVI-based filter is able to remove As and other pollutants filtered water, as compared with the original water source. Inside from drinking water, but its performance depends on the form of the KAF, non-galvanized iron nails are exposed to air and water, ZVI, filter design, water composition, and operating conditions. rusting quickly and producing ferric hydroxide on the iron nails’ KAFs use an upper bucket with ZVI in the form of commercial iron Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 surface. When As-containing water is poured into the filter, As is nails, followed by a sand filter, to remove As and pathogens. Wenk rapidly adsorbed onto the surface of the ferric hydroxide. This et al. (2014) evaluated factors that influence the removal of As mechanism is similar to As adsorption on ZVI and As adsorption and uranium (U) with laboratory columns containing iron nails on hydrous ferric oxides. Some of the As-loaded Fe particles are with six different synthetic groundwaters at pH 7.0 and 8.4 over flushed on to the sand layer below where they are trapped in the 30 days. During the first 10 days, As removal was 65%–95% and top few centimeters of the fine sand because of straining. As ferric strongly depended on the water composition. As removal at hydroxide particles “exfoliate” from the iron nails, new iron sur- pH 7.0 was better than at pH 8.4 and high P combined with low Ca faces are created, providing additional As adsorption capacity. A decreased As removal. From 10–30 days, As removal decreased to Kathmandu university study found that Fe and As do not migrate 45%–60% within all columns. Phosphate in combination with low through the sand media over time (Ngai et al. 2007). The filter Ca concentrations lowered As removal, but in combination with container can be constructed out of concrete or plastic. The high Ca concentrations a slightly positive effect was seen. The container is about 0.9 m tall by 0.3 m in diameter (Fig. 2). The drop in performance over time can be explained by a decreasing container is filled with layers of sieved and washed sand and release of Fe to solution because of the formation of layers of FeIII

Published by NRC Research Press 304 Environ. Rev. Vol. 25, 2017

phases and calcite covering the Fe surface. Mobile corrosion prod- Bhattacharya, P., Tandukar, N., Neku, A., Valero, A.A., Mukherjee, A.B., and ucts contained ferrihydrite, Si-containing hydrous ferric oxides, Jacks, G. 2003. Geogenic arsenic in groundwaters from Terai Alluvial Plain of Nepal. J. Phys. IV France, 107: 173–176. doi:10.1051/jp4:20030270. and amorphous Fe–Si–P phases. Comparisons with another type Bhowmick, S., Nath, B., Halder, D., Biswas, A., Majumder, S., Mondal, P., et al. of ZVI filter (SONO filter, see Hussam and Munir 2007) were used 2013. Arsenic mobilization in the aquifers of three physiographic settings of to evaluate filter design parameters. Higher ZVI surface areas and West Bengal, India: Understanding geogenic and anthropogenic influences. longer contact times should lead to satisfactory As removal with J. Hazard. Mater. 262: 915–923. doi:10.1016/j.jhazmat.2012.07.014. PMID:22999019. Kanchan-type filters. Economical and promising methods are co- Bhusal, S., and Paudyal, K. 2014. Status of arsenic contamination in groundwater of Makar VDC of Nawalparasi District, Nepal. Int. J. Environ. 3(3): 275–285. II III precipitation of As with naturally occurring or added Fe or Fe , doi:10.3126/ije.v3i3.11087. sorption or adsorption to inexpensive prefabricated sorbents or Brikowski, T.H., Smith, L.S., Shei, T.C., and Shrestha, S.D. 2004. Correlation of ion exchangers, or As removal with metallic Fe. Phosphate inter- electrical resistivity and groundwater arsenic concentration, Nawalparasi, acts strongly with precipitating FeIII-(hydr)oxides and outcom- Nepal. J. Nepal Geol. Soc. 30: 99–106. Brikowski, T.H., Neku, A., Shrestha, S.D., and Smith, L.S. 2014. Hydrologic con- petes As for sorption and incorporation, such that additional Fe is trol of temporal variability in groundwater arsenic on the Ganges floodplain necessary to remove both As and phosphate. Removal units using of Nepal. J. Hydrol. 