Journal of Contaminant Hydrology 161 (2014) 10–23

Contents lists available at ScienceDirect

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

Acidification of floodplains due to river level decline during drought Luke M. Mosley a,b,c,⁎, David Palmer a, Emily Leyden a, Freeman Cook e, Benjamin Zammit a, Paul Shand b,c,d, Andrew Baker b, Rob W. Fitzpatrick b,c a b c d e

Environment Protection Authority (South Australia), GPO Box 2607, Adelaide, SA 5001, Australia CSIRO Land and Water, Private Bag No. 2, Glen Osmond, SA 5064, Australia Acid Sulfate Soils Centre, The University of Adelaide, SA 5005, Australia School of the Environment, Flinders University, PO Box 2100, Adelaide, SA 5001, Australia Freeman Cook and Associates Ltd, PO Box 97, Glasshouse Mountains Q4518, Queensland, Australia

a r t i c l e

i n f o

Article history: Received 30 September 2013 Received in revised form 17 March 2014 Accepted 21 March 2014 Available online 31 March 2014 Keywords: Pyrite oxidation Acid sulfate soils Acid drainage Surface–groundwater interactions Climate change

a b s t r a c t A severe drought from 2007 to 2010 resulted in the lowest river levels (1.75 m decline from average) in over 90 years of records at the end of the Murray–Darling Basin in South Australia. Due to the low river level and inability to apply irrigation, the groundwater depth on the adjacent agricultural flood plain also declined substantially (1–1.5 m) and the alluvial clay subsoils dried and cracked. Sulfidic material (pH N 4, predominantly in the form of pyrite, FeS2) in these subsoils oxidised to form sulfuric material (pH b 4) over an estimated 3300 ha on 13 floodplains. Much of the acidity in the deeply cracked contaminated soil layers was in available form (in pore water and on cation exchange sites), with some layers having retained acidity (iron oxyhydroxysulfate mineral jarosite). Post drought, the rapid raising of surface and ground water levels mobilised acidity in acid sulfate soil profiles to the floodplain drainage channels and this was transported back to the river via pumping. The drainage water exhibited low pH (2–5) with high soluble metal (Al, Co, Mn, Fe, Mn, Ni, and Zn) concentrations, in exceedance of guidelines for ecosystem protection. Irrigation increased the short-term transport of acidity, however loads were generally greater in the non-irrigation (winter) season when rainfall is highest (0.0026 tonnes acidity/ha/day) than in the irrigation (spring–summer) season (0.0013 tonnes acidity/ha/day). Measured reductions in groundwater acidity and increases in pH have been observed over time but severe acidification persisted in floodplain sediments and waters for over two years post-drought. Results from 2-dimensional modelling of the river-floodplain hydrological processes were consistent with field measurements during the drying phase and illustrated how the declining river levels led to floodplain acidification. A modelled management scenario demonstrated how river level stabilisation and limited irrigation could have prevented, or greatly lessened the severity of the acidification. © 2014 Elsevier B.V. All rights reserved.

1. Introduction

⁎ Corresponding author at: Environment Protection Authority South Australia, GPO Box 2607, Adelaide, SA 5001, Australia. Tel.: +61 8 8463 7808; fax: +61 8 8124 4673. E-mail address: [email protected] (L.M. Mosley).

http://dx.doi.org/10.1016/j.jconhyd.2014.03.003 0169-7722/© 2014 Elsevier B.V. All rights reserved.

Understanding the landscape connectivity and relationships between rivers and floodplains is essential to understand the effects of water-resources management decisions in a basin (Sophoceleus, 2002). Groundwater systems that are hydraulically connected to rivers have been found to be very responsive to changing climates and surface water regimes (Lamontagne

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

et al., 2005; Wondzell and Swanson, 1996a). Droughts are predicted to increase over the next 30–90 years (Dai, 2013) which will place increased pressure on surface and groundwater resources. There is a high risk that declining surface water levels during droughts could alter hydrological and geochemical regimes beneath adjacent floodplains. Groundwater–surface water interactions can result in complex biogeochemical transformations (Banks et al., 2011; Winter et al., 1998), particularly in hyporheic and riverbank zones due to redox changes, mineral weathering and cation exchange processes (Boulton et al., 1998; Bourg and Bertin, 1993; Matsunaga et al., 1993; von Gunten et al., 1991, 1994). Iron sulfides such as pyrite (FeS2), build up during constant waterlogged conditions, especially where abundant organic material and sulfate are present (Dent and Pons, 1995; Pons, 1973). Upon exposure to oxygen caused by full or partial desaturation of the soil profile (Bronswijk et al., 1993), pyrite can react to release sulfuric acid and ferrous iron (Fe+2) via the following reaction: þ2

FeS2 þ 7=2O2 þ H2 O→Fe

−2

þ

þ 2SO4 þ 2H :

ð1Þ

The ferrous iron can subsequently be oxidised to Fe+3 which at low pH (b4) can rapidly oxidise additional pyrite, or at higher pH (N5) hydrolyse and precipitate as ferric iron oxide and oxyhydroxide minerals such as goethite and ferrihydrite. Such soils, termed acid sulfate soils, can turn sulfuric (pH b 4) if insufficient neutralising capacity (typically represented by their carbonate content) is present. The soil acidification can also release aluminium, manganese and other metals into pore water (Cook et al., 2000a; Österholm and Åström, 2002; Simpson et al., 2010). The causes of oxidation of acid sulfate soils have been found previously to be controlled by human-induced changes in drainage conditions (Cook et al., 2000a; Dent and Pons, 1995; MacDonald et al., 2007; Sammut et al., 1996), and post-glacial land uplift in some locations (Boman et al., 2010). There have also been reports of drought-influenced acidification of groundwater (Appleyard and Cook, 2009; Lucassen et al., 2002) and surface water (Mosley et al., 2013) following pyrite oxidation. However, decline of surface water levels during droughts resulting in severe pyrite oxidation and acidification on an adjacent (hydrologically-connected) floodplain, has not to our knowledge been reported previously. There are some reports of sulfide oxidation in connected surface:groundwater systems. Van den Berg et al. (1998) reported minor pH decline (pH 7.3 to 6.9) and metal precipitation (Fe and Mn oxides) when groundwater levels dropped and pyrite oxidised in an estuarine soil during summer, with sulfate reduction occurring following winter rewetting. Massmann et al. (2003) found zones of both sulfate reduction and oxidation (by O2 and/or NO2/NO3) in a river recharged floodplain aquifer but severe acidification was not present. Hydro-chemical models have proved useful for understanding dynamics between rivers and floodplains (Wondzell and Swanson, 1996b) and drainage processes from acid sulfate soils (Rassam and Cook, 2002). From 2007–2010, extremely low river inflows from the Murray–Darling Basin catchment led to large water level declines in the lower River Murray and Lakes in South Australia (Mosley et al., 2012). Following the return of water levels to pre-drought averages in the Lower Murray region in late 2010,

