ArticlePDF Available

Abstract and Figures

Reduction of phosphorus (P) inputs to surface waters may decrease eutrophication. Some researchers have proposed filtering dissolved P in runoff with P-sorptive byproducts in structures placed in hydrologically active areas with high soil P concentrations. The objectives of this study were to construct and monitor a P removal structure in a suburban watershed and test the ability of empirically developed flow-through equations to predict structure performance. Steel slag was used as the P sorption material in the P removal structure. Water samples were collected before and after the structure using automatic samples and analyzed for total dissolved P. During the first 5 mo of structure operation, 25% of all dissolved P was removed from rainfall and irrigation events. Phosphorus was removed more efficiently during low flow rate irrigation events with a high retention time than during high flow rate rainfall events with a low retention time. The six largest flow events occurred during storm flow and accounted for 75% of the P entering the structure and 54% of the P removed by the structure. Flow-through equations developed for predicting structure performance produced reasonable estimates of structure "lifetime" (16.8 mo). However, the equations overpredicted cumulative P removal. This was likely due to differences in pH, total Ca and Fe, and alkalinity between the slag used in the structure and the slag used for model development. This suggests the need for an overall model that can predict structure performance based on individual material properties.
Content may be subject to copyright.
Reduction of phosphorus (P) inputs to surface waters may decrease
eutrophication. Some researchers have proposed filtering dissolved
P in runoff with P-sorptive byproducts in structures placed
in hydrologically active areas with high soil P concentrations.
e objectives of this study were to construct and monitor a P
removal structure in a suburban watershed and test the ability of
empirically developed flow-through equations to predict structure
performance. Steel slag was used as the P sorption material in the
P removal structure. Water samples were collected before and
after the structure using automatic samples and analyzed for total
dissolved P. During the first 5 mo of structure operation, 25% of
all dissolved P was removed from rainfall and irrigation events.
Phosphorus was removed more efficiently during low flow rate
irrigation events with a high retention time than during high flow
rate rainfall events with a low retention time. e six largest flow
events occurred during storm flow and accounted for 75% of the P
entering the structure and 54% of the P removed by the structure.
Flow-through equations developed for predicting structure
performance produced reasonable estimates of structure “lifetime”
(16.8 mo). However, the equations overpredicted cumulative P
removal. is was likely due to differences in pH, total Ca and Fe,
and alkalinity between the slag used in the structure and the slag
used for model development. is suggests the need for an overall
model that can predict structure performance based on individual
material properties.
Trapping Phosphorus in Runo with a Phosphorus Removal Structure
Chad J. Penn,* Joshua M. McGrath, Elliott Rounds, Garey Fox, and Derek Heeren
R   (P) loading to surface
waters can help to prevent eutrophication. Previous
studies have suggested the use of certain industrial by-
products as P sorption materials (PSMs) for reducing P solu-
bility in high-P soils (Leader et al., 2008; Makris and Harris,
2006; Rhoton and Bigham, 2005). Although the addition of
PSMs to high-P soils has been shown to reduce water-soluble
P and therefore losses of dissolved P in runoff (Gallimore et
al., 1999), such reductions in P solubility can be temporary
(Penn and Bryant, 2006). In addition, such an approach does
not truly remove P from the watershed; P pools within the soil
solid phase are simply shifted to less soluble forms.
A potential modification to this approach is a P removal
structure. Such structures can be filled with PSMs and can be
strategically placed in “hot spots” or drainage ditches where
runoff with elevated concentrations of dissolved P regularly
occurs (Penn et al., 2010). e P removal structure is designed
to intercept runoff or subsurface drainage and channels flow
through contained PSMs. After the PSMs become saturated
with P, they can be replaced with new PSMs; using this
approach, P can be effectively removed from the watershed.
Some potential guidelines, theory, and approach for P removal
structure design are presented in Penn et al. (2010). Similarly,
previous studies have used various PSMs for removing P from
wastewaters (Koiv et al., 2010; Cucarella and Renman, 2009;
Wei et al., 2008) and subsurface drainage (McDowell et al.,
2008). A material that has shown tremendous promise as a
PSM in column studies is steel slag (Drizo et al., 2008, 2006,
2002), which is a by-product of the steel industry.
In a previous study, Penn and McGrath (2011) constructed
a pilot scale pond filter that used electric arc furnace steel slag
as the PSM. e authors developed empirical equations based
on laboratory flow-through experiments that predicted struc-
ture performance as a function of retention time (RT) (i.e.,
the time required for one pore volume to pass through the
structure) and inflow P concentration. At a RT of 10 min, the
pond filter removed 34% of the all P pumped into it (172 mg
kg-1 of PSM) at the point of P saturation (i.e., the point at
which P was no longer removed from passing water). e flow-
through equations reasonably predicted structure performance
(P removal and longevity), whereas the Langmuir equation
Abbreviations: DI, deionized; ICP–AES, inductively coupled plasma atomic
emission spectroscopy; PSM, phosphorus-sorbing material; PVC, polyvinyl
chloride; RhWT, rhodamine; RT, retention time.
