Indicators of phytoplankton response to particulate ...

Science of the Total Environment 636 (2018) 1416–1427

Contents lists available at ScienceDirect

Science of the Total Environment journal homepage: www.elsevier.com/locate/scitotenv

Indicators of phytoplankton response to particulate nutrient bioavailability in fresh and marine waters of the Great Barrier Reef Alexandra Garzon-Garcia a,b,⁎, Joanne Burton a,b, Hannah M. Franklin b, Philip W. Moody a, Robert W. De Hayr a, Michele A. Burford b,c a b c

Landscape Sciences, Department of Environment and Science, PO Box 5078, Brisbane, Queensland 4001, Australia Australian Rivers Institute, Nathan Campus, Griffith University, 170 Kessels Road, Nathan, Brisbane, Queensland 4111, Australia School of Environment and Science, Griffith University, 170 Kessels Road, Nathan, Queensland, 4111, Australia

H I G H L I G H T S

G R A P H I C A L

A B S T R A C T

• New approach linking sediment nutrient characteristics to phytoplankton response • Indicators of phytoplankton response developed for fresh and marine conditions • Indicators include carbon measurements highlighting the mediating role of bacteria. • Indicators performed better than traditional particulate nutrient measurements. • Bioavailability depends both on load and sediment characteristics linked to source.

a r t i c l e

i n f o

Article history: Received 15 February 2018 Received in revised form 24 April 2018 Accepted 24 April 2018 Available online xxxx Editor: Jay Gan Keywords: Microalgae Eutrophication Rivers Mineralization Sediment Soil

a b s t r a c t Sediments delivered to freshwater and marine environments can make important contributions to the aquatic bioavailable nutrient pool. In the Great Barrier Reef (GBR) catchments, particulate nutrients comprise an important fraction of the end of catchment loads; however, their contribution to the bioavailable nutrient pool is not well understood. This research determined which particulate nutrient parameters are the best indicators of the potential effect of fine sediment (b10 μm) on phytoplankton growth. Surface and subsurface sediments were lab-generated to cover a wide spectrum of particulate nutrient bioavailability from key soil types, land uses and erosion processes (hillslope and gully) in a wet and a dry tropics catchment of the GBR. Phytoplankton bioassays were used to assess freshwater and marine phytoplankton responses to sediments. The best indicators were selected by regressing measurements of phytoplankton growth against nutrient bioavailability parameters measured on the sediments. The selected indicator equations included organic carbon (C) pools for both fresh and marine water, highlighting the role of bacteria in mediating nutrient availability for phytoplankton. The equations also included various fractions of particulate nitrogen (N) (differentiating the adsorbed ammoniumN from the particulate organic N), and the ratios of C to N, which indicate the lability of the organic matter present in the sediment. Dissolved reactive phosphorus was also an important indicator in freshwater. The indicators performed better in assessing bioavailability potential than traditional methods to monitor particulate nutrients, e.g., particulate N and particulate phosphorus. Phytoplankton bioassays indicated that nutrients in sediment can promote phytoplankton growth, with nutrient bioavailability depending not only on sediment load, but

⁎ Corresponding author at: Landscape Sciences, Department of Environment and Science, PO Box 5078, Brisbane, Queensland 4001, Australia. E-mail address: [email protected] (A. Garzon-Garcia).

https://doi.org/10.1016/j.scitotenv.2018.04.334 0048-9697/© 2018 Elsevier B.V. All rights reserved.

A. Garzon-Garcia et al. / Science of the Total Environment 636 (2018) 1416–1427

1417

also sediment characteristics associated with its parent soil. These characteristics vary with soil type, land use and erosion process. Findings will help prioritize erosion control to catchment areas which are most likely to contribute large amounts of bioavailable particulate nutrients to the GBR. © 2018 Elsevier B.V. All rights reserved.

1. Introduction Sediments of terrestrial origin can contribute to the bioavailable nutrient pool of freshwater and marine environments (Keil et al., 1997; Mayer et al., 1998). Turbid rivers have been shown to contribute a significant fraction of bioavailable nitrogen (N) to coastal ecosystems via regenerated N from riverine particulates (Mayer et al., 1998). Coastal mineralization of organic matter is a major regeneration pathway, with an estimated ~55% to 80% of the riverine organic matter river flux mineralized along continental margins globally (Burdige, 2005). The complex biogeochemical mechanisms by which nutrients in sediment become bioavailable operate across multiple spatial and temporal scales. These mechanisms are largely associated with sediment transport dynamics which include the contribution from different terrestrial environments, fractionation during transport, deposition or sourcing from riverine habitats, and finally discharge into coastal waters. The source of sediment (i.e. geology, soil type, land use, vegetation cover type, surface versus subsurface soil) controls factors which have an effect on the bioavailability of particulate nutrients. These factors include silt and clay content, mineralogy, aggregation and organic matter content and type (Garten and Ashwood, 2002; Gregorich et al., 2006; Horowitz and Elrick, 1987; Keil et al., 1997). Fractionation and mixing of sediment and its associated organic matter during river transport is an important process controlling the bioavailability of particulate nutrients because it determines the relative mix of sediment fractions with different intrinsic bioavailabilities (Bianchi and Bauer, 2011; Mayer et al., 1998). Finer sediment particles (i.e. silt and clay) tend to be enriched in nutrients, Fe and Mn oxides and hydroxides, thereby enhancing the interaction of these particulate nutrients with water due to a larger surface area relative to volume (Horowitz and Elrick, 1987; Keil et al., 1997; Walling and Moorehead, 1989). Suspended sediment composition, including granulometry and plant litter content (Garzon-Garcia et al., 2017; Juarez et al., 2011), has been shown to vary both across the catchment and the hydrograph, and between wet and dry years (Garzon-Garcia et al., 2017; Kao and Liu, 2000; Ludwig and Probst, 1996; Rowland et al., 2017). Fractionation processes are also very relevant in estuaries, where river flood plumes are slowed down allowing for enhanced sediment settling and biogeochemical processing (Bauer et al., 2013). Additionally, the biogeochemical environment surrounding sediments will have an effect on the rate of conversion of particulate nutrients to their bioavailable forms. Stream water nutrient concentrations (Benstead et al., 2009), as well as temperature and oxygen concentrations (Gomez et al., 2003) are known to influence organic matter decomposition in streams (Tank et al., 2010). Salinity in estuaries affects the desorption of P from sediments as other anions compete with phosphate for sorption sites (Howarth et al., 1995), as well as the desorption of adsorbed ammonium by matrix cations (Mackin and Aller, 1984; Rosenfeld, 1979), and enhances the flocculation of fine suspended sediment and associated organic C (Bianchi and Bauer, 2011; Sholkovitz, 1976). In the Great Barrier Reef (GBR), Australia, an excess of bioavailable nutrients has been associated with a range of damaging impacts to the ecosystems including: an increase in the frequency of Crown of Thorns Starfish (COTS) outbreaks (Brodie et al., 2005; Fabricius et al., 2010); loss of seagrass and coral through reduced photic depth (Fabricius et al., 2014); an increase in susceptibility to coral bleaching (Wooldridge, 2009); reef degradation and reduced coral biodiversity (DeVantier et al., 2006; Fabricius, 2005); an increase in macroalgae

and subsequent competition with coral (De'ath and Fabricius, 2010); and possible links to coral disease (Haapkyla et al., 2011). Historically, management to improve water quality to the GBR has focused on reducing the loss of sediment and dissolved inorganic forms of nitrogen from catchments (Queensland Government, 2013a). It has been implicitly assumed that managing the largest sources of sediment would also target the largest sources of particulate nutrients. The contributions of particulate nutrients and organics to the bioavailable forms of nutrients (i.e. dissolved inorganic nitrogen, DIN) at the end of GBR catchments and in the GBR are potentially important but have received little attention to date. Very little is known of the sources, transformations/losses and impact of particulate nutrients as they are transported from terrestrial to marine systems. Phytoplankton in nutrient-limited environments rapidly assimilates bioavailable nutrients resulting in an increase in biomass, typically measured as chlorophyll a concentrations. This can therefore be used as an indicator of the GBR eutrophication status (Brodie et al., 2011). A fraction of the total particulate N (PN) and particulate phosphorus (PP) pools has the potential to become bioavailable to phytoplankton over a specified period of time, with flow-on effects in marine ecosystems. In this research, we used a standardized phytoplankton bioassay (Franklin et al., 2018) combined with data modelling to determine the best indicators of potential particulate nutrient bioavailability to phytoplankton growth in fresh and marine water of the GBR over a short timeframe (days). The indicators are equations that allow the quantification of algal response to bioavailable particulate nutrients in different sediment/soil types. The equations are composed of the bioavailable particulate nutrient properties (parameters) that were found to best estimate the response. The indicator parameters are measured on the sediment and have the potential to be linked to parent soil properties. The bioassay enabled the control of complex biogeochemical mechanisms involved in the transformation of particulate nutrients to bioavailable forms (sourcing, fractionation, mixing and biogeochemical environment) (Franklin et al., 2018). 2. Materials and methods 2.1. Selection of sampling sites The study was undertaken in two GBR catchments: one in the wet tropics, viz. Johnstone River catchment; and one in the dry tropics, the Bowen River sub-catchment (Supplementary material S1). In the Johnstone River catchment, the focus was predominantly in the South Johnstone sub-catchment which consists of the following dominant land uses: conservation (84%), cattle grazing (9.7%) and sugar cane (2.8%). The Bowen River sub-catchment consists predominantly of cattle grazing land. The Bowen River sub-catchment was selected as it is the sub-catchment of the Burdekin River that delivers the most sediment to the GBR, delivering around 45% of the annual fine (b63 μm) TSS Burdekin river load (Bainbridge et al., 2014). A range of major soil types and geologies were selected for parent soil sampling, covering the predominant catchment land uses and various erodibilities (Supplementary material S1) (Zund and Payne, 2014). This resulted in nine combinations of soil type, land use and erodibility for the Bowen River sub-catchment and six combinations of soil type and land use for the Johnstone River catchment (Supplementary material S2). Each combination was sampled across each catchment in triplicate, except in a few cases where circumstances prevented this collection. Identifying criteria for selecting test samples was important

