Speciation and bioavailability of particulate phosphorus in forested karst watersheds of southern Ontario during rain events
Introduction
Phosphorus is the principal driver of eutrophication in lakes (Elser and Bennett, 2011, Goyette et al., 2018), leading to significant efforts being made to reduce and control P inputs from anthropogenic sources. Controlling water quality issues and reversing the development of eutrophication in anthropogenically impacted aquatic environments depends on the delineation of point and non-point sources of P as well as characterization of its chemical speciation and bioavailability (Nausch et al., 2017, Bol et al., 2018, Bitschofsky and Nausch, 2019). Characterization of P sources is particularly important in agricultural watersheds, where non-point source P pollution may persist despite reduced P use reflecting net mobilization of legacy P pools (Powers et al., 2016). Owing to the presence of a subterranean network of conduits that facilitate P transport, agricultural karst watersheds are generally considered to be especially vulnerable to P losses from non-point sources during storm events (Kilroy and Coxon, 2005, Kilroy et al., 2005). Although, watershed P export driven by surface discharge of groundwater during storm events has been studied before (Kilroy and Coxon, 2005, Schilling and Helmers, 2008a, Schilling and Helmers, 2008b, Mellander et al., 2013, Jarvie et al., 2014), the dynamics of P chemical speciation during low and high flow regimes have received less attention. This is a significant omission because flushing of groundwater during storm events is associated with increased delivery of highly reactive particulate and dissolved P species, thereby negatively affecting the water quality of receiving aquatic systems (Mavrocordatos et al., 2000, Kilroy and Coxon, 2005). Moreover, considering increasing trends in extreme precipitation events and discharge in mid- and high-latitude regions (Bush and Lemmen, 2019, Alexander et al., 2006, Fischer and Knutti, 2015), climate change with more extreme events could promote P mobilization from non-point sources in agricultural watersheds (Wu et al., 2012, Mellander et al., 2015, Mellander et al., 2018, Bol et al., 2018).
P control programs under the auspices of the Great Lakes Water Quality Agreement (GLWQA) in Canada and the United States have led to a considerable reduction in the impact of anthropogenic point sources of P, which include municipal and industrial wastewater (Dolan and Chapra, 2012). While the substantial reduction of these sources has been credited with the notable decrease in P levels in the Great Lakes basin compared to the period before point source control, extensive phytoplankton blooms and associated impairment of beneficial uses are still a common occurrence in many aquatic systems. This invites the question of what drives lake eutrophication during periods of stable or declining anthropogenic point source loads. Recently, tributary non-point P inputs have been implicated as the key driver of algal resurgence (Baker et al., 2014, Kane et al., 2014, Jarvie et al., 2017). The temporal distribution of these watershed inputs is thought to be seasonal with generally elevated loads occurring during spring and winter (Stow et al., 2015, Stumpf et al., 2012, Michalak et al., 2013), and in some systems intermittent significant increases during summer associated with storm events (Long et al., 2015). Despite their importance, in many systems, non-point sources are poorly characterized and consequently challenging to mitigate. While many case studies have demonstrated that mitigation of lake eutrophication cannot be separated from the management of non-point P sources in the catchment, management efforts have seldom incorporated a comprehensive understanding of P mobilization processes and its chemical speciation which remains a significant impediment for effective non-point P control strategies (Bol et al., 2018).
Chemical speciation of riverine P determines its bioavailability and fate during biogeochemical cycling in the water column and sediments of receiving aquatic systems. Even if the total riverine P loading is reduced, changes in the relative availability of bioavailable forms of P can still promote eutrophication in receiving lakes. For instance, ongoing (re-)eutrophication of Lake Erie is attributed to an increase of bioavailable P loading (Maccoux et al., 2016, Baker et al., 2014, Kane et al., 2014, Jarvie et al., 2017). Therefore, development of effective and sustainable mitigation measures to reduce unaccounted for non-point P losses from the watershed involves identification of sources and transport pathways and necessitates quantification of P binding forms, including bioavailable phosphorus, and their temporal and spatial dynamics (Sharpley et al., 1993, Sharpley et al., 2015, Bol et al., 2018).