11

very low pH (b3) conditions and unusual orange–brown colourations were observed in drainage waters on agricultural floodplains adjacent to the river. It was hypothesised that: (a) the falling river levels led to groundwater declines and drying of the floodplain which caused oxidation of pyrite and subsequent acidification of the soil profile, and (b) post-drought, acidity in the dessicated and cracked soil profile was mobilised to drainage channels by rainfall and irrigation. The aim of this paper is to elucidate the combined hydro-geochemical and pedological processes that led to the observed acidification of soil, groundwater and drainage water. To help understand these processes integrated soil, groundwater, and drainage water sampling and analyses were undertaken at multiple sites over a 2 year period. The surface–groundwater interactions were also examined using a two dimensional (2D) saturated–unsaturated zone model. The outcomes of this study have broad relevance to other locations, which could experience decline in surface-water and groundwater levels under floodplains that contain acid sulfate soils. 2. Methods 2.1. Study area description The study site is part of a region of floodplains known locally as the Lower Murray Reclaimed Irrigation Area (LMRIA). The LMRIA comprises 5200 ha area of floodplain (consisting of 27 individual areas) reclaimed for agriculture in the lower Murray River region (Fig. 1). Historically, the floodplain was likely composed of reed beds (Phragmites sp.) with regular flooding under a natural river regime. Much of the floodplain was drained and developed for agriculture between 1880 and 1940 (Taylor and Poole, 1931). With the aim of improving water security and navigation, in the 1920s to 1940s a series of locks were constructed along the river, with barrages to keep the sea out near the river mouth (Fig. 1). The hydrological regime in the lower River Murray was more variable before these changes were initiated, with water levels rising and falling with seasonal and unregulated flow fluctuations. Once the locks and barrages were built, the river level was maintained at a relatively constant level, eliminating large seasonal fluctuations. This, along with the construction of the levee banks which contained and raised the level of the river in the main channel, benefited the development of agriculture in the LMRIA as it ensured agricultural water security and enabled low cost gravity fed irrigation. However, these infrastructure changes also altered the local hydrogeology, with the floodplain becoming the new discharge point for the highly saline (10,000–30,000 μS/cm conductivity) regional groundwater (Barnett et al., 2003). Discharge of river water also usually occurs to the upper floodplain (termed a “losing” system), due to the regulated river level now being mostly kept elevated (Barnett et al., 2003). To prevent the agricultural land from becoming water logged and salinised, a series of drainage channels were constructed to intercept and drain the saline regional groundwater and excess irrigation. The typical drain layout comprises shallow (0.5–1 m) lateral or side drains alongside graded irrigation bays and a deeper (1.5–2.5 m below ground level) salt drain at the rear of the bay that intercepts both the paddock drainage and regional groundwater (Figs. 1 and 2). The

12

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

13

Fig. 2. Hydrus 2D model domain showing the various boundaries (variable river level left-hand side, constant head at highland/regional groundwater right hand side, seepage face for salt drain, atmospheric boundary on the irrigation bay, no flux at bottom boundary).

drainage water is pumped back to the river to keep the local water table around 0.5–1 m below ground level (bgl). This range of groundwater levels was also present in the 1930s (Taylor and Poole, 1931). Historically, irrigation practises in the LMRIA were quite inefficient (Mosley and Fleming, 2009) and returned large volumes of polluted (high levels of nutrients and pathogens) drainage water that negatively impacted River Murray water quality (Mosley and Fleming, 2010). Maintaining water quality in this (approx. 75 km long) reach of the River Murray is critical as it contains several large drinking water off takes which service the city of Adelaide (population 1.23 million) and several regional areas (see Fig. 1 for locations). The river also has important recreational and tourism uses and immediately downstream is the internationally important (Ramsar-listed) Lower Lakes and Coorong aquatic ecosystems. 2.2. Soil sampling and analysis A push-tube corer was used to collect soil cores to a depth of 3 m at 3 sites at Long Flat irrigation area in June 2011 (Fig. 1). Wet soil samples were analysed for a range of parameters including pH in water and KCl (using a calibrated pH electrode), bulk density (gravimetric method), and particle size (% sand, silt, clay by hydrometer method from Gee and Bauder, 1986). Sub-samples of soil were dried (at 60 °C for 48 h) prior to crushing and analysis for full acid–base characteristics (methods from Ahern et al., 2004) including titratable actual acidity (TAA, measure of soluble and exchangeable acidity), chromium reducible sulfur (CRS, assumed to be pyrite), retained acidity (RA, typically comprised of iron oxyhydroxy sulfate minerals such as jarosite), and acid neutralising capacity (ANC, presumed to be mostly from carbonates). Exchangeable cations (Ca, K, M, Na, H, Al) were determined using a standard Ammonium Acetate extraction (Method 15D3, Rayment and Higginson, 1992) with no pretreatment for soluble salts. 2.3. Ground and drain water sampling and analysis A series of multi-level piezometers were installed at Long Flat irrigation area in June 2011 at the same sites as the soil cores (Fig. 1). The piezometers were installed in a transect from

the river, along the irrigation bay/floodplain towards the drainage channel, and highland (Fig. 1 inset). Sites 3, 5 and 7 each have three multilevel piezometers that are screened at depths of 0.3–0.5 m (A horizon), 0.75–1.25 m (B horizon) and 2.5–3.0 m (C horizon) bgl. Site 1 is a regional groundwater control site (single piezometer) that is above the floodplain on the immediately adjacent highland area and is screened between 2.5 and 3.0 m bgl (C horizon). Prior to groundwater samples being collected, the piezometers were purged using a 12-volt Solonist™ Peristaltic Pump. Three well volumes were typically pumped prior to sampling. Sampling was undertaken approximately fortnightly–monthly over a two year period with more intensive sampling at the time of managed irrigation events at the site. The LMRIA drain (surface) water was also sampled fortnightly– monthly at the various drainage water pumping stations that discharge to the river (Fig. 1). Samples were collected by grab sampling using a clean polyethylene groundwater bailer that was vertically lowered to obtain a representative sample of the water column in the drain. pH was measured in the drains and piezometers at the time of sample collection using a calibrated instrument (YSI Pro Plus), with a flow cell used for the groundwater samples to minimise exposure to air during measurement. New polyethylene bottles, washed and rinsed with deionised water, were used to collect groundwater and drain water samples for laboratory analyses of acidity, alkalinity, and conductivity. Acid-cleaned bottles were used to collect samples for trace metal (Al, As, Co, Cr, Cu, Mn, Fe, Mn, Ni, and Zn) analysis. Laboratory analyses were undertaken by the Australian Water Quality Centre's National Association of Testing Authorities (NATA) accredited laboratory within recommended holding times using standard methods (APHA, 2005). Total alkalinity was measured by titration to a pH 4.5 end-point. Acidity was measured by titration to pH 8.3 end point at 25 °C following hot peroxide digestion. “Dissolved” (b0.4 μm filtered) and total (following EPA 200.8 reflux digestion with nitric and hydrochloric acids) metals were measured by ICP-MS (Agilent 7500 series). Quality assurance and control checks were regularly carried out both in the field and laboratory and acceptable results were achieved for replicates (within 10%), blanks, and spikes (85–115% recovery).

Fig. 1. Location of the Lower Murray Reclaimed Irrigation Area (LMRIA) region in the lower River Murray of the Murray–Darling Basin South Australia. Also shown is the location of soil, groundwater (piezometers), drains (red = acidic and green = non-acidic floodplains/swamps), and meteorological (met) monitoring sites near Murray Bridge, and the long term monitoring station upstream at Lock 1.

