Spatial distribution of dissolved constituents in Icelandic river waters

Journal of Hydrology 397 (2011) 175–190

Contents lists available at ScienceDirect

Journal of Hydrology journal homepage: www.elsevier.com/locate/jhydrol

Spatial distribution of dissolved constituents in Icelandic river waters Sigrídur Magnea Oskarsdottir a,⇑, Sigurdur Reynir Gislason b, Arni Snorrason a, Stefanía Gudrún Halldorsdottir a,c, Gudrún Gisladottir b a

Icelandic Meteorological Office, Bustadavegur 9, 150 Reykjavík, Iceland Institute for Earth Sciences, Sturlugata 7, 101 Reykjavík, Iceland c HugurAx, Gudridarstigur 2-4, 113 Reykjavik, Iceland b

a r t i c l e

i n f o

Article history: Received 28 May 2010 Received in revised form 8 November 2010 Accepted 24 November 2010 Available online 30 November 2010 This manuscript was handled by L. Charlet, Editor-in-Chief, with the assistance of Peter Wolfgang Swarzenski, Associate Editor Keywords: GIS Chemical composition Chemical weathering of basalt Spatial analysis

s u m m a r y In this study we map the spatial distribution of selected dissolved constituents in Icelandic river waters using GIS methods to study and interpret the connection between river chemistry, bedrock, hydrology, vegetation and aquatic ecology. Five parameters were selected: alkalinity, SiO2, Mo, F and the dissolved inorganic nitrogen and dissolved inorganic phosphorus mole ratio (DIN/DIP). The highest concentrations were found in rivers draining young rocks within the volcanic rift zone and especially those draining active central volcanoes. However, several catchments on the margins of the rift zone also had high values for these parameters, due to geothermal influence or wetlands within their catchment area. The DIN/DIP mole ratio was higher than 16 in rivers draining old rocks, but lowest in rivers within the volcanic rift zone. Thus primary production in the rivers is limited by fixed dissolved nitrogen within the rift zone, but dissolved phosphorus in the old Tertiary catchments. Nitrogen fixation within the rift zone can be enhanced by high dissolved molybdenum concentrations in the vicinity of volcanoes. The river catchments in this study were subdivided into several hydrological categories. Importantly, the variation in the hydrology of the catchments cannot alone explain the variation in dissolved constituents. The presence or absence of central volcanoes, young reactive rocks, geothermal systems and wetlands is important for the chemistry of the river waters. We used too many categories within several of the river catchments to be able to determine a statistically significant connection between the chemistry of the river waters and the hydrological categories. More data are needed from rivers draining one single hydrological category. The spatial dissolved constituent distribution clearly revealed the difference between the two extremes, the young rocks of the volcanic rift zone and the old Tertiary terrain. Ó 2010 Elsevier B.V. All rights reserved.

1. Introduction Many studies have focused on river geochemistry in order to quantify the global silicate weathering flux and associated flux of consumed atmospheric CO2 (Stallard and Edmond, 1983; Négrel et al., 1993; Probst et al., 1994; Gaillardet et al., 1997, 1999; Galy and France-Lanord, 1999; Millot et al., 2002; Dessert et al., 2003). The importance of lithology has been considered by several authors who showed that basalts are easily weatherable compared to other silicate rocks (Meybeck, 1987; Bluth and Kump, 1994; Woelff-Boenisch et al., 2006). Recently, weathering of volcanic rocks has been addressed in several studies done on volcanic islands (Gislason et al., 1996, 2009; Louvat and Allegre, 1997, 1998; Cruz et al., 1999; Rad et al., 2006; Pogge von Strandmann et al., 2008b; Eiriksdottir et al., 2008; Gislason et al., 2009; Allegre et al., 2010), provinces of flood basalts (Benedetti et al., 1994; ⇑ Corresponding author. Tel.: +354 522 6000; fax: +354 5226001. E-mail addresses: [email protected] (S.M. Oskarsdottir), [email protected] (S.R. Gislason), [email protected], [email protected] (S.G. Halldorsdottir). 0022-1694/$ - see front matter Ó 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jhydrol.2010.11.028

Dessert et al., 2001; Das et al., 2005; Pokrovsky et al., 2005; Vigier et al., 2005; Wimpenny et al., 2007) and smaller continental volcanic regions (Aiuppa et al., 2000, 2001; Benedetti et al., 2003; Gaillardet et al., 2003; Riotte et al., 2003; Dessert et al., 2009; Goldsmith et al., 2010). According to Dessert et al. (2001) the chemical weathering rate of volcanic rocks is 5–10 times higher than the chemical weathering of granite and gneiss. Moreover, the atmospheric CO2 consumption flux derived from basalts weathering represents 30% of the global flux from all silicates and acts as an important regulator of Earth’s climate on a geological time scale (Dessert et al., 2003; Meybeck, 1986, 1987; Millot et al., 2003). The Icelandic rivers were chosen for the present study for a number of reasons. The river catchments studied are sparsely populated, thus minimizing the effect of human activity on the river chemistry. The catchments are mostly comprised of basalts. Furthermore, the postulated future warming with an increase in precipitation is expected to be greatest over land situated at high latitudes (Alley et al., 2007). It is therefore important to study the link between river chemistry, bedrock age and composition,

176

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

hydrology, vegetation and aquatic ecology in Iceland and its spatial variation. Iceland is a volcanic island in the North Atlantic Ocean largely formed from the volcanic activity along the North Atlantic mid-ocean ridge that runs through the island from SW to NE. The island is built up of volcanic rocks, primarily of basaltic composition (80–85%); the remaining 15–20% is mostly basaltic sediments and andesitic intrusions (Johannesson and Saemundsson, 1989). The youngest rocks, age 60.011 My, are confined to the volcanic rift zone (Fig. 1). The oldest rocks (8.5–16 My) are in the East and the West Fjords. Due to crustal accretion, the age of rocks increases with distance from the central volcanic rift zone (Oskarsson et al., 1982). Permeability is higher within the rift zone but lower in older rocks due to compaction and sealing by secondary minerals. Thus most of the precipitation that falls on ice-free land in the volcanic rift zone percolates down into the soil and

rock. From there it may either discharge as groundwater or emerge as springs forming spring-fed rivers (Gislason et al., 1996). Iceland is sparsely vegetated. Most of the vegetation is in the lowlands and along the coastline, but in some places it extends up to 700 m.a.s.l. as heaths and wetlands. The plateau in the middle of the country is almost free of vegetation. Rist (1956) evaluated the total annual discharge of rivers in Iceland, for the water years 1948–1955, by drawing contour lines of run-off based on average run-off from gauged watersheds and discrete measurements of discharge. The total annual discharge of rivers in Iceland for this period was 1690 mm/yr (equal to 5500 m3/s). The average run-off for Iceland (from 1961 to 1990) was 1460 mm/yr (equal to 4770 m3/s; Jonsdottir, 2008). Due to high precipitation and the presence of glaciers, there are rivers almost everywhere in the country, except in the volcanic rift zone. There the permeability of the young volcanic rocks is too

Fig. 1. Geological map of Iceland. Modified from the database of the Hydrological Service of the Iceland Meteorological Office. (From Johannesson and Saemundsson, 1989.)

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

high, resulting in no surface streams except those draining the glaciers. In this study Geographic Information Systems (GIS) were used to map the spatial distribution of selected dissolved constituents and the alkalinity of Icelandic river waters in order to study and interpret the connection between river chemistry, bedrock, hydrology, vegetation and aquatic ecology. The results of this study can be used for example to: (1) predict river water response to acidification and the spatial variation of the response and (2) predict the macronutrient ratio for primary production within the rivers and how this ratio varies spatially. On average there is a volcanic eruption every 2.5 years (Einarsson, 1985) in Iceland. Many of the eruptions produce volcanic ash that can ‘‘precipitate’’ over the nearby river catchments. The ash is formed because the magma is gash rich causing magma fragmentation at low pressure and several of the volcanoes are caped with glaciers, resulting in magma–meltwater interactions creating additional volcanic ash (Durant et al., 2010; Gislason and Alfredsson, 2010). The volcanic ash is coated with metaland proton-salts that condense out of the volcanic plume. The salts are highly soluble and dissolve much faster than the bulk volcanic ash (Frogner et al., 2001; Jones and Gislason, 2008). Most of the salts dissolve during the first few rain storms on the volcanic ash, often resulting in large but short-lived pollution of surface waters (Flaathen and Gislason, 2007). Rivers with high alkalinity are less affected by the pollution than rivers with low alkalinity (Flaathen and Gislason, 2007). It is therefore important to map the spatial variation of the alkalinity of the surface waters in volcanic terrain in order to prepare for volcanic hazards. Spatial distribution of constituents has been modeled before (Arnason, 1976; Sigurdsson, 1990) but GIS methods have not been used to map the distribution of dissolved constituents in Icelandic rivers until now. By applying the spatial analytical methods in GIS a clearer view of the distribution of dissolved constituents and alkalinity for the whole country can be achieved and the changes in the concentration of dissolved elements and alkalinity can be studied further with respect to bedrock age, hydrological classification, soil and vegetative cover. 1.1. Background Studies suggest that chemical weathering, sea spray, global pollution, admixture of geothermal water, local pollution from volcanoes and vegetation control the chemical composition of Icelandic rivers (Gislason and Arnórsson, 1993; Gislason et al., 1996; Gislason, 2005; Gislason and Torssander, 2006; Eiriksdottir et al., 2008; Sigfusson et al., 2008). Chemical weathering and denudation rates in Iceland are high due to the abundance of basaltic rocks, high glass content, high runoff and mechanical weathering (Gislason and Eugster, 1987a; Stefánsson and Gislason, 2001; Gislason et al., 1996; Louvat et al., 2008). The average chemical and denudation rate in Iceland (37 t/km2/yr) is higher than the average for the continents (26 t/km2/yr) (Gislason, 2005; Berner and Berner, 1996). The residence time of water in Icelandic river channels is short. In the main channels of the longest river in Iceland (the Thjorsá river) it is around 24 h. This results in only a short time for water–rock interaction to take place in the main river channels in Icelandic rivers (Gislason, 2005; Gislason et al., 1996). Glassy rocks are abundant within and close to the rift zone; glassy rocks dissolve up to 10 times faster than crystalline rocks of the same chemical composition, depending on the glass composition (Gislason and Eugster, 1987b; Stefánsson and Gislason, 2001; Woelff-Boenisch et al., 2006). Limited evaporation, high precipitation and rapid runoff lead to relatively low concentrations of dissolved constituents in Icelandic