518(C): 342–353. doi:10.1016/j.jhydrol.2013.09.021. metallic ZVI are promising for several reasons, including ZVI fil- British Geological Survey (BGS) Report. 2001. Groundwater Quality: Nepal, 4 pp. ters can be constructed with locally available materials (typically Diwakar, J., Johnston, S.G., Burton, E.D., and Shrestha, S.D. 2015. Arsenic mobi- sand and iron in various forms such as turnings, filings, nails, or lization in an alluvial aquifer of the Terai region, Nepal. J. Hydrol. Reg. Studies, 4(A): 59–79. doi:10.1016/j.ejrh.2014.10.001. cleaned scrap iron). Corroding Fe can potentially produce the larg- Emerman, S.H. 2005. Arsenic and other heavy metals in the rivers of central est amount of As-sorbing FeIII-(hydr)oxides per mass of starting Nepal. J. Nepal Geol. Soc. 31: 11–18. doi:10.3126/jngs.v31i0.249. material. Aerobic Fe corrosion leads to oxidation of AsIII to the Emerman, S.H., Prasai, T., Anderson, R.B., and Palmer, M.A. 2010. Arsenic con- more strongly sorbing AsV, without the need of added oxidants. tamination of groundwater in the Kathmandu Valley, Nepal, as a conse- Two measures that could improve the performance of KAFs are quence of rapid erosion. J. Nepal Geol. Soc. 40: 49–60. FAO. 2004. Arsenic threat and irrigation management in Nepal. Rome. (i) larger specific ZVI surface areas (e.g., by use of smaller nails) and Farrell, J., Wang, J., O’Day, P., and Conklin, M. 2001. Electrochemical and spec- (ii) increased contact times by more controlled and restricted flow troscopic study of arsenate removal from water using zero-valent iron media. from the upper diffuser bucket. As Singh et al. (2014) state in their Environ. Sci. Technol. 35(10): 2026–2032. doi:10.1021/es0016710. PMID:11393984. article, KAF efficacy in field conditions operating for a long period Fendorf, S., Michael, H.A., and van Green, A. 2010. Spatial and temporal varia- tions of groundwater arsenic in South and Southeast Asia. Science, 328: has been scarcely observed. They observed the efficacy of KAF 1123–1127. doi:10.1126/science.1172974. running over 6 months in highly As-affected households in Nawal- Guillot, S., Garçon, M., Weinman, B., Gajurel, A., Tisserand, D., parasi district. Of 62 tube wells, 41 had influent As concentration France-Lanord, C., et al. 2015. Origin of arsenic in Late Pleistocene to Holo- exceeding the Nepal drinking water quality standard value (50 ␮g/L). Of cene sediments in the Nawalparasi district (Terai, Nepal). Environ. Earth Sci. the 41 tube wells having unsafe As levels, KAF reduced As concen- doi:10.1007/s12665-015-4277-y. Gurung, J.K., Ishiga, H., and Khadka, M. 2005. Geological and geochemical ex- tration to the safe level for only 22 tube wells, an efficacy of 54%. amination of arsenic contamination in groundwater in the Holocene Terai In conclusion, they did not find a significantly high efficacy of Basin, Nepal. Environ. Geol. 49: 98–113. doi:10.1007/s00254-005-0063-6. KAFs in reducing unsafe influent As level to the safe level under Hagen, T. 1969. Report on geological survey of Nepal Preliminary Reconnais- the in situ field conditions. sance. Memoires de la Soc. Helvetique des Sci. Naturekkes, Zurich, Switzer- land. Halim, M.A., Majumder, R.K., Nessa, S.A., Hiroshiro, Y., Uddin, M.J., Shimada, J., Summary and future perspective and Jinno, K. 2009. Hydrogeochemistry and arsenic contamination of ground- As mentioned above, the factors that influence the removal of water in the Ganges Delta Plain, Bangladesh. J. Hazard. Mater. 