1418


to ensure that the indicators of particulate nutrient bioavailability approach can be applied across typical land uses, climate zones and soil types found within the GBR. Time-integrated sediment samplers (“rocket samplers”) (Phillips et al., 2000) were installed in the Johnstone River catchment (5) and Bowen River sub-catchment (10) to sample sediment transported during high-flow events during the 2016 wet season (Supplementary material S1). Sampling site locations were selected to sample sediment from drainage areas in homogeneous soil types and land uses (Supplementary material S2). In addition, some sites were selected to sample sediment on the main branch of each river where gauging sites were present. Two rocket samplers were installed at each sampling site. In the Bowen River sub-catchment the two rocket samplers were installed at different heights to capture sediments from different positions in the water column. In the Johnstone River catchment they were installed at the same height to obtain replicate samples.

2.2. Sample collection A total of 17 soil samples were collected during the first week of March 2016 from the Johnstone River catchment and a total of 41 soil samples were collected during the third week of April 2016 from the Bowen River sub-catchment. Surface soil samples (0–10 cm) were collected at all sampling points after removing vegetation, loose leaves and woody litter from the surface. Subsurface soil samples were taken at sampling points of high erodibility (Bowen River catchment only) by sampling all vertical strata differentiated by soil colour on an exposed gully bank. Samples from each stratum were integrated by scraping approximately 20 cm into the exposed face of the bank within each of the vertical strata with a spade. Surface soil was sampled from inter-rows at cane sites and rows at banana sites. Approximately 40 kg of soil was sampled at each sampling point and stored in clean bags. Sediment was sampled between February and March 2016 from four high-flow events in the Johnstone River catchment (total of 22 sediment samples, peak flows of 47, 63.2, 175.6 and 96 m3 s−1 at US Central Mill station) (Queensland Government, 2013b) and one highflow event in the Bowen River catchment (total of 17 sediment samples, peak flow of 1580.2 m3 s−1 at Myuna station) (Queensland Government, 2013b). In the Johnstone River catchment, high-flow event sediment samples were only obtained from the end of system site for Event 1 and from all sampling sites for Events 2 to 4. In the

Bowen River sub-catchment, high-flow event sediment samples were obtained from 9 out of the 10 rocket sampler sites installed (except for R8) for the sampled event. 2.3. Sample preparation and analysis Soil samples were dried at 40 °C until constant weight was achieved and then sieved to b2 mm or ground in a jaw crusher so that the soil would pass through a b2 mm sieve. Large litter fragments (N2 mm) were either removed manually or by the sieving process. A sample splitter was used to take representative subsamples of the sieved soil. All soil samples were analyzed for particulate nutrient bioavailability parameters (Table 1). For a detailed description of methods for each parameter in Table 1 see Supplementary material (S3). Additionally, sediments were generated from each soil sample by fractionation to b10 μm using settling columns. This size fraction (clays and fine silt) was selected as it is the dominant fraction transported to the GBR lagoon in large events (Bainbridge et al., 2012). Briefly, a soil subsample was suspended in deionised (DI) water and left undisturbed for the time necessary for a 10-μmdiameter particle to settle through the settling column height at 20 °C according to Stoke's Law. All sediment samples generated in the laboratory were subsequently analyzed for most of the particulate nutrient bioavailability parameters analyzed on their parent soil samples (see Table 1). High-flow event sediment samples were dried in the oven at 40 °C until a constant weight was achieved, sieved to b2 mm and analyzed for their total nutrient pools (PC, PN, POC, PP, 105 °C air-dry moisture) and particle size (Malvern Mastersizer 2000 laser sizer). High-flow event sediment samples selected for validation experiments were later analyzed for the necessary additional particulate nutrient bioavailability parameters to calculate the selected indicators. 2.4. Sediment selection for phytoplankton bioassays A reduced number of sediments covering a wide spectrum of variability in particulate nutrient bioavailability parameters were selected for the phytoplankton bioassays. Non-metric multi-dimensional scaling (NMDS) mapping was used to make this selection using the vegan package (Oksanen et al., 2017) in R statistical software version 3.4.0 (R Core Team, 2017). This statistical method of ordination utilises matrix

Table 1 Particulate nutrient bioavailability parameters analyzed for parent soil and laboratory-generated sediment.b Nitrogen (N)

Phosphorus (P)

Total particulate N (PN) (refers to the total N pool) Total particulate P (PP) (refers to the total P pool)

Total particulate organic N (PON) (refers to the total organic N pool = PN-Adsorbed Ammonium N) Adsorbed Ammonium-N (NH+ 4 -N) and Nitrate-N (NO− 3 -N) extracted by 2 M KCl Soluble organic N (SON) (water extractable) Potentially mineralizable N (PMN) under aerobic conditions (1d, 3d, 7d) Aqueous extractable NO3−-N

a b

Carbon (C)/organics

Other possible explanatory measures

Ratios

Total particulate C (PC)

Particle sizea (hydrometer and laser diffraction) R1/R2a (dispersion ratio)

POC/PN

Sorbed P (refers to the P sorbed to the soil/sediment surface that is extracted by the Colwell-P method)

Total particulate organic carbon (POC)

Mineral P (refers to P that is part of the soil/sediment mineral matrix. It is calculated as BSES-P minus Colwell-P) Phosphorus Buffer Index (PBI) (an indicator of how tightly sorbed P is bound to the soil/sediment surface) Dissolved reactive P (DRP) (calculated as Colwell-P/PBI)

Soluble organic carbon (SOC) (water extractable)

Clay activity ratioa

PN/PP

Potential production of soluble organic C (PPSOC) (1d, 3d, 7d)

Exchangeable cationsa

POC/PN/PP

Exchangeable aluminium and aciditya Effective cation exchange capacitya pH EC Chloride Oven dry moisture (105 °C)

SOC/mineral N SOC/SON

Only analyzed on parent soil. For detailed description of methods for each of the parameters see Supplementary material (S3).

POC/PON


algebra to derive unique, successive linear axes such that distances among objects (in this case, sampled parent soils or lab-generated sediment samples) in multivariate space (all particulate nutrient bioavailability parameters) are well represented in a low-dimensional ordination space e.g., two dimensional graphs. For the statistical analyses, particulate nutrient bioavailability parameters were log-transformed except for potentially mineralizable N (PMN), which was power-transformed (elevated to the 1/3 power) because it included negative values. All particulate nutrient bioavailability parameters were standardized for analysis. Four NMDS maps were generated, two for parent soils and two for lab-generated sediments, one for each of the two studied catchments. Using the sediment maps, 16 lab-generated sediments were selected for “indicator calibration” bioassay experiments to measure phytoplankton growth responses to sediments with varying particulate bioavailable nutrient concentrations and determine the indicators of the bioavailability of particulate nutrients (Fig. 1). Additionally, using similar maps, 13 high-flow event sediments were selected for “indicator validation” bioassay experiments. The “indicator validation” experiments were carried out to verify the validity of selected indicators for lab-generated sediments when using them for sediments sampled from rivers during high-flow events, which have mixed sources and particle sizes. The parent soil NMDS maps (not presented here) were used to verify that the selection of sediments covered the whole spectrum of variability in particulate nutrient bioavailability parameters measured on the parent soils of these sediments. The selection also aimed to include a representation of the main combinations of soil types, land uses and erosion processes (Supplementary material S4). Particulate nutrient bioavailability parameters for all parent soils, laboratory-generated sediments and high-flow event sampled sediments are presented in Supplementary material S5.