177

rivers (Gislason et al., 1996). However the chemical composition of rocks within each river catchment has a significant impact in terms of chemical weathering. Sea spray and seawater-derived aerosols have significant effects since Iceland is an island (Gislason et al., 1996). Overall, the concentrations of major dissolved constituents decline with increasing discharge (Gislason et al., 2004a,b). One of the largest studies on river chemistry in Iceland was undertaken in 1970–1974 (Armannsson, 1970a,b, 1971a,b, 1973; Rist, 1974, 1986) in rivers in South and West Iceland. The main objective was to study the chemistry of river waters, contamination and water quality (Gislason et al., 1996). This database, along with databases on rock and precipitation composition, was interpreted by Gislason et al. (1996) to show, for example, that the chemical denudation rate was much higher in Iceland than the world average and that the rates decreased with the age of the rocks in the catchment areas. In 1996 the Science Institute of the University of Iceland and the Hydrological Service of the Icelandic Meteorological Office, in collaboration with the Environment and Food Agency of Iceland and later Landsvirkjun, studied dissolved constituents, discharge and suspended load in rivers in South Iceland and have collected samples every year since then (Gislason et al., 2003, 2004a,b, 2006a; Gislason and Torssander, 2006). The same factors were studied by the same collaborators for the River Skaftá catchment and rivers in East Iceland in 1998–2003 (Gislason et al., 2003; Gislason et al., 2004a,b) and Northwest Iceland in 2004–2005 (Gislason et al., 2006a). In 1997–1998 Kristmansdottir et al. (2006) studied rivers draining from sub-glacial central volcanoes within the Vatnajökull and My´rdalsjökull glaciers in 1997–1998. From 1995 to 1998 Adalsteinsson and Gislason (1998) collected samples from rivers all over the country. Only one sample was collected from each sampling station in each river. A similar study was undertaken from rivers in North, South and East Iceland (Louvat et al., 2008). Studies of isotopes have also been undertaken. Gislason and Torssander (2006) studied the response of Icelandic river sulphate concentration and isotope composition to the decline in man-made atmospheric SO2 emissions in the North Atlantic region. Studies of other isotopes in Icelandic river waters have also been undertaken, as can be seen in Vigier et al. (2006, 2009), Abdelmouchine et al. (2006), Pogge von Strandmann et al. (2006, 2008a,b, 2010), Georg et al. (2007) and Pearce et al. (2010). These studies have mostly aimed at understanding the influence of basaltic weathering processes on land and in estuaries on dissolved and particulate riverine Mg, Li, Mo, Si, Os, and U-series isotopes in basaltic terrain. The U-series have shown a maximum weathering timescale of 10,000 years for the overall weathering of Icelandic riverine material (Vigier et al., 2006). Detrital silicates delivered to the estuarine mixing zone are relatively unweathered with respect to major element composition; however, the (234U/238U) activity ratios and Li and Mo isotope composition of the suspended material indicating ongoing weathering of detrital material and the formation of secondary minerals in seawater (Pogge von Strandmann et al., 2008b; Pearce et al., 2010). The fractionation of riverine Si isotopes suggests that the total amount of Si being dissolved by primary weathering of Iceland is 3 million tons per yr. However, only 50% of this Si is released in dissolved form into the rivers and contributes to denudation, the rest left behind in secondary minerals (Georg et al., 2007). 1.2. River classification In 1945 Kjartansson classified Icelandic rivers as: spring-fed rivers, direct run-off rivers, glacier-fed rivers, rivers draining lakes and mixtures thereof, based on the origin of their discharge. In Iceland spring-fed rivers are located within the margin of the

178

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

permeable bedrock of the volcanic zone and their origin can be traced by fissure zones (Sigurdsson and Einarsson, 1988; Sigurdsson, 1990). Glacier-fed rivers drain from glaciers due to glacial melting. Direct run-off rivers mostly drain areas with little groundwater capacity and are thus abundant within the Tertiary and Quaternary zones in the West Fjords, the East Fjords and North Iceland (Egilsson et al., 1991; Fig. 1). In terms of temperature, discharge and concentration of dissolved constituents, spring-fed rivers are more stable than direct runoff and glacier-fed rivers because the discharge of the last two varies throughout the year and thus also the concentration of dissolved constituents (Fig. 2). The concentration is higher during winter than summer, when the discharge is lower (Gislason, 1993; Gislason et al., 1996, 2006a). However, the concentration does change with temperature in spring-fed rivers (Gislason et al., 1996). The Hydrological Service of the National Energy Authority (now Iceland Meteorological Office) has been using the river classification by Kjartansson (1945) for many years but has recently carried out a study aimed at a more detailed classification of Icelandic rivers based on a catchment scale and area response to precipitation, including a description of the characteristics of the

Table 1 Categories of the hydrological classification. ID No. 1 2 3 4 5 6 7 8 9 10 11 12

Category Braided rivers Direct run-off rivers in the highlands with considerable soil storage effect Direct run-off rivers in the lowland with considerable soil storage effect Groundwater-fed rivers in very permeable areas Groundwater-fed rivers in less permeable areas Glaciers on a highly permeable surface Glaciers on an impermeable surface Wetlands in the lowland Direct run-off rivers with sediment storage effects Direct run-off rivers, high altitude watersheds, continuing snowmelts well into the summer Direct run-off rivers in the highlands with wetland storage effect Direct run-off rivers with no storage effect

catchments (Halldorsdottir et al., 2006). The catchments are divided into two main categories, sub-glacial catchment areas and non-glacier catchment areas, with the former category further divided into areas with low permeability and areas with high permeability (Halldorsdottir et al., 2006). The non-glacier classification is more complex since it reflects the geographical variability within the country with regard to groundwater and snow storage reservoirs and to wetland distribution, as well as vegetation patterns. Thus the non-glacier catchment areas category has been further divided into 10 more categories as illustrated in Table 1 and Fig. 3 (Halldorsdottir et al., 2006). Extensive chemical data were not used for the hydrological classification; it is thus very interesting to add the spatial distribution of dissolved constituents to the hydrological classification. However, the chemical composition of these reservoirs has been studied and can be used to describe the chemistry of river waters. Kristmansdottir et al. (2006) classified rivers according to their chemical composition but they neither used GIS in their research nor studied the spatial distribution of dissolved constituents in the rivers. 1.3. Spatial distribution

Fig. 2. Differences in discharge and concentration between river types. Brúará is a spring-fed river, Nordurá is a direct run-off river and Jökla, at Hjardarhagi, is a glacial-fed river. It is striking that Brúará had the most stable conditions compared to the other two rivers. The scale of the X and Y axes has been standardized for all three river types for comparison.

On a global scale GIS is an important tool in mapping spatial distribution of dissolved constituents. Monsef et al. (2004) used GIS techniques, for example to map the distribution of rocks, mainly sulphide-rich rocks in Egypt. Grunwald and Qi (2006) studied the water quality within the Sandusky watershed in Ohio, USA, and used the SWAT tool (Soil and Water Assessment Tool) in ArcMap to map the spatial distribution of dissolved constituents like N, P and others. Wang and Wang (2007) studied the spatial distribution of dissolved pH, Hg, Cd, Cu and As in the Bohai Sea, China. Studies of spatial distribution of dissolved constituents have been rather limited in Iceland. Arnason (1976) mapped the spatial distribution of hydrogen isotope ratios in precipitation and groundwater in Iceland and Sigurdsson and Einarsson (1988) have studied the spatial distribution of chlorine in Icelandic groundwaters and precipitation. Sigurdsson (1990) studied the spatial distribution of glacier-derived groundwater and Hreinsson and Sigurdsson (2004) studied the spatial distribution of selected constituents in the River Skaftá catchment in South-east Iceland. However, very few studies of spatial distribution of river chemistry in Iceland have been carried out using GIS. Kardjilov et al. (2006) and Kardjilov (2008) studied present and past carbon fluxes in eight catchments in East Iceland and the influence of land degradation and future climate change on those fluxes. GIS was used to calculate, for example, the net age of catchment rock and the gross/

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

179

Fig. 3. Categories of the hydrological classification and the catchments of the sampled rivers. Large catchments belong to several categories while others belong to only one category. Categories are listed in Table 1.