Dep. of Plant and Soil Science, Oklahoma State Univ., 367 Agricultural Hall,
Stillwater, OK, 47078-1020. Assigned to Associate Editor Gerwin F. Koopmans.
Copyright © 2011 by the American Society of Agronomy, Crop Science Society
of America, and Soil Science Society of America. All rights reserved. No part of
this periodical may be reproduced or transmitted in any form or by any means,
electronic or mechanical, including photocopying, recording, or any information
storage and retrieval system, without permission in writing from the publisher.
J. Environ. Qual.
Posted online 1 Nov. 2011.
Received 14 Feb. 2011.
*Corresponding author (
5585 Guilford Rd., Madison, WI 53711 USA
Journal of Environmental Quality SPECIAL SUBMISSIONS
Journal of Environmental Quality • Volume 41 • X–X 2012
developed from a batch isotherm experiment with the same
PSM material failed.
Other studies have shown potential for the development
of P removal structures. Penn et al. (2007) constructed a P
removal structure in a drainage ditch located on the Eastern
Shore of Maryland. is structure was filled with 226 kg of
acid mine drainage residual, and the PSM was able to remove
99% of the P, Zn, and Cu that flowed into it during a 24-h
rainfall event that produced 30 cm of precipitation. However,
the structure soon thereafter failed as a result of flow becom-
ing restricted through it (i.e., clogging). Agrawal et al. (2011)
tested a cartridge filtration system on a golf course green sub-
surface drainage system for removing P and several pesticides
using a mixture of slag, zeolite, cement kiln dust, silica sand,
and coconut shell–activated carbon. Although the system was
effective for removing certain pesticides, it was ineffective at
removing P, likely due to the small amount of slag used in the
filtration system (3.5 L).
ere are no published studies on monitoring of a P
removal structure. erefore, the objectives of this study were
to construct and monitor a P removal structure in a subur-
ban watershed and to test the ability of previously constructed
flow-through equations for predicting structure performance.
Materials and Methods
Site Description
e P removal structure was placed at the outlet of a 320-ha
suburban watershed in Stillwater, Oklahoma. e watershed
land use consisted of approximately 35, 50, and 15% residen-
tal, undeveloped, and gof course, respectively. Two irrigated
golf greens were located within 130 to 150 m from the struc-
ture. e greens were regularly irrigated by golf course per-
sonnel as necessary, and this irrigation produced runoff that
reached the P removal structure. e structure was located
in a drainage ditch immediately on the downstream side of a
drainage culvert (Fig. 1) where all water exited the watershed
via a concrete trapezoidal bar ditch maintained by the city
of Stillwater. e bar ditch drained directly into Stillwater
Creek. Some runoff entered the structure by flowing along
the side of the culvert into the structure inlet (Fig. 1).
Structure Construction and Runo Sampling
e P removal structure was 2.4 m wide × 3 m long × 0.2 m
deep and was constructed using 0.63-cm-thick carbon steel with
all joints welded to be water tight. e structure was welded in
situ along with two 3-m steel support pipes (5 cm diameter).
e bottom of the structure was set to a 3% slope toward the
outlet. irteen inlet pipes (5 cm diameter) were welded into
the front plate of the structure, and then each pipe was adapted
to polyvinyl chloride (PVC) pipe of the same diameter inside the
structure. Each PVC pipe was 2.3 m long and perforated (four
rows of 0.635-mm diameter holes at 5 cm apart) to evenly dis-
tribute inflow water throughout the surface of the structure. e
perforated distribution manifold is not visible in Fig. 1 because
the pipes are buried immediately below the surface. A 10-cm-
diameter steel drainage pipe was welded at the bottom center of
the downstream side of the structure; this pipe was adapted to a
15.2-cm-diameter PVC pipe fitted with a shutoff valve. All steel
was treated with two coats of primer and paint.
Two Isco 6712 (Teledyne Isco Inc., Lincoln, NE) automatic
samplers were housed on site in a small plastic building to take
runoff samples at the structure inlet and outlet (drainage pipe)
during flow events. In addition, the automatic sampler for the
outflow side of the structure was fitted with an Isco 730 flow
module (“bubbler”), which was connected to a 15.2-cm-flow
orifice insert placed in the structure drainage pipe (outflow).
e 730 flow module was programmed to take a flow rate mea-
surement every minute. e automatic sampler for the outflow
water was programmed as the “primary” and began sampling
when flow was detected; the inlet sampler was programmed as
the “slave” to the outflow sampler and therefore was triggered to
sample at the same time as the outflow sampler. Discrete (not
composited) samples (800 mL) were taken using two programs;
from 0 to 34.5 L min-1 samples were taken every 30 min, and
at flow rates >34.5 L min-1 samples were taken every 45 min.
Regarding potential “overflow” runoff events, an Isco 2112
ultrasonic probe was fitted near the downstream side of the
structure to monitor the depth of water on top of the structure.
e Isco 2112 could provide the flow rate of untreated overflow
water during events that exceeded the capacity of the structure.
erefore, outlet flow volume plus overflow volume equals total
ditch flow volume.