Fig. 1. NMDS maps of lab-generated sediments for particulate nutrient bioavailability parameters in the Johnstone River catchment (a), and Bowen River sub-catchment (b). Numbers in maps represent each sediment sample. Blue axes represent the direction each particulate nutrient bioavailability parameter increases in the NMDS space. Selected sediments for phytoplankton bioassays are in red circles (see Supplementary material S4 for identification). (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

1419

2.5. Bioassay methodology A sediment exposure bioassay was used to determine growth responses of freshwater and marine phytoplankton to sediments. The bioassay methodology is described briefly below and in full in Franklin et al. (2018). 2.5.1. Sediment slurry preparation A slurry of each sediment was prepared 2 h prior to each bioassay to a target concentration of 3 g L−1 (equivalent to 200 mg L−1 TSS accounting for a 1:15 dilution in bioassay bottles). This concentration was representative of that experienced during high-flow in the Johnstone River. Fine suspended sediment concentrations of 134 mg L−1 and 324 mg L−1 were estimated under flows of 325 m3 s−1 (median of historic highflow) and 750 m3 s−1 (mean of peak flows 2014–2016) respectively, using a TSS concentration-flow relationship from 2016 high-flow events. No particle size/flow relationship was available for the Bowen sub-catchment, so 3.0 g L−1 was also used for sediments from this river. This concentration would likely be in the lower range of highflow event concentrations for the Bowen River, considering that the Bowen tends to have much higher TSS and fine sediment concentrations than the Johnstone River (Garzon-Garcia et al., 2015; Turner et al., 2013; Turner et al., 2012; Wallace et al., 2015; Wallace et al., 2014). Laboratory-generated sediment slurries were prepared directly from whole soil (b2 mm) rather than dried pre-fractionated sediments to avoid biogeochemical changes during drying and fractionation. Firstly, a 200 g soil subsample was placed in a 2 L clean settling tube with 400 mL of DI water, sonicated in a bath (2 min) and then DI water added to make 1800 mL. The sample was agitated (60 s) and then left to settle for 48 min (based on Stoke's Law). All liquid above the 400 mL mark was carefully removed through a side opening probe leaving the settled particles (N10 μm) behind. The removed suspension was sieved (63 μm) to remove floating litter and centrifuged at 4500 rpm (30 min). The supernatant which contained the rapidly soluble fraction of particulate nutrients (produced during the 48 min fractionation) was discarded as the aim of the study was to assess the effect of bioavailable nutrients associated with sediment already in transport to freshwater and coastal phytoplankton communities. The centrifuged particulate fraction was diluted with DI water at a predetermined volume (based on sediment dry weight post-centrifuging in preliminary experiments) to make a 3 g L−1 suspension. To prepare sediment slurries of the high-flow events, 3 g of dry sediment was suspended in 1 L of DI water. DI water was used to facilitate the slurry preparation in the laboratory and later incubated in dialysis tubes surrounded by stream or marine water (1:15 dilution) (see Section 2.5.3). The soluble fraction was retained for these slurries as the biogeochemical processes releasing rapidly soluble nutrients had already occurred during transport downstream. 2.5.2. Water source selection and characterisation The South Johnstone River (Supplementary material S1) was selected as a representative freshwater source as preliminary experiments indicated its potential to respond to nutrient and sediment treatments was greater than water from the Bowen River [see Franklin et al. (in press) for more details]. Although preliminary trials suggested coastal GBR water had the capacity to respond to treatments, it proved logistically difficult to collect and freight this to the laboratory in a timely manner without compromising the phytoplankton. Therefore, water from central Moreton Bay (closer to Brisbane, Supplementary material S6) was selected as a representative marine source and provided a reasonable proxy for conditions in the GBR coastal waters. The community in Moreton Bay has many genera in common with the GBR (Davies et al., 2016) and both waters have been shown to be N-limited (Furnas et al., 2005; O'Donohue et al., 2000). Briefly, bulk surface water samples for the bioassays were collected 5 h prior to commencing the bioassays for marine water and 24 h for

1420


freshwater. Bulk water conductivity/salinity, pH, temperature and turbidity were measured in the field or on return to laboratory, using a calibrated data logger (HYDROLAB, Quanta, Hydrolab Corporation, USA). Water temperature was estimated from average monthly data when not measured in the field (http://www.bom.gov.au and https://watermonitoring.information.qld.gov.au). The photosynthetic yield (Fv/Fm) of the bulk water was measured in duplicate on return to the laboratory on 3 replicate 100 mL sub-samples that were dark-adapted for 20 min using a PHYTO-PAM System II Emitter-Detector (PHYTO-ED) unit (Heinz Walz GmbH, 2003) with PHYTO-WIN software version 2.10. To assess background nutrient concentrations, triplicate samples for DIN − (as NOx-N and NH+ 4 -N), soluble reactive phosphorus (SRP, PO4 -P), dissolved organic carbon (DOC), total dissolved nitrogen (TDN) and phosphorus (TDP) analyses were filtered through 0.45 μm membrane filters. Samples were also collected for total nitrogen (TN) and phosphorus (TP) analyses. Samples were stored frozen prior to analysis by standard colorimetric methods (APHA/AWWA/WPCF, 2012). For chlorophyll a analysis, 0.2 L was filtered through glass fibre filters (GF/F Whatman, 0.7 μm) and filters stored at −80 °C for subsequent acetone extraction using a sonicator followed by fluorometry (limit of detection 0.01 μg L−1) (APHA/AWWA/WPCF, 2012). Background source water conditions, physicochemical conditions and phytoplankton community composition are presented in Supplementary material S6. 2.5.3. Sediment exposure bioassays Four experimental bioassays were carried out: two “indicator calibration” experiments using the 16 selected lab-generated sediments in marine (Experiment 1) and freshwater (Experiment 3) conditions, and two “indicator validation” experiments using the 13 selected high-flow event sediments in marine (Experiment 2) and freshwater (Experiment 4) conditions. To set up each bioassay, 300 mL subsamples of bulk water were placed into four replicate clear plastic bottles for each treatment. A 20 mL aliquot of each sediment slurry was transferred into preprepared dialysis tubing (0.25 m lengths), sealed at each end and placed in the bottles of bulk water (Supplementary material S7) (1:15 dilution). It is assumed that this very low dilution caused by the use of DI water in the dialysis tubes is not a significant change in the simulated ambient conditions experienced by the microbial community and phytoplankton. For example in Experiment 1 this dilution reduced the salinity from 35 PSU to 33 PSU, a range typically experienced by the community in Moreton Bay (Saeck et al., 2016). This ensured that the particulate material did not directly interact with the algae (e.g. causing flocculation) and did not interfere with the PHYTOPAM methods. High molecular weight cut-off (14,000 Da, 16 mm diameter, Sigma Aldrich) cellulose membrane dialysis tubing was used to allow diffusion of larger molecules into the bulk water. Tubing was pre-treated as per manufacturer instructions to remove compounds present in more than trace amounts. In addition to the sediment treatments, control treatments with and without nutrients were included to assess whether N and/or P were limiting phytoplankton growth in the source water. In these, 20 mL of Milliq water (control) or nutrient solution were placed inside dialysis tubing as per the sediment treatments. Nitrogen was added as NH4Cl at 10 times the estimated background DIN, after accounting for dilution in 300 mL bottles (final concentrations 0.07 and 0.3 mg N L−1 for marine and freshwater, respectively). Phosphorus was added as KH2PO4 at a stoichiometrically balanced rate with N (approximately Redfield molar ratio of 16:1, final concentrations 4 and 19 μg L−1 for marine and freshwater, respectively). Nitrogen and phosphorus were also added in combination to the same concentrations as individually. A trace metal and iron solution was added to all nutrient control treatments to ensure these elements were not limiting. Bioassays were incubated for 7 days (168 h) on a light table fitted with fluorescence bulbs [160–180 μmol (PAR) m−2 s−1, 36 W Cool White Lumilux, Osram], under 12 h total dark and 12 h total light, at

the ambient water temperature (22–23 and 28 °C for marine and freshwater experiments respectively, Supplementary material S6). Each day, bottles were gently inverted to re-suspend the sediment. After 72 h the bottles were dark-adapted for 20 min before Fv/Fm was measured on two 5 mL sub-samples as described previously. Fv/Fm was measured at multiple time points (24, 48, 72, and 168 h), however 72 h was selected as suitable to collect Fv/Fm data to develop indicators of bioavailability. At this exposure period, the phytoplankton community had sufficient time to respond to treatments, but was not artificially limited, i.e. no significant decline in Fv/Fm of the control treatment. Chlorophyll a concentrations in each bioassay bottle were measured after 7 days by filtering 0.1–0.175 L through glass fibre filters (GF/F Whatman) and analyzed as per background samples.