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

Lower Murray River water (Murray Bridge, Site ID 4261003 and downstream of Lock 1, Site ID A4260903) and floodplain groundwater (Mobilong site 032/01A, 5 m bgl piezometer; Mobilong site 034/02A, 3 m bgl piezometer) levels were obtained from the Department of Environment, Water and Natural Resources South Australia (available online at www. waterconnect.sa.gov.au, see Fig. 1 for locations). Piezometric level data were also collected at 15-minute intervals at the Long Flat piezometers over the monitoring period using In-Situ Level TROLL™ 500 vented water level loggers. Ground and surface water levels were corrected to metres Australian Height Datum (m AHD) from survey data collected at the time of piezometer installation (zero m AHD corresponds approximately to mean sea level). Recording of drainage pump run time was undertaken using an Agenti™ logger (at 15 minute intervals) connected to the pump switch. Estimates of the drainage volume were obtained by gauging the pump inlet channel/culvert flow using an ultrasonic raft and multiplying by the pump run time. 2.5. Modelling of surface–ground water interactions A two dimensional (2D) saturated–unsaturated zone model (Hydrus 2D/3D, Šimůnek et al., 2012) was constructed to assess the river-floodplain groundwater interactions and zone of soil drying and pyrite oxidation during the drought. The 2D domain comprised a cross section from the river, across the LMRIA floodplain to the salt drain, and to the beginning of the elevated highland area (Fig. 2). The 2D model cross section geometry and model parameters were based on surveys and soil measurements at the Long Flat research site (just downstream of Mobilong irrigation area, see Fig. 1). The measured river level at Murray Bridge was used as the left hand side boundary condition (variable head) for the model. The highland groundwater level was kept constant (−1.5 m AHD, informed by data collected from piezometer 1C) for the right hand side of the model. The salt drain was represented by a seepage face in the model. Variable atmospheric boundary conditions were included in the model based on daily rainfall at Murray Bridge obtained from the Australian Bureau of Meteorology (available online at www.bom.gov.au) and potential evapo-transpiration data (calculated using the Penman–Monteith equation for a short/pasture crop from method in ASCE, 2005) at Mypolonga from the South Australian Murray–Darling Basin Natural Resources Management Board (available online at www.aws-samdbnrm.sa.gov. au, see Fig. 1 for general locations). An observation node was placed at the same distance and depth in the model as the Mobilong Mob032/01A piezometer (see Fig. 1) to enable comparison of the model with measurements. Particle size and bulk density results were used in pedotransfer functions (Rawls and Brakensiek, 1985) to estimate initial soil hydraulic parameters which were subsequently optimised using inverse modelling in Hydrus (Table 1). The saturated hydraulic conductivities (Ks) also shown in Table 1 were estimated using the Hvorslev slug (rising head) test. Water mass balance outputs for the Hydrus model showed satisfactory results (b 0.2% error). Examples of the Hydrus 2D model files are available on request.

Table 1 Soil physical and hydraulic properties used in the HYDRUS 2D model. Depth m

θr

θs

α

n

Ks m/day

I

Bulk D.

0–3

0.10

0.590

2.117

1.709

0.1

0.5

1.09

3. Results 3.1. Surface and ground water hydrology From 2007 to 2010 the lower Murray River (downstream of Lock 1) experienced an extreme water level decline (approx. 1.75 m, Fig. 3). These were the lowest water levels measured in over 90 years of records from the long term monitoring station downstream from Lock 1 at Blanchetown (see Fig. 1 for location). The groundwater levels on the floodplain at Mobilong declined between 1 and 1.5 m (Fig. 3). The trends in the groundwater levels were similar to the Murray Bridge surface water levels, with a lag time of a few months (Fig. 3). The magnitude of groundwater decline was similar at both monitoring sites with the lower absolute level at the Mob034/02A site likely due to this piezometer being much further (560 m) from the river than the Mob032/01A piezometer (62 m, see Fig. 1).

9

7

Water Level (m AHD)

2.4. Hydrological data

5

3

1

−1 1920 1.5

Level (m AHD)

14

Drought period

1940

1960

1980

2000

2020

Jan−09

Jan−11

Jan−13

0

−1.5

River Groundwater Mob032(01A) Groundwater Mob034(02A)

−3 Jan−03

Jan−05

Jan−07

Fig. 3. (top) long term (1921 to 2013) dataset of Lower Murray River levels measured upstream of the LMRIA at Lock 1, (bottom) River level (at Murray Bridge) and groundwater level at Mobilong (Bores Mob01A/032 and Mob02A/034) from 2003 to 2012. See Fig. 1 for locations.

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

15

0

3.2. Soil acid–base accounting As a consequence of the ground water decline, the heavy clay soils dried and developed deep (0.5-2 m) desiccation cracks (see photos in Supplementary material S1). Acid–base accounting data of the soil profile at Long Flat post-drought are presented in Fig. 4. The upper (0–1 m soil layer) had a circum-neutral pH with acid neutralising capacity (ANC) present at Sites C and B. This upper layer is above the lateral drain depth in the zone of the typical pre-drought water table level (0.5–1.0 m bgl). Hence this upper 1 m of soil has been regularly wet and dried during normal irrigation activities and rainfall events, which would not have allowed significant buildup of sulfidic materials. There was a low pH (b4.5) zone at a depth of 1–2.5 m bgl (Fig. 4) with high amounts of soil acidity in available form (TAA 100–200 mol H+/tonne). There was also retained acidity (RA) present in a few layers, which was likely in the form of the iron oxyhydroxysulfate mineral natrojarosite (see Fitzpatrick et al., 2013). There is also still unoxidised pyrite (CRS) in the 1–2.5 m bgl acidic layer suggesting only partial oxidation occurred during the drought. Acidic exchangeable cation (H+, Al+3) concentrations increased 1–2 orders of magnitude in the 1–2.5 m bgl layer, with a decrease in base cation (Mg+2, Na+, K+) concentrations (Fig. 5). The deeper (≈ 3 m bgl) soil layer had a circum-neutral pH and a large proportion of unoxidised pyrite (CRS, Fig. 4). This suggests that this layer did not dry and oxidise during

Soil depth (cm)

50 100 150

Exch. Al Exch. H Exch. Ca Exch. Mg Exch. K Exch. Na

200 250 300 −2 10

10−1

100

101

102

Exchangeable Cations (meq/100g) Fig. 5. Exchangeable acidic (H+, Al+3) and base (Ca+2, Mg+2, Na+, K+) cations in the soil profile at Long Flat (site DS 01 B, see Figs. 1 and 4) in 2011.

the drought. More details of the nature and properties of the acid sulfate soils are provided in Fitzpatrick et al. (2013). 3.3. Groundwater quality Groundwater quality trends over a two year period are shown in Fig. 6 for the multi-level piezometer transect at

DS 01 C

DS 01 B

DS 01 A

0

Soil depth (cm)

50 100 150 200 250 300 2

4

6

8

2

4

pH

6

8

2

4

pH

6

8

500

1000

pH

0

Soil depth (cm)

50 100 150 200 250 300 −250

0

250

−50

mol H+/tonne

0

125

250

−500

mol H+/tonne TAA

CRS

0

mol H+/tonne RA

ANC

Fig. 4. (Top) pHKCl and (bottom) acid–base accounting, TAA = titratable actual acidity, CRS = chromium reducible sulfur (assumed to be pyrite), RA = retained acidity (likely comprised of the iron oxyhydroxy sulfate mineral jarosite) and ANC = acid neutralising capacity (presumed to be from carbonates) in the soil profile at Long Flat (see Fig. 1 for site map).

16

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

8 7 6 5 4

8 7 6 5 4

3000

2000

2000

1500

1000

1000

0

0

0

1000

1000

2000

500

500

1000

Alkalinity

Acidity

pH

8 7 6 5 4

0

0

SEC (muS/cm)

4

3

0 4

x 10

2

4

x 10

2

x 10

2 1

1

1

0 Jun−11

Jun−12

Jun−13

0 Jun−11

7A 7B 7C

Jun−12

Jun−13

0 Jun−11

5A 5B 5C

Jun−12

Jun−13

3A 3B 3C 1C

Fig. 6. Groundwater quality (pH, acidity, alkalinity and specific electrical conductivity) at Long Flat from June 2011 to June 2013 (see Fig. 1 for piezometer locations).