Riverine P as a component of watershed P losses exists in both the dissolved/colloidal phase and as various mineral and organic particulate P (PP) phases or binding forms. The former is generally considered to be the most bioavailable, while the latter often dominates P transport and cycling in riverine systems (Reynolds and Davies, 2001). Given the crucial role of PP in riverine P transport, its chemical speciation needs to be monitored to estimate its impact on eutrophication of receiving lakes (Boström et al., 1988b, Bol et al., 2018). While many phosphorus management and monitoring programs in watersheds of eutrophic lakes routinely estimate the proportion of total PP, few studies have comprehensively considered PP chemical speciation and accounted for the relative significance of bioavailable PP forms (e.g., Pacini and Gachter, 1999, Shinohara et al., 2018). This lack of comprehensive understanding of chemical speciation of PP, its transport and transformation mechanisms and the relative proportions of bioavailable forms negatively impacts nutrient management in the watershed and leads to inaccurate assessment of P loading reduction that is needed to combat eutrophication in receiving lakes (Sharpley et al., 1993, Bol et al., 2018). Especially in case of P control efforts in persistently eutrophic lakes, complex interaction of often poorly constrained non-point PP external inputs, recycling of legacy phosphorus in sediments, changes of food web structure and climate patterns frustrate extensive remedial efforts focussed on point source P loading reduction and result in continuing degradation of water quality (e.g., Baker et al., 2014, Kane et al., 2014, Jarvie et al., 2017, Bol et al., 2018). Accurate assessment and monitoring of chemical speciation and bioavailable P loading from non-point sources in such systems is necessary to improve understanding of transport mechanisms and the fate of P and design effective remedial strategies in such systems.
Elevated water column total phosphorus (TP) levels and nuisance algal blooms are recurrent water quality issues in the Bay of Quinte, one of the Areas of Concern (AOC) under GLWQA (e.g., Shimoda et al., 2016, Markovic et al., 2019). Tributary P inputs in the Bay of Quinte are presently dominated by non-point sources in predominantly rural and forested watersheds (Kim et al., 2016, Kim et al., 2017) situated in a region with extensive development of karst features and generally thin soil cover. Past modeling approaches, in the Bay of Quinte, have been used to identify catchments with high annual TP loadings, and these models were calibrated with data collected during low precipitation conditions (e.g., baseflows) because event-based monitoring of nutrients was not available previously (Kim et al., 2016, Kim et al., 2017). Currently, planned and implemented P management strategies and practices designed to reduce P loss from non-point sources are reliant on modeling estimates of TP export from the watershed (Kim et al., 2016, Kim et al., 2017). However, this approach cannot determine the portion of TP that is bioavailable to phytoplankton, and consequently, there is residual uncertainty about the long-term effectiveness of mitigation measures. This is particularly important in the context of climate change which is predicted to increase bioavailable P exports from the watersheds due to combination of factors including temperature increase, elevated frequency and intensity of storm events leading to amplified erosion and leaching of P from soils in conjunction with enhanced organic P mobilization mediated by increased microbial metabolic activity at higher temperatures (Jeppesen et al., 2009, Mellander et al., 2015, McDowell et al., 2017).
In this context, our study aimed to test the hypothesis that bioavailable PP chemical forms dominate riverine PP loading in predominantly karst watersheds, which can be related to increased groundwater-surface water interactions during storm events. Furthermore, we hypothesized that increased delivery of riverine reactive bioavailable PP species accompanies active regeneration of P from particulate to dissolved phase in the water column and sediments of river outlets and thus exacerbates ongoing eutrophication in the Bay of Quinte AOC.
We investigated the dynamics of PP binding forms and the relative contributions of bioavailable P phases in two predominantly forested and agricultural karst watersheds in the Bay of Quinte AOC (Napanee river and Wilton Creek) which have been previously identified as nutrient hotspots regarding non-point P sources (Kim et al., 2016, Kim et al., 2017). The aims of this study were: 1) to quantify distribution of different PP binding forms using sequential extraction techniques and elemental microanalysis, 2) to determine spatio-temporal changes in the PP binding forms during periods of low and high flow, with emphasis on changing contributions of bioavailable and inert PP species and 3) to compare distribution of binding forms of riverine PP and surface sediment in a receiving lake near a river inflow in order to evaluate PP transformation pathways in the water column and sediment and its impact on ongoing eutrophication processes in the Bay of Quinte.
Understanding the processes of PP mobilization and transport remains a crucial challenge for nutrient management strategies targeting non-point P sources. This study investigates how PP binding forms vary temporally and spatially and under contrasting hydrological regimes and provides fresh insights into processes controlling PP mobilization and transport in the karst watersheds. We show that bioavailable PP loading in karst watersheds increases during storm events which has important implications for assessment of eutrophication risk associated with changing climate in similar systems worldwide, particularly in areas that are expected to experience increased intensity and frequency of extreme precipitation events such as Great Lakes region.