2.6. Statistical methods The best indicators of the bioavailability of particulate nutrients were selected by all-subsets step-up regressions using the leaps package in R (Lumley, 2017) to determine which combination of parameters in Table 1 best explained phytoplankton growth/yield in the bioassay “calibration experiments” for marine conditions (Experiment 1) and freshwater conditions (Experiment 3). This type of regression tests all the possible combinations of parameters and reports on the best subsets for each size (number of explanatory variables used in the regression). Step-up regressions were run for phytoplankton yield as measured by Fv/Fm after 3 days, and chlorophyll a at the end of the incubation as dependent variables and bioavailable nutrient parameters as independent variables. Marine and freshwater indicators were considered separately. The selection of “indicators”, using the analysis of regression outputs (significance of independent variables and R2 ) was based on the criteria that (a) the addition of further bioavailable nutrient parameters in the equation did not significantly improve the R2, and (b) the combination of parameters and the type of relationship with phytoplankton response (positive or negative) made sense in terms of our conceptual understanding of the biogeochemical processes that may be driving it. A comparative analysis was carried out to quantify the error in estimating chlorophyll a when using the multiple linear equations obtained for the proposed marine indicators versus using a linear equation with only particulate nitrogen (PN) (data from Experiment 1). Firstly, for each of the linear equations to compare, the absolute error in estimating chlorophyll a for each sediment used in the experiment was calculated by subtracting the modelled (as per the corresponding equation) minus the measured chlorophyll a value. The cumulative error was calculated for the experiment as a whole by adding this figure for all sediments used in the experiment. Secondly, the cumulative errors and errors by sediment type were compared between linear equations. To validate the indicators, the selected multiple linear model equations (“indicator equations”) were applied to the 13 high-flow event sediments used in the “validation experiments” for marine conditions (Experiment 2) and freshwater conditions (Experiment 4) and the predicted phytoplankton and diatom growth/yield results were regressed against the actual phytoplankton and diatom growth/yield measured in the bioassay experiments. Additionally, the best performing indicator equation was applied to the total pool of lab-generated sediments to estimate the bioavailability of particulate nutrients generated from sediment sourced from different soil types, land uses and erosion processes (surface versus subsurface samples to simulate hillslope and gully or streambank erosion, respectively) in marine water. This exercise illustrates the potential use of the indicator equations to rate the risk of different sediment types in their contribution to bioavailable particulate nutrients to the GBR.


3. Results 3.1. Indicators of phytoplankton response to particulate nutrient bioavailability in marine conditions In marine water, two indicator equations of phytoplankton response to particulate nutrient bioavailability in marine conditions (PNBm) were selected (Eqs. (1) and (2)). These two indicator equations were the best combination of bioavailable particulate nutrient parameters1 measured on the sediment to explain phytoplankton photosynthetic yield after 72 h (photosynthetic yield: Fv/Fm) using multiple linear equations (for detailed methods see Section 2.6). Parameters are used on a weight basis (mass unit, i.e. g), which means that photosynthetic yield depends not only on the sediment quality (nutrient concentration in the sediment), but also on the amount of sediment present. The equations can also be calculated for an arbitrary sediment mass (e.g., bioavailability of 1 kg of sediment) (Fig. 2a, Table 2). Similar equations (including the same combination of bioavailable particulate nutrient parameters measured on the sediment) best explained chlorophyll a concentration measured at 168 h (Fig. 2b, Table 2). PNBm1 ¼ 0:198 þ 0:457 PN þ 0:004

POC 2 ; R ¼ 0:88 PN

PNBm2 ¼ 0:221 þ 9:638 NH4 þ −N þ 0:402 PON þ 0:001 SOC ; R2 NH 4 þ −N ¼ 0:93

ð1Þ

ð2Þ

The first indicator equation (PNBm1), which had a slightly lower explanatory power, was selected for future use because it includes bioavailable nutrient parameters measured on sediment that can be easily estimated from source soil properties using pedo-transfer functions (see conceptual diagram in Fig. 3). These source soil properties (viz. TOC and TN) are routinely measured and mapped for soils in most soil mapping programs. Particulate N alone had a good explanatory power (R2) for both measures of phytoplankton response, but slightly lower explanatory power than the two selected indicator equations (Fig. 2a,b, Table 2). The added reduction in error for estimated chlorophyll a from all the 16 sediments used in the bioassay was close to 0 μg L−1 for PNBm1, when compared to using PN as a single parameter. This indicator equation (PNBm1) is not very different to using PN alone on average. In particular, for two of the sediments, the reduction in error using PNBm1 for estimated chlorophyll a was significant, around 0.13 μg L−1 of chlorophyll a. The added reduction in error for estimated chlorophyll a from all the 16 sediments used in the bioassay was close to 0.34 μg L−1 for PNBm2, when compared to using PN as a single parameter. This reduction in error is important considering the values of chlorophyll a monitored in the Great Barrier Reef, may range from 0.23 to 0.54 μg L−1 (Brodie et al., 2007). For one sediment in particular, the reduction in error was very high (0.48 μg L−1). 3.2. Validations of indicators for marine conditions A regression model between the indicator equations of phytoplankton response to particulate nutrient bioavailability in marine conditions (PNBm) calculated for sediment sampled in high-flow events (fitted) and the phytoplankton yield (Fv/Fm) after 72 h measured in the “validation experiment” for marine conditions (Experiment 2) (observed) accounted for 75% of the variance (R2 ~ 0.75). The indicator equations performed very well in estimating the outcomes of this experiment. Particulate organic carbon (POC); 2 M KCl extractable ammonium (NH+ 4 N); particulate organic nitrogen (PON); soluble organic carbon (SOC). The units of all equation parameters are based on a weight basis to be calculated from the sediment content and sediment concentration e.g., mg POC. 1

1421

The same was the case when applying the chlorophyll a equations (Fig. 2c, d). PNBm2 had a lower R2 than PNBm1 because there was not enough sample left for one of the sediments (the one that caused the highest phytoplankton yield) to quantify the former. When removing that same sample for PNBm1, the R2 was similar for both indicator equations. This implies that similar R2 values to PNBm1 would be expected for PNBm2 if the omitted sample had been included in the analysis. Sampled high-flow event sediment in both catchments was predominantly fine-grained, with an average of 85–88% of the particles in each sample below b63 μm in size and 36–58% below b16 μm. Highflow event sediments at the end of catchments have been reported to have N90% of the particles in the b63 μm fraction (Turner et al., 2013). It is important to note that although the high-flow event sediment had different particle sizes to the lab-generated sediments (Fig. 4), which were fractionated to b10 μm, the indicator equations still performed well. 3.3. Indicators of phytoplankton response to particulate nutrient bioavailability in freshwater conditions In freshwater, there was evidence of co-limitation of phytoplankton by N and P [see Franklin et al. (2018)], so two sets of indicator equations were selected. In the case of N limitation only, there were two indicator equations to estimate the phytoplankton response to particulate nutrient bioavailability (PNBfw) (Eqs. (3) and (4)) and when both N and P were limiting, there was one indicator equation to estimate the phytoplankton response to particulate nutrient bioavailability (PNPBfw) (Eq. (5)). These three indicator equations were the best combinations of bioavailable particulate nutrient parameters2 measured on the sediment to explain phytoplankton yield (Fv/Fm) after 72 h using multiple linear equations (Fig. 5a, Table 3). Similar equations (including the same combination of bioavailable particulate nutrient parameters measured on the sediment) provided the best explanation of chlorophyll a concentrations measured at 168 h (Fig. 5b, Table 3). PNBfw1 ¼ 0:467 þ 0:015 POC−0:001

POC 2 ; R ¼ 0:51 PN

ð3Þ

SOC 2 ; R ¼ 0:72 SON

ð4Þ

PNBfw2 ¼ 0:438 þ 3:387 SOC þ 0:001

PNPBfw1 ¼ 0:451 þ 13:28 NH 4 þ N þ 451:7 DRP þ 0:0005 SOC 2 ;R SON ¼ 0:58

ð5Þ

The first indicator equation (PNBfw1) was selected because it includes bioavailable particulate nutrient parameters measured on sediment that can be easily estimated from routinely measured source soil properties using pedo-transfer functions (see conceptual diagram in Fig. 3). In comparison to the three selected indicator equations, particulate N as a single parameter in the equations accounted for less variance in phytoplankton response measurements in freshwater conditions (compare R2 values; Table 3). Particulate P did not have a significant explanatory power for phytoplankton response measurements (R2 = 0.05 for yield at 72 h). 3.4. Validation of indicators for freshwater conditions Linear regressions between the phytoplankton yield (Fv/Fm) at 72 h calculated for high-flow event sampled sediment using the indicator 2 Particulate organic carbon (POC); 2 M KCl extractable ammonium N (NH+ 4 N); particulate nitrogen (PN); soluble organic carbon (SOC); soluble organic nitrogen (SON); dissolved reactive phosphorus (DRP). The units of all equation parameters are based on a weight basis to be calculated from the sediment content and sediment concentration e.g., mg POC

