- Email: [email protected]
Anthropogenic forcing of estuarine hypoxic events in sub-tropical catchments: Landscape drivers and biogeochemical processes
Science of the Total Environment 409 (2011) 5368–5375
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Anthropogenic forcing of estuarine hypoxic events in sub-tropical catchments: Landscape drivers and biogeochemical processes Vanessa N.L. Wong a, b,⁎, Scott G. Johnston a, Edward D. Burton a, Richard T. Bush a, Leigh A. Sullivan a, Peter G. Slavich c a
Southern Cross GeoScience, Southern Cross University, PO Box 157, Lismore New South Wales 2480, Australia School of Geography and Environmental Science, Monash University, Victoria 3800, Australia Wollongbar Primary Industries Institute, New South Wales Department of Trade and Investment, Regional Infrastructure and Services, 1243 Bruxner Highway, Wollongbar NSW 2477, Australia
b c
a r t i c l e
i n f o
Article history: Received 17 May 2011 Received in revised form 23 August 2011 Accepted 25 August 2011 Available online 21 September 2011 Keywords: Estuary Hypoxia Anoxia Deoxygenation Floodplain drainage Labile carbon
a b s t r a c t Episodic hypoxic events can occur following summer floods in sub-tropical estuaries of eastern Australia. These events can cause deoxygenation of waterways and extensive fish mortality. Here, we present a conceptual model that links key landscape drivers and biogeochemical processes which contribute to post-flood hypoxic events. The model provides a framework for examining the nature of anthropogenic forcing. Modification of estuarine floodplain surface hydrology through the construction of extensive drainage networks emerges as a major contributing factor to increasing the frequency, magnitude and duration of hypoxic events. Forcing occurs in two main ways. Firstly, artificial drainage of backswamp wetlands initiates drier conditions which cause a shift in vegetation assemblages from wetland-dominant species to drylanddominant species. These species, which currently dominate the floodplain, are largely intolerant of inundation and provide abundant labile substrate for decomposition following flood events. Decomposition of this labile carbon pool consumes oxygen in the overlying floodwaters, and results in anoxic conditions and waters with excess deoxygenation potential (DOP). Carbon metabolism can be strongly coupled with microbiallymediated reduction of accumulated Fe and Mn oxides, phases which are common on these coastal floodplain landscapes. Secondly, artificial drainage enhances discharge rates during the flood recession phase. Drains transport deoxygenated high DOP floodwaters rapidly from backswamp wetlands to the main river channel to further consume oxygen. This process effectively displaces the natural carbon metabolism processes from floodplain wetlands to the main channel. Management options to reduce the impacts of post-flood hypoxia include i) remodifying drainage on the floodplain to promote wetter conditions, thereby shifting vegetation assemblages towards inundation-tolerant species, and ii) strategic retention of floodwaters in the backswamp wetlands to reduce the volume and rate during the critical post-flood recession phase. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Coastal and estuarine hypoxia has a significant effect on aquatic organisms and ecosystem function (Conley et al., 2007; Vaquer-Sunyer and Duarte, 2008). Hypoxia is defined as low concentrations of dissolved oxygen (i.e. b30% saturation; b2 mg L − 1; b62.5 mM L − 1) and can cause mortality, reduced growth rates and altered distributions of aquatic life (Breitburg, 2002). Dissolved oxygen concentrations in surface waters are determined by inputs from the atmosphere and photosynthesis, and outputs, which are dominated by respiration and other processes that consume oxygen. Globally, the frequency, magnitude and spatial distribution of anthropogenically-induced
⁎ Corresponding author at: School of Geography and Environmental Science, Monash University, Victoria 2800, Australia. Tel.: +61 3 9905 2930; fax: +61 3 9905 2948. E-mail address: [email protected] (V.N.L. Wong). 0048-9697/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2011.08.065
coastal hypoxic events are increasing (Diaz, 2001). Such events can occur episodically, periodically or seasonally, impacting on benthic and pelagic fauna, nutrient cycling and other biogeochemical processes (Diaz and Rosenberg, 2008). The estuaries and waterways of sub-tropical eastern Australia are prone to episodic hypoxic events following summer floods. Flood events increase connectivity between floodplains and river channels to deliver pulses of sediments, nutrients and organic matter (Amoros and Bornette, 2002; Junk et al., 1989). These events result in a flush of microbial activity which can consume available oxygen in surface waters, thereby causing hypoxia. Large-scale fish kill events have occurred in catchments in the sub-tropical region of eastern Australia during summer as a result of post-flood hypoxia, resulting in the closure of estuaries to commercial and recreational fishing activities for weeks to several months (Eyre, 1997; Johnston et al., 2003; Wong et al., 2010). Similar hypoxic events and fish kills have also been recorded in tropical coastal (Townsend and Edwards, 2003; Townsend et al., 1992) and
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
5369
semi-arid inland (Hladyz et al., 2011; Howitt et al., 2007) river systems following high-flow events. This study presents a conceptual model that incorporates the key processes, major drivers and temporal dynamics of estuarine hypoxic events on the Holocene coastal floodplains of eastern Australia following floods. The conceptual model is used to examine how anthropogenic modification of floodplain hydrology has increased post-flood floodplain-estuary connectivity and altered the nature of carbon pools and fluxes on the floodplains. The findings provide the basis for discussion of possible management and mitigation measures.
Table 2 Flood peaks during the 2008 and 2009 flood events (1:10 ARI) at selected sites on the Richmond River floodplain.
2. Methods
the period immediately following the flood peak during which water levels recede to baseflow levels. Surface water chemistry was sourced from drains in the backswamp wetlands of the Clarence River floodplain and the main Richmond River channel. Methods are described in detail in Johnston et al. (2003) and Wong et al. (2010). Briefly, unfiltered samples were analysed for chemical oxidation demand (COD) by dichromate oxidation (COD APHA 5220; APHA, 2005). Samples filtered through 0.45 μm filters were analysed for dissolved organic carbon (DOC) and dissolved Fe and speciation. Fe speciation was determined using 1,10-phenanthroline (APHA 3500; APHA, 2005). Fe3+ was determined from the difference between total Fe and Fe2+ following reduction with hydroxylammonium chloride. Dissolved organic carbon (DOC) was analysed by wet oxidation with an Aurora 1030 Total Carbon/Total Organic Carbon Analyser (APHA 5310 C; APHA, 2005). Estimates of the daily flux of deoxygenation potential (DOP) from drained wetlands to the main channel were determined from measured drain discharge rates and geochemical data according to Eq. (2).
Data used in this analysis are derived from the Richmond and Clarence River catchments. The Richmond and Clarence River catchments are located in the sub-tropical region of eastern Australia and discharge at Ballina (153.56°E 28.84°S) and Yamba (153.36°E 29.43°S), respectively. The region experiences wet, humid summers with rainfall falling predominantly in summer (November–April) and mild, dry winters (May– October). Mean annual rainfall is 1778 mm for Ballina and 1456 mm for Yamba. Mean maximum temperatures are 28.2 °C at Ballina and 26.7 °C at Yamba in January and mean minimum temperatures are 8.5 °C at Ballina and 9.7 °C at Yamba in July (BOM, 2011). The conceptual model is developed based on observations from four flood events. Raw data is sourced and compiled from Johnston et al. (2003) and Wong et al. (2010). Overbank flooding occurred in the Clarence River in February and March 2001, resulting in floods of approximately 1:10 and 1:14 year average return interval (ARI), respectively (Table 1). Two floods of magnitude 1:10 ARI occurred in the Richmond River in January 2008 and May 2009. The May 2009 flood was of a similar magnitude, with comparable rainfall and water level peaks in similar parts of the catchment (Table 2). Data from the 2009 flood event was collected and analysed using the same techniques and in the same locations as described in Wong et al. (2010). Briefly, surface water samples were collected for approximately one month following the flood peak. Samples were collected in the main channel directly downstream from the main backswamp wetland basins and at the junction where the backswamps discharge into the main channel. Backswamp wetland basins are low-lying depressions commonly found on coastal floodplains with a seasonal or permanent watertable at or above the surface (Naylor et al., 1998). Water level data and discharge rates from two backswamp wetlands on the Clarence River floodplain (Everlasting Swamp and Shark Creek; Johnston et al., 2003), and one backswamp wetland on the Richmond River floodplain (Tuckean Swamp; Wong et al., 2010), were used to develop the model. These data were obtained from data loggers located within the backswamp wetlands. The normalised discharge rate was calculated according to Eq. (1). Q norm ¼ Q t =Q tot
ð1Þ
where Qnorm is the normalised discharge, Qt is the cumulative discharge (m3 s− 1) at time t during the flood recession phase and Qtot is the total discharge that occurred during approximately 30 days of the flood recession phase following the flood peak. The flood recession phase is
Table 1 Flood peaks during the February (1:10 ARI) and March (1:14 ARI) 2001 flood events at selected sites on the Clarence River floodplain. Site
February 2001 river level peak (m AHD)
March 2001 river level peak (m AHD)
Grafton Ulmarra Maclean
6.7 5.3 2.7
7.7 6.1 3.2
Site
2008 river level peak (m AHD)
2009 river level peak (m AHD)
Coraki Bungawalbyn Junction Woodburn
5.8 4.9 3.1
6.6 5.6 4.0
−1 F ðCODÞt kgday ¼ Q t Ct
ð2Þ
where F(COD)t is the daily flux of chemical oxygen demand (COD), Q is the discharge volume (m3 s − 1) and C is the concentration of the COD (mg O2 L − 1) at time t. For further details, see Johnston et al. (2003). 3. Results and discussion 3.1. Coastal floodplain morphology While the Holocene coastal floodplains of sub-tropical eastern Australia are at an advanced stage of evolution, contemporary sediment supply is constrained and the floodplains are still in the process of actively infilling (Lin and Melville, 1993). This leads to distinctive geomorphic features with floodplains characterised by large areas of low-elevation backswamp wetland basins which were previously barrier estuary embayments (Wilson, 2005). The backswamp wetland basins are occluded from the main river channel by extensive natural levee systems. Levees typically range in height up to approximately 4.5 m Australian Height Datum (AHD, where 0 m AHD ≈ sea level). An example of this morphology is provided in Fig. 1a. When intact, these natural barriers typically impound surface waters within backswamp wetland basins for periods of up to 4 months following flood events (White et al., 1997). 3.2. Modification of floodplain surface hydrology The construction of extensive drainage networks for flood mitigation and agricultural expansion on coastal floodplains began around the 1900s and peaked in the 1950s–70s (Pressey, 1982). These drainage networks were designed to remove floodwaters rapidly and efficiently from the floodplains into the main river channels. For example, surface hydrology on the lower Richmond River floodplain (1070 km2) is controlled by N400,000 km of open ditch drains, including farm drains which feed into secondary drains (Tulau, 1999). These drainage networks facilitate rapid
5370
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
Fig. 1. a) The Rocky Mouth Creek backswamp wetland on the Richmond River floodplain provides a typical example of backswamp-levee toe-levee morphology commonly found on sub-tropical floodplains in eastern Australia (Wong et al., 2010); (AHD is Australian Height Datum where 0 m AHD ≈ sea level) and b) increasing the connectivity between the floodplain to the main river channel by dissecting natural levees with artificial drains.