164(2–3): 1335– For personal use only. As with laboratory columns containing iron nails were evaluated. 1345. doi:10.1016/j.jhazmat.2008.09.046. PMID:18977593. Hug, S.J., Gaertner, D., Roberts, L.C., Schirmer, M., Ruettimann, T., Rosenberg, T.M., et al. As stated, the drop in performance over time could be explained 2011. Avoiding high concentrations of arsenic, manganese and salinity in by a decreasing release of Fe to solution because of the formation deep tubewells in Munshiganj District, Bangladesh. Appl. Geochem. 26(7): of layers of FeIII phases and calcite covering the iron surface. In- 1077–1085. doi:10.1016/j.apgeochem.2011.03.012. spection of operating filters in Nawalparasi district during a field Hussam, A., and Munir, A.K.M. 2007. A simple and effective arsenic filter based campaign in October 2015 often revealed corrosion products cov- on composite iron matrix: Development and deployment studies for ground- water of Bangladesh. J. Environ. Sci. Health Part A, 42: 1869–1878. doi:10.1080/ ering the nails as well as insufficent contact time with the nails. 10934520701567122. Higher ZVI surface areas and longer contact times should lead to Johnston, S.G., Diwakar, J., and Burton, E.D. 2015. Arsenic solid-phase speciation satisfactory As removal with Kanchan-type filters. Three critical in an alluvial aquifer system adjacent to the Himalayan forehills, Nepal. measures that could improve the performance of KAFs include Chem. Geol. 419: 55–66. doi:10.1016/j.chemgeo.2015.10.035. Kanel, S.R., Manning, B., Charlet, L., and Choi, H. 2005. Removal of arsenic(III) larger specific ZVI surface areas (e.g., by use of smaller nails), from groundwater by nanoscale zero-valent iron. Environ. Sci. Technol. increased contact times by more controlled and restricted flow 39(5): 1291–1298. doi:10.1021/es048991u. PMID:15787369. from the upper diffuser bucket, and immersed and anoxic condi- Kansakar, D.R. 2004. Geologic and geomorphologic characteristics of arsenic tions in the nailbed under no-flow conditions. Further improve- contaminated groundwater area in Terai, Nepal. In Arsenic testing and final- ization of groundwater legislation project, summary project report. Edited by ments concerning these questions are under investigation. D.R. Kansakar. Department of Irrigation, Lalitpur, Nepal, HMG/Nepal, pp. 85–96. Acknowledgements MacDonald, A.M., Bonsor, H.C., Ahmed, K.M., Burgess, W.G., Basharat, M., Calow, R.C., et al. 2016. Groundwater quality and depletion in the Indo-

Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 I am grateful for the assistance of Stephan Hug, Eawag, Düben- Gangetic Basin mapped from in situ observations. Nat. Geosci. 9: 762–766. dorf; Christian de Capitani, Department of Environmental Sci- doi:10.1038/ngeo2791. ence, University of Basel; and Marcel Guillong, Earth Sciences Maharjan, M., Watanabe, C., Ahmad, S.A., and Ohtsuka, R. 2005. Short report: Department, ETH Zürich. My great appreciation for support is also Arsenic contamination in drinking water and skin manifestations in lowland expressed to Tommy Ngai, Candice Young-Rojanschi, Finn Macdon- Nepal: The first community-based survey. Am. J. Trop. Med. Hyg. 73(2): 477– ald, and Laura MacDonald from CAWST, Calgary, Canada; Bipin Dangol 479. PMID:16103627. Maharjan, M., Shrestha, R.R., Ahmad, S.A., Watanabe, C., and Ohtsuka, R. 2006. and Hari Boudhatoki, ENPHO, Kathmandu, Nepal; Gyan Prakash Prevalence of arsenicosis in Terai, Nepal. J. Health Popul. Nutr. 24(2): 246– Yadav, Parasi, Nepal; and last but not least to Shankar Rai and Som 252. PMID:17195566. Rai, my loyal expedition and trekking guides in Nepal, responsi- Mahat, R.K., and Kharel, R.P. 2009. Status of arsenic contamination and assess- ble for logistics over many years. ment of other probable heavy metal contamination in groundwater of Dang district in Nepal. Sci. World, 7: 33–36. doi:10.3126/sw.v7i7.3821. Mahat, R.K., and Shrestha, R. 2008. Metal contamination in ground water of References Dang district. Nepal J. Sci. Technol. 