1422


Fig. 2. Linear regressions between modelled and measured phytoplankton response (PNBm) (as measured by phytopam in phytoplankton bioassay) (a); and chlorophyll a (b); to particulate nutrient bioavailability in the marine “calibration experiment” (Experiment 1), and between modelled and measured phytoplankton response (PNBm) (as measured by phytopam in phytoplankton bioassay) (c); and chlorophyll a (d); to particulate nutrient bioavailability in the marine “validation experiment” (Experiment 2). The first indicator + equation includes PN and POC/PN (black diamonds) and the second indicator equation includes NH+ 4 -N, PON and SOC/NH4 -N (dark grey circles). The linear model using PN alone is presented for reference for Experiment 1 results (light grey triangles). Lines represent a linear regression (y = ax + b) with b = 0. Multiple linear equation parameters and R2 for (a) and (b) are presented in Table 2.

equations for freshwater conditions (PNBfw, PNPBfw) and Fv/Fm measured in the “validation experiment” for freshwater conditions (Experiment 4) did not have a good fit (Fig. 5c). This poor fit may be related to the large difference between the base flow water used in Experiment 3 (used to select the indicators) and the water used in Experiment 4 (used to validate the indicators) (see Supplementary material S6). The water used for the freshwater “validation experiment” had higher turbidity and lower initial Fv/Fm compared to the water used for the other three experiments and did not have a positive Fv/Fm response to N addition after 72 h see Fig. 2 in Franklin et al. (2018). These observations indicate that this water may not have been N-limited and thus did not have the capacity to respond to the river suspended sediment treatments. When applying the chlorophyll a equations, linear regressions had a relatively good fit, though estimated chlorophyll a values were different to the measured ones (Fig. 5d). This means that the equations performed well in indicating the relative magnitude of the chlorophyll a response, but were imprecise at predicting actual chlorophyll a values (note, however, that this is not their intended use).

3.5. Particulate nutrient bioavailability to phytoplankton The bioavailability of particulate nutrients to phytoplankton varied with sediment type (Figs. 2 and 5). When the best performing bioavailable particulate nutrient indicator equation (PNBm2) was applied under marine water conditions to the 58 laboratory-generated sediments, differences in particulate nutrient bioavailability were found between sediments sourced from different land uses, soil types and erosion processes (surface and subsurface) (Fig. 6). 4. Discussion 4.1. Indicators of bioavailable particulate nutrients We have developed a new approach linking sediment nutrient characteristics to phytoplankton response in marine and freshwater conditions. Using this novel approach, we have demonstrated that particulate nutrients associated with sediments from both surface

Table 2 Multiple linear equation parameters and fit for selected indicators to estimate phytoplankton response to particulate nutrient bioavailability in marine conditions (PNBm) and chlorophyll a using phytoplankton bioassay results (Experiment 1) and bioavailable particulate nutrient parameters. Dependent variable (indicator)

Parameter 1

Parameter 2

Parameter 3

PNBm1 (yield at 72 h) PNBm2 (yield at 72 h) Yield at 72 h Chlorophyll a Chlorophyll a Chlorophyll a

+0.457 PN*** +9.638 NH+ 4 -N** +0.469 PN *** +3.131 PN*** +64.61 NH+ 4 -N** +3.210 PN***

+0.004 POC/PN +0.402 PON***

+0.001 SOC/NH+ 4 -N

+0.025 POC/PN +2.714 PON***

−0.0005 SOC/NH+ 4 -N

(***) p b 0.001, (**) p b 0.01, (*) p b 0.05. The units of all equation parameters are based on a w basis e.g., mg POC.

b

R2

+0.198 +0.221 +0.233 +0.334 +0.554 +0.562

0.88 0.93 0.87 0.86 0.91 0.85


1423

Fig. 3. Conceptual diagram of the link between soil properties, sediment properties and the effect of different sediments on aquatic algae response. Bioavailable particulate nutrient parameters explaining algae response and identified in the indicator equations can be measured on the sediment or can be predicted from soil properties through the use of pedotransfer functions.

and subsurface sources, as well as multiple land uses and soil types are potentially bioavailable to freshwater and marine phytoplankton of the GBR. This implies that eroded sediment poses an important risk to the eutrophication of these waters. We found that the phytoplankton response to fine particulates is not only dependent on sediment load, but also on the nutrient characteristics of the sediment. In the GBR, as well as in many other end of catchment and marine monitoring programs worldwide, total pools of particulate N and P have traditionally been used as indicators of water quality. This study demonstrated that while PN had a good explanatory power for phytoplankton response in marine water, it performed less well than the bioavailability indicator equations developed using several other particulate nutrient parameters in marine and fresh water. PN may explain phytoplankton response in marine water because it integrates various relevant bioavailable particulate nutrient parameters, including fractions of PN with different bioavailability potential (viz. adsorbed ammonium, PON) and organic C content. However, the N and P colimitation for freshwater phytoplankton may lessen its explanatory power in freshwater conditions. When selecting parameters to indicate a biological or ecosystem health response it is important that the indicators, or in this case the indicator equations, not only are statistically significant but also that the selection process is supported by a conceptual understanding of the underlying mechanisms that cause this response (Udy et al., 2006), in this case biogeochemical. In this study, the application of indicator equations to estimate phytoplankton response to bioavailable particulate nutrients is based on the conceptual understanding that there is a biogeochemical link between source soil properties, sediment properties and their effect on phytoplankton. Sediment properties are linked to parent soil properties and understanding fractionation and enrichment processes as the finer sediment is transported downstream is necessary to understand this link (Horowitz and Elrick, 1987; Mayer et al., 1998). The bioavailability of particulate nutrients in fine sediment to

Fig. 4. Fine particle size (b63um and b 16 um) distribution statistics for high-flow event sampled sediment in the Bowen River sub-catchment and Johnstone River catchment during the wet season 2015–2016.

phytoplankton is mediated by two main processes: mineralization of the PON (Mayer et al., 1998; Tank et al., 2010) and desorption of the adsorbed ammonium-N (Mackin and Aller, 1984; Rosenfeld, 1979). Because these two processes differ in timeframes and controls, the effect of sediments on phytoplankton may be better explained by specific nutrient pools associated with these processes rather than by total pools alone (PN and PP) (Fig. 2). This justifies the inclusion of these two fractions, PON and adsorbed ammonium-N, in the indicator equations. The second indicator Eq. (2) for marine conditions (PNBm2), which had a slightly higher explanatory power than indicator Eq. (1) (PNBm1), includes these two fractions. Understanding the differences in bioavailability of the PON and adsorbed ammonium-N across different sediment types (soil type, land use and erosion process) is necessary to assess the bioavailability of sediment to phytoplankton. It is important to mention here that the sediments with higher PON bioavailability as measured by the potential mineralization of the PON (PMN), are not necessarily the sediments with higher adsorbed ammonium-N content (see Supplementary material S5). The bioavailability of adsorbed ammonium-N to phytoplankton is determined by adsorption/desorption physico-chemical processes (Mackin and Aller, 1984; Rosenfeld, 1979). The potential of sediment to adsorb ammonium is linked to the sediment source because it primarily depends on the cation exchange capacity of the sediment which is determined by organic matter content, humic content and clay mineralogy (Boatman and Murray, 1982; Zhang et al., 2016). Most of the adsorbed ammonium-N would desorb in short timeframes (hours) when terrestrial sediment enters estuaries because of the large concentration of cations in seawater (Na+, K+, Ca+2, Mg+2) though adsorption/desorption processes will also depend on the ammonium-N concentration (Boatman and Murray, 1982). In contrast, PON bioavailability is determined by microbial mineralization processes that depend on external factors such as temperature and moisture, and the ease of degradation (lability) of the organic matter in the sediments (See conceptual diagram in Fig. 7). This lability is also linked to sediment source but depends on characteristics of the sediment intricately associated with organic matter as explained below. The mineralization of PON is a continuous process with longer timeframes (days) that occurs in broader spatial scales as sediment is transported through catchments and continues to occur in the marine environment. Based on this conceptual understanding it is not surprising to find that all of the proposed indicator equations include organic C, which is not traditionally assessed to quantify the effect of nutrients on phytoplankton. The C and nutrient (N and P) cycles are tightly coupled in terrestrial, freshwater and marine systems (Gruber and Galloway, 2008). This is a direct consequence of the presence of life, which links these cycles at the molecular level. Carbon appears to be even more important in freshwater than in marine conditions. The first parameter in the freshwater indicator Eqs. (3) and (4) relates to the existing organic C pool, whereas in marine conditions, the first parameter is related to N. The C to N ratios, POC/PN and SOC/NH+ 4 -N were selected as parameters in the indicator equations for marine conditions, while POC/PN