decreases in water levels following rainfall events, which fall to the mean low water (MLW) level and enhance the removal of surface and shallow groundwater (Johnston et al., 2004; Wilson et al., 1999). Connectivity between the floodplain and main channel has greatly increased as the drains dissect the natural levee system to remove surface water (Fig. 1b). Prior to the implementation of drainage networks, backswamp wetlands were natural water storage basins on the floodplain. Following flood events, recession of floodwaters occurred slowly under the natural drainage conditions due to high channel sinuosity and roughness and low hydraulic gradients (White et al., 1997). When surface water levels decreased to below the minimum height of the natural levees, recession of floodwaters then proceeded very slowly via evapotranspiration, infiltration and lateral leakage (Fig. 2a). Under the current artificially augmented hydrological regime, floodwaters are now removed from the backswamp wetlands at least an order of magnitude faster (Sammut et al., 1996; White et al., 1997). Backswamp water levels decrease rapidly to the MLW level with the current artificial regime. During the 2001 and 2008 flood events, backswamp water levels had decreased to below the height of the levees within 3 to 9 days after the flood peak (Fig. 2b). Rapidly decreasing backswamp water levels on modified floodplains is a function of higher discharge rates compared to unmodified floodplains (Fig. 3a). During the 2001 and 2008 flood events, N50% of backswamp wetland discharge had occurred within 4 to 9 days after the flood peak, and 80% within 8 to 16 days (Fig. 3b).
Floodplains are important overbank organic C stores (Battin et al., 2008). Unmodified wetlands in backswamp basins can be important C sinks on coastal floodplains as a result of relatively slow decomposition rates due to high soil water content and limited oxygen supply (Delaune et al., 1981; Gregorich and Janzen, 2000). However, backswamp wetlands are now substantial net C sources during flood events. Modification of floodplains, rivers and their connectivity has substantially altered fluxes of DOC (Robertson et al., 1999). Extensive drainage networks not only facilitate the rapid removal of surface waters, but also transport of highly labile DOC from floodplains to streams and rivers via drains during high flow events (Vidon et al., 2008; Warrner et al., 2009). During high flow events, fresh particulate organic carbon (POC) is mobilised and hydrolysed rapidly to DOC (Battin et al., 2008). These modified floodplains can act as potential sources of large quantities of organic material during floods (Cuffney, 1988; Valett et al., 2005). 3.3. Altered vegetation assemblages—a source of labile carbon Prior to European settlement of eastern Australia in the 19th century, backswamp wetlands on these coastal floodplains were semipermanently inundated. Vegetation was dominated by wetland species such as reeds (Phragmites australis), rushes (Juncus and Eleocharis spp), Swamp Oak (Casuarina glauca), Swamp Mahogany (Eucalyptus robusta) and Broad-leaved Paperbark (Melaleuca quinquenervia) (Tulau, 1999). Changes in water regime, as a result of changes to the rate, magnitude
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
Fig. 2. a) Conceptual representation of backswamp water levels during the post-flood recession phase in a drained and undrained backswamp wetland and b) backswamp water levels during the post-flood recession phase from Everlasting Swamp (E1, E2), Shark Creek (S1, S2) during the two floods in 2001, and from the Tuckean Swamp (T1) during the flood in 2008. Horizontal lines indicate the range in levee height.
and duration of flooding, significantly alter wetland vegetation assemblages (Barrett et al., 2010). Where drier floodplain conditions dominate, vegetation assemblages can shift from predominantly wetland species to dryland species (Blanch et al., 1999; Pressey, 1982). Extensive anthropogenic modification of the surface hydrology on sub-tropical coastal floodplains has allowed for year-round agriculture and has resulted in the dominance of dryland pasture species such as common couch (Cynodon dactylon), carpet grass (Axonopis fissifolius), kikuyu (Pennisetum clandestium) and water couch (Paspalum distichum) (Tulau, 1999). The alteration of vegetation assemblages to a predominance of dryadapted grass species is of vital importance. Vegetation assemblages that currently dominate on coastal floodplains are largely intolerant of waterlogging and flooding. Therefore, when floodplains are rapidly inundated as a result of flooding, vegetation dies and provides an abundant source of labile C substrate for decomposition. A series of studies has shown that the pasture species that are currently commonly found on these coastal floodplains generally have a faster decomposition rate and higher oxygen demand during decomposition compared to the original, endemic wetland vegetation species (Eyre et al., 2006; Johnston et al., 2005). Rates of decomposition, and hence, oxygen demand, are a function of the lability of organic material rather than its total concentration (Brunke and Gonser, 1997). Therefore, organic material derived from grasses, which have a high cellulose content relative to lignin, is decomposed more rapidly than material derived from wetland plants, which typically have a higher lignin content (Benner et al., 1985; Maccubbin and Hodson, 1980). Simple chain C structures are
5371
Fig. 3. a) Conceptual representation of discharge rates during the post-flood recession phase on a drained and undrained floodplain, and b) normalised cumulative discharge during the post-flood recession phase from Everlasting Swamp (E1, E2), Shark Creek (S1, S2) during the two floods in 2001, and from the Tuckean Swamp (T1) during the flood in 2008.
usually associated with carbohydrates and fresh organic material, and are easily decomposable (Baldock et al., 1997; Skjemstad et al., 1998). Fresh organic material can release highly labile DOC which is consumed rapidly (Baldwin, 1999; Bourbonniere and Creed, 2006). Decomposition of floodplain vegetation consumes oxygen due to microbial respiration and organic C oxidation, as reflected in the strong positive correlation between DOC and COD during several different flood events (Fig. 4). Low molecular weight components of DOC with simple structure are available for decomposition almost immediately and result in a rapid increase in bacterial numbers (McDonald et al., 2007; O'Connell et al., 2000), which can further drive decomposition processes mediated by redox transformations under anaerobic conditions. Similarly, under anoxic conditions, more labile organic material with smaller C:N ratios is decomposed preferentially and more rapidly compared to more recalcitrant organic material (Burdige, 1991). Dissolved organic matter (DOM) exported from floodplains which experience prolonged wetter conditions are generally more aromatic in structure (Wilson and Xenopoulos, 2009). Similarly, DOM sourced from wetland sediments frequently have higher molecular weights, are more aromatic and refractory, and are, therefore, more difficult to decompose (Chin et al., 1998; Petrone et al., 2011). Aromatic structures are generally associated with lignin and dominate in peat soils where there is an accumulation of more recalcitrant organic C (Baldock et al., 1997). Due to their high lignin content, anaerobic decomposition of organic material sourced from sedges, rushes and mangroves can occur an order of magnitude slower compared to decomposition of material sourced from grasses (Benner et al., 1984).
5372
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
oxygen sink (Van Cappellen and Wang, 1996). For example, oxidation of Fe (II) has a high deoxygenation potential and forms reactive Fe (III) minerals (Eq. (8)) which are then available for microbiallymediated anaerobic decomposition (Miles and Brezonik, 1981). 2þ
4Fe
þ
þ O2 þ 10H2 O→4FeðOHÞ3 þ 8H
ð8Þ
3.5. The effects of temperature
Fig. 4. Increasing COD with increasing concentrations of labile organic carbon (i.e. DOC) where Clarence 2001a is sourced from Everlasting Swamp and Shark Creek during the February 2001 flood event, Clarence 2001b is sourced from Everlasting Swamp and Shark Creek during the March 2001 flood event, and Richmond 2008 is sourced from Wardell during the January 2008 flood event.
3.4. Accumulation of alternative terminal electron acceptors Under oxic conditions, decomposition of organic material occurs with O2 as the terminal electron acceptor to produce CO2. In the absence of oxygen, decomposition occurs with alternative terminal electron acceptors via an approximate thermodynamic sequence with increasing energy requirements in the order O2 N NO3− N Mn4+ N Fe3+ N SO42− N CO2 according to Eqs. (3)–(7) where CH2O approximates labile organic material for decomposition (Capone and Kiene, 1988; Ponnamperuma, 1972). −
þ
4NO3 þ CH2 O þ 4H →2N2 þ 5CO2 þ H2 O
ð3Þ
þ
ð4Þ
2þ
2MnO2 þ CH2 O þ 4H →2Mn
þ
2þ
4FeðOHÞ3 þ CH2 O þ 4H →4Fe
2þ
þ
þ CO2 þ 3H2 O
þ CO2 þ 11H2 O
The occurrence and intensity of hypoxic conditions is strongly associated with seasonal temperature. Lower dissolved oxygen concentrations in surface waters are frequently found during warmer months with higher air and water temperatures (e.g. Conley et al., 2007; MacPherson et al., 2007; Townsend, 1999). Surface water temperatures on coastal sub-tropical floodplains in eastern Australia vary by approximately 15– 20 °C (Fig. 5a). Deoxygenation events generally occur during summer, when ambient temperatures and water temperatures are higher. Higher temperatures provide optimal environmental conditions for increased microbial metabolism, enhanced rates of oxygen consumption (Kirschbaum, 1995) and lower dissolved oxygen saturation potential (Stumm and Morgan, 1996). In highly modified drainage canals, seasonal hypoxia during summer months can lead to complete deoxygenation of the water column (Luther et al., 2004). COD in the drains of backswamp wetlands on the Clarence River floodplain is positively correlated with antecedent drain water temperatures (Fig. 5b), most likely as a result of increasing decomposition rates of floodplain-derived organic material. DOC is known to become more labile with higher temperatures as the rate of microbial breakdown of larger, less labile DOC compounds into smaller, more labile compounds increases (Marschner and Kalbitz, 2003). A comparison of a summer and winter flood on the Richmond River estuary provides a clear example of the critical threshold of temperature as a key component of anaerobic metabolism and the development of estuarine hypoxia. Hypoxic conditions established in the Richmond
ð5Þ
SO4 þ 2CH2 O þ 2H →H2 S þ 2CO2 þ 2H2 O
ð6Þ
2CH2 O→CH4 þ CO2
ð7Þ
The coastal floodplains of eastern Australia are frequently underlain by acid sulfate soils (ASS) (Walker, 1972). These soils can develop abundant surface and near-surface accumulations of bioavailable Fe, Mn and SO42−, especially in backswamp wetlands (Dent, 1986; Sullivan and Bush, 2004; Willett et al., 1992). Under anaerobic conditions, these redox sensitive species can be an important source of alternative terminal electron acceptors for decomposition of organic material under anaerobic conditions (Burton et al., 2008; Lovley and Phillips, 1986). Extended inundation of floodplain ASS facilitates microbially-mediated reduction of Fe (III) and Mn (IV) coupled with decomposition of labile organic material. This process results in enhanced mobilisation of Fe (II) and Mn (II) in surface and porewaters (Johnston et al., 2003; van Breemen, 1975). Furthermore, reductive dissolution of poorly crystalline Fe (oxy)hydroxide minerals can release additional sorbed DOM, providing additional substrate for decomposition (Chin et al., 1998). Abiotic processes can further consume oxygen in surface waters. Oxidation of reduced inorganic species such as Fe (II) and Mn (II) and the coupling of the corresponding redox cycles can act as a potential
Fig. 5. a) Seasonal variation in mean daily surface water temperature from the Clarence River in a backswamp wetland drain (backbridge) and the drain's junction with the main river channel (floodgates) and b) COD correlation with antecedent surface water temperature as measured in drains in Everlasting Swamp and Shark Creek following the two flood events in 2001.