Fig. 4. River catchments and sampling stations.

180

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

primary and net production of vegetation within the catchments, and the soil organic carbon storage during the last 10,000 years.

2. Material and methods 2.1. Material 2.1.1. Databases The catchments included in the database along with their sampling stations are shown in Fig. 4 and Table 2. The chemical databases used in this study are listed in Table 3. They were of various sizes and thus the data are mapped according to the following hierarchy: 1. The largest databases were from South Iceland, the Eastern part, NW Iceland, the River Skaftá and the River Laxá at Lake My´vatn (Gislason et al., 2003, 2004a,b, 2006b {South Iceland}; Gislason et al., 2004a,b {East Iceland}; Gislason et al., 2006c {River Skaftá}; Gislason et al., 2006d {North-Western Iceland}, Eiriksdottir, 2007 {East Iceland}) along with data from the 1970’s for West and South Iceland (Armannsson et al., 1973; Rist, 1974, 1986). Samples were collected 5–12 times each year for individual rivers. Up to 50 constituents were analysed, along with measurements of pH/T, discharge, electrical conductivity and air and water temperatures. These databases therefore give the best information on variation in both discharge and the concentration of dissolved constituents. 2. Database from Kristmansdottir et al. (2006). This is a large database where samples were collected each month for a year in 1997–1998 in 10 rivers in North-east and South-east Iceland where 31 chemical constituents were analysed and the pH/T, discharge, electrical conductivity and temperature were measured. 3. Databases from Adalsteinsson and Gislason (1998) and Louvat et al. (2008) where only single samples were collected in 47

Table 2 Rivers and sampling stations. Sampling station

River catchment

Sampling station

River catchment

V502 V144 V328 V4026 V54 V164 V108 V150 V19 V330 V301 V206 V83 V109 V4017 V200 V277 V159 V339 V4015 V106 V102 V265 V51 V110 V71 V4013 V66 V252 V163 V17

Andakilsa A-Jokulsa Asa-Eldvatn Berjadalsa Blanda Bru Bruara Djupa Dynjandi Eldvatn M Ellidaar Fellsa Fjardara Fljotsdal Flokadalsa Fnjoska Geithellnaa Gigja Grenlaekur Grimsa L Grimsa S Grimsstadir Hamarsa Hjaltadalsa Hjardarhagi Hverfisfljot Hvita S Hvita W Kaldakvisl Kreppa Laqarfliot

V105 V4020 V1203 V401 V2605 V89 V128 V64 V2639 V4016 V26 V70 V299 V88 V238 V402 V271 V2201 V230 V297 V4014 V4019 V98 V68 V45 V204 V470 V486 V145 V321

Laxa A Laxa V Leira Langisjor Loni Mulakvisl Nordura Olfusa Osa Reykjadalsa Sanda Skaftardalur Sveinstindur Skeidara Skjalfandafljot Skötufjardara Sog Sula Thjorsa U Thjorsa S Thvera N Thvera S Tungnaa Tungufljot Vatnsdalsa Vatnsdalsa B Vesturkvisl Vididalsa W-Jokulsa Ytri-Ranga

rivers, mainly within the old Tertiary bedrocks. Up to 31 dissolved constituents were analysed along with measurements of discharge, pH/T, temperature and electrical conductivity. The samples were collected close to the hydrometric stations run by the Hydrological Service of the Iceland Meteorological Office (Fig. 4). Discharge, water and air temperature were measured at the time of sampling. Discharge at each station was calculated by applying a rating curve that describes the relationship between discharge and water level (Gislason et al., 2003). Sometimes the water level can be interrupted by ice and snow in the channel during wintertime. In these cases the discharge was estimated by comparing the air temperature and the amount of precipitation at the time with the discharge in nearby rivers. The discharge at the time of sampling is given for all samples (e.g. Gislason et al., 2003). 2.1.2. GIS data and spatial distribution maps The data were imported into the relational Chemical Database of the Iceland Meteorological Office, where they were stored in database tables. The data can either be viewed via an interface or they can be exported from the database in a spread sheet and imported into ArcGIS where they can be mapped in ArcMap. In the Excel spreadsheet the exported average concentration and long-term average concentration were calculated and the results were then imported into GIS in text tables and mapped. Chemical data were imported into ArcMap from the Chemical Database of the Iceland Meteorological Office as Excel sheets and transformed into a shape file, which is a non-topological data structure that stores the geometry and attributes information for geographic features in a dataset. The geometry for each feature is stored as a shape comprising a set of vector co-ordinates and is linked to their attributes (DeMers, 2005). After entering the data, they were then edited for the correct watershed and classified according to the concentration of the constituent in question. Because bedrock age in Iceland differs and catchments can cover bedrocks that vary in age the weighted average bedrock age for each catchment was calculated. The age of bedrock in Iceland and the geological periods are shown in Table 4. 2.1.3. Methods The original databases used in the present study used various units for dissolved constituents, for example, ppm (mg/kg), mmole/kg, meq/kg. All the databases were made compatible by using [mmol/kg], [lmol/kg], [nmol/kg], and [meq/kg] for alkalinity. The data can be classified based on the number of samples in each river, i.e., the data from Gislason et al. from 1996 to 2005, from Ármannsson et al. in 1972–1973, and Rist (1986) in 1973– 1974 include most of the samples. Thus the relationship between discharge and concentration of each element within each river can be plotted as shown in Fig. 2. Five dissolved constituents were selected for this study, as listed below: 1. Alkalinity: Precipitation as it occurs has no alkalinity; thus alkalinity is a direct measure of how much rock weathering a unit mass of water has caused at the early stage of weathering (Gislason, 2005). Furthermore, the alkalinity indicates how the river waters will respond to external acidification. The higher the alkalinity, the more acidification the water can sustain without major pH changes (Gislason, 2005; Flaathen and Gislason, 2007). 2. Silicon (Si) is the second most abundant element in the basalt bedrock but with low concentrations in the oceans; hence almost all the SiO2 dissolved in rivers originates from the rocks and is often a good criterion for weathering (Gislason et al., 1996). According to the relative mobility of the major elements

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

181

Table 3 Periods of sampling and sampled rivers used in the present study. The grey-scale represent the various studies of each river at a given period. Black: Armannsson et al. (1973) and Rist (1974, 1986); dark-grey: Gislason et al. (2003, 2004a,b, 2006a); light-grey: Kristmansdottir et al. (2006); dotted: Adalsteinsson and Gislason (1998); dashed-lines: Louvat et al. (2008).

during basaltic weathering in Iceland, significant amount of silica is left behind in secondary weathering products (e.g. Gislason et al., 1996). On averaged, about half is left behind in

weathering products within the Icelandic river catchments, according to the riverine Si isotopes (Georg et al., 2007). Silica is a very important nutrient for aquatic life and primary

182

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

Table 4 Average bedrock age in Iceland and its geological periods. Age of bedrock (My)

Geological period

0.0–0.011 0.011–0.8 0.8–3.3 3.3–8.5 8.5–15.0

Holocene lava flows Upper Pliocene bedrock Late Pliocene and Upper Pleistocene bedrock Upper Miocene and Lower Pleistocene bedrock Miocene bedrock

production. Thus, a river draining high-productivity lakes, like Lake My´vatn in NE Iceland, can be low in silica in the summer, despite high weathering rates in the catchment (Olafsson, 1979; Thorbergsdottir and Gislason, 2004; Eiriksdottir, 2007). 3. Molybdenum is an important element of nitrogen-fixing enzymes, nitrogenases. Mo and Fe nitrogenases are the most common and efficient nitrogen-fixing enzymes (White, 1999). Gislason and Eiriksdottir (2004) suggested that Mo may govern the N fixation in surface waters within the rift zone in Iceland. The concentrations of Mo in igneous rocks increase with increasing silica content. Mo is mobile during water/rock interactions and reflects the extent of such interaction (Arnórsson and Óskarsson, 2006). 4. Fluorine is not consumed by a secondary mineral, in most surface waters, and is therefore an indicator of how much rock weathering a unit mass of water has caused. Its presence also indicates some geothermal activity. It is a conservative constituent in most river waters (Flaathen and Gislason, 2007). 5. The mole ratio of Dissolved Inorganic Nitrogen over Dissolved Inorganic Phosphate (DIN/DIP) can indicate which of these important nutrients could be rate determining for primary productions within the rivers and perhaps the river catchments (Olafsson, 1979; Thorbergsdottir and Gislason, 2004; Gislason and Eiriksdottir, 2004). When the so called DIN/DIP Redfield mole ratio is >16 (Eq. (1)), it is likely that DIP is limiting, and vice versa when the mole ratio is <16.