Section snippets
Study area
Napanee River and Wilton Creek are the principal tributaries in the middle section of the Bay of Quinte, a Z shaped embayment on the northern shore of Lake Ontario. These watersheds are characterized by predominantly agricultural and forested land cover (Fig. 1).
Napanee River originates in the Depot Lakes system, draining a catchment area of 814 km2. The main tributaries are Varty Creek and Hardwood Creek. The average discharge is ∼ 9 m3/s. Wilton Creek has a catchment area of ∼ 223 km2 and an
Sample collection and processing
Sample collecting stations were installed at three locations along the Napanee River: Springside Park (Nap), Newburgh (NewB), Camden East (CamE), and one station was located at Wilton Creek (WiltC) in Morven (Fig. 1.). All sampling locations are located downstream from headwaters, and thus the effect of various processes of P mobilization in headwater sub-watersheds is represented in the aggregate. However, as a result of the proximity of sampling sites to river mouths, we were able to sample
Determination of P binding forms and chemical analyzes
Sediment P binding forms were extracted following the widely used sequential P leaching technique of Psenner and Pucsko (1988), with modifications by Hupfer et al. (2009). The separation of P binding pools is operational, and the respective extraction solutions are meant to mimic natural physicochemical conditions under which specific PP binding forms could be mobilized. Because each P fraction is defined by an extraction solution representing specific pH and redox conditions, the extracted P
Scanning electron microscopy and electron microprobe analyses
Scanning electron microscopy and electron probe X-ray microanalyses were conducted on at least two samples per site (one representative of high flow regime and one representative of baseflow). For scanning electron microscopy, freeze-dried bulk suspended sediment samples were fixed on magnetic tape and carbon-coated prior to analyses on Hitachi SU-3500 SEM (20 kV). Energy-dispersive X-ray spectroscopy (EDS) spectra were collected from selected areas of interest using an Oxford Instruments X-Max
Data analysis and PP load calculations
The Kruskal–Wallis test and multiple pairwise comparisons with the Dunn-Sidak method (Sokal and Rohlf, 1995) was used to evaluate whether PP binding forms and suspended solids content of samples collected at different sites were statistically different. All statistical analyses were conducted using Origin 2018.
We estimated instantaneous PP loading during different hydrological events as a product of measured PP concentrations and average riverine discharge during a 24 h period of sampling. In
Trends of PP concentrations and suspended sediment during low and high flow events
PP concentrations during low flow events in November 2016 varied between 8 and 35 µg P/L and 25–35 µg P/L, at Napanee river and Wilton Creek, respectively. During low flow events in July 2017, PP concentrations were 23–52 µg P/L and 27–35 µg P/L at Napanee and Wilton Creek, respectively. In contrast, during high flow events in July 2017, PP concentrations were 7–40 µg P/L and 25–50 µg P/L at Napanee and Wilton Creek. There is no significant difference in P binding forms and suspended solids at
Dynamics of particulate P binding forms during rain events
Bioavailable P binding forms (NH4Cl-P, BD-P, and NaOH-NRP) dominate PP at all stations and sampling dates (62–75% of total PP, Fig. 4, Fig. 5). Total PP concentrations in suspended particulates were constant during low flow sampling events (2–2.5 mg P/g d.w.) (Fig. 4, Fig. 5). BD-P and NaOH-NRP forms varied between 0.5 and 0.75 mg P/g d.w. and 0.6–0.8 mg P/g d.w. Concentrations of HCl-P varied between 0.2 and 0.35 mg P/g d.w. In contrast to low flow events, concentrations of PP in suspended
SEM-EDS and electron microprobe analysis: Linking mineralogy with potential sediment sources
SEM coupled with EDS elemental analysis and electron microprobe analysis suggests that suspended sediment is composed of aggregates (flocs) with high Al, Si, Fe, and P content likely incorporating iron oxyhydroxides, clay minerals and carbonates (Fig. 6 and ESM Fig. S2). The aggregates were 10–200 μm in diameter with the individual mineral clasts were 1–2 μm (Fig. S2 and Fig. 6). Sporadic individual particles larger than 10 μm mostly represent identifiable diatom frustule (Fig. S2). Electron
PP loading during rain events
PP loading varies between 1.25 and 1.8 kg P/d at Wilton Creek during low flow (Fig. 7). At Napanee river station proximal to the river mouth (Nap) baseflow PP loading varies between 3.1 and 4 kg P/d. Most of the PP load is bioavailable P, which is estimated to be 0.85–1.25 kg P/d and 1.9–2.6 kg P/d (Fig. 7) at Wilton Creek and Napanee stations, respectively.