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
5373
River estuary during the summer 2008 flood event when surface water temperatures in the main channel were between 23 and 27 °C (Wong et al., 2010). Conversely, the main river channel remained oxic during a similar magnitude flood event in winter 2009 due to lower surface water temperatures (i.e. 16–18 °C), reduced microbial activity and higher dissolved oxygen saturation potential. Temperature also influences the intensity of anoxia and, therefore, the dominant terminal electron acceptor processes that are coupled with anaerobic C metabolism. During both the 2008 and 2009 flood events, dissolved Fe concentrations peaked between 6 and 14 days after the main channel flood peak in the main channel of the Richmond River (Fig. 6). However, while Fe2+ is the dominant Fe species during the post-flood recession period during the summer 2008 flood event (Fig. 6a), Fe 3+ dominates during the post-flood recession during the winter 2009 flood event (Fig. 6b). Similarly, the total dissolved Fe concentration mobilised during the 2008 flood event was more than twice the peak Fe concentration mobilised during the 2009 flood event. These trends suggest that decomposition of organic material is more strongly coupled with microbially-mediated Fe reduction at higher temperatures, such as those occurring during summer flood events. 3.6. Synthesis It is likely that the magnitude, duration and frequency of post-flood estuarine hypoxic events have substantially increased as a result of changes to the surface hydrology and vegetation assemblages on coastal floodplains. The flux of DOP from an individual drained floodplain wetland to the main river channel can be expressed according to Eq. (2) and is a function of a) the rate of drainage discharge from the floodplain, and b) the DOP of floodplain waters. Deoxygenation potential is a complex function of many different factors and is strongly influenced by temperature and the concentration and lability of C. Eq. (2) provides a simple framework to qualitatively evaluate anthropogenic forcing, as changes to either the discharge rate or DOP will shift both the timing and magnitude of the DOP flux (Fig. 7a), which can act to compound or diminish its effects. We estimated a high and low post-flood COD scenario from data collected from two
Fig. 6. Dissolved Fe concentrations and speciation during the flood recession phase at Wardell following a) the summer 2008 flood event, and b) winter 2009 flood event.
Fig. 7. a) Conceptual model of deoxygenation potential flux during the flood recession phase, and b) temporal dynamics of COD in drains at Everlasting Swamp and Shark Creek during the 2001 floods (Johnston et al., 2003). Curves represent a high COD scenario and a low COD scenario.
flood events (Fig. 7b). For example, a high post-flood COD scenario, such as observed in Everlasting Swamp in 2001 (Fig. 7b), is driven by the dominance of dryland pasture species in this backswamp wetland (Johnston et al., 2003). In contrast, the low COD values at Shark Creek (Fig. 7b) are largely the result of M. quinquenervia forest, comprised of highly recalcitrant C (Johnston et al., 2005) covering approximately two thirds of the backswamp wetland. Therefore, a change in land use from native forest to dryland pasture will shift the DOP flux from Scenario 3 in the direction of Scenario 1 (Fig. 7a) due to increased C lability. Conversely, a period of lower temperatures during the flood recession phase, such as during the 2009 flood event, will decrease COD and therefore, the magnitude and duration of the DOP flux. Similarly, lower concentrations of alternative terminal electron acceptors or a decrease in the biomass of pasture species will also shift the DOP from Scenario 3 towards Scenario 1 (Fig. 7b). In sub-tropical estuaries of eastern Australia, the onset of hypoxic conditions in the main estuarine channel following flood events typically occurs approximately 4–6 days following the flood peak (Eyre and Twigg, 1997; Johnston et al., 2003; Wong et al., 2010). Peak COD concentrations in drains usually occur during this period (Fig. 7b) and coincide with high discharge rates (Fig. 3b). Maximum DOP flux coincides with the falling limb of the flood hydrograph, when flow volumes in the main channel are decreasing sharply. Therefore, the capacity of the main channel water column to assimilate high oxygen demand inputs of floodplain-derived surface waters during this period is also decreasing. Transported deoxygenated surface waters can then exert additional oxygen demand to further consume oxygen in the main channel (Lin et al., 2004).
5374
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
3.7. Influence of land management There are a variety of land management actions which may attenuate the impact of the DOP flux scenarios represented in Fig. 7a, thereby reducing the frequency and magnitude of hypoxic events in estuarine waterways. According to Eq. (2), DOP flux can be reduced by 1) retaining a portion of floodwaters in the backswamp wetland basins, or 2) slowing the drainage rate from the floodplain during the flood recession phase. COD flux can be used as a proxy for DOP flux. Analysis of COD flux from drained wetlands during past events relative to time and backswamp wetland water levels provides a useful tool to identify the relative attenuation effect that might be achieved by retaining floodwaters at a given time or water level threshold. Based on this analysis, retention of floodwaters at the onset of main channel hypoxic conditions (i.e. approximately 4–6 days after the flood peak) can retain approximately 65–90% of the COD flux (at Day 5) in the backswamp wetlands (Fig. 8a). The retention of COD flux within this range corresponds with maintaining backswamp water levels at approximately 0.5–3.0 m AHD, which is dependent on the wetland basin and flood event (Fig. 8b). By the time that backswamp water levels had decreased to 0.5 m AHD, approximately 80% of the COD flux had been discharged. Analysis of these flood events in three backswamp wetlands suggests that maintaining backswamp water levels at 50% of the maximum water level can retain approximately 80% of the COD flux, and hence, DOP. In addition, a shift in vegetation assemblages to wetland-dominant species on approximately two-thirds of the area in the backswamp wetland may have a similar effect.
Managing the surface hydrology is a vital key to minimising the magnitude and frequency of estuarine hypoxic events. Prolonged retention of floodwaters within floodplain wetlands will decrease the discharge rates and shift the DOP flux towards Scenario 1 (Fig. 7b). The current infrastructure on the floodplain may provide the basis for floodwater retention. Retention of floodwaters can be undertaken through modification and retro-fitting existing drainage infrastructure such as concrete headworks and floodgates. Retention of floodwaters in the backswamp wetlands will allow for the natural decomposition of labile organic C to occur within the wetland confines. This process will decrease the COD concentrations of surface waters over time (e.g. Fig. 7b). Slowing the rate of discharge by gradually releasing floodwaters, ideally to a rate at which the assimilation capacity of the main channel can buffer the input of hypoxic water, will also attenuate the DOP flux. Partial or full restoration of the natural floodplain surface hydrology by in-filling drains will result in wetter conditions on the floodplains by retaining water in the backswamp wetlands. Wetter conditions will encourage the gradual recolonisation of wetland species, which are tolerant of waterlogging and inundation, lignin-rich and hence, less easily decomposable. Wetter conditions may also promote reducing conditions and inhibit surface accumulations of oxidised Fe, S and Mn species. The biogeochemical processes driving the development of hypoxia following high-flow events remain only partially understood. These processes are part of the natural cycling of organic material in floodplain environments. Further research is required to address these knowledge gaps to minimise the impacts of post-flood hypoxia on aquatic ecosystems and develop effective management strategies. These knowledge gaps include an understanding of i) the interactions between the coupling of redox cycles with labile organic material and identification of their partial decomposition products; ii) the processes which occur during mixing when hypoxic floodplain-derived waters are discharged into oxic main channel waters; and iii) the processes and temporal dynamics of re-aeration through exchange with the atmosphere via wind mixing or tidal exchange. 4. Summary Estuarine hypoxic events can have devastating effects on aquatic ecosystems, water quality and local economies. While such events may occur naturally, the frequency, magnitude and duration of these events has been subject to substantial anthropogenic forcing by 1) extensively altered surface hydrology on coastal floodplains, which has increased the rate and volume of discharging floodwaters during the flood recession phase, and resulted in 2) a shift in vegetation assemblages, from wetland-dominant species, which are tolerant of wetter conditions and inundation, to dryland-dominant species intolerant of waterlogging. Management strategies which promote natural surface hydrology on coastal floodplains and a shift to wetter regimes can attenuate the forced component of these estuarine hypoxic events. Acknowledgements The authors would like to thank Chrisy Clay, Yasmin Carbot, Graeme Robertson and Michael Wood for assisting with sampling following the flood events, the Environmental Analysis Laboratory, Southern Cross University for sample analysis and two anonymous reviewers for improving this manuscript. This research was funded by the Australian Research Council (LP0882141), Richmond River County Council and the Northern Rivers Catchment Management Authority. References
Fig. 8. a) COD flux retained following the flood peak in the Clarence, shaded area indicates the typical duration of post-flood hypoxic period; and b) the relationship between backswamp water levels and COD flux during the flood recession, where COD flux retained (F(COD)retainedt) is calculated by F(COD)retainedt (%) = 100 − [(F(COD)t/∑F(COD)30d) × 100].
Amoros C, Bornette G. Connectivity and biocomplexity in waterbodies of riverine floodplains. Freshw Biol 2002;47:761–76. APHA. Standard methods for the examination of water and waste water. Washington DC, USA: American Public Health Association, American Wastewater Association, World Environment Fund; 2005.