106CO2 þ 16NO3 þ HPO2 4 þ 122H2 O

(Fig. 2). The discharge and concentration were fitted by a second order power function. The function and the regression coefficient (R2) were superimposed on the diagrams (Fig. 2). The fit functions as shown in Fig. 2, sometimes referred to as rating curves, can be used along with the long-term average discharge for each river to calculate a long-term average concentration for each constituent. After computation the long-term average concentration values were then mapped in ArcMap. Rating curves were made for each river from which samples had been collected for at least a year. A striking difference was found between the rating curves for a spring-fed river, on the one hand, and a direct runoff or glacial river, on the other, due to the greater variability in discharge in glacial and direct run-off rivers than in spring-fed rivers. This difference is illustrated in Fig. 2. The conditions of the spring-fed rivers were the most stable for these three river types, both in terms of discharge and concentration (Gislason et al., 1996). The rivers covered by Adalsteinsson and Gislason (1998) and Louvat et al. (2008) were only sampled once each. If the rivers were spring-fed, the data could be used without much concern, as demonstrated in Fig. 2. However, if they were direct runoff or glacier-fed rivers, only samples that were collected at close to the long-term-average discharge could be used. 2.1.1.1. Advantages and disadvantages. The disadvantages of the dataset are mainly that samples had not been collected at the same time in each river. Datasets for some rivers were large while those of others were smaller. That is why the distribution maps show either average values or a single value. Table 3 illustrates the times at which sample collection took place. The dataset for the glacierfed rivers along with large direct run-off rivers in southern and eastern Iceland was the largest, together with those for several rivers on the west coast. The smallest datasets represented rivers in the West Fjords and some rivers in the northern part. The main advantage of the dataset, however, is that the largest datasets were from within the volcanic rift zone, the margins of the rift zone and the older Tertiary rocks. In Table 3 all the rivers used in this study, along with sampling times, are listed.

þ 18Hþ ðþtrace elements and sun energyÞ () Respiration and decayC106 H263 O110 N16 P þ 138O2

ð1Þ

2.1.4. Discharge and concentration The discharge and concentration of dissolved elements in direct run-off rivers and glacial rivers vary considerably over the year

2.1.5. Water and river classification Halldorsdottir et al. (2006) divided river catchments into several sub-basins where sub-basins were classified into categories, representing the hydrological condition of the region. The main findings of that report suggest that the catchments can be divided into the categories listed in Table 1 where each number identifies

Fig. 5. Zonal analysis and the hydrological classification. Some cathcments belong to two or more categories in the hydrological classification.

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

183

This means that the value for each dissolved constituent was used in calculating the mean for each category; thus the value applies to all four categories. 3. Results Results are shown on the scatter plot in Fig. 6 and the maps in Figs. 7–11. Specific rivers and river catchments will be referred to in the text. Individual rivers and the station numbers are listed alphabetically in Table 2 and the station numbers are shown in Fig. 4. 3.1. Spatial distribution maps Fig. 6. Average alkalinity and the river categories of the hydrological classification. The error bars represent the standard deviation for each category. The table describes the categories on the X axis.

each category. Table 4 lists the average bedrock age and geological periods of the bedrock in Iceland. Fig. 3 shows the hydrological classification, as described in Table 1 and the catchments used in this study. In this study the chemistry component is added to the hydrological classification in order to get some correlation between the classification and the concentration of each constituent by applying GIS methods. The zonal analysis method was used to assign the values of the concentration layer (a raster dataset) to the categories of the hydrological classification layer (Fig. 5). The result is a zonal statistical table containing calculated values of mean and standard deviation for each dissolved constituent. The values with 95% confidence intervals were then plotted as shown for example for alkalinity in Fig. 6 The main disadvantage of this method is that each catchment can cover several hydrological categories in Fig. 6, as is the case for example with the River Jökulsá á Fjöllum (Grímsstadir), which covers four categories (Table 1).

3.1.1. Alkalinity Fig. 7 illustrates the distribution of alkalinity in Icelandic rivers based on data collected from several rivers during the periods 1972–1974 and 1996–2005. The influence of the volcanic rift zone on the alkalinity distribution is striking. The concentration was highest in the following rivers in declining order (the station number refers to Fig. 4 and Table 2): Ytri-Rangá (V321), Skeidará (V88), Múlakvísl (V89), Leirá (V1203), Jökulsá á Fjöllum (V102) and Laxá (V105), which is the outlet of Lake My´vatn in North-east Iceland. All these rivers drain water from central volcanoes, such as Mt Hekla (Ytri-Rangá), the Grímsvötn caldera in the Vatnajökull Glacier (Skeidará), Kverkfjöll (Jökulsá á Fjöllum), the Katla caldera in My´rdalsjökull Glacier (Múlakvísl and Leirá), and the Krafla caldera (Laxá at Lake My´vatn). Alkalinity was lowest within the old Tertiary rocks in the West and East Fjords, followed by catchments in the Tröllaskagi Peninsula in northern Iceland. The rivers that are closest to the active volcanoes have the highest alkalinity. These rivers are most likely to contain major airborne pollution, such as volcanic ash, because of their closeness to the volcanoes and they can sustain higher

Fig. 7. Spatial distribution of alkalinity in Icelandic river waters.

184

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

Fig. 8. Spatial distribution of dissolved silica in Icelandic river waters.

Fig. 9. Spatial distribution of dissolved fluorine in Icelandic river waters.

pollution than the rivers further away that drain the Tertiary formation. Alkalinity was high in several catchments draining water from relatively old rocks, such as the rivers Vatnsdalsá (V45), Vídidalsá (V468), West Jökulsá (V145) and Reykjadalsá

(V4016). The reason was most likely the effect of the wetlands in the catchment areas in the case of the Vatnsdalsá (V45) and Vídidalsá (V468) but admixture with geothermal water in the Reykjadalsá (V4016) and the spring-water and geothermal

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

185

Fig. 10. Spatial distribution of dissolved molybdenum in Icelandic river waters.

Fig. 11. The spatial distribution of the ratio of dissolved inorganic nitrogen to dissolved inorganic phosphorus. In dark brown catchments the molar ratio is >16 and there primary production is probably DIP limited while primary production is DIN limited within the volcanic rift zone.

component in the West Jökulsá (V145). Higher alkalinity was found in water from rivers that drain from younger rocks, e.g. those draining Holocene lava flows and rocks from late Pliocene and upper Pleistocene.

3.1.2. SiO2 The spatial distribution of silica is shown in Fig. 8 for rivers sampled in 1972–1974 and 1996–2005. The silica concentration was highest in waters from the catchments that drain the youngest

186

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

rocks and are within the volcanic rift zone. In contrast, silica concentrations were low within catchments draining lakes, such as the Ellidaár (V301), Andakílsá (V502) and North Thverá (V4019) rivers and the outlet of Lake Langisjór (V401). This was due to the primary production of diatoms in the lakes. The Laxá at Lake My´vatn (V105) was the exception. Despite the fact that it drains the highly productive, Lake My´vatn (Olafsson, 1979; Thorbergsdottir and Gislason, 2004) its silica concentration was high due to high concentrations of silica in the inflow to the lake (Fig. 8). Several catchments outside the rift zone had high concentrations of silica, probably because they drain areas of wetlands where chemical weathering rates are high (Kardjilow et al., 2006). This was the case for the rivers Vatnsdalsá (V45), Vídidalsá (V468), Fnjóská (V200) and Sandá (V26) since all of them drain wetland and heath areas, either totally or partly (Fig. 4 and Table 2). Despite the fact that most rivers draining young rocks were high in silica, the highest silica concentration was in the River Sandá (V26) (0.565 [mmol/kg] and 3.3 [My]), which drains relatively old rocks and is located outside the volcanic rift zone. This catchment is rich in heath vegetation and wetlands, resulting in silica accumulating in the system and leading to high concentrations. The same applies to the Rivers Vatnsdalsá (V45) and Vídidalsá (V486), which also drain wetlands. The Reykjadalsá (V4016) was also very high in silica, despite draining relatively old rocks (3–4 My), but in this case the silica content was due to geothermal influence. Other rivers high in silica were located within the volcanic rift zone with bedrock of relatively young age. 3.1.3. Fluorine The distribution of dissolved fluorine in Icelandic rivers is illustrated in Fig. 9. In relation to alkalinity the fluorine distribution in Icelandic river waters is striking. It was completely dominated by the volcanic rift zone where the concentrations were highest in rivers draining central volcanoes and high temperature geothermal fields. The F concentration in rivers with runoff from wetlands was also relatively high compared to the age of rock within their catchments. Fluorine has been shown to be mobile during weathering of rocks in Iceland; it is not consumed by secondary minerals (Gislason et al., 2006a, 1992). The highest F concentrations were found in the rivers Ytri-Rangá (V321) and Múlakvísl (V89), rivers that drain relatively young rocks (0.011–0.8 My) and run from central volcanoes. Interestingly, the River Reykjadalsá (V4016) drains relatively old rocks but its F concentration was still high. That was because of admixture with geothermal waters. 3.1.4. Molybdenum The distribution of Mo is illustrated in Fig. 10. Molybdenum content has only been measured by Gislason et al. (2003, 2004, 2006) and Louvat et al. (2008). The data are therefore not as extensive as for alkalinity, silica and fluorine (Figs. 7–9). The effects of the volcanic zone and the bedrock age on the distribution of Mo is interesting, i.e. rivers draining young rocks contain the highest concentration of Mo compared to rivers draining older rocks outside the volcanic rift zone such as the Fjardará (V2856) in the east. Mo concentrations were also high in rivers draining rhyolitic terrain, even outside the volcanic zone. To name a few, the River Jökulsá in Fljótsdalur (V109) had a strikingly high Mo concentration due to runoff from Mt Snaefell, an old central volcano containing mostly rhyolitic and andesitic rocks (Johannesson and Saemundsson, 1989). The Nordurá (V128) in West Iceland is another example, as it drains water from the rhyolitic Mt Baula. The Rivers Ytri-Rangá (V321) and Laxá (V105) contained the highest Mo concentrations, rivers that drain water from the Hekla and Krafla central volcanoes, respectively.