Long Flat irrigation area (see Fig. 1 for locations). pH in the upper level piezometers (‘A’ series, 0.3–0.5 m bgl) was circum-neutral at all sites (note data series is non-continuous as groundwater was only present post rainfall or irrigation). This is consistent with the soil pH results in this layer (Fig. 4). Minor amounts of acidity were present in the upper piezometers, but alkalinity was typically also present at higher concentrations (Fig. 6). Salinity in piezometer 3A was higher than the other ‘A’ level piezometers, likely due to its location at lower elevation near the salt drain (where drainage water is retained for longer). In contrast, groundwater in the mid-level (‘B’ series, 0.75–1.25 m bgl) and lower level (‘C’ series, 2.5–3 m bgl) piezometers was highly acidic (Fig. 6, pH 3–5, acidity 200– 2000 mg/L as CaCO3). Over the 2 year monitoring period, there appears to be a trend of increasing pH and decreasing acidity in the groundwater. The pH increase was greatest near the river (Site 7) with alkalinity appearing in the mid and lower level (7B & C) piezometers (Fig. 6). Groundwater salinity also decreased substantially over the monitoring period.

floodplain area with acid drainage was estimated to be 3300 ha, spread over 13 irrigation areas (Fig. 1). The drainage water is returned via pumps to the River Murray and has dissolved metal (Al, Co, Mn, Fe, Mn, Ni, and Zn) concentrations that also consistently exceed guidelines for aquatic ecosystem protection (Table 2). The orange-brown coloration in many of the drain water's (see photo in Supplementary material S3) was identified as the iron oxyhydroxysulfate mineral schwertmannite (identified using X-Ray Diffraction, see Supplementary material S4 and Fitzpatrick et al., 2012, 2013). The drainage water quality is quite variable between and within different sites, but over the two year monitoring period pH and alkalinity increased at most sites, while acidity and salinity decreased (Fig. 7). Drain acidity appears to peak in spring and early summer, which is the beginning of the irrigation season. Mobilong, which has low pH values, and very high acidity and metal concentrations (Table 2), has a different seasonal pattern than the other sites. This site was retired from active farming and irrigation a decade ago, and has only received very few irrigations since then.

3.4. Drainage water quality and loads 3.5. Modelling and acidity loads There was a persistent low pH (3–6) in the LMRIA salt drains (measured at pump station returning water to the river) over the two year monitoring period (Fig. 7). Median pH levels (Table 2) were below the guidelines (pH 6.5–9.0) for aquatic ecosystem protection (ANZECC, 2000) in all but one area (Kilsby). The total

Rewetting of the dried, cracked and acidified soil occurred in late 2010 when the river and groundwater returned to pre-drought levels (Fig. 2). Shortly after, irrigation and drainage activities recommenced in the LMRIA. Study of

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

pH

Acidity 1500

mg L as CaCO3

8

pH

6

4

750

2

0

Alkalinity

4

4

µS/cm at 25C

mg L as CaCO3

350

175

0

17

Jun−11

Jun−12 Jervois Sth

Jun−13 Long Flat

x 10

Conductivity

2

0

Jun−11 Mobilong

Jun−12

Jun−13

Riverglen

Fig. 7. Drain (surface) water quality (pH, acidity, alkalinity, conductivity) in the LMRIA (at point of discharge to the river) for selected sites (see Supplementary material S2 for all sites) from 2011–2013. Refer to Fig. 1 for site locations.

hydrological processes at the Long Flat site showed that flood irrigation (1st post drought) resulted in a rapid rise in groundwater levels followed by a rapid recession/drainage over the next few days (Fig. 8). A large (200–300 mm) irrigation water depth was required due to the poor postdrought condition of the floodplain (cracking, heaving, slumping). Rainfall events (maximum of 54 mm on 18 December 2011) resulted in similar but lesser magnitude trends in groundwater levels (Fig. 8). Slight decreases in acidity occurred during irrigation (Fig. 8). Following export of acidity from the paddocks to the drains, pumping of drainage water to the river occurs. Pumping increases post-irrigation and rainfall, but also occurs periodically in between these periods (Fig. 8), presumably due to the regional groundwater input (Barnett et al., 2003). Average acidity loads pumped from the drains to the river were greater in the non-irrigation (winter) season when rainfall is highest (0.0026 tonnes acidity/ha/day) than in the irrigation (spring– summer) season (0.0013 tonnes acidity/ha/day) (Table 3). Drainage volumes (2–10 ML/day depending on irrigation area) were also highest on average in the non-irrigation season (Table 3). To assess the river-floodplain interaction processes that led to the acidification, the Hydrus 2D model was run for a period of 2121 days from 23/10/2005 to 13/8/2011 (comprising the period of drought induced water level decline, Fig. 2). The lower depth of drought-induced groundwater decline (to 2.5 m bgl) in the model compared quite well to measurements at Mobilong (Fig. 9). However, the periodic spring rise in measured groundwater levels (driven by increases in river level, Fig. 3) was much greater than in the model. The model

outputs for the whole river-floodplain system show that as the river level fell, the groundwater level also fell 1–1.5 m, with the soil desaturating to a similar depth (Fig. 10). 4. Discussion 4.1. Acidity generation processes The declining groundwater levels on the LMRIA floodplain during the drought period followed trends in the river level. The depth of the acidic soil layer that developed was very similar to the 1–1.5 m depth of groundwater decline (below the pre-drought groundwater level). This strongly supports our hypothesis that the river–groundwater level decline led to floodplain subsoil drying, which resulted in pyrite oxidation and acidification (Eq. (1)). Artificially maintained high river levels probably kept these subsoils continually saturated, conducive (along with the presence of sulfate in saline regional groundwater and high amounts of organic matter in the soil) to building up pyrite concentrations (Fitzpatrick et al., 2009), and preventing any major release/oxidation for more than 70 years (Fig. 3). Pre-river regulation, the more frequent and variable wetting and drying cycles would have resulted in some formation of sulfides (potential acidity) in the floodplain subsoils during wet periods with the oxidation and release of these small amounts of acidity during the drier periods. The amount of pyrite in the LMRIA subsoil that was exposed during the drought was very high, in the order of hundreds of mol H+/tonne (compare unoxidised/CRS in 3 m bgl layer in Fig. 3). Pyrite oxidation processes produced high levels of soluble

18 Table 2 Summary statistics (median, interquartile range = IQR, and number of sample = n) for pH, acidity, alkalinity, pH, specific electric conductivity (Cond.), sulfate (SO4), and dissolved (diss., b0.45 μm) and total (tot.) metal (Al, As, Cr, Co, Cu, Fe, Mn, Ni, Zn) concentrations at selected sites the LMRIA (refer to Fig. 1 for locations). Bold values indicate exceedances of water quality guidelines (WQG) for aquatic ecosystem protection (95 percentile values, ANZECC, 2000). Site

WQG Burdett

Jervois (Nth)

Kilsby

Long Flat

Mobilong

Monteith

Pompoota

Riverglen

Toora

Westbrook Pk

Woods Pt

All sites combined

Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n Median IQR n

pH mg/L

Acidity mg/L CaCO3

Alkalinity mg/L CaCO3

Cond. uS/cm

SO4 mg/L

Al (diss.) mg/L

Al (tot.) mg/L

As (diss.) mg/L

As (tot.) mg/L

Cr (tot.) mg/L

Co (tot.) mg/L

Cu (tot.) mg/L

Fe (diss.) mg/L

Fe (tot.) mg/L

Mn (diss.) mg/L

Mn (tot.) mg/L

Ni (tot.) mg/L

Zn (tot.) mg/L

6.5 4.0 2.6 13 3.5 0.9 22 3.8 2.1 22 6.8 0.6 20 5.4 1.0 22 3.5 0.7 22 3.9 1.8 23 4.3 1.1 23 6.8 0.8 23 5.4 2.8 22 6.1 1.3 19 6.4 0.8 21 5.1 2.7 252

n/a 286 157 15 130 76 24 103 90 24 135 74 22 156 139 24 331 438 24 120 69 25 107 122 25 68 52 25 82 118 22 57 104 20 43 35 23 112 133 273

n/a 0 10 15 0 1 24 0 6 24 417 295 22 16 50 24 0 0 24 1 9 25 3 8 25 145 168 25 8 52 24 31 150 20 55 89 23 6 67 275

n/a 8820 3630 13 3760 1298 22 5805 2583 22 2660 2938 20 3870 2463 22 21,250 12,000 22 5280 3135 23 8430 4115 23 6660 12,590 23 10,495 9365 22 18,400 2750 19 5710 3080 21 6635 8068 252