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375 Baldock JA, Oades JM, Nelson PN, Skene TM, Golchin A, Clarke P. Assessing the extent of decomposition of natural organic materials using solid-state 13C NMR spectroscopy. Aust J Soil Res 1997;35:1061–83. Baldwin DS. Dissolved organic matter and phosphorus leached from fresh and ‘terrestrially’ aged river red gum leaves: implications for assessing river-floodplain interactions. Freshw Biol 1999;41:675–85. Barrett R, Nielsen DL, Croome R. Associations between the plant communities of floodplain wetlands, water regime and wetland type. River Res Appl 2010;26:866–76. Battin TJ, Kaplan LA, Findlay S, Hopkinson CS, Marti E, Packman AI, et al. Biophysical controls on organic carbon fluxes in fluvial networks. Nat Geosci 2008;1:95-100. Benner R, Maccubbin AE, Hodson RE. Anaerobic biodegradation of the lignin and polysaccharide components of lignocellulose and synthetic lignin by sediment microflora. Appl Environ Microbiol 1984;47:998-1004. Benner R, Moran MA, Hodson RE. Effects of pH and plant source on lignocellulose biodegradation rates in two wetland ecosystems, the Okefenokee Swamp and a Georgia salt marsh. Limnol Oceanogr 1985;31:89-100. Blanch SJ, Ganf GG, Walker KF. Tolerance of riverine plants to flooding and exposure indicated by water regime. Regul Rivers Res Manage 1999;15:43–62. BOM. Australian Bureau of Meteorology; 2011. Bourbonniere RA, Creed IF. Biodegradability of dissolved organic matter extracted from a chronosequence of forest-floor materials. J Plant Nutr Soil Sci 2006;169:101–7. Breitburg D. Effects of hypoxia, and the balance between hypoxia and enrichment, on coastal fishes and fisheries. Estuaries Coasts 2002;25:767–81. Brunke M, Gonser T. The ecological significance of exchange processes between rivers and groundwater. Freshw Biol 1997;37:1-33. Burdige DJ. The kinetics of organic matter mineralization in anoxic marine sediments. J Mar Res 1991;49:727–61. Burton ED, Bush RT, Sullivan LA, Mitchell DRG. Schwertmannite transformation to goethite via the Fe(II) pathway: reaction rates and implications for iron-sulfide formation. Geochim Cosmochim Acta 2008;72:4551–64. Capone DG, Kiene RP. Comparison of microbial dynamics in marine and freshwater sediments: contrasts in anaerobic carbon catabolism. Limnol Oceanogr 1988;33: 725–49. Chin YP, Traina SJ, Swank CR, Backhus D. Abundance and properties of dissolved organic matter in porewaters of a freshwater wetland. Limnol Oceanogr 1998;43: 1287–96. Conley DJ, Carstensen J, Ãrtebjerg G, Christensen PB, Dalsgaard T, Hansen Jr LS, et al. Long-term changes and impacts of hypoxia in Danish coastal waters. Ecol Appl 2007;17:S165–84. Cuffney TF. Input, movement and exchange of organic matter within a subtropical coastal black water river-flood plain system. Freshw Biol 1988;19:305–20. Delaune RD, Reddy CN, Patrick WH. Organic matter decomposition in soil as influenced by pH and redox conditions. Soil Biol Biochem 1981;13:533–4. Dent D. Acid sulphate soils: a baseline for research and development, Vol. 39. International Institute for Land Reclamation and Improvement ILRI: Wageningen, The Netherlands; 1986. Diaz RJ. Overview of hypoxia around the world. J Environ Qual 2001;30:275. Diaz RJ, Rosenberg R. Spreading dead zones and consequences for marine ecosystems. Science 2008;321:926–9. Eyre B. Water quality changes in an episodically flushed sub-tropical Australian estuary: a 50 year perspective. Mar Chem 1997;59:177–87. Eyre B, Twigg C. Nutrient behaviour during post-flood recovery of the Richmond River Estuary, Northern NSW, Australia. Estuar Coast Shelf Sci 1997;44:311–26. Eyre BD, Kerr G, Sullivan LA. Deoxygenation potential of the Richmond River Estuary floodplain, northern NSW, Australia. River Res Appl 2006;22:981–92. Gregorich EG, Janzen HH. Microbially mediated processes: decomposition. In: Sumner ME, editor. Handbook of Soil Science. Boca Raton: CRC Press; 2000. p. C107–20. Hladyz S, Watkins SC, Whitworth KL, Baldwin DS. Flows and hypoxic blackwater events in managed ephemeral river channels. J Hydrol 2011;401:117–25. Howitt JA, Baldwin DS, Rees GN, Williams JL. Modelling blackwater: predicting water quality during flooding of lowland river forests. Ecol Model 2007;203:229–42. Johnston SG, Slavich P, Hirst P. The acid flux dynamics of two artificial drains in acid sulfate soil backswamps on the Clarence River floodplain, Australia. Aust J Soil Res 2004;42:623–37. Johnston SG, Slavich PG, Hirst P. Changes in surface water quality after inundation of acid sulfate soils of different vegetation cover. Aust J Soil Res 2005;43:1-12. Johnston SG, Slavich PG, Sullivan LA, Hirst P. Artificial drainage of floodwaters from sulfidic backswamps: effects on deoxygenation in an Australian estuary. Mar Freshw Res 2003;54:781–95. Junk WJ, Bayley PB, Sparks RE. The flood pulse concept in river-floodplain systems. In: Dodge DP, editor. Proceedings of the International Large River Symposium, 106. Canadian Special Publication for Fish and Aquatic Science; 1989. p. 110–27. Kirschbaum MUF. The temperature dependence of soil organic matter decomposition and the effect of global warming on soil organic C storage. Soil Biol Biochem 1995;27:753–60. Lin C, Melville MD. Control of soil acidification by fluvial sedimentation in an estuarine floodplain, eastern Australia. Sediment Geol 1993;85:271–84. Lin C, Wood M, Haskins P, Ryffel T, Lin J. Controls on water acidification and deoxygenation in an estuarine waterway, eastern Australia. Estuar Coast Shelf Sci 2004;61:55–63. Lovley DR, Phillips EJP. Availability of ferric iron for microbial reduction in bottom sediments of the freshwater tidal Potomac River. Appl Environ Microbiol 1986;52: 751–7.
5375
Luther G, Ma S, Trouwborst R, Glazer B, Blickley M, Scarborough R, et al. The roles of anoxia, H2S, and storm events in fish kills of dead-end canals of Delaware inland bays. Estuaries Coasts 2004;27:551–60. Maccubbin AE, Hodson RE. Mineralization of detrital lignocelluloses by salt marsh sediment microflora. Appl Environ Microbiol 1980;40:735–40. MacPherson T, Cahoon L, Mallin M. Water column oxygen demand and sediment oxygen flux: patterns of oxygen depletion in tidal creeks. Hydrobiologia 2007;586: 235–48. Marschner B, Kalbitz K. Controls of bioavailability and biodegradability of dissolved organic matter in soils. Geoderma 2003;113:211–35. McDonald S, Pringle JM, Prenzler PD, Bishop AG, Robards K. Bioavailability of dissolved organic carbon and fulvic acid from an Australian floodplain river and billabong. Mar Freshw Res 2007;58:222–31. Miles CJ, Brezonik PL. Oxygen consumption in humic-colored waters by a photochemical ferrous–ferric catalytic cycle. Environ Sci Technol 1981;15:1089–95. Naylor SD, Chapman GA, Atkinson G, Murphy CL, Tulau MJ, Flewin TC, et al. Guidelines for the use of acid sulfate soil risk maps. Sydney: Department of Land and Water Conservation; 1998. O'Connell M, Baldwin DS, Robertson AI, Rees G. Release and bioavailability of dissolved organic matter from floodplain litter: influence of origin and oxygen levels. Freshw Biol 2000;45:333–42. Petrone KC, Fellman JB, Hood E, Donn MJ, Grierson PF. The origin and function of dissolved organic matter in agro-urban coastal streams. J Geophys Res (Biogeosci) 2011;116:G01028. Ponnamperuma FN. The chemistry of submerged soils. Adv Agron 1972;24:29–96. Pressey RL. Impacts of flood mitigation works on coastal wetlands in New South Wales. Wetlands (Australia) 1982;2:27–44. Robertson AI, Bunn SE, Walker KF, Boon PI. Sources, sinks and transformations of organic carbon in Australian floodplain rivers. Mar Freshw Res 1999;50:813–29. Sammut J, White I, Melville MD. Acidification of an estuarine tributary in eastern Australia due to drainage of acid sulfate soils. Mar Freshw Res 1996;47:669–84. Skjemstad JO, Janik LJ, Taylor JA. Non-living organic matter: what do we know about it? Aust J Exp Agric 1998;38:667–80. Stumm W, Morgan JJ. Aquatic chemistry: chemical equilibria and rates in natural waters. New York: John Wiley and Sons; 1996. Sullivan LA, Bush RT. Iron precipitate accumulations associated with waterways in drained coastal acid sulfate landscapes of eastern Australia. Mar Freshw Res 2004;55:727–36. Townsend SA. The seasonal pattern of dissolved oxygen, and hypolimnetic deoxygenation, in two tropical Australian reservoirs. Lakes Reserv Res Manage 1999;4: 41–53. Townsend SA, Boland KT, Wrigley TJ. Factors contributing to a fish kill in the Australian wet/dry tropics. Water Res 1992;26:1039–44. Townsend SA, Edwards CA. A fish kill event, hypoxia and other limnological impacts associated with early wet season flow into a lake on the Mary River floodplain, tropical northern Australia. Lakes Reserv Res Manage 2003;8:169–76. Tulau MJ. Acid sulfate soil management priority areas in the lower Richmond floodplain. Sydney: Department of Land and Water Conservation; 1999. Valett HM, Baker MA, Morrice JA, Crawford CS, Molles MC, Dahm CN, et al. Biogeochemical and metabolic responses to the flood pulse in a semiarid floodplain. Ecology 2005;86:220–34. van Breemen N. Acidification and deacidification of coastal plain soils as a result of periodic flooding. Soil Sci Soc Am J 1975;39:1153–7. Van Cappellen P, Wang Y. Cycling of iron and manganese in surface sediments; a general theory for the coupled transport and reaction of carbon, oxygen, nitrogen, sulfur, iron, and manganese. Am J Sci 1996;296:197–243. Vaquer-Sunyer R, Duarte CM. Thresholds of hypoxia for marine biodiversity. Proc Natl Acad Sci 2008;105:15452–7. Vidon P, Wagner L, Soyeux E. Changes in the character of DOC in streams during storms in two Midwestern watersheds with contrasting land uses. Biogeochemistry 2008;88:257–70. Walker PH. Seasonal and stratigraphic controls in coastal floodplain soils. Aust J Soil Res 1972;10:127–42. Warrner T, Royer T, Tank J, Griffiths N, Rosi-Marshall E, Whiles M. Dissolved organic carbon in streams from artificially drained and intensively farmed watersheds in Indiana, USA. Biogeochemistry 2009;95:295–307. White I, Melville MD, Wilson BP, Sammut J. Reducing acidic discharges from coastal wetlands in eastern Australia. Wetlands Ecol Manage 1997;5:55–72. Willett IR, Crockford RH, Milnes AR. In: Skinner HCW, Fitzpatrick RW, editors. Transformations of iron, manganese and aluminium during oxidation of a sulfidic material from an acid sulfate soil, 21. Catena Supplement-Biomineralization: processes of iron and manganese-modern and ancient environments; 1992. p. 287–302. Wilson BP. Elevations of sulfurous layers in acid sulfate soils: what do they indicate about sea levels during the Holocene in eastern Australia? Catena 2005;62:45–56. Wilson BP, White I, Melville MD. Floodplain hydrology, acid discharge and change in water quality associated with a drained acid sulfate soil. Mar Freshw Res 1999;50:149–57. Wilson HF, Xenopoulos MA. Effects of agricultural land use on the composition of fluvial dissolved organic matter. Nat Geosci 2009;2:37–41. Wong VNL, Johnston SG, Bush RT, Sullivan LA, Clay C, Burton ED, et al. Spatial and temporal changes in estuarine water quality during a post-flood hypoxic event. Estuar Coast Shelf Sci 2010;87:73–82.