The Mo concentrations of other river waters within the volcanic rift zone were high, which was to be expected due to their high alkalinity (Fig. 7) which reflects the amount of rock weathering per unit mass of water and admixture of geothermal water. Mo is thought to play a role in nitrogen fixation but Gislason and Eiriksdottir (2004) conclude that Mo concentration may govern primary production in Icelandic lakes, within the rift zone, where primary production is limited by the amount of fixed nitrogen  þ (NO 3 ; NO2 ; NH4 ). 3.1.5. DIN/DIP This mole ratio of ‘‘Dissolved Inorganic Nitrogen’’ over ‘‘Dissolved Inorganic Phosphate’’ can indicate which of these important nutrients could be rate determining for primary productions within the rivers and even for vegetation within the river catchments. Fig. 11 illustrates the spatial distribution of this ratio. The highest ratios, some of them higher than the critical molar Redfield molar ratio of 16 (Eq. (1)), were in rivers draining rocks older than 2 My. The DIN/DIP distribution shows that phosphate was the main limiting factor of primary production in rivers draining old rocks and highly vegetated areas, whereas dissolved inorganic nitrate was the main limiting component within the volcanic rift zone. 3.1.6. Hydrological classification of Icelandic rivers and their chemistry The categories of the hydrological classification and the river catchments of the rivers sampled are illustrated in Fig. 3. The relationship between each category of the hydrological classification (Table 1) and the average concentrations of the selected dissolved constituents is shown in Fig. 6. According to Fig. 6 braided rivers (category 1) had the highest alkalinity and the largest standard deviation, as can also be seen in the spatial distribution map (Fig. 7). The rivers in this category are the Skeidará (V88), Múlakvísl (V89) and Leirá (V1203). Rivers draining glaciers on an impermeable surface (category 7) had the second highest alkalinity and standard deviation. This category includes rivers like the Skeidará, Múlakvísl and Leirá, but as shown in Fig. 7, the alkalinity of these three rivers, along with Ytri-Rangá which also drains a central volcano, (Fig. 4 and Table 2), was the highest. Direct run-off rivers at high altitude with snowmelt well into the summer (category 10) had the lowest alkalinity. Rivers that fall into this category are all the rivers in the Western Fjords (V19, V204, V2639, V4020), along with the Rivers Hjaltadalsá (V59) and Fjardará (V83) (Fig. 4 and Table 2) whose catchments all drain old Tertiary rocks. The alkalinity of direct run-off rivers in the highlands with considerable soil storage effect (category 2) and direct run-off rivers with no storage effect (category 12) was low. These were rivers such as the Thjórsá, Olfusa and Kaldakvísl (category 2) and the Nordurá and Skjálfandafljót (category 12) (Fig. 4 and Table 2). Direct run-off rivers in the lowland with considerable soil storage effect (category 3) had a higher alkalinity than similar category 2 rivers in the highlands. The same applied to direct run-off rivers with sediment storage effects (category 9) and direct run-off rivers in the highlands with wetland storage (category 11). The weathering rate (Kardjilov et al., 2006) increased with increasing wetland cover in Icelandic catchments. The temporal effect was greater in the lowlands than in the highlands, probably because of higher temperatures in the lowlands. The dissolution rates of rocks are temperature dependent: the higher the temperature the faster the dissolution (e.g. Gislason and Oelkers, 2003). The concentrations are also dependent on runoff, i.e. the higher the runoff in a given catchment, the lower the concentration (Gislason et al., 2006a). Silica concentrations were highest in braided rivers (category 1) with the Skeidará, Múlakvísl and Leirá rivers (Fig. 4 and Table 2)

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

classified as braided. The lowest silica concentration was found in direct run-off rivers with no storage effect (category 12), with the Nordurá and Skjálfandafljót rivers in that category (Fig. 4 and Table 2). Other categories seemed to fall in between. However the difference between the highest and the lowest values was very small. As for silica and alkalinity, the fluorine concentration of braided rivers (category 12) (the Skeidará, Múlakvísl and Leirá) was highest and lowest in rivers with no storage effect (category 12) (the Nordurá and Skjálfandafljót). Molybdenum concentrations were highest in category 3 (direct run-off rivers in the lowlands with considerable soil storage effect). There was only one river that partly belongs to this category, the Lagarfljót (Table 2 and Fig. 4). The Mo concentration, however, was lowest in braided rivers (category 12) (the Skeidará, Múlakvísl and Leirá). This finding could be due to scarcity of data for these types of rivers or precipitation of sulphides consuming Mo in these rivers, since they are known to contain some H2S. The standard deviation for category 11 reached a DIN/DIP mole ratio of 220. The highest ratio was found for category 11 (direct run-off rivers in the highlands with wetland storage effect). Rivers within this category were the Vatnsdalsá (V45), Vídidalsá (V468), West Hvítá (V66), Reykjadalsá (V4016) and North Thverá (V4014). These data were consistent with the DIN/DIP distribution as P originates in the rocks whereas N originates from the decay of organic matter, global pollution and nitrogen fixation (Gislason et al., 1996; Gislason and Eiriksdottir, 2004). These results showed that for all dissolved constituents studied, the standard deviation was so high that one can assert, with 95% confidence, that the mean concentration for each category is statistically insignificant. Therefore it can be claimed that the chemistry of the Icelandic rivers used in this study cannot be used as examples of hydrological classification without further studies. The catchments of the rivers studied were too large and cover terrain that belongs to several categories: for example, the Jökulsá á Fjöllum catchment at Grímsstadir (V102), (Table 2 and Fig. 4) that belongs to four categories; glaciers on an impermeable surface (category 7), glaciers on a permeable surface (category 6); groundwater-fed rivers in a very permeable area (category 4); and groundwater-fed rivers in a less permeable area (category 5), where the greatest part (around 70%) of the total catchment area belongs to category 4. The alkalinity was 0.811 [meq/kg]. This value was assigned to all the four categories within the river’s catchment areas and was used to calculate the mean and standard deviation for each single category performed by the GIS zonal analysis calculations. Thus chemical data from each category are needed and samples from all four categories are needed to add the consistent chemistry to the classification. This study revealed that more specific data from small catchments covering a single category are needed in order to study whether the hydrological classification and the river chemistry can be combined into one classification.

4. Discussion 4.1. Spatial distribution maps The results for the spatial distribution are quite remarkable. All of the five parameters, except DIN/DIP, had high values in the volcanic rift zone. The results for each constituent will be discussed in the following sections. The distribution of alkalinity (Fig. 7) is rather interesting because there are few rivers outside the volcanic rift zone that have high alkalinity concentrations. This was most likely due to the effect of geothermal activity, as was the case for the River Reykjadalsá. The rivers Vatnsdalsá, Vídidalsá and Sandá drain

187

relatively old rocks (0.8–3.3 My) but wetlands cover a relatively large part of their catchments. The concentrations of dissolved constituents were high in wetlands and therefore in the rivers draining those wetlands (Sigfússon et al., 2006; Sigfusson et al., 2008; Kardjilov et al., 2006). The alkalinity of the Asa-Eldvatn river water is conspicuous since the River Skaftá, along with the Grenlaekur and Eldvatn, were lower in alkalinity. However, groundwater rich in dissolved minerals flows into the River Asa-Eldvatn resulting in a higher alkalinity than in the Skaftá, where the mixing of groundwater and river water had not taken place (Hreinsson and Sigurdsson, 2004). The reason for the high alkalinity within the volcanic rift zone was the influence of geothermal and volcanic activity and the resulting abundance of young, crystalline and glassy basalt. Pristine rock is more reactive than older rocks (Gislason et al., 1996). The most reactive minerals and glass have been weathered in the old rocks (Gislason et al., 1996). Furthermore, there was an abundance of spring-fed rivers that had a higher alkalinity than direct run-off rivers. The distribution of silica concentrations was consistent with the alkalinity distribution. The volcanic rift zone had similar effects on the silica concentration. Rivers draining young rocks had higher silica concentrations. In contrast the silica concentrations were lower in rivers draining lakes, such as the Andakílsá (V502), North Thverá (V4014), Sog (V271) and Langisjór (V401). This was due to the primary production of diatoms in the lakes which consumed dissolved silica (Olafsson, 1979; Thorbergsdottir and Gislason, 2004). The River Laxá, at the outlet of Lake My´vatn in NE Iceland, was the exception as its silica concentrations were extremely high due to the influx of geothermal water to the lake. However, the silica concentration in the River Sog was also rather high (0.211 [mmol/kg]) compared to that of other rivers draining lakes, e.g. the Andakílsá (0.111 [mmol/kg]). The Sog is primarily a springfed river, and the springs’ silica concentrations were similar to those of the head springs of the River Brúará, just east of the Sog (Table 2 and Figs. 4 and 8). The lower the concentration in the Sog was caused by the primary production of diatoms. The silica concentrations of the River Tungufljót (V68) (Table 2 and Figs. 4 and 8) was high. The Geysir geothermal field is within the river’s drainage area and could affect the river’s composition in a similar manner to the geothermal admixture in the River Reykjadalsá (V4016) in West Iceland (Table 2 and Figs. 4 and 8). The same considerations apply to fluorine as to the other two constituents. Age of bedrock and volcanic and geothermal activity influence the distribution. Thus the concentration decreased with increasing age of bedrock within the catchment and with increasing distance from the volcanic rift zone. Since fluorine is a weatheringdependent constituent its concentrations were highest where chemical weathering was fastest, i.e. in the volcanic rift zone. Mo concentrations were also highest in rivers within the volcanic rift zone and especially if they drained from central volcanoes, i.e. the Ytri-Rangá river draining Mt Hekla and the Laxá at Lake My´vatn that drains the Krafla central volcano (Table 2 and Figs. 4 and 10). Dissolved inorganic nitrogen and phosphate are important to vegetation growth in rivers and their catchments. These constituents can limit primary production and their concentration is low due to fixation in vegetation and/or algae in the rivers (Olafsson, 1979; Gislason et al., 1996; Adalsteinsson and Gislason, 1998) as shown in Eq. (1) the ratio is 16/1 for N/P. Therefore for rivers with a DIN/DIP mole ratio higher than ca. 16 the DIP may limit primary production in the river. If the N/P mole ratio is lower than 16, dissolved inorganic nitrogen (DIN) might limit the primary production within the river. The average age of rocks within the catchments with a DIN/DIP ratio higher than 16 was rather high (0.8–9.0 My) and the catchments were also rather highly vegetated. In terms of DIN