3240 1712.5 14 1330 570 23 2245 1110 24 463 710 21 1160 340 23 3750 1860 23 1470 1095 24 2905 1315 24 1445 1245 24 2830 2255 23 3030 1665 19 1240 705 22 2010 1882.5 264

0.055 9.61 18.23 15 1.25 3.30 24 3.69 10.03 25 0.01 0.04 22 0.05 0.19 24 10.02 17.97 24 0.26 0.86 25 0.95 1.32 25 0.01 0.04 25 0.77 2.20 24 0.03 0.76 19 0.01 0.38 22 0.45 2.82 274

0.055 11.84 17.60 16 2.23 2.90 25 4.68 10.88 26 0.20 1.53 23 0.25 0.31 25 12.90 21.04 24 0.63 1.36 26 1.64 1.68 26 0.13 0.70 26 1.53 2.45 25 0.79 9.23 20 0.35 1.13 23 1.22 4.02 286

0.013 0.003 0.001 13 0.002 0.002 22 0.002 0.002 23 0.003 0.001 20 0.004 0.011 22 0.004 0.005 22 0.003 0.001 23 0.003 0.001 23 0.003 0.000 23 0.003 0.000 22 0.003 0.000 19 0.003 0.002 21 0.003 0.001 253

0.013 0.006 0.004 16 0.003 0.001 25 0.003 0.001 26 0.006 0.008 23 0.018 0.011 25 0.005 0.006 24 0.004 0.003 26 0.005 0.007 26 0.004 0.002 26 0.003 0.004 24 0.003 0.000 20 0.003 0.001 23 0.004 0.004 285

0.001 0.001 0.003 13 0.001 0.001 24 0.001 0.001 23 0.001 0.001 22 0.001 0.000 23 0.002 0.003 22 0.001 0.000 24 0.001 0.001 24 0.001 0.000 24 0.001 0.001 24 0.001 0.000 19 0.001 0.001 22 0.001 0.001 265

0.001 0.297 0.247 13 0.162 0.121 24 0.136 0.208 23 0.008 0.130 22 0.083 0.049 23 0.526 0.434 22 0.125 0.182 24 0.265 0.153 24 0.049 0.104 24 0.269 0.390 24 0.072 0.132 19 0.030 0.071 22 0.136 0.239 265

0.001 0.004 0.006 13 0.003 0.002 24 0.006 0.009 23 0.001 0.003 22 0.003 0.002 23 0.006 0.010 22 0.002 0.002 24 0.004 0.004 24 0.003 0.003 24 0.002 0.004 24 0.002 0.002 19 0.002 0.003 22 0.003 0.004 265

n/a 18.1 37.1 13 7.1 19.6 22 2.3 6.5 23 0.7 5.5 20 2.5 13.1 22 63.3 86.6 22 6.9 15.9 22 8.4 11.4 23 0.6 1.6 23 2.9 19.1 22 0.5 3.7 19 0.7 1.1 21 3.2 16.9 252

n/a 73.9 116.1 16 32.7 30.6 25 13.9 21.5 26 33.5 42.7 23 34.2 16.2 25 104.5 100.6 24 34.4 46.2 26 53.7 64.4 26 9.4 19.9 26 13.4 71.1 25 8.3 15.1 20 7.0 10.8 22 27.5 50.7 285

1.9 5.6 3.5 13 4.1 1.9 22 3.8 2.8 23 1.0 1.1 20 2.6 0.9 22 8.3 5.9 22 2.2 3.1 23 5.6 2.7 23 1.2 1.4 23 2.9 4.0 22 9.9 9.6 19 1.6 1.3 21 3.5 4.7 253

1.9 6.3 3.7 16 4.3 2.5 25 4.0 4.0 26 1.2 4.1 22 2.8 1.8 25 9.2 6.6 24 2.7 3.6 26 6.5 3.7 26 1.5 1.7 26 4.4 4.7 25 9.9 7.8 20 1.7 1.6 23 4.2 5.1 285

0.011 0.366 0.365 13 0.158 0.119 24 0.186 0.248 23 0.011 0.132 22 0.077 0.058 23 0.477 0.516 22 0.123 0.185 24 0.294 0.157 24 0.069 0.096 24 0.304 0.472 24 0.100 0.225 19 0.046 0.066 22 0.161 0.254 265

0.008 0.214 0.271 13 0.085 0.065 24 0.119 0.173 23 0.014 0.071 22 0.053 0.042 23 0.276 0.372 22 0.086 0.105 24 0.187 0.119 24 0.026 0.062 24 0.132 0.289 24 0.047 0.131 19 0.042 0.075 19 0.086 0.163 262

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

Jervois (Sth)

Statistic

19

Head above drain (m)

Acidity

Measured

0.5 500 0

−0.5 Aug−2011

0 Feb−2012

Nov−2011

1000

10000

Pump flow (kL)

Acidity

Pump Flow

500

5000

0 Aug−2011

0 Feb−2012

Nov−2011

Acidity (mg/L as CaCO3

1000

1

Acidity (mg/L as CaCO3

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

Fig. 8. Measured groundwater level (head above drain), and acidity, in response to irrigation in piezometer 5B at Long Flat (top), and acidity and pump flow at the Long Flat drainage discharge point to the River Murray (bottom) (see Fig. 1 for locations). The blue dotted lines indicate irrigation events and the red dashed lines indicate rainfall events (N5 mm).

acidity (TAA and RA, Fig. 4) that could be readily mobilised upon rewetting. Displacement of a significant portion of the nonacidic/base cations (Ca2+, Mg2+, Na+, K+) on the soil exchange sites (X(s)) with acidic cations (H+, Al3+) also occurred (Fig. 5), likely via the following process (shown for Al uncomplexed and hydrolysed species, from Essington, 2004): 2þ

2þ

þ

þ

3þ

2þ

þ

ð2Þ

fCa ; Mg ; K ; Na gXðsÞ þ fAl ; AlOH ; AlðOHÞ2 gðaqÞ 3þ

→fAl ; AlOH

2þ

þ ; AlðOHÞ2 gXðsÞ

2þ

2þ

þ

þ

þ fCa ; Mg ; K ; Na gðaqÞ :

Acidic drain waters were not observed in a pre- (2005–2007) drought study in the LMRIA (EPA, 2008 and unpublished data).