Contents lists available at SciVerse ScienceDirect
Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv
Anthropogenic forcing of estuarine hypoxic events in sub-tropical catchments: Landscape drivers and biogeochemical processes Vanessa N.L. Wong a, b,⁎, Scott G. Johnston a, Edward D. Burton a, Richard T. Bush a, Leigh A. Sullivan a, Peter G. Slavich c a
Southern Cross GeoScience, Southern Cross University, PO Box 157, Lismore New South Wales 2480, Australia School of Geography and Environmental Science, Monash University, Victoria 3800, Australia Wollongbar Primary Industries Institute, New South Wales Department of Trade and Investment, Regional Infrastructure and Services, 1243 Bruxner Highway, Wollongbar NSW 2477, Australia
b c
a r t i c l e
i n f o
Article history: Received 17 May 2011 Received in revised form 23 August 2011 Accepted 25 August 2011 Available online 21 September 2011 Keywords: Estuary Hypoxia Anoxia Deoxygenation Floodplain drainage Labile carbon
a b s t r a c t Episodic hypoxic events can occur following summer floods in sub-tropical estuaries of eastern Australia. These events can cause deoxygenation of waterways and extensive fish mortality. Here, we present a conceptual model that links key landscape drivers and biogeochemical processes which contribute to post-flood hypoxic events. The model provides a framework for examining the nature of anthropogenic forcing. Modification of estuarine floodplain surface hydrology through the construction of extensive drainage networks emerges as a major contributing factor to increasing the frequency, magnitude and duration of hypoxic events. Forcing occurs in two main ways. Firstly, artificial drainage of backswamp wetlands initiates drier conditions which cause a shift in vegetation assemblages from wetland-dominant species to drylanddominant species. These species, which currently dominate the floodplain, are largely intolerant of inundation and provide abundant labile substrate for decomposition following flood events. Decomposition of this labile carbon pool consumes oxygen in the overlying floodwaters, and results in anoxic conditions and waters with excess deoxygenation potential (DOP). Carbon metabolism can be strongly coupled with microbiallymediated reduction of accumulated Fe and Mn oxides, phases which are common on these coastal floodplain landscapes. Secondly, artificial drainage enhances discharge rates during the flood recession phase. Drains transport deoxygenated high DOP floodwaters rapidly from backswamp wetlands to the main river channel to further consume oxygen. This process effectively displaces the natural carbon metabolism processes from floodplain wetlands to the main channel. Management options to reduce the impacts of post-flood hypoxia include i) remodifying drainage on the floodplain to promote wetter conditions, thereby shifting vegetation assemblages towards inundation-tolerant species, and ii) strategic retention of floodwaters in the backswamp wetlands to reduce the volume and rate during the critical post-flood recession phase. © 2011 Elsevier B.V. All rights reserved.
1. Introduction Coastal and estuarine hypoxia has a significant effect on aquatic organisms and ecosystem function (Conley et al., 2007; Vaquer-Sunyer and Duarte, 2008). Hypoxia is defined as low concentrations of dissolved oxygen (i.e. b30% saturation; b2 mg L − 1; b62.5 mM L − 1) and can cause mortality, reduced growth rates and altered distributions of aquatic life (Breitburg, 2002). Dissolved oxygen concentrations in surface waters are determined by inputs from the atmosphere and photosynthesis, and outputs, which are dominated by respiration and other processes that consume oxygen. Globally, the frequency, magnitude and spatial distribution of anthropogenically-induced
⁎ Corresponding author at: School of Geography and Environmental Science, Monash University, Victoria 2800, Australia. Tel.: +61 3 9905 2930; fax: +61 3 9905 2948. E-mail address: [email protected] (V.N.L. Wong). 0048-9697/$ – see front matter © 2011 Elsevier B.V. All rights reserved. doi:10.1016/j.scitotenv.2011.08.065
coastal hypoxic events are increasing (Diaz, 2001). Such events can occur episodically, periodically or seasonally, impacting on benthic and pelagic fauna, nutrient cycling and other biogeochemical processes (Diaz and Rosenberg, 2008). The estuaries and waterways of sub-tropical eastern Australia are prone to episodic hypoxic events following summer floods. Flood events increase connectivity between floodplains and river channels to deliver pulses of sediments, nutrients and organic matter (Amoros and Bornette, 2002; Junk et al., 1989). These events result in a flush of microbial activity which can consume available oxygen in surface waters, thereby causing hypoxia. Large-scale fish kill events have occurred in catchments in the sub-tropical region of eastern Australia during summer as a result of post-flood hypoxia, resulting in the closure of estuaries to commercial and recreational fishing activities for weeks to several months (Eyre, 1997; Johnston et al., 2003; Wong et al., 2010). Similar hypoxic events and fish kills have also been recorded in tropical coastal (Townsend and Edwards, 2003; Townsend et al., 1992) and
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
5369
semi-arid inland (Hladyz et al., 2011; Howitt et al., 2007) river systems following high-flow events. This study presents a conceptual model that incorporates the key processes, major drivers and temporal dynamics of estuarine hypoxic events on the Holocene coastal floodplains of eastern Australia following floods. The conceptual model is used to examine how anthropogenic modification of floodplain hydrology has increased post-flood floodplain-estuary connectivity and altered the nature of carbon pools and fluxes on the floodplains. The findings provide the basis for discussion of possible management and mitigation measures.
Table 2 Flood peaks during the 2008 and 2009 flood events (1:10 ARI) at selected sites on the Richmond River floodplain.
2. Methods
the period immediately following the flood peak during which water levels recede to baseflow levels. Surface water chemistry was sourced from drains in the backswamp wetlands of the Clarence River floodplain and the main Richmond River channel. Methods are described in detail in Johnston et al. (2003) and Wong et al. (2010). Briefly, unfiltered samples were analysed for chemical oxidation demand (COD) by dichromate oxidation (COD APHA 5220; APHA, 2005). Samples filtered through 0.45 μm filters were analysed for dissolved organic carbon (DOC) and dissolved Fe and speciation. Fe speciation was determined using 1,10-phenanthroline (APHA 3500; APHA, 2005). Fe3+ was determined from the difference between total Fe and Fe2+ following reduction with hydroxylammonium chloride. Dissolved organic carbon (DOC) was analysed by wet oxidation with an Aurora 1030 Total Carbon/Total Organic Carbon Analyser (APHA 5310 C; APHA, 2005). Estimates of the daily flux of deoxygenation potential (DOP) from drained wetlands to the main channel were determined from measured drain discharge rates and geochemical data according to Eq. (2).
Data used in this analysis are derived from the Richmond and Clarence River catchments. The Richmond and Clarence River catchments are located in the sub-tropical region of eastern Australia and discharge at Ballina (153.56°E 28.84°S) and Yamba (153.36°E 29.43°S), respectively. The region experiences wet, humid summers with rainfall falling predominantly in summer (November–April) and mild, dry winters (May– October). Mean annual rainfall is 1778 mm for Ballina and 1456 mm for Yamba. Mean maximum temperatures are 28.2 °C at Ballina and 26.7 °C at Yamba in January and mean minimum temperatures are 8.5 °C at Ballina and 9.7 °C at Yamba in July (BOM, 2011). The conceptual model is developed based on observations from four flood events. Raw data is sourced and compiled from Johnston et al. (2003) and Wong et al. (2010). Overbank flooding occurred in the Clarence River in February and March 2001, resulting in floods of approximately 1:10 and 1:14 year average return interval (ARI), respectively (Table 1). Two floods of magnitude 1:10 ARI occurred in the Richmond River in January 2008 and May 2009. The May 2009 flood was of a similar magnitude, with comparable rainfall and water level peaks in similar parts of the catchment (Table 2). Data from the 2009 flood event was collected and analysed using the same techniques and in the same locations as described in Wong et al. (2010). Briefly, surface water samples were collected for approximately one month following the flood peak. Samples were collected in the main channel directly downstream from the main backswamp wetland basins and at the junction where the backswamps discharge into the main channel. Backswamp wetland basins are low-lying depressions commonly found on coastal floodplains with a seasonal or permanent watertable at or above the surface (Naylor et al., 1998). Water level data and discharge rates from two backswamp wetlands on the Clarence River floodplain (Everlasting Swamp and Shark Creek; Johnston et al., 2003), and one backswamp wetland on the Richmond River floodplain (Tuckean Swamp; Wong et al., 2010), were used to develop the model. These data were obtained from data loggers located within the backswamp wetlands. The normalised discharge rate was calculated according to Eq. (1). Q norm ¼ Q t =Q tot
ð1Þ
where Qnorm is the normalised discharge, Qt is the cumulative discharge (m3 s− 1) at time t during the flood recession phase and Qtot is the total discharge that occurred during approximately 30 days of the flood recession phase following the flood peak. The flood recession phase is
Table 1 Flood peaks during the February (1:10 ARI) and March (1:14 ARI) 2001 flood events at selected sites on the Clarence River floodplain. Site
February 2001 river level peak (m AHD)
March 2001 river level peak (m AHD)
Grafton Ulmarra Maclean
6.7 5.3 2.7
7.7 6.1 3.2
Site
2008 river level peak (m AHD)
2009 river level peak (m AHD)
Coraki Bungawalbyn Junction Woodburn
5.8 4.9 3.1
6.6 5.6 4.0
−1 F ðCODÞt kgday ¼ Q t Ct
ð2Þ
where F(COD)t is the daily flux of chemical oxygen demand (COD), Q is the discharge volume (m3 s − 1) and C is the concentration of the COD (mg O2 L − 1) at time t. For further details, see Johnston et al. (2003). 3. Results and discussion 3.1. Coastal floodplain morphology While the Holocene coastal floodplains of sub-tropical eastern Australia are at an advanced stage of evolution, contemporary sediment supply is constrained and the floodplains are still in the process of actively infilling (Lin and Melville, 1993). This leads to distinctive geomorphic features with floodplains characterised by large areas of low-elevation backswamp wetland basins which were previously barrier estuary embayments (Wilson, 2005). The backswamp wetland basins are occluded from the main river channel by extensive natural levee systems. Levees typically range in height up to approximately 4.5 m Australian Height Datum (AHD, where 0 m AHD ≈ sea level). An example of this morphology is provided in Fig. 1a. When intact, these natural barriers typically impound surface waters within backswamp wetland basins for periods of up to 4 months following flood events (White et al., 1997). 3.2. Modification of floodplain surface hydrology The construction of extensive drainage networks for flood mitigation and agricultural expansion on coastal floodplains began around the 1900s and peaked in the 1950s–70s (Pressey, 1982). These drainage networks were designed to remove floodwaters rapidly and efficiently from the floodplains into the main river channels. For example, surface hydrology on the lower Richmond River floodplain (1070 km2) is controlled by N400,000 km of open ditch drains, including farm drains which feed into secondary drains (Tulau, 1999). These drainage networks facilitate rapid
5370
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
Fig. 1. a) The Rocky Mouth Creek backswamp wetland on the Richmond River floodplain provides a typical example of backswamp-levee toe-levee morphology commonly found on sub-tropical floodplains in eastern Australia (Wong et al., 2010); (AHD is Australian Height Datum where 0 m AHD ≈ sea level) and b) increasing the connectivity between the floodplain to the main river channel by dissecting natural levees with artificial drains.
decreases in water levels following rainfall events, which fall to the mean low water (MLW) level and enhance the removal of surface and shallow groundwater (Johnston et al., 2004; Wilson et al., 1999). Connectivity between the floodplain and main channel has greatly increased as the drains dissect the natural levee system to remove surface water (Fig. 1b). Prior to the implementation of drainage networks, backswamp wetlands were natural water storage basins on the floodplain. Following flood events, recession of floodwaters occurred slowly under the natural drainage conditions due to high channel sinuosity and roughness and low hydraulic gradients (White et al., 1997). When surface water levels decreased to below the minimum height of the natural levees, recession of floodwaters then proceeded very slowly via evapotranspiration, infiltration and lateral leakage (Fig. 2a). Under the current artificially augmented hydrological regime, floodwaters are now removed from the backswamp wetlands at least an order of magnitude faster (Sammut et al., 1996; White et al., 1997). Backswamp water levels decrease rapidly to the MLW level with the current artificial regime. During the 2001 and 2008 flood events, backswamp water levels had decreased to below the height of the levees within 3 to 9 days after the flood peak (Fig. 2b). Rapidly decreasing backswamp water levels on modified floodplains is a function of higher discharge rates compared to unmodified floodplains (Fig. 3a). During the 2001 and 2008 flood events, N50% of backswamp wetland discharge had occurred within 4 to 9 days after the flood peak, and 80% within 8 to 16 days (Fig. 3b).