At Wilton creek, PP load increases by one order of magnitude (11.3–21 kg P/d) in response to a 6-fold increase of water discharge.
Influence of hydrological regimes on concentrations and transport pathways of suspended sediment and PP
At all stations, concentrations of total PP, bioavailable PP forms, and HCl-P on particle dry weight basis decrease with increased suspended solids concentrations (Fig. 3b,d,f). This is consistent with previous findings by Pacini and Gachter (1999) who showed that PP is primarily associated with surface coatings of suspended particles, and PP concentrations on sediment dry mass basis increase with decreasing suspended sediment because lower sediment loads are characterized by the prevalence of
Impact of the dominance of redox-sensitive and organic P forms on P export from Napanee and Wilton Creek watersheds
The prominent feature of PP binding form dynamics at all sites is the dominance of BD-P and NaOH-NRP, as indicated by strong linear trends between BD-P and NaOH-NRP and total PP content of the suspended particulate matter (Fig. 8, see also Fig. 4, Fig. 5).
The distribution of PP binding forms in suspended solids from Napanee and Wilton Creek is similar with earlier water quality monitoring studies in 1990s in adjacent Trent and Moira river tributaries of the Bay of Quinte which showed that
Eutrophication risk related to riverine PP loading
Comparison between binding forms of PP with concentrations of P species in surface sediment near river outlets reveals that PP is readily cycled during deposition with concentrations (on dry mass basis) of most bioavailable forms (NH4Cl-P, BD-P, and NaOH-NRP) decreasing significantly from Napanee river and Wilton creek to receiving lake sediment (Fig. 9). PP in Napanee river and Wilton Creek is predominantly composed of these potentially bioavailable forms, which constitute 62–75% of total PP (
Conclusions
We quantified dynamics of different PP binding forms in suspended sediment from Napanee River and Wilton Creek, major tributaries draining karst watersheds in the Bay of Quinte AOC. We show that highly bioavailable P species (NH4Cl-P, BD-P, and NaOH-NRP) dominate PP loading and comprise 62–75% of total PP. Furthermore, we show that concentrations of bioavailable PP binding forms in suspended solids increase 2–3-fold on a dry mass basis during periods of high river flow. Primary carriers of PP
Acknowledgments
This project was funded by the Great Lakes Sustainability Fund, Great Lakes Areas of Concern, Environment and Climate Change Canada (GCXE19P033) awarded to MD, University of Toronto Scarborough. We are thankful for the generous support by the Environment and Climate Change Canada, Lower Trent Conservation Authority, Bay of Quinte Remediation Action Plan Office, Ontario Ministry of the Environment and Climate Change, and Quinte Conservation Association. We are grateful to Dr. Jay Guo
References (109)
- et al.
Phosphorus loading to Lake Erie from the Maumee, Sandusky and Cuyahoga rivers: The importance of bioavailability
J. Great Lakes Res.
(2014) - et al.
Contrasting controls on the phosphorus concentration of suspended particulate matter under baseflow and storm event conditions in agricultural headwater streams
Sci. Total Environ.
(2015) - et al.
Phosphorus retention in a mesotrophic lake under transient loading conditions: Insights from a sediment phosphorus binding form study
Water Res.
(2013) - et al.
Great Lakes total phosphorus revisited: 1. Loading analysis and update (1994–2008)
J. Great Lakes Res.
(2012) - et al.
Release of dissolved phosphorus from riparian wetlands: Evidence for complex interactions among hydroclimate variability, topography and soil properties
Sci. Total Environ.
(2017) - et al.
Re-eutrophication of Lake Erie: Correlations between tributary nutrient loads and phytoplankton biomass
J. Great Lakes Res.
(2014) - et al.
Evaluating the relationships between watershed physiography, land use patterns, and phosphorus loading in the Bay of Quinte, Ontario, Canada
J. Great Lakes Res.
(2016) - et al.
Variations in chemical speciation and reactivity of phosphorus between suspended-particles and surface-sediment in seasonal hypoxia-influenced Green Bay
J. Great Lakes Res.
(2018) - et al.
Estimation of tributary total phosphorus loads to Hamilton Harbour, Ontario, Canada, using a series of regression equations
J. Great Lakes Res.
(2015) - et al.
Total and soluble reactive phosphorus loadings to Lake Erie: A detailed accounting by year, basin, country, and tributary
J. Great Lakes Res.
(2016)