9: 143–148. doi:10.3126/njst.v9i0.3178. Adhikari, H.J., and Ghimire, T.R. 2009. Prevalence of arsenicosis in Ramgram McArthur, J.M., Nath, B., Banerjee, D.M., Purohit, R., and Grassineau, N. 2011. municipality, Nawalparasi, Nepal. Int. J Health Res. 2: 183–188. Palaeosol control on groundwater flow and pollutant distribution: The example

Published by NRC Research Press Mueller 305

of arsenic. Environ. Sci. Technol. 45(4): 1376–1383. doi:10.1021/es1032376. PMID: Sharma, C.K. 1995. Shallow (phreatic) aquifers of Nepal. Sangeeta Publishing 21268629. Kathmandu, Nepal, 1st Ed. 272 pp. Mukherjee, A., Fryar, A.E., and O’Shea, B.M. 2009. Major occurrences of elevated Sharma, R.M. 1999. Research study on possible contamination of groundwater arsenic in groundwater and other natural waters. In Arsenic. Edited by K.R. with arsenic in Jhapa, Morang, and Sunsari districts of Eastern Terai of Nepal. Henke. John Wiley & Sons, Ltd. pp. 303–350. Report of WHO Project, DWSS Government of Nepal. Mukherjee, A., Scanlon, B.R., Fryar, A.E., Saha, D., Ghosh, A., Chowdhuri, S., and Shrestha, B.R., and Shrestha, K.B. 2004. Spatial distribution of arsenic concen- Mishra, R. 2012. Solute chemistry and arsenic fate in aquifers between the tration in groundwater in the Terai, Nepal. In Summary Project Report. Himalayan foothills and Indian craton (including central Gangetic plain): Edited by D.R. Kansakar. Department of Irrigation, Lalitpur, Nepal, HMG/ Influence of geology and geomorphology. Geochim. Cosmochim. Acta, 90: Nepal, pp. 85–96. 283–302. doi:10.1016/j.gca.2012.05.015. Shrestha, B.R., Whitney, J.W., and Shrestha, K.B. 2004. The state of arsenic in Nakano, A., Kurosawa, K., Shamim, U.M., and Tani, M. 2014. Geochemical assess- Nepal 2003. The National Arsenic Steering Committee and Environment and ment of arsenic contamination in well water and sediments from several Public Health Organization: Kathmandu. communities in the Nawalparasi District of Nepal. Environ. Earth Sci. 72: Shrestha, R. 2007. Report on Impact Study of Nawalparasi District Sunwal- 3269–3280. doi:10.1007/s12665-014-3231-8. Swathi Cluster. Groundwater Resources Development Project, Babarmahal, NASC-NRCS. 2011. The state of arsenic in Nepal - 2011. Nepal Arsenic Steering Kathmandu. Committee/Nepal Red Cross Society. Kathmandu, Nepal. Shrestha, R.R., Shrestha, M.P., Upadhyay, N.P., Pradhan, R., Khadka, R., Maskey, Neku, A., and Tandukar, N. 2003. An overview of arsenic contamination in A., et al. 2003. Groundwater arsenic contamination in Nepal: A new challenge groundwater of Nepal and its removal at household level. J. Phys. IV France, for water supply sector. In Arsenic Exposure and Health Effects. Edited by 107: 941. doi:10.1051/jp4:20030453. W. R. Chappell, C. O. Abernathy, R. L. Calderon, and D. J. Thomas. Elsevier Neumann, A., Kaegi, R., Voegelin, A., Hussam, A., Munir, A.K.M., and Hug, S.J. B.V., pp. 25–37. 2013. Arsenic removal with composite iron matrix filters in Bangladesh: A Shrestha, R.K., Regmi, D., and Kafle, B.P. 2014. Seasonal variation of arsenic field and laboratory study. Environ. Sci. Technol. 47(9): 4544–4554. doi:10. concentration in groundwater of Nawalparasi district of Nepal. Int. J Appl. 1021/es305176x. PMID:23647491. Sci. Biotechnol. 2(1): 59–63. doi:10.3126/ijasbt.v2i1.9477. Ngai, T.K.K., and Walewijk, S. 2003. The Arsenic Biol. Sand Filter (ABF) Project: Shrestha, S.D., Brikowski, T., Smith, L., and Shei, T.