Floodplains are important overbank organic C stores (Battin et al., 2008). Unmodified wetlands in backswamp basins can be important C sinks on coastal floodplains as a result of relatively slow decomposition rates due to high soil water content and limited oxygen supply (Delaune et al., 1981; Gregorich and Janzen, 2000). However, backswamp wetlands are now substantial net C sources during flood events. Modification of floodplains, rivers and their connectivity has substantially altered fluxes of DOC (Robertson et al., 1999). Extensive drainage networks not only facilitate the rapid removal of surface waters, but also transport of highly labile DOC from floodplains to streams and rivers via drains during high flow events (Vidon et al., 2008; Warrner et al., 2009). During high flow events, fresh particulate organic carbon (POC) is mobilised and hydrolysed rapidly to DOC (Battin et al., 2008). These modified floodplains can act as potential sources of large quantities of organic material during floods (Cuffney, 1988; Valett et al., 2005). 3.3. Altered vegetation assemblages—a source of labile carbon Prior to European settlement of eastern Australia in the 19th century, backswamp wetlands on these coastal floodplains were semipermanently inundated. Vegetation was dominated by wetland species such as reeds (Phragmites australis), rushes (Juncus and Eleocharis spp), Swamp Oak (Casuarina glauca), Swamp Mahogany (Eucalyptus robusta) and Broad-leaved Paperbark (Melaleuca quinquenervia) (Tulau, 1999). Changes in water regime, as a result of changes to the rate, magnitude
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
Fig. 2. a) Conceptual representation of backswamp water levels during the post-flood recession phase in a drained and undrained backswamp wetland and b) backswamp water levels during the post-flood recession phase from Everlasting Swamp (E1, E2), Shark Creek (S1, S2) during the two floods in 2001, and from the Tuckean Swamp (T1) during the flood in 2008. Horizontal lines indicate the range in levee height.
and duration of flooding, significantly alter wetland vegetation assemblages (Barrett et al., 2010). Where drier floodplain conditions dominate, vegetation assemblages can shift from predominantly wetland species to dryland species (Blanch et al., 1999; Pressey, 1982). Extensive anthropogenic modification of the surface hydrology on sub-tropical coastal floodplains has allowed for year-round agriculture and has resulted in the dominance of dryland pasture species such as common couch (Cynodon dactylon), carpet grass (Axonopis fissifolius), kikuyu (Pennisetum clandestium) and water couch (Paspalum distichum) (Tulau, 1999). The alteration of vegetation assemblages to a predominance of dryadapted grass species is of vital importance. Vegetation assemblages that currently dominate on coastal floodplains are largely intolerant of waterlogging and flooding. Therefore, when floodplains are rapidly inundated as a result of flooding, vegetation dies and provides an abundant source of labile C substrate for decomposition. A series of studies has shown that the pasture species that are currently commonly found on these coastal floodplains generally have a faster decomposition rate and higher oxygen demand during decomposition compared to the original, endemic wetland vegetation species (Eyre et al., 2006; Johnston et al., 2005). Rates of decomposition, and hence, oxygen demand, are a function of the lability of organic material rather than its total concentration (Brunke and Gonser, 1997). Therefore, organic material derived from grasses, which have a high cellulose content relative to lignin, is decomposed more rapidly than material derived from wetland plants, which typically have a higher lignin content (Benner et al., 1985; Maccubbin and Hodson, 1980). Simple chain C structures are
5371
Fig. 3. a) Conceptual representation of discharge rates during the post-flood recession phase on a drained and undrained floodplain, and b) normalised cumulative discharge during the post-flood recession phase from Everlasting Swamp (E1, E2), Shark Creek (S1, S2) during the two floods in 2001, and from the Tuckean Swamp (T1) during the flood in 2008.
usually associated with carbohydrates and fresh organic material, and are easily decomposable (Baldock et al., 1997; Skjemstad et al., 1998). Fresh organic material can release highly labile DOC which is consumed rapidly (Baldwin, 1999; Bourbonniere and Creed, 2006). Decomposition of floodplain vegetation consumes oxygen due to microbial respiration and organic C oxidation, as reflected in the strong positive correlation between DOC and COD during several different flood events (Fig. 4). Low molecular weight components of DOC with simple structure are available for decomposition almost immediately and result in a rapid increase in bacterial numbers (McDonald et al., 2007; O'Connell et al., 2000), which can further drive decomposition processes mediated by redox transformations under anaerobic conditions. Similarly, under anoxic conditions, more labile organic material with smaller C:N ratios is decomposed preferentially and more rapidly compared to more recalcitrant organic material (Burdige, 1991). Dissolved organic matter (DOM) exported from floodplains which experience prolonged wetter conditions are generally more aromatic in structure (Wilson and Xenopoulos, 2009). Similarly, DOM sourced from wetland sediments frequently have higher molecular weights, are more aromatic and refractory, and are, therefore, more difficult to decompose (Chin et al., 1998; Petrone et al., 2011). Aromatic structures are generally associated with lignin and dominate in peat soils where there is an accumulation of more recalcitrant organic C (Baldock et al., 1997). Due to their high lignin content, anaerobic decomposition of organic material sourced from sedges, rushes and mangroves can occur an order of magnitude slower compared to decomposition of material sourced from grasses (Benner et al., 1984).
5372
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
oxygen sink (Van Cappellen and Wang, 1996). For example, oxidation of Fe (II) has a high deoxygenation potential and forms reactive Fe (III) minerals (Eq. (8)) which are then available for microbiallymediated anaerobic decomposition (Miles and Brezonik, 1981). 2þ
4Fe
þ
þ O2 þ 10H2 O→4FeðOHÞ3 þ 8H
ð8Þ
3.5. The effects of temperature
Fig. 4. Increasing COD with increasing concentrations of labile organic carbon (i.e. DOC) where Clarence 2001a is sourced from Everlasting Swamp and Shark Creek during the February 2001 flood event, Clarence 2001b is sourced from Everlasting Swamp and Shark Creek during the March 2001 flood event, and Richmond 2008 is sourced from Wardell during the January 2008 flood event.
3.4. Accumulation of alternative terminal electron acceptors Under oxic conditions, decomposition of organic material occurs with O2 as the terminal electron acceptor to produce CO2. In the absence of oxygen, decomposition occurs with alternative terminal electron acceptors via an approximate thermodynamic sequence with increasing energy requirements in the order O2 N NO3− N Mn4+ N Fe3+ N SO42− N CO2 according to Eqs. (3)–(7) where CH2O approximates labile organic material for decomposition (Capone and Kiene, 1988; Ponnamperuma, 1972). −
þ
4NO3 þ CH2 O þ 4H →2N2 þ 5CO2 þ H2 O
ð3Þ
þ
ð4Þ
2þ
2MnO2 þ CH2 O þ 4H →2Mn
þ
2þ
4FeðOHÞ3 þ CH2 O þ 4H →4Fe
2þ
þ
þ CO2 þ 3H2 O
þ CO2 þ 11H2 O
The occurrence and intensity of hypoxic conditions is strongly associated with seasonal temperature. Lower dissolved oxygen concentrations in surface waters are frequently found during warmer months with higher air and water temperatures (e.g. Conley et al., 2007; MacPherson et al., 2007; Townsend, 1999). Surface water temperatures on coastal sub-tropical floodplains in eastern Australia vary by approximately 15– 20 °C (Fig. 5a). Deoxygenation events generally occur during summer, when ambient temperatures and water temperatures are higher. Higher temperatures provide optimal environmental conditions for increased microbial metabolism, enhanced rates of oxygen consumption (Kirschbaum, 1995) and lower dissolved oxygen saturation potential (Stumm and Morgan, 1996). In highly modified drainage canals, seasonal hypoxia during summer months can lead to complete deoxygenation of the water column (Luther et al., 2004). COD in the drains of backswamp wetlands on the Clarence River floodplain is positively correlated with antecedent drain water temperatures (Fig. 5b), most likely as a result of increasing decomposition rates of floodplain-derived organic material. DOC is known to become more labile with higher temperatures as the rate of microbial breakdown of larger, less labile DOC compounds into smaller, more labile compounds increases (Marschner and Kalbitz, 2003). A comparison of a summer and winter flood on the Richmond River estuary provides a clear example of the critical threshold of temperature as a key component of anaerobic metabolism and the development of estuarine hypoxia. Hypoxic conditions established in the Richmond
ð5Þ
SO4 þ 2CH2 O þ 2H →H2 S þ 2CO2 þ 2H2 O
ð6Þ
2CH2 O→CH4 þ CO2
ð7Þ
The coastal floodplains of eastern Australia are frequently underlain by acid sulfate soils (ASS) (Walker, 1972). These soils can develop abundant surface and near-surface accumulations of bioavailable Fe, Mn and SO42−, especially in backswamp wetlands (Dent, 1986; Sullivan and Bush, 2004; Willett et al., 1992). Under anaerobic conditions, these redox sensitive species can be an important source of alternative terminal electron acceptors for decomposition of organic material under anaerobic conditions (Burton et al., 2008; Lovley and Phillips, 1986). Extended inundation of floodplain ASS facilitates microbially-mediated reduction of Fe (III) and Mn (IV) coupled with decomposition of labile organic material. This process results in enhanced mobilisation of Fe (II) and Mn (II) in surface and porewaters (Johnston et al., 2003; van Breemen, 1975). Furthermore, reductive dissolution of poorly crystalline Fe (oxy)hydroxide minerals can release additional sorbed DOM, providing additional substrate for decomposition (Chin et al., 1998). Abiotic processes can further consume oxygen in surface waters. Oxidation of reduced inorganic species such as Fe (II) and Mn (II) and the coupling of the corresponding redox cycles can act as a potential
Fig. 5. a) Seasonal variation in mean daily surface water temperature from the Clarence River in a backswamp wetland drain (backbridge) and the drain's junction with the main river channel (floodgates) and b) COD correlation with antecedent surface water temperature as measured in drains in Everlasting Swamp and Shark Creek following the two flood events in 2001.