1424


Fig. 5. Linear regressions between modelled and measured phytoplankton response with N as a limiting element (PNBfw) and with both N and P as limiting elements (PNPBfw) as measured by phytopam in phytoplankton bioassay (a) and chlorophyll a (b) to particulate nutrient bioavailability in the “freshwater calibration” experiment (Experiment 3), and between modelled and measured phytoplankton response with N as a limiting element (PNBfw) and with both N and P as limiting elements (PNPBfw) as measured by phytopam in phytoplankton bioassay (c) and chlorophyll a (d) to particulate nutrient bioavailability in the freshwater “validation experiment” (Experiment 4). The first indicator equation (PNBfw1) or regression includes POC and POC/PN (black diamonds), the second (PNBfw2) includes SOC and SOC/SON (dark grey circles) and the third (PNPBfw1) includes NH+ 4 -N, DRP and SOC/ SON (light grey triangles). Lines represent a linear regression (y = ax + b) with b = 0. Multiple linear equation parameters and R2 for (a) and (b) are presented in Table 3.

and SOC/SON were selected for freshwater conditions. The C to N ratio is commonly used as a simple indicator of the quality of organic matter for mineralization (Manzoni and Porporato, 2009; Sollins et al., 1984; Whalen et al., 2000); this lability (ease of degradation) ultimately determines the potential mineralization of the PON (conversion to inorganic N which is readily bioavailable). The selection of soluble organic matter fractions (SOC, SON) and their ratios (SOC:SON) as components of the indicator equations in freshwater is related to the important role of soluble fractions in microbial metabolism. Both particulate and soluble organic matter (POM, SOM) are metabolized microbially, but enzymatic action is necessary for POM to become bioavailable as dissolved compounds (Pusch et al., 1998). Soluble organic matter is an important intermediate in stream C cycling (Battin et al., 2009) and its production and composition are

affected by that of POM, which may play a role in stream N processing (Stelzer et al., 2014; Yoshimura et al., 2010). Additionally, phytoplankton is capable of directly using simple forms of dissolved organic N (urea and dissolved free amino acids) to fulfil its nutrient requirements (Bronk et al., 2007). Indicator Eq. (5) for freshwater conditions (PNPBfw1) includes parameters that relate to the bioavailability of both sediment N and P (adsorbed ammonium, SOC:SON and DRP). These parameters are associated with the release of mineral N and P with time through the mineralization of PON and POP by microorganisms and the desorption of chemically bound ammonium-N and P. Indicator Eq. (5) includes the adsorbed ammonium-N and a measure of the fraction of PP that is released into solution (DRP). The DRP index (Colwell-P/PBI) has been previously proposed as an environmental risk indicator for soil P status

Table 3 Multiple linear equation parameters and fit for selected indicators to estimate phytoplankton response (yield and chlorophyll a) to particulate nutrient bioavailability (PNBfw, PNPBfw) in freshwater conditions (Experiment 4). Dependent variable (indicator)

Parameter 1

Parameter 2

PNBfw1 (yield at 72 h) PNBfw2 (yield at 72 h) PNPBfw1 (yield at 72 h) Yield at 72 h Chlorophyll a Chlorophyll a Chlorophyll a Chlorophyll a

+0.015 POC** +3.387 SOC +13.28 NH+ 4 -N** +0.148 PN +0.677 POC*** +59.592 SOC +352.4 NH+ 4 -N*** +6.834 PN

−0.001 POC/PN +0.001 SOC/SON +451.7 DRP

+0.0005 SOC/SON

−0.055 POC/PN −0.015 SOC/SON −9052 DRP.

−0.018 SOC/SON.

(***) p b 0.001, (**) p b 0.01, (*) p b 0.05, (.) p b 0.1. The units of all equation parameters are based on a weight basis e.g., mg POC.

Parameter 3

b

R2

+0.467 +0.438 +0.451 +0.458 +2.764 +2.625 +2.95 +2.289

0.51 0.72 0.58 0.42 0.76 0.37 0.87 0.64


1425

the b63 μm fraction even though they were derived from laboratory experiments conducted on the b10 μm fraction. 4.2. Implications for management

Fig. 6. Indicator equations of phytoplankton response to particulate nutrient bioavailability in marine water (PNBm) applied to lab-generated fine sediment (b10 μm). Values are grouped by erosion source [surface (surf) versus subsurface (sub)] for Bowen River sub-catchment sediments (a) and erosion source for all soil type x land use combinations in both the Johnstone and Bowen River catchments (b). Parent soil types are abbreviated as in Supplementary material S2. Subdivisions in the x axis for Figure b, indicate land use: g (grazing), b (banana), c (cane), f (forest), d (dairy).

(Moody, 2011) and the current research points to the importance of accounting for the particulate P bioavailability to phytoplankton in freshwater. The indicators performed well when applied to sediment from highflow events, which had a different range of particle sizes to the labgenerated sediments (b10 μm) used to determine the indicators. This finding supports the use of the indicator equations to evaluate the potential bioavailability of fine sediment samples that have been transported and collected under natural conditions. Sampled highflow event sediment in both catchments was predominantly fine, consistent with previous findings (Turner et al., 2013). Although further testing around particle size is required, the results above provide a reasonable level of confidence that the indicator equations can be used on

The results of this work suggest that the measurement of the bioavailable particulate nutrient pools selected for the indicator equations, including the different metrics of organic matter lability (i.e. C:N ratios), help explain the relative bioavailability of particulate nutrients and improve our ability to estimate phytoplankton response to eroded sediments in freshwater and marine conditions compared to measurements of PN and PP alone. Separating PN and PP into bioavailable pools also enables us to understand which components of the total N and P pools the phytoplankton are responding to, thereby giving us additional insights when determining appropriate management actions. For these reasons it can be argued that bioavailable nutrient parameters should be included in future water quality monitoring and research in the GBR catchments and other sensitive systems globally. Future work on the role of sediment C content in nutrient bioavailability to phytoplankton would aid in understanding the role of rivers in global carbon budgets (Cole et al., 2007; Lal, 2003); these are of particular importance in regions of the world where high erosion rates control riverine C inputs to the ocean (Bauer et al., 2013). Assessing the bioavailability of different sediment types is the first step to target the sediments (and corresponding parent soils) posing the higher risk to the aquatic environment and the GBR as part of erosion management programs. This knowledge needs to be combined with information on the quantity of sediment generated from different soil type/land use combinations and erosion processes in catchments to develop bioavailable nutrient budgets to support targeting and prioritisation. The sources of sediments with the highest bioavailability do not necessarily correspond to the sources producing the greatest yields of sediment. For example, bioavailability per mass unit of eroded sediment tended to be higher if the sediment was sourced from surface soils than if it was sourced from subsurface soils (Fig. 6a). In the Bowen River catchment, tracing studies have demonstrated that subsurface sources dominate sediment budgets, with between 83 and 95% contribution to catchment export (Wilkinson et al., 2015), whilst in the Johnstone River catchment surface sources dominate (Bartley et al., 2017). Additionally, subsurface sediment sourced from soil types with high erodibility potential (Sodosols and Chromosols) had the lowest particulate nutrient bioavailability of all sediment types in the Bowen River catchment. Nutrient bioavailability also varied depending on the parent soil type and land use. Sediment sourced from Ferrosols tended to have the highest particulate nutrient bioavailability of all parent source soil types, and a higher bioavailability than Dermosols in the Johnstone River catchment (Fig. 6b). Sediments sourced from Dermosols tended to have the highest particulate nutrient bioavailability potential irrespective of erosion process in the Bowen River catchment (Fig. 6b). Sediments sourced from soils under forest land use tended to have the highest particulate nutrient bioavailability potential per unit mass of

Fig. 7. Conceptual diagram of the processes and controls that determine the bioavailability of the different fractions of particulate nitrogen in sediment.

1426


eroded sediment in the Bowen catchment; in the Johnstone catchment, sediment sourced from soils under dairy land use had similarly high values to soils under forest land use, followed by sediment sourced from soil under cultivated banana and sugar cane land uses (Fig. 6b). The contribution of PN to DIN generation as sediment is transported through rivers to the GBR has not been previously quantified, nor its contribution to DIN generation in estuaries and coastal waters. Future research should investigate the contribution of PN from anthropogenic erosion to DIN at end of catchments. Understanding this contribution would assist in achieving water quality targets, i.e., a 50% reduction in anthropogenic DIN at end of catchments by 2018 (Queensland Government, 2013a). Quantifying DIN generation from different sediment types will also help managers to prioritize erosion control and to report on reductions to DIN generation when investing in these programs. 5. Conclusions Particulate nutrients associated with fine sediment are bioavailable to both freshwater and marine phytoplankton of the GBR and its catchments. This implies that eroded sediment is contributing to GBR eutrophication. The magnitude of this bioavailability depends not only on the sediment load, but on sediment characteristics associated with its parent soil. These characteristics vary with soil type, land use and erosion process. It is possible to assess the potential bioavailability of fine sediments of different provenance using indicator equations composed of various particulate nutrient bioavailability parameters that were developed in this study. The indicator equations proposed in this research perform better than traditional methods to monitor and measure the potential bioavailability of particulate nutrients (i.e., PN and PP). Monitoring the selected particulate nutrient bioavailability parameters would also give insight into the underlying biogeochemical processes driving particulate nutrient bioavailability which may in turn help to refine management strategies and prioritisation. These parameters include organic carbon, different particulate nutrient fractions (adsorbed ammonium, particulate organic N and dissolved reactive P) and ratios of C to N, which relate to the lability of sediment organic matter. The outcomes of this research will assist in the management of nutrient pollution in the GBR. The proposed particulate nutrient bioavailability indicators combined with information on the quantity of sediment produced from different soil types and the future development of pedo-transfer functions will enable the identifications and prioritisation of catchment areas with higher risk of promoting phytoplankton growth. These areas may not correspond to the main sediment sources and should be prioritised for erosion management. Additionally, we foresee the use of the indicators to compare the potential impact of sediment on freshwater and marine phytoplankton primary productivity between different catchments, sampling sites, sediment types, gully outlets or positions in a flood plume. Supplementary data to this article can be found online at https://doi. org/10.1016/j.scitotenv.2018.04.334. Author contributions JB, MB, PM, and RD conceived the ideas, AG and HF further developed the methodology and undertook the technical work and data analysis. AG, JB and HF wrote the majority of the manuscript, PM, RD and MB contributed to editing of the manuscript. Acknowledgements This research was funded by the Reef Programs Science Program and the Chemistry Centre, Landscape Sciences, Department of Environment and Science (DES) funded through grant RP128G Phase 2. For assistance with field work and soil and water sampling we thank Peter Zund and Luke Finn, DES; Benjamin Hall and Stephen Faggotter, Griffith University;