Despite our Hydrus 2D model having some obvious limitations (e.g. inability to represent the deep soil desiccation cracks that progressively developed during the drought period), it adequately represented the observed groundwater decline and soil desaturation during the drought. This suggests that these models could be useful for risk assessment in other river–floodplain systems. The cracks that developed would have allowed oxygen to penetrate deeper into the profile and through the sides of clay peds, affecting the rate and amount of pyrite oxidation (Bronswijk et al., 1993). The under representation of the spring rise in groundwater level (Fig. 9) may relate to preferential groundwater flow

Table 3 Average acidity, drainage volumes and acidity loads discharged to the lower River Murray from the LMRIA (major irrigation areas where loggers installed). Data is separated into the irrigation season (1 September–30 April) and non-irrigation season (1 May–31 August). The acidity in tonnes as CaCO3 is approximately equivalent to tonnes sulphuric acid (H2SO4). Site (area)

Season

Acidity average (mg/L as CaCO3)

Drainage volume average (ML/day)

Acidity load average (t/day)

Load average per unit area (t/ha/day)

Jervois Nth (540 ha) Jervois Sth (950 ha) Long Flat (129 ha) Monteith (386 ha) Pompoota (160 ha) River Glen (163 ha) Toora (143 ha) Woods Point (262 ha) All areas average

Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation Irrigation Non-irrigation

113 304 205 289 267 214 163 257 226 453 104 193 154 272 66 151 162 267

2.38 3.59 5.45 10.91 1.69 1.75 2.98 5.98 0.41 0.90 1.17 1.94 0.62 0.67 1.67 1.60 2.05 3.42

0.62 1.09 1.19 3.27 0.38 0.44 0.53 1.70 0.11 0.33 0.24 0.38 0.09 0.29 0.15 0.25 0.41 0.97

0.0012 0.0020 0.0013 0.0034 0.0030 0.0034 0.0014 0.0044 0.0007 0.0021 0.0015 0.0023 0.0007 0.0020 0.0006 0.0010 0.0013 0.0026

20

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

Groundwater level (m bgl)

−0.5

Measured Mob032(01A) groundwater level Modelled groundwater level

−1

−1.5

−2

−2.5

Jan−06

Jan−07

Jan−08

Jan−09

Jan−10

Fig. 9. Measured (at the Mob032/01A site, 63 m from river) and modelled (Hydrus 2D, Long Flat irrigation area at same distance from the river) groundwater levels over a 5 year period (November 2006 to November 2011).

through deep cracks in the soil profile that were not accounted for in our model. In the future it would be useful to extend the Hydrus 2D model to include dual porosity (to represent cracks), changing permeability as a function of crack formation and healing, and key geochemical components (e.g. pyrite oxidation, cation exchange) but further process-level research would be required to satisfactorily parameterise the model. 4.2. Acidity transport processes The LMRIA drainage channels were mostly dry during the drought and the acidity was retained within the unsaturated

floodplain soil. In the late 2010 to early 2011, local recharge and rising groundwater levels (Fig. 3) were mixed through the acidic soil profile. The pH of the rising groundwater was likely buffered to the acidic soil pH (Simpson et al., 2010), and began to be mobilised to drainage channels. The first rising (“first flush”) of groundwater levels through an acidic profile has been shown to mobilise large amounts of acidity and dissolve secondary sulfate minerals in areas with acid mine drainage (Gzyl and Banks, 2007). Our measurements indicate that irrigation transports acidity in soil and groundwater to drainage channels but average acidity loads and drainage volumes were greater in the non-irrigation winter season when rainfall was highest. This pattern differs from previous nutrient load findings (Mosley and Fleming, 2010) and likely reflects the more limited post-drought irrigation and deeper subsoil source of acidity relative to more surficial nutrient loads associated with pre-drought dairy farming. Load estimates were consistent with previous estimates of acid sulfate soil drainage in other locations (Cook et al., 2000b; Wilson et al., 1999). Drainage water quality was quite variable on spatial and temporal time scales which may relate to differences in irrigation and drainage regimes (as noted above for Mobilong site) but also variable soil and regional groundwater hydrogeology. Saturated hydraulic conductivity has been noted to be very variable in acid sulfate soils (Johnston et al., 2009). This is likely also the case in the LMRIA and further research on soil physical properties and the role of preferential flow paths in acidity transport would be beneficial. Groundwater quality (salinity, pH and acidity) slowly improved over a period of two years. These changes were likely due to the gradual flushing and export of groundwater acidity during irrigation and rainfall (Rassam and Cook, 2002).

Modelled groundwater levels (GWL) Pre-drought GWL 1 November 2005

River level stabilisation & irrigation Drought GWL (lowest river level) 1 May 2009 Drought GWL (lowest) 16 July 2010

Vertical Scale, 1m

Modelled soil moisture (16 July 2010)

0.6 0.5 0.4 0.3 0.2

Fig. 10. Modelled (Hydrus 2D) groundwater level (top) at the Long Flat irrigation area at the beginning (1 November 2005) and end of the drought period (lowest river level in 1 May 2009 and lowest groundwater level on 16 July 2010). A management scenario (river level stabilisation at 0.5 m AHD and five irrigations applied, see text for further details) is also shown. Soil moisture is also shown (bottom) at the date the lowest groundwater levels were recorded, 0.6 represents fully saturated conditions (see θs in Table 1) with less saturated conditions indicating where oxygen would be present for pyrite oxidation. The vertical dimension has been stretched for ease of visualisation.

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

It is also likely due to flushing of acidity from the surfaces of soil peds and cracks which gradually reduces over time. The greatest improvement in groundwater quality at Long Flat appears to be occurring nearest the river (see 7 B and C piezometers, Fig. 6). This suggests that lateral transport of alkalinity from the river water to the floodplain may be neutralising some of the soil and groundwater acidity. There is also more ANC in the soil at this site (Fig. 4) which may have dissolved and neutralised some acidity. Sulfate reduction processes could also eventually reverse the acidification over time, particularly now that groundwater pH values are increasing above 4.5 (non-limiting to sulfate-reducing bacteria, Jong and Parry, 2006). Importantly, our Hydrus-2D modelling results did not indicate seepage of groundwater acidity back to the river when river levels were at their lowest (Fig. 10). However, this process could present a risk if a “losing” groundwater system switched to a “gaining” system following soil and groundwater acidification. 4.3. Management implications Our study highlights the vulnerability of floodplains to acidification driven by declining surface–ground water levels during droughts. The acidification in the LMRIA was severe and has now persisted for several years following the return of water levels to pre-drought levels. This discharge of the acid drainage poses risks to an important surface water body. Given these issues and difficulties in treating acidity located 1–3 m below the ground, it is important that prevention of similar events occurs in the future via local and basin scale management options. At the local scale, creating more variable river levels to prevent large-scale pyrite accumulation in the floodplain soil would be beneficial. The application of irrigation (even if limited due to reduced allocations) and environmental water during the drought would also be beneficial by maintaining a greater degree of sub-soil saturation and lessening the degree of acidification. Anecdotal evidence suggests several of the non-acidic areas (see Fig. 1) related to sites where limited water was applied during the drought. While there were practical difficulties (low river level and infrastructure limitations provided insufficient head for flood irrigation), solutions such as building of coffer dams and pumping were successfully used by some irrigators. The 2-D model was used to assess a management scenario where river levels were stabilised at 0.5 m AHD (relating to a management target to maintain water levels N0.4 m AHD immediately downstream in the Lower Lakes 90% of the time, MDBA, 2010) and five irrigations (1.5 ML/ha, 150 mm water depth per watering, Mosley and Fleming, 2009) applied at monthly intervals from Oct–Feb. The number of irrigations was chosen to represent the worst case irrigation/environmental water allocations during the drought (32% of full allocation occurred in 2008–2009, a full allocation typically enables 16–18 irrigations, Mosley and Fleming, 2009). The groundwater level under the floodplain in this scenario remained at a much higher level (Fig. 10), which would have resulted in much less pyrite oxidation. Based on these results, maintaining environmental water allocations to irrigators during drought and providing government support for emergency irrigation measures should be considered during future