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
5373
River estuary during the summer 2008 flood event when surface water temperatures in the main channel were between 23 and 27 °C (Wong et al., 2010). Conversely, the main river channel remained oxic during a similar magnitude flood event in winter 2009 due to lower surface water temperatures (i.e. 16–18 °C), reduced microbial activity and higher dissolved oxygen saturation potential. Temperature also influences the intensity of anoxia and, therefore, the dominant terminal electron acceptor processes that are coupled with anaerobic C metabolism. During both the 2008 and 2009 flood events, dissolved Fe concentrations peaked between 6 and 14 days after the main channel flood peak in the main channel of the Richmond River (Fig. 6). However, while Fe2+ is the dominant Fe species during the post-flood recession period during the summer 2008 flood event (Fig. 6a), Fe 3+ dominates during the post-flood recession during the winter 2009 flood event (Fig. 6b). Similarly, the total dissolved Fe concentration mobilised during the 2008 flood event was more than twice the peak Fe concentration mobilised during the 2009 flood event. These trends suggest that decomposition of organic material is more strongly coupled with microbially-mediated Fe reduction at higher temperatures, such as those occurring during summer flood events. 3.6. Synthesis It is likely that the magnitude, duration and frequency of post-flood estuarine hypoxic events have substantially increased as a result of changes to the surface hydrology and vegetation assemblages on coastal floodplains. The flux of DOP from an individual drained floodplain wetland to the main river channel can be expressed according to Eq. (2) and is a function of a) the rate of drainage discharge from the floodplain, and b) the DOP of floodplain waters. Deoxygenation potential is a complex function of many different factors and is strongly influenced by temperature and the concentration and lability of C. Eq. (2) provides a simple framework to qualitatively evaluate anthropogenic forcing, as changes to either the discharge rate or DOP will shift both the timing and magnitude of the DOP flux (Fig. 7a), which can act to compound or diminish its effects. We estimated a high and low post-flood COD scenario from data collected from two
Fig. 6. Dissolved Fe concentrations and speciation during the flood recession phase at Wardell following a) the summer 2008 flood event, and b) winter 2009 flood event.
Fig. 7. a) Conceptual model of deoxygenation potential flux during the flood recession phase, and b) temporal dynamics of COD in drains at Everlasting Swamp and Shark Creek during the 2001 floods (Johnston et al., 2003). Curves represent a high COD scenario and a low COD scenario.
flood events (Fig. 7b). For example, a high post-flood COD scenario, such as observed in Everlasting Swamp in 2001 (Fig. 7b), is driven by the dominance of dryland pasture species in this backswamp wetland (Johnston et al., 2003). In contrast, the low COD values at Shark Creek (Fig. 7b) are largely the result of M. quinquenervia forest, comprised of highly recalcitrant C (Johnston et al., 2005) covering approximately two thirds of the backswamp wetland. Therefore, a change in land use from native forest to dryland pasture will shift the DOP flux from Scenario 3 in the direction of Scenario 1 (Fig. 7a) due to increased C lability. Conversely, a period of lower temperatures during the flood recession phase, such as during the 2009 flood event, will decrease COD and therefore, the magnitude and duration of the DOP flux. Similarly, lower concentrations of alternative terminal electron acceptors or a decrease in the biomass of pasture species will also shift the DOP from Scenario 3 towards Scenario 1 (Fig. 7b). In sub-tropical estuaries of eastern Australia, the onset of hypoxic conditions in the main estuarine channel following flood events typically occurs approximately 4–6 days following the flood peak (Eyre and Twigg, 1997; Johnston et al., 2003; Wong et al., 2010). Peak COD concentrations in drains usually occur during this period (Fig. 7b) and coincide with high discharge rates (Fig. 3b). Maximum DOP flux coincides with the falling limb of the flood hydrograph, when flow volumes in the main channel are decreasing sharply. Therefore, the capacity of the main channel water column to assimilate high oxygen demand inputs of floodplain-derived surface waters during this period is also decreasing. Transported deoxygenated surface waters can then exert additional oxygen demand to further consume oxygen in the main channel (Lin et al., 2004).
5374
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375
3.7. Influence of land management There are a variety of land management actions which may attenuate the impact of the DOP flux scenarios represented in Fig. 7a, thereby reducing the frequency and magnitude of hypoxic events in estuarine waterways. According to Eq. (2), DOP flux can be reduced by 1) retaining a portion of floodwaters in the backswamp wetland basins, or 2) slowing the drainage rate from the floodplain during the flood recession phase. COD flux can be used as a proxy for DOP flux. Analysis of COD flux from drained wetlands during past events relative to time and backswamp wetland water levels provides a useful tool to identify the relative attenuation effect that might be achieved by retaining floodwaters at a given time or water level threshold. Based on this analysis, retention of floodwaters at the onset of main channel hypoxic conditions (i.e. approximately 4–6 days after the flood peak) can retain approximately 65–90% of the COD flux (at Day 5) in the backswamp wetlands (Fig. 8a). The retention of COD flux within this range corresponds with maintaining backswamp water levels at approximately 0.5–3.0 m AHD, which is dependent on the wetland basin and flood event (Fig. 8b). By the time that backswamp water levels had decreased to 0.5 m AHD, approximately 80% of the COD flux had been discharged. Analysis of these flood events in three backswamp wetlands suggests that maintaining backswamp water levels at 50% of the maximum water level can retain approximately 80% of the COD flux, and hence, DOP. In addition, a shift in vegetation assemblages to wetland-dominant species on approximately two-thirds of the area in the backswamp wetland may have a similar effect.
Managing the surface hydrology is a vital key to minimising the magnitude and frequency of estuarine hypoxic events. Prolonged retention of floodwaters within floodplain wetlands will decrease the discharge rates and shift the DOP flux towards Scenario 1 (Fig. 7b). The current infrastructure on the floodplain may provide the basis for floodwater retention. Retention of floodwaters can be undertaken through modification and retro-fitting existing drainage infrastructure such as concrete headworks and floodgates. Retention of floodwaters in the backswamp wetlands will allow for the natural decomposition of labile organic C to occur within the wetland confines. This process will decrease the COD concentrations of surface waters over time (e.g. Fig. 7b). Slowing the rate of discharge by gradually releasing floodwaters, ideally to a rate at which the assimilation capacity of the main channel can buffer the input of hypoxic water, will also attenuate the DOP flux. Partial or full restoration of the natural floodplain surface hydrology by in-filling drains will result in wetter conditions on the floodplains by retaining water in the backswamp wetlands. Wetter conditions will encourage the gradual recolonisation of wetland species, which are tolerant of waterlogging and inundation, lignin-rich and hence, less easily decomposable. Wetter conditions may also promote reducing conditions and inhibit surface accumulations of oxidised Fe, S and Mn species. The biogeochemical processes driving the development of hypoxia following high-flow events remain only partially understood. These processes are part of the natural cycling of organic material in floodplain environments. Further research is required to address these knowledge gaps to minimise the impacts of post-flood hypoxia on aquatic ecosystems and develop effective management strategies. These knowledge gaps include an understanding of i) the interactions between the coupling of redox cycles with labile organic material and identification of their partial decomposition products; ii) the processes which occur during mixing when hypoxic floodplain-derived waters are discharged into oxic main channel waters; and iii) the processes and temporal dynamics of re-aeration through exchange with the atmosphere via wind mixing or tidal exchange. 4. Summary Estuarine hypoxic events can have devastating effects on aquatic ecosystems, water quality and local economies. While such events may occur naturally, the frequency, magnitude and duration of these events has been subject to substantial anthropogenic forcing by 1) extensively altered surface hydrology on coastal floodplains, which has increased the rate and volume of discharging floodwaters during the flood recession phase, and resulted in 2) a shift in vegetation assemblages, from wetland-dominant species, which are tolerant of wetter conditions and inundation, to dryland-dominant species intolerant of waterlogging. Management strategies which promote natural surface hydrology on coastal floodplains and a shift to wetter regimes can attenuate the forced component of these estuarine hypoxic events. Acknowledgements The authors would like to thank Chrisy Clay, Yasmin Carbot, Graeme Robertson and Michael Wood for assisting with sampling following the flood events, the Environmental Analysis Laboratory, Southern Cross University for sample analysis and two anonymous reviewers for improving this manuscript. This research was funded by the Australian Research Council (LP0882141), Richmond River County Council and the Northern Rivers Catchment Management Authority. References
Fig. 8. a) COD flux retained following the flood peak in the Clarence, shaded area indicates the typical duration of post-flood hypoxic period; and b) the relationship between backswamp water levels and COD flux during the flood recession, where COD flux retained (F(COD)retainedt) is calculated by F(COD)retainedt (%) = 100 − [(F(COD)t/∑F(COD)30d) × 100].
Amoros C, Bornette G. Connectivity and biocomplexity in waterbodies of riverine floodplains. Freshw Biol 2002;47:761–76. APHA. Standard methods for the examination of water and waste water. Washington DC, USA: American Public Health Association, American Wastewater Association, World Environment Fund; 2005.