-C. 2004. Grain size con- Design of an Appropriate Household Drinking Water Filter for Rural Nepal. straints on arsenic concentration in shallow wells of Nawalparasi, Nepal. J. Report Prepared for RWSSSP and ENPHO, Kathmandu, Nepal. Nepal Geol. Soc. 30: 93–98. Ngai, T.K.K., Dangol, B., Murcott, S., and Shrestha, R.R. 2005. Kanchan Arsenic Singh, A., Smith, L.S., Shrestha, S., and Maden, N. 2014. Efficacy of arsenic filtra- Filter. Massachusetts Institute of Technology (MIT) and Environment and tion by Kanchan Arsenic Filter in Nepal. J. Water Health, 12(3): 596–599. Public Health Organization (ENPHO). Kathmandu, Nepal. Booklet published doi:10.2166/wh.2014.148. PMID:25252363. by: Environment and Public Health Organization (ENPHO). Smedley, P.L., and Kinniburgh, D.G. 2002. A review of the source, behaviour and Ngai, T.K.K., Murcott, S.E., Shrestha, R.R., Dangol, B., and Maharjan, M. 2006. distribution of arsenic in natural waters. Appl. Geochem. 17: 517–568. doi:10. Development and dissemination of KanchanTM arsenic filter in rural Nepal. 1016/S0883-2927(02)00018-5. Water Sci. Technol. 6: 137–146. Smith, A.H., Lingas, E.O., and Rahman, M. 2000. Contamination of drinking- Ngai, T.K.K., Shrestha, R.R., Dangol, B., Maharjan, M., and Murcott, S.E. 2007. water by arsenic in Bangladesh: a public health emergency. Bull. World Design for sustainable development – Household drinking water filter for Health Organ. 78(9): 1093–1103. PMID:11019458. arsenic and pathogen treatment in Nepal. J. Environ. Sci. Health – A. Tox. Smith, L.S.S., Shrestha, S.D., Brikowski, T.H., and Shei, T.C. 2004. Abundant Hazard. Subst. Environ. Eng. 42: 1879–1888. doi:10.1080/10934520701567148. arsenic sources are discovered in the Nepal Himalayas. Nepal J. Geol. 8. PMID:17952789. Stanger, G. 2005. A palaeo-hydrogeological model for arsenic contamination in Nickson, R.T., McArthur, J.M., Ravenscroft, P., Burgess, W.G., and Ahmed, K.M. southern and south-east Asia. Environ. Geochem. Health 27: 359–367. doi:10. 2000. Mechanism of arsenic release to groundwater, Bangladesh and West 1007/s10653-005-7102-9. Bengal. Appl. Geochem. 15(4): 403–413. doi:10.1016/S0883-2927(99)00086-4. Tandukar, N., Bhattacharya, P., Jacks, G., and Valero, A.A. 2005. Naturally occur- Nordstrom, D.K. 2002. Worldwide occurrences of arsenic in groundwater. Sci- ring arsenic in groundwater of Terai region in Nepal and mitigation options. ence, 296: 143–145. doi:10.1126/science.1072375. In Natural Arsenic in Groundwater: Occurrence, Remediation and Manage- NRCS–ENPHO. 2002. Report on the household survey on the health impact of ment. Edited by J. Bundschuh, P. Bhattacharya, and D. Chandrasekharam. A.A. arsenic contaminated ground water in Bara district. A mimeographed report. Balkema Publishers, Leiden, pp. 41–48. Drinking water quality improvement program, Kathmandu, Nepal. Thakur, J.K., Thakur, K.R., Ramanathan, A., Kumar, M., and Singh, S.K. 2011.

For personal use only. NRCS–ENPHO. 2003. An overview of arsenic contamination and its mitigation in Arsenic contamination of groundwater in Nepal — An overview. Water, 3: Nepal Red Cross Society Program Areas (Jhapa, Sarlahi, Saptari, Bara, Parsa, 1–20. doi:10.3390/w3010001. Rautahat, Nawalparasi, Rupandehi, Kapilvastu, Banke and Bardiya) Drinking Upreti, B.N. 2001. The physiography and geology of Nepal and their bearing on Water Quality Improvement Program. Nepal Red Cross Society/Japanese Red landslide problem. In Landslide hazard mitigation in the Hindu Kush- Cross Society/ENPHO, Kathmandu, Nepal. Himalayas. Edited by L. Tianchi, S.R. Chalise, and B.N. Upreti. International Panthi, R.S., Sharma, S., and Mishra, K.A. 2006. Recent status of arsenic contam- Centre for Integrated Mountain Development (ICIMOD), Kathmandu, Nepal, ination in groundwater of Nepal — A review. Kathmandu Univ. J. Sci. Eng. pp. 31–49. Technol. 