21

events. To better inform future management options it would be advantageous to undertake further research and modelling on subsurface hydrological processes (in particular the influence of macropores/preferential flow paths), and potentially extend the 2D model to 3D to enable simultaneous simulation of river level decline with irrigation and lateral drainage processes. The basin-scale drivers of the acidification events in the LMRIA were a drought in a river basin with a very high level of water extraction (on average approximately 50% of available water is diverted annually for use, predominantly for irrigated agriculture, CSIRO, 2008). The scarcity of environmental water availability, coupled with the downstream location of the lower River Murray in the river basin, led to an inability to maintain surface water levels during drought. In the next few decades, drought and extreme low flow periods will likely become more frequent and intense in the Lower Murray (13% decrease in median flows is predicted by 2030, CSIRO, 2008) and globally (Dai, 2013) due to climate change. Hence, events such as those described in this paper will likely occur with more frequency. However, a new whole of Murray–Darling Basin Plan is being implemented to reduce irrigation water extractions and increase provision of water for the environment (MDBA, 2010). This should reduce the likelihood of similar events reoccurring in the LMRIA floodplains. 5. Conclusions Our study elucidated for the first time the combined hydro-chemical and pedological processes that led to severe and sustained acidification impacts on a river floodplain post-drought. Multiple lines of evidence (soil, groundwater, drainage water data; modelling) indicated that the key driver of the observed acidification was the decrease in connected river–groundwater levels during the 2007–2010 drought. Soil acid-base accounting and cation exchange analyses indicated that a previously submerged sulfidic (pH N 4) subsoil layer had dried, oxidised and acidified. The depth of the acidic layer (approximately 1–2.5 m bgl) corresponded to the measured decline in the floodplain groundwater table. Post-drought, the return of water levels mobilised soil acidity to the groundwater resulting in low pH, high acidity and soluble metal concentrations. Acidity in the groundwater was exported to the drainage channels, during irrigation but on average the loads were greater in the non-irrigation (winter rainfall) season. The LMRIA drainage water is pumped to the Murray River with a low pH and high soluble metal (Al, Co, Mn, Fe, Mn, Ni, and Zn) concentrations, exceeding guidelines for ecosystem protection. Some gradual improvements in groundwater and drainage water quality have been noted (increased pH, appearance of net alkalinity, decreased acidity and salinity) but the soil and water acidification has persisted for over two years post-drought. Hydrus-2D modelling provided further evidence to support the hypothesis that declining river levels led to groundwater declines and drying of the floodplain and that maintaining higher river levels and applying limited irrigation/environmental water may have prevented much of the problem. Additional assessment and model simulation of floodplain geochemistry and irrigation application are warranted to improve future management of the LMRIA and other floodplains. The risks that have emerged in the Lower Murray River floodplains could

22

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23

occur in other areas vulnerable to river level decline and should be carefully considered by environmental managers of those systems. Acknowledgements The assistance of EPA field staff in sample collection and analysis is gratefully acknowledged. The part funding contribution of the Murray–Darling Basin Authority is also gratefully acknowledged as is the assistance of Rob Kingham. We thank various LMRIA irrigators who provided access to the sample sites and field observations that assisted our understanding of this issue. We thank the two anonymous reviewers for their excellent and constructive comments that enabled us to improve the manuscript. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx.doi.org/10.1016/j.jconhyd.2014.03.003. References Ahern, C.R., McElnea, A.E., Sullivan, L.A., 2004. Acid Sulfate Soils Laboratory Methods Guidelines. Queensland Department for Natural Resources Mines and Energy, Australia (Available at: www.nrm.qld.gov.au/land/ ass/pdfs/lmg.pdf). ANZECC, 2000. Australian and New Zealand Guidelines for Fresh and Marine Water Quality. Australian and New Zealand Environment and Conservation Council and Agriculture and Resource Management Council of Australia and New. Zealand, Canberra. APHA, 2005. Standard Methods for the Examination of Water and Wastewater, 21st edn. American Public Health Association, American Water Works Association and Water Environment Federation, Washington, DC. Appleyard, S., Cook, T., 2009. Reassessing the management of groundwater use from sandy aquifers with population increase and climate change: acidification and calcium depletion exacerbated by drought, groundwater withdrawal and land use practices on the Gnangara Mound, Western Australia. Hydrogeol. J. 17, 579–588. ASCE, 2005. ASCE Standardized Reference Evapotranspiration Equation. American Society of Civil Engineers (260 pp.). Banks, E.W., Simmons, C., Love, A., Shand, P., 2011. Assessing spatial and temporal connectivity between surface water and groundwater in a regional catchment: Implications for regional scale water quantity and quality. J. Hydrol. 404, 30–49. Barnett, S., Cresswell, D., Marsden, Z., Yan, W., 2003. Regional Salt and Water Balances for the Lower Murray in South Australia. Department of Water, Land and Biodiversity (2003/27. Available at: www.waterconnect.sa. gov.au/Content/Publications/DEWNR/ki_dwlbc_report_2003_27.pdf). Boman, A., Fröjdö, S., Backlund, K., Åström, M., 2010. Impact of isostatic land uplift and artificial drainage on oxidation of brackish-water sediments rich in metastable iron sulfide. Geochim. Cosmochim. Acta 74, 1268–1281. Boulton, A.J., Findlay, S., Marmonier, P., Stanley, E.H., Valett, H.M., 1998. The functional significance of the hyporheic zone in streams and rivers. Ann. Rev. Ecol. Evol. Syst. 29, 59–81. Bourg, A.C.M., Bertin, C., 1993. Biogeochemical processes during the infiltration of river water into an alluvial aquifer. Environ. Sci. Technol. 27, 661–666. Bronswijk, J.J.E., Nugroho, K., Aribawa, I.B., Groenenberg, J.E., Ritsema, C.J., 1993. Modeling of oxygen transport and pyrite oxidation in acid sulphate soils. J. Environ. Qual. 22, 544–554. Cook, F.J., Hicks, W., Gardner, E.A., Carlin, G.D., Froggatt, D.W., 2000a. Export of acidity in drainage water from acid sulphate soils. Mar. Pollut. Bull. 41, 319–326. Cook, F.J., Rassam, D.W., Blunden, B., Gardner, E.A., Carlin, G.D., 2000b. Irrigation and drainage effects on acidity export from acid sulfate soils. In: Slavich, P. (Ed.), Proceedings of Workshop on Remediation and Assessment of Broadacre Acid Sulfate Soils. ASSMAC, Wollonbar, NSW, pp. 78–87. CSIRO, 2008. Water availability in the Murray–Darling Basin: a report to the Australian Government from the CSIRO Murray–Darling Basin Sustainable Yields Project. CSIRO, Australia (67 pp. Available at http://www.