V.N.L. Wong et al. / Science of the Total Environment 409 (2011) 5368–5375 Baldock JA, Oades JM, Nelson PN, Skene TM, Golchin A, Clarke P. Assessing the extent of decomposition of natural organic materials using solid-state 13C NMR spectroscopy. Aust J Soil Res 1997;35:1061–83. Baldwin DS. Dissolved organic matter and phosphorus leached from fresh and ‘terrestrially’ aged river red gum leaves: implications for assessing river-floodplain interactions. Freshw Biol 1999;41:675–85. Barrett R, Nielsen DL, Croome R. Associations between the plant communities of floodplain wetlands, water regime and wetland type. River Res Appl 2010;26:866–76. Battin TJ, Kaplan LA, Findlay S, Hopkinson CS, Marti E, Packman AI, et al. Biophysical controls on organic carbon fluxes in fluvial networks. Nat Geosci 2008;1:95-100. Benner R, Maccubbin AE, Hodson RE. Anaerobic biodegradation of the lignin and polysaccharide components of lignocellulose and synthetic lignin by sediment microflora. Appl Environ Microbiol 1984;47:998-1004. Benner R, Moran MA, Hodson RE. Effects of pH and plant source on lignocellulose biodegradation rates in two wetland ecosystems, the Okefenokee Swamp and a Georgia salt marsh. Limnol Oceanogr 1985;31:89-100. Blanch SJ, Ganf GG, Walker KF. Tolerance of riverine plants to flooding and exposure indicated by water regime. Regul Rivers Res Manage 1999;15:43–62. BOM. Australian Bureau of Meteorology; 2011. Bourbonniere RA, Creed IF. Biodegradability of dissolved organic matter extracted from a chronosequence of forest-floor materials. J Plant Nutr Soil Sci 2006;169:101–7. Breitburg D. Effects of hypoxia, and the balance between hypoxia and enrichment, on coastal fishes and fisheries. Estuaries Coasts 2002;25:767–81. Brunke M, Gonser T. The ecological significance of exchange processes between rivers and groundwater. Freshw Biol 1997;37:1-33. Burdige DJ. The kinetics of organic matter mineralization in anoxic marine sediments. J Mar Res 1991;49:727–61. Burton ED, Bush RT, Sullivan LA, Mitchell DRG. Schwertmannite transformation to goethite via the Fe(II) pathway: reaction rates and implications for iron-sulfide formation. Geochim Cosmochim Acta 2008;72:4551–64. Capone DG, Kiene RP. Comparison of microbial dynamics in marine and freshwater sediments: contrasts in anaerobic carbon catabolism. Limnol Oceanogr 1988;33: 725–49. Chin YP, Traina SJ, Swank CR, Backhus D. Abundance and properties of dissolved organic matter in porewaters of a freshwater wetland. Limnol Oceanogr 1998;43: 1287–96. Conley DJ, Carstensen J, Ãrtebjerg G, Christensen PB, Dalsgaard T, Hansen Jr LS, et al. Long-term changes and impacts of hypoxia in Danish coastal waters. Ecol Appl 2007;17:S165–84. Cuffney TF. Input, movement and exchange of organic matter within a subtropical coastal black water river-flood plain system. Freshw Biol 1988;19:305–20. Delaune RD, Reddy CN, Patrick WH. Organic matter decomposition in soil as influenced by pH and redox conditions. Soil Biol Biochem 1981;13:533–4. Dent D. Acid sulphate soils: a baseline for research and development, Vol. 39. International Institute for Land Reclamation and Improvement ILRI: Wageningen, The Netherlands; 1986. Diaz RJ. Overview of hypoxia around the world. J Environ Qual 2001;30:275. Diaz RJ, Rosenberg R. Spreading dead zones and consequences for marine ecosystems. Science 2008;321:926–9. Eyre B. Water quality changes in an episodically flushed sub-tropical Australian estuary: a 50 year perspective. Mar Chem 1997;59:177–87. Eyre B, Twigg C. Nutrient behaviour during post-flood recovery of the Richmond River Estuary, Northern NSW, Australia. Estuar Coast Shelf Sci 1997;44:311–26. Eyre BD, Kerr G, Sullivan LA. Deoxygenation potential of the Richmond River Estuary floodplain, northern NSW, Australia. River Res Appl 2006;22:981–92. Gregorich EG, Janzen HH. Microbially mediated processes: decomposition. In: Sumner ME, editor. Handbook of Soil Science. Boca Raton: CRC Press; 2000. p. C107–20. Hladyz S, Watkins SC, Whitworth KL, Baldwin DS. Flows and hypoxic blackwater events in managed ephemeral river channels. J Hydrol 2011;401:117–25. Howitt JA, Baldwin DS, Rees GN, Williams JL. Modelling blackwater: predicting water quality during flooding of lowland river forests. Ecol Model 2007;203:229–42. Johnston SG, Slavich P, Hirst P. The acid flux dynamics of two artificial drains in acid sulfate soil backswamps on the Clarence River floodplain, Australia. Aust J Soil Res 2004;42:623–37. Johnston SG, Slavich PG, Hirst P. Changes in surface water quality after inundation of acid sulfate soils of different vegetation cover. Aust J Soil Res 2005;43:1-12. Johnston SG, Slavich PG, Sullivan LA, Hirst P. Artificial drainage of floodwaters from sulfidic backswamps: effects on deoxygenation in an Australian estuary. Mar Freshw Res 2003;54:781–95. Junk WJ, Bayley PB, Sparks RE. The flood pulse concept in river-floodplain systems. In: Dodge DP, editor. Proceedings of the International Large River Symposium, 106. Canadian Special Publication for Fish and Aquatic Science; 1989. p. 110–27. Kirschbaum MUF. The temperature dependence of soil organic matter decomposition and the effect of global warming on soil organic C storage. Soil Biol Biochem 1995;27:753–60. Lin C, Melville MD. Control of soil acidification by fluvial sedimentation in an estuarine floodplain, eastern Australia. Sediment Geol 1993;85:271–84. Lin C, Wood M, Haskins P, Ryffel T, Lin J. Controls on water acidification and deoxygenation in an estuarine waterway, eastern Australia. Estuar Coast Shelf Sci 2004;61:55–63. Lovley DR, Phillips EJP. Availability of ferric iron for microbial reduction in bottom sediments of the freshwater tidal Potomac River. Appl Environ Microbiol 1986;52: 751–7.
5375
Luther G, Ma S, Trouwborst R, Glazer B, Blickley M, Scarborough R, et al. The roles of anoxia, H2S, and storm events in fish kills of dead-end canals of Delaware inland bays. Estuaries Coasts 2004;27:551–60. Maccubbin AE, Hodson RE. Mineralization of detrital lignocelluloses by salt marsh sediment microflora. Appl Environ Microbiol 1980;40:735–40. MacPherson T, Cahoon L, Mallin M. Water column oxygen demand and sediment oxygen flux: patterns of oxygen depletion in tidal creeks. Hydrobiologia 2007;586: 235–48. Marschner B, Kalbitz K. Controls of bioavailability and biodegradability of dissolved organic matter in soils. Geoderma 2003;113:211–35. McDonald S, Pringle JM, Prenzler PD, Bishop AG, Robards K. Bioavailability of dissolved organic carbon and fulvic acid from an Australian floodplain river and billabong. Mar Freshw Res 2007;58:222–31. Miles CJ, Brezonik PL. Oxygen consumption in humic-colored waters by a photochemical ferrous–ferric catalytic cycle. Environ Sci Technol 1981;15:1089–95. Naylor SD, Chapman GA, Atkinson G, Murphy CL, Tulau MJ, Flewin TC, et al. Guidelines for the use of acid sulfate soil risk maps. Sydney: Department of Land and Water Conservation; 1998. O'Connell M, Baldwin DS, Robertson AI, Rees G. Release and bioavailability of dissolved organic matter from floodplain litter: influence of origin and oxygen levels. Freshw Biol 2000;45:333–42. Petrone KC, Fellman JB, Hood E, Donn MJ, Grierson PF. The origin and function of dissolved organic matter in agro-urban coastal streams. J Geophys Res (Biogeosci) 2011;116:G01028. Ponnamperuma FN. The chemistry of submerged soils. Adv Agron 1972;24:29–96. Pressey RL. Impacts of flood mitigation works on coastal wetlands in New South Wales. Wetlands (Australia) 1982;2:27–44. Robertson AI, Bunn SE, Walker KF, Boon PI. Sources, sinks and transformations of organic carbon in Australian floodplain rivers. Mar Freshw Res 1999;50:813–29. Sammut J, White I, Melville MD. Acidification of an estuarine tributary in eastern Australia due to drainage of acid sulfate soils. Mar Freshw Res 1996;47:669–84. Skjemstad JO, Janik LJ, Taylor JA. Non-living organic matter: what do we know about it? Aust J Exp Agric 1998;38:667–80. Stumm W, Morgan JJ. Aquatic chemistry: chemical equilibria and rates in natural waters. New York: John Wiley and Sons; 1996. Sullivan LA, Bush RT. Iron precipitate accumulations associated with waterways in drained coastal acid sulfate landscapes of eastern Australia. Mar Freshw Res 2004;55:727–36. Townsend SA. The seasonal pattern of dissolved oxygen, and hypolimnetic deoxygenation, in two tropical Australian reservoirs. Lakes Reserv Res Manage 1999;4: 41–53. Townsend SA, Boland KT, Wrigley TJ. Factors contributing to a fish kill in the Australian wet/dry tropics. Water Res 1992;26:1039–44. Townsend SA, Edwards CA. A fish kill event, hypoxia and other limnological impacts associated with early wet season flow into a lake on the Mary River floodplain, tropical northern Australia. Lakes Reserv Res Manage 2003;8:169–76. Tulau MJ. Acid sulfate soil management priority areas in the lower Richmond floodplain. Sydney: Department of Land and Water Conservation; 1999. Valett HM, Baker MA, Morrice JA, Crawford CS, Molles MC, Dahm CN, et al. Biogeochemical and metabolic responses to the flood pulse in a semiarid floodplain. Ecology 2005;86:220–34. van Breemen N. Acidification and deacidification of coastal plain soils as a result of periodic flooding. Soil Sci Soc Am J 1975;39:1153–7. Van Cappellen P, Wang Y. Cycling of iron and manganese in surface sediments; a general theory for the coupled transport and reaction of carbon, oxygen, nitrogen, sulfur, iron, and manganese. Am J Sci 1996;296:197–243. Vaquer-Sunyer R, Duarte CM. Thresholds of hypoxia for marine biodiversity. Proc Natl Acad Sci 2008;105:15452–7. Vidon P, Wagner L, Soyeux E. Changes in the character of DOC in streams during storms in two Midwestern watersheds with contrasting land uses. Biogeochemistry 2008;88:257–70. Walker PH. Seasonal and stratigraphic controls in coastal floodplain soils. Aust J Soil Res 1972;10:127–42. Warrner T, Royer T, Tank J, Griffiths N, Rosi-Marshall E, Whiles M. Dissolved organic carbon in streams from artificially drained and intensively farmed watersheds in Indiana, USA. Biogeochemistry 2009;95:295–307. White I, Melville MD, Wilson BP, Sammut J. Reducing acidic discharges from coastal wetlands in eastern Australia. Wetlands Ecol Manage 1997;5:55–72. Willett IR, Crockford RH, Milnes AR. In: Skinner HCW, Fitzpatrick RW, editors. Transformations of iron, manganese and aluminium during oxidation of a sulfidic material from an acid sulfate soil, 21. Catena Supplement-Biomineralization: processes of iron and manganese-modern and ancient environments; 1992. p. 287–302. Wilson BP. Elevations of sulfurous layers in acid sulfate soils: what do they indicate about sea levels during the Holocene in eastern Australia? Catena 2005;62:45–56. Wilson BP, White I, Melville MD. Floodplain hydrology, acid discharge and change in water quality associated with a drained acid sulfate soil. Mar Freshw Res 1999;50:149–57. Wilson HF, Xenopoulos MA. Effects of agricultural land use on the composition of fluvial dissolved organic matter. Nat Geosci 2009;2:37–41. Wong VNL, Johnston SG, Bush RT, Sullivan LA, Clay C, Burton ED, et al. Spatial and temporal changes in estuarine water quality during a post-flood hypoxic event. Estuar Coast Shelf Sci 2010;87:73–82.