2(1): 1–11. van Geen, A., Radloff, K., Aziz, Z., Cheng, Z., Huq, M.R., Ahmed, K.M., et al. 2008. Paudyal, K.R. 2011. Arsenic in groundwater of Terai region of Nepal: Possible Comparison of arsenic concentrations in simultaneously-collected ground- geological sources and health impacts. News Bull. Nepal Geol. Soc. 28: 87–92. water and aquifer particles from Bangladesh, India, Vietnam and Nepal. Pokhrel, D., Bhandari, B.S., and Viraraghavan, T. 2009. Arsenic contamination of Appl. Geochem. 23: 3244–3251. doi:10.1016/j.apgeochem.2008.07.005. groundwater in the Terai region of Nepal: An overview of health concerns Wenk, C., Kaegi, R., and Hug, S.J. 2014. Factors affecting arsenic and uranium and treatment options. Environ. Int. 35: 157–161. doi:10.1016/j.envint.2008.06. removal with zero-valent iron: laboratory tests with Kanchan-type iron nail 003. PMID:18723222. filter columns with different groundwaters. Environ. Chem. 11: 547–557. Polizzotto, M.L., Harvey, C.F., Sutton, S.R., and Fendorf, S. 2005. Processes con- doi:10.1071/EN14020. ducive to the release and transport of arsenic into aquifers of Bangladesh. WHO. 2001. Arsenic contamination in groundwater affecting some countries in Proc. Natl. Acad. Sci. 102: 18819–18823. doi:10.1073/pnas.0509539103. PMID: the South-East Asia region. WHO: Washington, DC, U.S.A. 16357194. Yadav, I.C., Dhuldhaj, U.P., Mohan, D., and Singh, S. 2011. Current status of Polizzotto, M.L., Harvey, C.F., Li, G., Badruzzman, B., Ali, A., Newville, M., et al. groundwater arsenic and its impacts on health and mitigation measures in 2006. Solid-phases and desorption processes of arsenic within Bangladesh the Terai basin of Nepal: An overview. Environ. Rev. 19: 55–67. doi:10.1139/ sediments. Chem. Geol. 228(1–3): 97–111. doi:10.1016/j.chemgeo.2005.11.026. a11-002. Pradhan, B. 2006. Arsenic contaminated drinking water and nutrition status of Yadav, I.C., Singh, S., Devi, N.L., Mohan, D., Pahari, M., Tater, P.S., and Shakya, Environ. Rev. Downloaded from www.nrcresearchpress.com by ETH Zuerich Gruene Bibliothek on 10/13/17 the rural communities in Bagahi village, Rautahat district, Nepal. J. Inst. Med. B.M. 2012. Spatial distribution of arsenic in groundwater of southern Nepal. Nepal, 28(2): 47–51. In Reviews of Environmental Contamination and Toxicology. Edited by D.M. Roberts, L.C., Hug, S.J., Voegelin, A., Dittmar, J., Kretzschmar, R., Wehrli, B., et al. Whitacre. Springer Science + Business Media, pp. 125–140. 2011. Arsenic dynamics in porewater of an intermittently irrigated paddy Yadav, I.C., Devi, N.L., and Sing, S. 2014. Spatial and temporal variation in arsenic field in Bangladesh. Environ. Sci. Technol. 45(3): 971–976. doi:10.1021/es102882q. in the groundwater of upstream of Ganges River Basin, Nepal. Environ. Earth PMID:21166387. Sci. doi:10.1007/s12665-014-3480-6. Sah, R.B., Thakur, P.K., Gurmaita, H.N., and Paudyal, K.R. 2003. Studies for Yadav, I.C., Devi, N.L., and Singh, S. 2015. Reductive dissolution of iron- possible natural sources of arsenic poisoning of groundwater in Terai Plain of oxyhydroxides directs groundwater arsenic mobilization in the upstream of Nepal. Final Report submitted to Ministry of Physical Planning and Works, Ganges River basin, Nepal. J. Geochem. Explor. 148: 150–160. doi:10.1016/j. Department of Water Supply and Sewerage, 69 pp. gexplo.2014.09.002. Shah, B.A. 2008. Role of Quaternary stratigraphy on arsenic-contaminated Zakhaznova-Herzog, V.P., and Seward, T.M. 2006. Antimonous acid protonation/ groundwater from parts of Middle Ganga Plain, UP-Bihar, India. Environ. deprotonation equilibria in hydrothermal solutions to 300 °C. Geochim. Geol. 53: 1553–1561. doi:10.1007/s00254-007-0766-y. Cosmochim. Acta 70(9): 2298–2310. doi:10.1016/j.gca.2006.01.029.

Published by NRC Research Press