188

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

concentrations, however, Fig. 11 illustrates that the DIN was the limiting factor in catchments draining younger rocks that are located near or within the volcanic rift zone. The dissolved inorganic nitrogen was most likely fixed in vegetation or consumed by vegetation and/or algae. By comparing Figs. 10 and 11 the correlation between the high concentration of Mo and the low DIN/DIP ratio is apparent. A high concentration of Mo probably aids the fixation of nitrogen through Mo–Fe–enzymes in nitrogen fixing bacteria, and therefore primary production in surface waters within and close to the rift zones in Iceland (Gislason and Eiriksdottir, 2004), especially in rivers like the Laxá and Ytri-Rangá (Table 2 and Figs. 4 and 10). Overall it seems that the volcanic rift zone, active central volcanoes’ geothermal activity outside the rift zone, and wetlands had significant effects on the distribution of the parameters studied. This applied to all the parameters except perhaps the DIN/DIP since it is, to a smaller extent, related to weathering, volcanic or geothermal activity. Concentrations in rivers draining Tertiary rocks in the West and East Fjords were probably more dependent on precipitation since the rock-derived concentrations were low. 4.2. Hydrological classification On the other hand, the hydrological classification is perhaps too general for river chemistry because river chemistry depends on bedrock type, geothermal, and volcanic activity, and vegetation. Therefore river chemistry might better be classified separately, independent of the hydrological classification. Then the zonal analysis method could be applied to the ‘new’ river chemistry classification and some other data that affect the chemical content of rivers, like precipitation or bedrock data. It is however worth emphasizing that rivers have been classified according to river chemistry based on different discharge qualities (i.e. spring-fed rivers, glacier-fed rivers and direct run-off rivers) (see, e.g., Gislason (1993) and Gislason and Arnorsson (1988)). This classification could be improved with the objective of establishing a river chemistry classification based on geology, vegetation, precipitation distribution and weathering and denudation rates, to name a few factors. Whether or not this classification will be undertaken, more data are needed to improve both the hydrological and the river chemistry classifications. 5. Conclusion The spatial distributions of alkalinity, dissolved SiO2, Mo, F and the DIN/DIP mole ratio have been mapped by applying GIS methods. The results indicated that there was a large difference in all parameters between rivers in the volcanic rift zone and the ones draining older rocks of Tertiary age. Values for rivers in between these two extremes were variable and were at times hard to explain. Therefore, each river has to be assessed individually and local attributes need to be used to explain the anomaly; e.g. the River Reykjadalsá drains a low-temperature geothermal field. More samples are needed from small river catchments belonging to one river category in order to fit the river chemistry to the hydrological classification. Acknowledgements We are grateful to Hákon Adalsteinsson at the Landsvirkjun power company, Iceland, Hrefna Kristmannsdóttir at University of Akureyri Iceland, and many friends and colleagues at the Icelandic Meteorological Office and the Institute of Earth Sciences of the University of Iceland, along with the owners of the data used in this study. We also would like to thank Terry Lacy for proof-reading the article. This study was funded by the Energy Resources

Department of the National Energy Authority, the Hydrological Service of the National Energy Authority, and RANNÍS, the Icelandic Science Foundation and we are thankful for their contribution. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jhydrol.2010.11.028. References Abdelmouchine, Gannoun, Burton, Kevin W., Vigier, N., Gislason, S.R., Rogers, N., Mokadem, F., Sigfússon, B., 2006. The influence of weathering process on riverine osmium isotopes in a basaltic terrain. Earth Planet. Sci. Lett. 243 (3–4), 732–748. Adalsteinsson, H., Gislason, G.M., 1998. Impact of terrestrial factors on life in rivers. Summary in English. Náttúrufræðingurinn 68 (2), 97–112. Aiuppa, A., Allard, P., D’Allessandro, W., Michel, A., Parello, F., Treuil, M., Valenza, M., 2000. Mobility and fluxes of major, minor and trace metals during basalt weathering and groundwater transport at Mt. Etna volcano (Sicily). Geochim. Cosmochim. Acta 64, 1827–1841. Aiuppa, A., Bonfanti, P., Brusca, L., D’Alessandro, W., Federico, C., Parello, F., 2001. Evaluation of the environmental impact of volcanic emissions from the chemistry of rainwater: Mount Etna area (Sicily). Appl. Geochem. 16, 985–1000. Allegre, C., Louvat, P., Gaillardet, J., Meynadier, L., Rad, S., Campas, F., 2010. The fundamental role of island arc weathering in the oceanic Sr isotope budget. Earth Planet. Sci. Lett. 292, 51–56. Alley, R. et al., 2007. Climate Change 2007: The Physical Sciece Basis, Summary for Policymakers. Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Intergovernmental Panel on Climate Change, Geneva. Armannsson, H., 1970a. Chemical Analysis of Elliðaar Rivers and their Runoff. Report No. 26. Industrial Research and Development Institute. Armannsson, H., 1971a. Chemical Analysis of Elliðaar Rivers and their Runoff. Report No. 35. Industrial Research and Development Institute. Armannsson, H., 1971b. Chemical Analysis of Elliðaar Rivers and their runoff II. Period May 1970–January 1971. Report No. 35. Industrial Research and Development Institute. Armannsson, H., Magnusson, H., Sigurðsson, P., Rist, S., 1973. Chemical Analysis of Water: Hvita-Olfusa Catchments; also Thjorsa at Urriðafoss, Waterfall. National Energy Authority, Industrial Research and Development Institute, OS-RI, Reykjavik. Arnason, B., 1976. Groundwater Systems in Iceland. Reykjavik. Arnórsson, S., Óskarsson, N., 2006. Molybdenum and tungsten in volcanic rocks and in surface and <100 °C ground waters in Iceland. Geochim. Cosmochim. Acta 71, 284–304. Benedetti, M.F., Menard, O., Noack, Y., Carvalho, A., Nahon, D., 1994. Water rock interactions in tropical catchments: field rates of weathering and biomass impact. Chem. Geol. 118, 203–220. Benedetti, M.F., Dia, A., Riotte, J., Chabaux, F., Gérard, M., Bouleuge, J., Fritz, B., Chauvel, C., Bulourde, M., Déruelle, B., Ildefonse, P., 2003. Chemical weathering of basaltic lava flows undergoing extreme climatic conditions: the water geochemistry record. Chem. Geol. 201, 1–17. Berner, E.K., Berner, R.A., 1996. Global Environment: Air, Water and Geochemical Cycles. New Jersey. Bluth, G.J.S., Kump, L.R., 1994. Lithologic and climatic control of river chemistry. Geochim. Cosmochim. Acta 58, 2341–2359. Cruz, J.V., Coutinho, R., Carvalho, M.R., Oskarsson, N., Gislason, S.R., 1999. Chemistry of waters from the Furnas volcano, Sao Miguel, Azores: fluxes of volcanic carbon dioxide and leached material. J. Volcanol. Geotherm. Res. 92, 151–167. Das, A., Krisnaswami, S., Sarin, M.M., Pande, K., 2005. Chemical weathering in the Krishna Basin had Western Ghats of the Deccan Traps, India: rates of basalt weathering and their controls. Geochim. Cosmochim. Acta 69, 2067–2084. DeMers, M.N., 2005. Fundamentals of Geographic Information Systems. John Wiley and Sons Inc., New Jersey, USA. Dessert, C., Dupré, B., Francois, L.M., Schott, J., Gaillardet, J., Chakrapani, G.J., Bajpai, S., 2001. Erosion of Deccan Traps determined by river geochemistry: impact on the global climate and the 87Sr/86Sr ratio of seawater. Earth Planet. Sci. Lett. 188, 459–474. Dessert, C., Dupré, B., Gaillardet, J., Francois, L.M., Allegre, C.J., 2003. Basalt weathering laws and impact of basalt weathering on the global carbon cycle. Chem. Geol. 202, 257–273. Dessert, C., Gaillardet, J., Dupré, B., Schott, J., Pokrovsky, O.S., 2009. Fluxes of highversus low-temperature water–rock interactions in aerial volcanic areas: example from the Kamchatka Peninsula, Russia. Geochim. Cosmochim. Acta 73 (1), 148–169. Durant, A.J., Bonadonna, C., Horwell, C.J., 2010. Atmospheric and environmental impacts of volcanic particulates. Elements 6, 235–240. Egilsson, D., Sigurdsson, F., Johannesson, H., Sigurdsson, P., Gudjónsson, S., Einarsson, S.M., Sigfússon, S.H., 1991. Rivers and Geomorphology (Report in Icelandic). Soil Conservation Service of Iceland, Council for Nature Protection, National Energy Authority of Iceland, Icelanic Road Administration, Institute of Freshwater Fisheries.