csiro.au/Organisation-Structure/Flagships/Water-for-a-Healthy-CountryFlagship/Sustainable-Yields-Projects/MDBSY.aspx). Dai, A., 2013. Increased drought under global warming in observations and models. Nat. Clim. Chang. 3, 52–58. Dent, D.L., Pons, L.J., 1995. A world perspective on acid sulfate soils. Geoderma 67, 263–276. EPA, 2008. Lower Murray Reclaimed Irrigation Area (LMRIA) Rehabilitation Project Environmental Monitoring Report: Phase 1. South Australian Environment Protection Authority, Adelaide (Available at: www.epa.sa. gov.au/xstd_files/Water/Report/lm_irrigation.pdf). Essington, M.E., 2004. Soil and Water Chemistry: An Integrative Approach. CRC Press (534 pp.). Fitzpatrick, R.W., Shand, P., Merry, R.H., 2009. Acid sulfate soils. In: Jennings, J.T. (Ed.), Natural History of the Riverland and Murraylands. Royal Society of South Australia (Inc.), Adelaide, South Australia, pp. 65–111. Fitzpatrick, R.W., Raven, M.D., Self, P.G., Shand, P., Mosley, L., Simpson, S., Baker, A., 2012. Occurrence, genesis and environmental significance of schwertmannite in re-flooded acid sulfate soils in the Lower Murray Reclaimed Irrigation Area. Proceedings Combined Australian Regolith Geoscientists Association and Australian Clay Minerals Society Conference: Mildura 7–10 February, pp. 147–152 (Available at: http://www. smectech.com.au/ACMS/ACMS_Conferences/ACMS22/Proceedings/Papers_ by_Author/Papers_Authors.htm). Fitzpatrick, R.W, Shand, P, Simpson, S., Grocke, S, Raven, M., Baker, A.K.M., Mosley, L.M., Self, P.G, 2013. Assessment of re-flooded Acid Sulfate Soil environments at Long Flat, Jervois, Toora and Pompoota in the Lower Murray Reclaimed Irrigation Area (LMRIA). CSIRO Land and Water Science Report. Gee, G.W., Bauder, J.W., 1986. Particle size analysis, In: Klute, A. (Ed.), Methods of Soil Analysis. Part 1, 2nd ed. Agron. Monogr., 9. American Soc. of Agronomy Inc., Soil Sci. Soc. America Inc., Madison WI, pp. 383–411. Gzyl, G., Banks, D., 2007. Verification of the “first flush” phenomenon in mine water from coal mines in the Upper Silesian Coal Basin, Poland. J. Contam. Hydrol. 92, 66–86. Johnston, S.G., Hirst, P., Slavich, P.G., Bush, R.T., Aaso, T., 2009. Saturated hydraulic conductivity of sulfuric horizons in coastal floodplain acid sulfate soils: variability and implications. Geoderma 151, 387–394. Jong, T., Parry, D.L., 2006. Microbial sulfate reduction under sequentially acidic conditions in an upflow anaerobic packed bed bioreactor. Water Res. 40, 2561–2571. Lamontagne, S., Leaney, R., Herczeg, A., 2005. Groundwater–surface water interactions in a large semi-arid floodplain: implications for salinity management. Hydrol. Process. 19, 3063–3080. Lucassen, E.C.H.E.T., Smolders, A.J., Roelofs, J.G.M., 2002. Potential sensitivity of mires to drought, acidification and mobilisation of heavy metals: the sediment s/(Ca + Mg) ratio as diagnostic tool. Environ. Pollut. 120, 635–646. Macdonald, B.C.T., White, I., Åström, M.E., Keene, A.F., Melville, M.D., Reynolds, J.K., 2007. Discharge of weathering products from acid sulfate soils after a rainfall event, Tweed River, eastern Australia. Appl. Geochem. 22, 2695–2705. Massmann, G., Tichomirowa, M., Merz, C., Pekdeger, A., 2003. Sulfide oxidation and sulfate reduction in a shallow groundwater system (Oderbruch aquifer, Germany). J. Hydrol. 278, 231–243. Matsunaga, T., Karametaxas, G., von Gunten, H.R., Lichtner, P.C., 1993. Redox chemistry of iron and manganese minerals in river-recharged aquifers: a model interpretation of a column experiment. Geochim. Cosmochim. Acta 57, 1691–1704. MDBA, 2010. Guide to the proposed basin plan. Overview, vol. 1. Murray Darling Basin Authority, Canberra (publication no. 60/10. Available at: http://www.mdba.gov.au/kid/guide/). Mosley, L.M., Fleming, N., 2009. Reductions in water use following rehabilitation of a flood-irrigated area on the Murray River in South Australia. Agric. Water Manag. 96, 1679–1682. Mosley, L.M., Fleming, N., 2010. Pollutant Loads returned to the Lower Murray River from flood-irrigated agriculture. Water Air Soil Pollut. 211, 475–487. Mosley, L.M., Zammit, B., Leyden, E., Heneker, T.M., Hipsey, M.R., Skinner, D., Aldridge, K.T., 2012. The impact of extreme low flows on the water quality of the Lower Murray River and Lakes (South Australia). Water Resour. Manag. 26, 3923–3946. Mosley, L.M., Zammit, B., Jolley, A., Barnett, L., Fitzpatrick, R., 2013. Monitoring and assessment of surface water acidification following rewetting of oxidised acid sulphate soils. Environ. Monitor. Assess. http://dx.doi.org/10.1007/s10661-013-3350-9. Österholm, P., Åström, M., 2002. Spatial trends and losses of major and trace elements in agricultural acid sulphate soils distributed in the artificially drained Rintala area, W. Finland. Appl. Geochem. 17, 1209–1218.

L.M. Mosley et al. / Journal of Contaminant Hydrology 161 (2014) 10–23 Pons, L.J., 1973. Outline of the genesis, characteristics, classification and improvement of acid sulphate soils. In: Dost, H. (Ed.), Proceedings of the 1972 (Wageningen, Netherlands) International Acid Sulphate Soils Symposium, Volume 1, 18. International Land Reclamation Institute Publication, pp. 3–27 (Wageningen, Netherlands). Rassam, D.W., Cook, F.J., 2002. Numerical simulations of water flow and solute transport applied to acid sulfate soils. J. Irrig. Drain. Eng. 128, 107–115. Rawls, W.J., Brakensiek, D.L., 1985. Prediction of soil water properties for hydrologic modelling. In: Jones, E., Ward, T.J. (Eds.), Watershed Management in the Eighties. Proc. Symp. Irrig. Drain. Div. ASCE, Denver, CO, pp. 293–299 (ASCE, New York). Rayment, G.E., Higginson, F.R., 1992. Australian Laboratory Handbook of Soil and Water Chemical Methods. Inkata Press. Sammut, J., White, I., Melville, M.D., 1996. Acidification of an estuarine tributary in eastern Australia due to drainage of acid sulfate soils. Mar. Freshw. Res. 47, 669–684. Simpson, S.L., Fitzpatrick, R.W., Shand, P., Angel, B.M., Spadaro, D.A., Mosley, L., 2010. Climate-driven mobilisation of acid and metals from acid sulfate soils. Mar. Freshw. Res. 61, 129–138. Šimůnek, J., van Genuchten, M.Th, Šejna, M., 2012. The Hydrus software package for simulating two- and three dimensional movement of water, heat, and multiple solutes in variably-saturated media. Technical Manual, Version 2.0. PC Progress, Prague, Czech Republic (258 pp.). Sophoceleus, M., 2002. Interactions between groundwater and surface water: the state of the science. Hydrogeol. J. 10, 52–67.

23

Taylor, J.K., Poole, H.G., 1931. A soil survey of the swamps of the Lower Murray River. CSIRO Bulletin, No. 51 CSIRO, Melbourne, Australia. Van den Berg, G.A., Loch, J.P.G., Winkels, H.J., 1998. Effect of fluctuating hydrological conditions on the mobility of heavy metals in soils of a freshwater estuary in the Netherlands. Water Air Soil Pollut. 102, 377–388. von Gunten, H.R., Karametaxas, G., Krahebnbuhl, U., Kuslys, M., Giovanoli, R., Hoehn, E., Keil, R., 1991. Seasonal biogeochemical cycles in riverborn groundwater. Geochim. Cosmochim. Acta 55, 3597–3609. von Gunten, H.R., Karametaxas, G., Keil, R., 1994. Chemical processes in infiltrated riverbed sediment. Environ. Sci. Technol. 28, 2087–2093. Wilson, B.P., White, I., Melville, M.D., 1999. Floodplain hydrology, acid discharge and change in water quality associated with a drained acid sulfate soil. Mar. Freshw. Res. 50, 149–157. Winter, T.C., Harvey, J.W., Franke, O.L., Alley, W.M., 1998. Ground water and surface water — a single resource. Geological Survey Circular 1139 United States Geological Survey, Denver, Colorado (Available at: http:// pubs.usgs.gov/circ/circ1139/pdf/circ1139.pdf). Wondzell, S., Swanson, F.J., 1996a. Seasonal and storm dynamics of the hyporheic zone of a 4th-order mountain stream. I: hydrologic processes. J. N. Am. Benthol. Soc. 15, 3–19. Wondzell, S., Swanson, F.J., 1996b. Seasonal and storm dynamics of the hyporheic zone of a 4th-order mountain stream. II: nitrogen cycling. J. N. Am. Benthol. Soc. 15, 20–34.