Bronwyn Masters, Nikita Tahir and Christina Mortimore DNRM; John Armour; Steve Lewis, JCU; Jack Roberston and Stewart Lindsay, Department of Agriculture Fisheries; Debra Telford, Cane Growers; and Peter Rothlisberg. For assistance with bioassay set up and sampling we thank Jianyin Huang and Jing Lu (Griffith University) and for help with chlorophyll a analysis we thank Yolande Burford (Griffith University). A special thankyou to Ann Chuang for her time and expertise in analysing the algal community samples, and her assistance with the bioassay set-up and sampling. Soil and sediment sample processing and analysis was conducted by the Chemistry Centre and Soil Processes DES, Queensland Government. In particular we thank Benjamin Hall, Wayne Hall, David Wardle, Siok Yo, Kate Dolan, Luke Kitchens, Dan Yousaf, Cathy McCombes, Sonya Mork, Joshua Hansen and Angus McElnea. Water sample processing and analysis was conducted by the Chemistry Centre and Soil Processes DES. In particular we thank Lisa Finocchiaro, Gail Zerk, Nan Lian and Fred Oudyn. We also thank Dr. Jon Brodie, Centre of Excellence for Coral Reef Studies, James Cook University and Dr. Reinier Mann, DES, for the scientific review of the research project and Dr. Di Allen, DES, and two anonymous reviewers for their review of this manuscript. References APHA/AWWA/WPCF, 2012. Standard Methods for the Examination of Water and Wastewater. 21st and 22nd edition. American Public Health Association, Washington D.C. Bainbridge, Z.T., Wolanski, E., Alvarez-Romero, J.G., Lewis, S.E., Brodie, J.E., 2012. Fine sediment and nutrient dynamics related to particle size and floc formation in a Burdekin River flood plume, Australia. Mar. Pollut. Bull. 65, 236–248. Bainbridge, Z.T., Lewis, S.E., Smithers, S.G., Kuhnert, P.M., Henderson, B.L., Brodie, J.E., 2014. Fine suspended sediment and water budgets for a large, seasonally dry tropical catchment: Burdekin River catchment, Queensland, Australia. Water Resour. Res. 50, 9067–9087. Bartley, R., Waters, D., Turner, R., Kroon, F., Wilkinson, S., Garzon-Garcia, A., et al., 2017. Sources of sediment, nutrients, pesticides and other pollutants to the Great Barrier Reef. Scientific Consensus Statement 2017: A Synthesis of the Science of Land-based Water Quality Impacts on the Great Barrier Reef. State of Queensland, Queensland. Battin, T.J., Kaplan, L.A., Findlay, S., Hopkinson, C.S., Marti, E., Packman, A.I., et al., 2009. Biophysical controls on organic carbon fluxes in fluvial networks (vol 1, pg 95, 2008). Nat. Geosci. 2 (595–595). Bauer, J.E., Cai, W.J., Raymond, P.A., Bianchi, T.S., Hopkinson, C.S., Regnier, P.A.G., 2013. The changing carbon cycle of the coastal ocean. Nature 504, 61–70. Benstead, J.P., Rosemond, A.D., Cross, W.F., Wallace, J.B., Eggert, S.L., Suberkropp, K., et al., 2009. Nutrient enrichment alters storage and fluxes of detritus in a headwater stream ecosystem. Ecology 90, 2556–2566. Bianchi, T.S., Bauer, J.E., 2011. Particulate organic carbon cycling and transformation. Treatise on estuarine and coastal science. Biogeochemistry 5, 69–117. Boatman, C.D., Murray, J.W., 1982. Modeling exchangeable NH+ 4 adsorption in marine sediments - process and controls of adsorption. Limnol. Oceanogr. 27, 99–110. Brodie, J., Fabricius, K., De'ath, G., Okaji, K., 2005. Are increased nutrient inputs responsible for more outbreaks of crown-of-thorns starfish? An appraisal of the evidence. Mar. Pollut. Bull. 51, 266–278. Brodie, J., De'ath, G., Devlin, M., Furnas, M., Wright, M., 2007. Spatial and temporal patterns of near-surface chlorophyll a in the Great Barrier Reef lagoon. Mar. Freshw. Res. 58, 342–353. Brodie, J.E., Devlin, M., Haynes, D., Waterhouse, J., 2011. Assessment of the eutrophication status of the Great Barrier Reef lagoon (Australia). Biogeochemistry 106, 281–302. Bronk, D.A., See, J.H., Bradley, P., Killberg, L., 2007. DON as a source of bioavailable nitrogen for phytoplankton. Biogeosciences 4, 283–296. Burdige, D.J., 2005. Burial of terrestrial organic matter in marine sediments: a reassessment. Glob. Biogeochem. Cycles 19. Cole, J.J., Prairie, Y.T., Caraco, N.F., McDowell, W.H., Tranvik, L.J., Striegl, R.G., et al., 2007. Plumbing the global carbon cycle: integrating inland waters into the terrestrial carbon budget. Ecosystems 10, 171–184. Davies, C.H., Coughlan, A., Hallegraeff, G., Ajani, P., Armbrecht, L., Atkins, N., et al., 2016. A database of marine phytoplankton abundance, biomass and species composition in Australian waters. Sci. Data 3. De'ath, G., Fabricius, K., 2010. Water quality as a regional driver of coral biodiversity and macroalgae on the Great Barrier Reef. Ecol. Appl. 20, 840–850. DeVantier, L.M., De'ath, G., Turak, E., Done, T.J., Fabricius, K.E., 2006. Species richness and community structure of reef-building corals on the nearshore Great Barrier Reef. Coral Reefs 25, 329–340. Fabricius, K.E., 2005. Effects of terrestrial runoff on the ecology of corals and coral reefs: review and synthesis. Mar. Pollut. Bull. 50, 125–146. Fabricius, K.E., Okaji, K., De'ath, G., 2010. Three lines of evidence to link outbreaks of the crown-of-thorns seastar Acanthaster planci to the release of larval food limitation. Coral Reefs 29, 593–605. Fabricius, K.E., Logan, M., Weeks, S., Brodie, J., 2014. The effects of river run-off on water clarity across the central Great Barrier Reef. Mar. Pollut. Bull. 84, 191–200. Franklin, H.M., Garzon-Garcia, A., Joanne, B., Moody, P., De Hayr, R., Burford, M.A., 2018. A novel bioassay to assess phytoplankton responses to soil-derived particulate nutrients. Sci. Total Environ. https://doi.org/10.1016/j.scitotenv.2018.04.195 (in press).