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190 Einarsson, Th., 1985. Geology, fifth ed. Mál og menning, Reykjavik (in Icelandic). Eiriksdottir, E.S., 2007. Chemical and Mechanical Weathering of Basalt: Measured and Modelled Data from North-east Iceland. M.Sc. Thesis in Geology, University of Iceland. Eiriksdottir, E.S., Louvat, P., Gislason, S.R., Óskarsson, N., Hardardottir, J., 2008. Temporal variation of chemical and mechanical weathering in NE Iceland: evaluation of a steady-state model of erosion. Earth Planet. Sci. Lett. 272, 78–88. Flaathen, T.K., Gislason, S.R., 2007. The effect of volcanic eruptions on the chemistry of surface waters: the 1991 and 2000 eruptions of Mt. Hekla, Iceland. J. Volcanol. Geotherm. Res. 164 (4), 293–316. Frogner, P., Gislason, S.R., Oskarsson, N., 2001. Fertilizing potential of volcanic ash in ocean surface water. Geology 29, 487–490. Gaillardet, J., Dupré, B., Allegre, C.J., Négrel, P., 1997. Chemical and physical denudation in the Amazon River Basin. Chem. Geol. 142, 141–173. Gaillardet, J., Dupré, B., Louvat, P., Allegre, C.J., 1999. Global silicate weathering and CO2 consumption rates deduced from the chemistry of the large rivers. Chem. Geol. 159, 3–30. Gaillardet, J., Millot, R., Dupré, B., 2003. Chemical denudation rates of the western Canadian orogenic belt: the Strikine terrane. Chem. Geol. 201, 257–279. Galy, A., France-Lanord, C., 1999. Weathering processes in the Ganges–Brahmaputra basin and the riverine alkalinity budget. Chem. Geol. 159, 31–60. Georg, R.B., Reynolds, B.C., West, A.J., Burton, K.W., Halliday, A.N., 2007. Silicone isotope variations accompanying basalt weathering in Iceland. Earth Planet. Sci. Lett. 261, 476–490. Gislason, S.R., 1993. Chemistry of precipitation, glaciers, rivers, lakes and groundwater in Iceland. Náttúrufræðingurinn 63, 219–236 (in Icelandic). Gislason, S.R., 2005. Chemical weathering, chemical denudation and the CO2 budget for Iceland. In: Caselding, C., Russel, A., Hardardottir, J., Knudsen, O. (Eds.), Developments in Quaternary Sciences. Iceland Modern Processes and Past Environments. Elsevier B, Amsterdam, pp. 289–307. Gislason, S.R., Alfredsson, H.A., 2010. Sampling the volcanic ash from the Eyjafjallajökull Volcano, Iceland – a personal account. Elements 6, 269–270. Gislason, S.R., Arnorsson, S., 1988. Chemistry of rivers in Iceland and chemical weathering rates. Náttúrufræðingurinn 58, 183–197 (summary in English). Gislason, S.R., Arnórsson, S., 1993. Dissolution of primary basaltic minerals in natural waters: saturation state and kinetics. Chem. Geol. 105, 117–135. Gislason, S.R., Eiriksdottir, E.S., 2004. Molybdenum control of primary production in the terrestrial environment. In: Wantly, Seal II (Eds.), Water–Rock Interaction, pp. 1119–1122. Gislason, S.R., Eugster, H.P., 1987a. Meteoric water–basalt interactions I: a laboratory study. Geochim. Cosmochim. Acta 51, 2827–2840. Gislason, S.R., Eugster, H.P., 1987b. Meteoric water–basalt interactions II: a field study in NE Iceland. Geochim. Cosmochim. Acta 51, 2841–2855. Gislason, S.R., Oelkers, H.E., 2003. Mechanism, rates and consequences of basaltic glass dissolution II: an experimental study of the dissolution rates of basaltic glass as a function of pH and temperature. Geochim. Cosmochim. Acta 67, 3817–3832. Gislason, S.R., Torssander, P., 2006. The response of Icelandic river sulfate concentration and isotope composition to the decline in global atmospheric SO2 emission to the North Atlantic region. Environ. Sci. Technol. 40, 680–686. Gislason, S.R., Andrésdottir, A., Sveinbjörnsdottir, A.E., Oskarsson, N., Thordarson, Th., Torssander, P., Novak, M., Zak, K., 1992. Local effects of volcanoes on the hydrosphere: example from Hekla, southern Iceland. In: Kharaka, Y.K., Maest, A.S. (Eds.), Water–Rock Interaction, vol. 1. Balkema, Rotterdam, pp. 477–481. Gislason, S.R., Arnórsson, S., Ármannsson, H., 1996. Chemical weathering of basalt in southwest Iceland: effects of run-off, age of rocks and vegetative/glacial cover. Am. J. Sci. 296, 837–907. Gislason, S.R., Snorrason, A., Eiriksdottir, E.S., Sigfússon, B., Elefsen, S.O., Hardardottir, J., Gunnarsson, A., Torssander, P., 2003. Chemical Composition, Discharge and Suspended Load in Rivers in South Iceland VI. Report RH-032003. Database of the Science Institute and the Hydrological Service of the National Energy Authority. Gislason, S.R., Snorrason, A., Sigfússon, B., Eiriksdottir, E.S., Elefsen, S.O., Hardardottir, J., Gunnarsson, A., Hreinsson, E.O., Torssander, P., Oskarsson, N.O., Oelkers, E., 2004a. Chemical Composition, Discharge and Suspended Load in Rivers in East Iceland V. Report RH-05-2004. Database of the Science Institute and the Hydrological Service of the National Energy Authority. Gislason, S.R., Eiriksdottir, E.S., Olafsson, J.S., 2004b. Chemical Composition of the Interstitial Water and Diffusive Fluxes within the Diatomaceous Sediment in Lake My´vatn, Iceland. Aquat. Ecol. 38, 163–175. Gislason, S.R., Oelkers, E.H., Snorrason, A., 2006a. Role of river-suspended material in the global carbon cycle: reply to critical comment. Geology, 112–113. Gislason, S.R., Snorrason, A., Ingvarsson, G.B., Camargo, L.G.Q., Eiriksdottir, E.S., Elefsen, S.O., Hardardottir, J., Thorlaksdottir, S.B., Torssander, P., 2006b. Chemical Composition, Discharge and Suspended Load in Rivers in South Iceland IX. Report RH-05-2006. Database of the Science Institute and the Hydrological Service of the National Energy Authority. 46 pp. Gislason, S.R., Snorrason, A., Ingvarsson, G.B., Eiriksdottir, E.S., Sigfússon, B., Flaathen, T.K., Camargo, L.G.Q., Elefsen, S.O., Hardardottir, J., Thorlaksdottir, S.B., Torssander, P., 2006c. Chemical Composition, Discharge and Suspended Load in the River Skaftá 2002 to 2006. Report RH-04-2006, pp. 61. Gislason, S.R., Snorrason, A., Ingvarsson, G.B., Sigfússon, B., Eiriksdottir, E.S., Elefsen, S.O., Hardardottir, J., Thorlaksdottir, S.B., Torssander, P., 2006d. Chemical Composition, Discharge and Suspended Load in Rivers in North-Western Iceland. Report RH-07-2006. Database of the Science Institute and the Hydrological Service of the National Energy Authority. 51 pp.

189

Gislason, S.R., Oelkers, E.H., Eiriksdottir, E.S., Kardjilov, M.I., Gisladottir, G., Sigfusson, B., Snorrason, A., Elefsen, S.O., Hardardottir, J., Torssander, P., Oskarsson, N., 2009. Direct evidence of the feedback between climate and weathering. Earth Planet. Sci. Lett. 277, 213–222. Goldsmith, S.T., Carey, A.E., Johnson, B.M., Welch, S.A., Lyons, W.B., McDowell, W.H., Pigott, J.S., 2010. Stream geochemistry, chemical weathering and CO2 consumption potential of andesitic terrains, Dominica Lesser Antilles. Geochim. Cosmochim. Acta 74 (1), 85–103. Grunwald, S., Qi, C., 2006. GIS based water quality modeling in the Sandusky Watershed, Ohio, USA. J. Am. Water Resour. Assoc. (JAWRA) 42 (4), 957–973. Halldorsdottir, S.G., Sigurdsson, F., Jonsdottir, J.F., Jóhannsson, Th., 2006. Hydrological Classification for Icelandic Waters. Nordic Water 2006: Experience and Challenges in Implementation of the EU Water Framework Directive, Vingsted Denmark, August 6th–9th 2006. Jens Christian Refsgaard, Anker Lager Hojberg (Eds.), pp. 219–236. Hreinsson, E.O, Sigurdsson, F., 2004. The Origins of the River Skaftá. Report from the National Energy Authority. OS-2004/028. Johannesson, H., Saemundsson, K., 1989. Geological Map of Iceland, Bedrock Geology. Reykjavík Icelandic Museum of Natural History and Icelandic Geodetic Survey. Scale: 1:500000. Jones, M.T., Gislason, S.R., 2008. Rapid releases of metal salts and nutrients following the deposition of volcanic ash into aqueous environments. Geochim. Cosmochim. Acta 72 (15), 3661–3680. Jonsdottir, J.F., 2008. A runoff map of Iceland based on numerically simulated precipitation and a projection of future runoff in Iceland. Hydrol. Sci. J. 53, 100– 111. Kardjilov, M.I., 2008. Riverine and Terrestrial Carbon Fluxes in Iceland. PhD Thesis, University of Iceland. Kardjilov, M.I., Gisladottir, G., Gislason, S.R., 2006. The effect of gross primary production, net primary production and net ecosystem exchange on the carbon fixation by chemical weathering of basalt in northeastern Iceland. J. Geochem. Explor. 88, 292–295. Kardjilow, M.I., Gisladottir, G., Gislason, S.R., 2006. Land degradation and carbon fluxes in Iceland. Land Degrad. Develop. 17, 401–417. Kjartansson, G., 1945. Icelandic Rivers. Fold og votn. Culture fund, Reykjavik. pp. 43–55 (in Icelandic). Kristmansdottir, H., Arnorsson, S., Sveinbjornsdottir, A.E., Armannsson, H., Bogason, E., 2006. The water resource in Iceland project. In: Energy Conference in Reykjavik, October 12th–13th 2006, pp. 195–203. Louvat, P., Allegre, C.J., 1997. Present denudation rates at Réunion Island determined by river geochemistry: basalt weathering and mass budget between chemical and mechanical erosions. Geochim. Cosmochim. Acta 61, 3645–3669. Louvat, P., Allegre, C.J., 1998. Riverine erosion rates on Sao Miguel volcanic island, Azores archipelago. Chem. Geol. 148, 177–200. Louvat, P., Gislason, S.R., Allégre, C.J., 2008. Chemical and mechanical erosion rates in Iceland as deduced from river dissolved and solid material. Am. J. Sci. 308 (May), 679–726. Meybeck, M., 1986. Composition chimique des ruisseaux non pullués de France. Sci. Geol. 39, 3–77. Meybeck, M., 1987. Global chemical weathering of surficial rocks estimated from river dissolved loads. Am. J. Sci. 287, 401–428. Millot, R., Gaillardet, J., Dupré, B., Allegre, C.J., 2002. The global control of silicate weathering rates and the coupling with physical erosion: new insights from rivers of the Canadian Shield. Earth Planet. Sci. Lett. 196, 83–98. Millot, R., Gaillardet, J., Dupré, B., Allegre, C.J., 2003. Northern latitude chemical weathering rates: clues from the Mackenzie River Basin, Canada. Geochim. Cosmochim. Acta 67, 1305–1329. Monsef, A.E.H., El-Ghawabi, M., Smith, S., 2004. GIS analysis system for investigating sulphide mineralization in South Sinai, Egypt. Survey. Land Inform. Sci. 64 (4), 243–251. Négrel, P., Allegre, C.J., Dupré, B., Lewin, E., 1993. Erosion sources determined by inversion of major and trace elements ratios and strontium isotopic ratios in river water: the Congo Basin case. Earth Planet. Sci. Lett. 120, 59–76. Olafsson, J., 1979. The chemistry of the Lake My´vatn and the River Laxá. Oikos 32, 82–112. Oskarsson, N., Sigvaldason, G.E., Steinthórsson, S., 1982. A dynamic model of rift zone petrogenesis and the regional petrology of Iceland. J. Petrol. 23 (1), 28–74. Pearce, C.R., Burton, K.W., Pogge von Strandmann, P.A.E., James, R.H., Gislason, S.R., 2010. Molybdenum isotope behaviour accompanying weathering and riverine transport in a basaltic terrain. Earth Planet. Sci. Lett. 295, 104–114. Pogge von Strandmann, P.A.E., Burton, K.W., James, R.H., van Calsteren, P., Gislason, S.R., Mokadem, F., 2006. Riverine behaviour of uranium and lithium isotopes in an actively glaciated basaltic terrain. Earth Planet. Sci. Lett. 251, 134–147. Pogge von Strandmann, P.A.E., Burton, Kevin W., James, Rachael H., Peter van Calsteren, P., Gislason, S.R., Sigfusson, B., 2008a. The influence of weathering processes on riverine magnesium isotopes in a basaltic terrain. Earth Planet. Sci. Lett. 276, 187–197. Pogge von Strandmann, P.A.E., James, R.H., van Calsteren, P., Gislason, S.R., Burton, K.W., 2008b. Lithium, magnesium and uranium isotope behaviour in the estuarine environment of basaltic islands. Earth Planet. Sci. Lett. 274, 462–471. Pogge von Strandmann, P.A.E., Burton, Kevin W., James, R.H., van Calsteren, P., Gislason, S.R., 2010. Assessing the role of climate on uranium and lithium isotope behaviour in rivers draining basaltic terrain. Chem. Geol. 270, 227–239. Pokrovsky, O., Schott, J., Kudryavtzev, D.I., Dupré, B., 2005. Basalt weathering in Central Siberia under permafrost conditions. Geochim. Cosmochim. Acta 69, 5659–5680.

190

S.M. Oskarsdottir et al. / Journal of Hydrology 397 (2011) 175–190

Probst, J.L., Mortatti, J., Tardy, Y., 1994. Carbon river fluxes and weathering CO2 consumption in the Congo and Amazon river basins. Appl. Geochem. 9, 1–13. Rad, S., Louvat, P., Gorge, C., Gaillardet, J., Allegre, C.J., 2006. River dissolved and solid loads in the Lesser Antilles: new insight into basalt weathering processes. J. Geochem. Explor. 88, 308–312. Riotte, J., Chabaux, F., Benedetti, M.F., Dia, A., Gérard, M., Boulegue, J., Etamé, J., 2003. U colloidal transport and origin of the 234U/238U fractionation in surface waters: new insights from Mount Cameroon. Chem. Geol., 202. Rist, S., 1956. Íslenzk vötn: Icelandic Fresh Waters. Text and Summary in English. The Hydrological Service of the National Authority of Iceland. Rist, S., 1974. Chemical Analysis of Water: Hvita-Olfusa River Catchments; also Thjorsa at Urridafoss, Waterfall, 1973. National Energy Authority, Reykjavik. OSV7405. Rist, S., 1986. Chemical Analysis of Water: Borgarfjordur, also Ellidaar River in Reykjavik. National Energy Authority, Reykjavik. OS-86070/VOD-03. Sigfússon, B., Gislason, S.R., Paton, G.I., 2006. The effect of soil solution chemistry on the weathering rates of Histic Andosol. J. Geochem. Explor. 88, 321–324. Sigfusson, B., Gislason, S.R., Paton, G.I., 2008. Pedogenesis and weathering rates of a Histic Andosol in Iceland: field and experimental soil solution study. Geoderma 144, 572–592. Sigurdsson, F., 1990. Groundwater from glacial areas in Iceland. Jökull 40, 119–146. Sigurdsson, F., Einarsson, K., 1988. Groundwater resources of Iceland – availability and demand. Jökull 38, 35–52. Stallard, R.F., Edmond, J.M., 1983. Geochemistry of the Amazon 2. the influence of geology and weathering environment on the dissolved load. J. Geophys. Res. 88, 9671–9688.

Stefánsson, A., Gislason, S.R., 2001. Chemical weathering of basalts, SW Iceland: effect of rock crystallinity and secondary minerals on chemical fluxes to the ocean. Am. J. Sci. 301, 513–556. Thorbergsdottir, I.M., Gislason, S.R., 2004. Internal loading of nutrients and certain metals in the shallow eutrophic lake My´vatn, Iceland. Aquat. Ecol. 38, 191–207. Vigier, N., Bourdon, B., Dupré, B., Turner, S., Chakrapani, G.J., van Calsteren, P., Allegre, C.J., 2005. Mobility of U-series nuclides during basalt weathering: an example from the Deccan Traps (India). Chem. Geol. 219, 69–91. Vigier, N., Burton, K.W., Gislason, S.R., Rogers, N.W., Duchene, S., Thomas, L., Thomas, E., Schaefer, B., 2006. The relationship between riverine U-series disequilibrium and erosion rates in a basaltic terrain. Earth Planet. Sci. Lett. 249 (3–4), 258–273. Vigier, N., Gislason, S.R., Burton, K.W., Millot, R., Mokadem, F., 2009. The relationship between riverine lithium isotope composition and silicate weathering rates in Iceland. Earth Planet. Sci. Lett. 287 (3–4), 434–441. Wang, C.Y., Wang, X.L., 2007. Spatial distribution of dissolved Pb, Hg, Cd and As in the Bohai Sea. J. Environ. Sci. 19, 1061–1066. White, D., 1999. The Physiology and Biochemistry prokaryotes, second ed. Oxford University Press, Oxford. Wimpenny, J., Gannoun, A., Burton, K.W., Widdowson, M., James, R.H., Gíslason, S.R., 2007. Rhenium and osmium isotope and elemental behaviour accompanying laterite formation in the Deccan region of India. Earth Planet. Sci. Lett. 261, 239– 258. Woelff-Boenisch, D., Gislason, S.R., Oelkers, E.H., 2006. The effects of crystallinity on dissolution rates and CO2 consumption capacity of silicates. Geochem. Acta. 70, 858–870.