A. Garzon-Garcia et al. / Science of the Total Environment 636 (2018) 1416–1427 Furnas, M., Mitchell, A., Skuza, M., Brodie, J., 2005. In the other 90%: phytoplankton responses to enhanced nutrient availability in the Great Barrier Reef lagoon. Mar. Pollut. Bull. 51, 253–265. Garten, C.T., Ashwood, T.L., 2002. Landscape level differences in soil carbon and nitrogen: implications for soil carbon sequestration. Glob. Biogeochem. Cycles 16, 61–1-61-14. Garzon-Garcia, A., Wallace, R., Huggins, R., Turner, R.D.R., Smith, R.A., Orr, D., et al., 2015. Total suspended solids, nutrient and pesticide loads (2013–2014) for rivers that discharge to the Great Barrier Reef. Great Barrier Reef Catchment Loads Monitoring Program. Water Quality and Investigations, Environmental Monitoring and Assessment Science, Science Division, Department of Science, Information Technology and Innovation. Garzon-Garcia, A., Laceby, J.P., Olley, J.M., Bunn, S.E., 2017. Differentiating the sources of fine sediment, organic matter and nitrogen in a subtropical Australian catchment. Sci. Total Environ. 575, 1384–1394. Gomez, B., Trustrum, N.A., Hicks, D.M., Rogers, K.M., Page, M.J., Tate, K.R., 2003. Production, storage, and output of particulate organic carbon: Waipaoa River basin, New Zealand. Water Resour. Res. 39, 1161–1168. Gregorich, E.G., Beare, M.H., McKim, U.F., Skjemstad, J.O., 2006. Chemical and biological characteristics of physically uncomplexed organic matter. Soil Sci. Soc. Am. J. 70, 975–985. Gruber, N., Galloway, J.N., 2008. An Earth-system perspective of the global nitrogen cycle. Nature 451, 293–296. Haapkyla, J., Unsworth, R.K.F., Flavell, M., Bourne, D.G., Schaffelke, B., Willis, B.L., 2011. Seasonal rainfall and runoff promote coral disease on an inshore reef. PLoS One 6. Horowitz, A.J., Elrick, K.A., 1987. The relation of stream sediment surface area, grain size and composition to trace element chemistry. Appl. Geochem. 2, 437–451. Howarth, R.W., Jensen, H.S., Marino, R., Postma, H., 1995. Transport to and processing of P in near-shore and oceanic waters. In: Tiessen, H. (Ed.), Phosphorus in the Global Environment: Transfers, Cycles and Management. John Wiley and Sons, New York. Juarez, S., Rumpel, C., MChunu, C., Chaplot, V., 2011. Carbon mineralization and lignin content of eroded sediments from a grazed watershed of South-Africa. Geoderma 167–168, 247–253. Kao, S.J., Liu, K.K., 2000. Stable carbon and nitrogen isotope systematics in a human disturbed watershed (Lanyang-Hsi) in Taiwan and the estimation of biogenic particulate organic carbon and nitrogen fluxes. Glob. Biogeochem. Cycles 14, 189–198. Keil, R.G., Mayer, L.M., Quay, P.D., Richey, J.E., Hedges, J.I., 1997. Loss of organic matter from riverine particles in deltas. Geochim. Cosmochim. Acta 61, 1507–1511. Lal, R., 2003. Soil erosion and the global carbon budget. Environ. Int. 29, 437–450. Ludwig, W., Probst, J., 1996. Predicting the oceanic input of organic carbon by continental erosion. Glob. Biogeochem. Cycles 10, 23–41. Lumley, T., 2017. leaps: regression subset selection. R Package. Mackin, J.E., Aller, R.C., 1984. Ammonium adsorption in marine-sediments. Limnol. Oceanogr. 29, 250–257. Manzoni, S., Porporato, A., 2009. Soil carbon and nitrogen mineralization: theory and models across scales. Soil Biol. Biochem. 41, 1355–1379. Mayer, L.M., Keil, R.G., Macko, S.A., Joye, S.B., Ruttenberg, K.C., Aller, R.C., 1998. Importance of suspended particulates in riverine delivery of bioavailable nitrogen to coastal zones. Glob. Biogeochem. Cycles 12, 573–579. Moody, P.W., 2011. Environmental risk indicators for soil phosphorus status. Soil Res. 49, 247–252. O'Donohue, M., Glibert, P., Dennison, W., 2000. Utilization of nitrogen and carbon by phytoplankton in Moreton Bay, Australia. Mar. Freshw. Res. 51, 703–712. Oksanen, J., Blanchet, F.G., Friendly, M., Kindt, R., Lagendre, P., McGlinn, D., et al., 2017. vegan: Community Ecology Package. Phillips, J.M., Russell, M.A., Walling, D.E., 2000. Time-integrated sampling of fluvial suspended sediment: a simple methodology for small catchments. Hydrol. Process. 14, 2589–2602. Pusch, M., Fiebig, D., Brettar, I., Eisenmann, H., Ellis, B.K., Kaplan, L.A., et al., 1998. The role of micro-organisms in the ecological connectivity of running waters. Freshw. Biol. 40, 453–495. Queensland Government, 2013a. Reef water quality protection plan 2013. Reef Water Quality Protection Plan Secretariat, Brisbane.

1427

Queensland Government, 2013b. Water Monitoring Information Portal. 2017. R Core Team, 2017. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Rosenfeld, J.K., 1979. Ammonium adsorption in nearshore anoxic sediments. Limnol. Oceanogr. 24, 356–364. Rowland, R., Inamdar, S., Parr, T., 2017. Evolution of particulate organic matter (POM) along a headwater drainage: role of sources, particle size class, and storm magnitude. Biogeochemistry 133, 181–200. Saeck, E.A., O'Brien, K.R., Burford, M.A., 2016. Nitrogen response of natural phytoplankton communities: a new indicator based on photosynthetic efficiency F-v/F-m. Mar. Ecol. Prog. Ser. 552, 81–92. Sholkovitz, E.R., 1976. Flocculation of dissolved organic and inorganic matter during mixing of river water and seawater. Geochim. Cosmochim. Acta 40, 831–845. Sollins, P., Spycher, G., Glassman, C.A., 1984. Net nitrogen mineralization from lightfraction and heavy-fraction forest soil organic-matter. Soil Biol. Biochem. 16, 31–37. Stelzer, R.S., Scott, J.T., Bartsch, L.A., Parr, T.B., 2014. Particulate organic matter quality influences nitrate retention and denitrification in stream sediments: evidence from a carbon burial experiment. Biogeochemistry 119, 387–402. Tank, J.L., Rosi-Marshall, E.J., Griffiths, N.A., Entrekin, S.A., Stephen, M.L., 2010. A review of allochthonous organic matter dynamics and metabolism in streams. J. N. Am. Benthol. Soc. 29, 118–146. Turner, R.D.R., Huggins, R.L., Wallace, R.M., Smith, R.A., Vardy, S., J Warne, M.J., 2012. Sediment, Nutrient and Pesticide Loads: Great Barrier Reef Catchment Loads Monitoring 2009–2010. Department of Science, Information Technology, Innovation and the Arts, Brisbane. Turner, R., Huggins, R., Wallace, R., Smith, R., Vardy, S., MSJ, Warne, 2013. Total suspended solids, nutrient and pesticide loads (2010–2011) for rivers that discharge to the Great Barrier Reef. Great Barrier Reef Catchment Loads Monitoring. Department of Science, Information Technology, Innovation and the Arts, Brisbane. Udy, J.W., Fellows, C.S., Bartkow, M.E., Bunn, S.E., Clapcott, J.E., Harch, B.D., 2006. Measures of nutrient processes as indicators of stream ecosystem health. Hydrobiologia 572, 89–102. Wallace, R., Huggins, R., Smith, R., Turner, R., Vardy, S., Warne, M.S.J., 2014. Total suspended solids, nutrient and pesticide loads (2011–2012) for rivers that discharge to the Great Barrier Reef. Great Barrier Reef Catchment Loads Monitoring Program. Department of Science, Information Technology, Innovation and the Arts, Brisbane. Wallace, R., Huggins, R., Smith, R., Turner, R., Garzon-Garcia, A., Warne, M.S.J., 2015. Total suspended solids, nutrient and pesticide loads (2012–2013) for rivers that discharge to the Great Barrier Reef. Great Barrier Reef Catchment Loads Monitoring Program. Department of Science, Information Technology, Innovation and the Arts, Brisbane. Walling, D.E., Moorehead, P.W., 1989. The particle size characteristics of fluvial suspended sediment: an overview. Hydrobiologia 176–177, 125–149. Whalen, J.K., Bottomley, P.J., Myrold, D.D., 2000. Carbon and nitrogen mineralization from light- and heavy-fraction additions to soil. Soil Biol. Biochem. 32, 1345–1352. Wilkinson, S.N., Olley, J.M., Furuichi, T., Burton, J., Kinsey-Henderson, A.E., 2015. Sediment source tracing with stratified sampling and weightings based on spatial gradients in soil erosion. J. Soils Sediments 15, 2038–2051. Wooldridge, S.A., 2009. Water quality and coral bleaching thresholds: formalising the linkage for the inshore reefs of the Great Barrier Reef, Australia. Mar. Pollut. Bull. 58, 745–751. Yoshimura, C., Fujii, M., Omura, T., Tockner, K., 2010. Instream release of dissolved organic matter from coarse and fine particulate organic matter of different origins. Biogeochemistry 100, 151–165. Zhang, L., Wang, S.R., Jiao, L.X., Li, Y.P., Yang, J.C., Zhang, R., et al., 2016. Effects of organic matter content and composition on ammonium adsorption in lake sediments. Environ. Sci. Pollut. Res. 23, 6179–6187. Zund, P.R., Payne, J.E., 2014. In: Department of Science ITaI (Ed.), Erodible Soils Map for Burdekin Dry Tropics Grazing Lands (Brisbane).