Electric arc furnace steel slag was obtained from a steel mill
in Ft. Smith, Arkansas (Tube City IMS). Slag was sieved at
a nearby gravel quarry to achieve a size of 6.35 to 11 mm in
diameter. Previous experiments showed that the nonsieved slag
had a limited saturated hydraulic conductivity (Penn et al.,
2011). Approximately 2712 kg of the sieved slag was placed in
the P removal structure on 10 July 2010.
Dye Test
A rhodamine WT (RhWT) dye test was conducted to quan-
tify hydraulic RT in the structure. A constant water flow rate
was discharged into a pool of water at the inlet of the struc-
ture (Fig. 1) for approximately 1 h to achieve steady state
flow before initiating the dye test. e dye was injected into
the inflow solution and monitored in the inflow and outflow
Fig. 1. Picture of the phosphorus (P) removal structure with runo
inlets, drain for treated water, and overow weir. The P sorption mate-
rial in the structure is 2712 kg of 6.3- to 11-mm-diameter steel slag.
over time. e dye test was simulated using CXTFIT (ver-
sion 2.1) (Toride et al., 1999), a model used extensively for
solving the one-dimensional convective–dispersion equation
for solute transport through soils (e.g., Baumann et al., 2002;
Lee et al., 2002). Fate and transport parameters in the model,
such as pore velocity, hydrodynamic dispersion, and retar-
dation coefficient, were optimized to the observed RhWT
concentrations. is process is also known as “inverse estima-
tion” of model parameters, as opposed to forward modeling,
where parameters are input and concentrations are predicted.
From these fate and transport parameters, various character-
istics of the flow and contaminant transport system can be
measured, such as the RT and Peclet number. Physical and
chemical equilibrium of RhWT was assumed. e input
boundary condition for the dye was modeled in CXTFIT
as multiple pulse inputs based on measured inflow concen-
trations. CXTFIT used a nonlinear least-squares parameter
optimization method to derive the dye transport parameters
(i.e., velocity and dispersion coefficient) that best predicted
the outflow RhWT concentrations. e inversely estimated
velocity from CXTFIT was used to estimate the average RT
of the dye in the structure.
Analysis of Water Samples and Slag
All water samples were collected within 12 h of a runoff
event, filtered through a 0.45-µm membrane, and refriger-
ated. Samples were analyzed within 3 d for P, copper (Cu),
zinc (Zn), chromium (Cr), and boron (B) by inductively cou-
pled plasma–atomic emission spectroscopy (ICP–AES). A pH
probe was used to measure pH in all samples. Alkalinity was
determined by automatic titration (TitriLab 865; Radiometer
Analytical, Villeurbanne Cedex, France) to pH 4.5.
All analyses of steel slag used in the P removal structure
were conducted in triplicate. Slag pH was determined with a
pH meter using a solid/deionized (DI) water ratio of 1:5 (w/v).
Alkalinity was determined as previously described using 2 g
of material suspended in 20 mL of DI water. Slag was ground
before analysis of total elements by the EPA 3051 nitric acid
digestion method (USEPA, 1997). Digestion solutions were
analyzed for Ca, Mg, S, Fe, and Al by ICP–AES. Samples were
also extracted with DI water at a 1:10 (w/v) solid/solution ratio
for 1 h, followed by filtration with a 0.45-µm filter and analysis
for Ca, Mg, S, Fe, and Al by ICP–AES.
A standard batch isotherm was conducted for the slag using
2 g of sample and 16 h equilibration (shaking) in 30-mL
solutions of 0, 1, 10, 25, 50, and 100 mg P L-1. Phosphorus
solutions were made using KH2PO4, and the matrix solution
consisted of 5.6, 132, 110, 10, and 17 mg L-1 of Mg, Ca, S,
Na, and K, respectively, adjusted to a pH of 7. Reagent-grade
magnesium sulfate, calcium sulfate, sodium chloride, and
potassium chloride were used to make the matrix. is matrix
was chosen because it was found to be representative of agricul-
tural runoff measured in a previous study (Penn et al., 2007).
After equilibration, solutions were centrifuged for 15 min and
filtered through a 0.45-mm filter before P analysis by ICP–AES.
Phosphorus sorption was quantified by the difference
between P concentrations added and the final equilibrated con-
centrations. ese values were applied to a nonlinear Langmuir
using the following equation:
=+ [1]
where S is the sorbed P concentration (mg kg-1), Smax is the
maximum sorption capacity of the soil (mg kg-1), K is the
Langmuir binding strength coefficient (L mg-1), and C is the
equilibrium concentration (mg L-1). e best fit model param-
eters for the nonlinear equation were obtained by finding the
combinations of parameters that provided the best fit to the
observed data. is was done by using an Excel spreadsheet
as prepared and described by Bolster and Hornberger (2007).
is program was designed to provide K and Smax values in
addition to the “goodness-of-fit” indicator, model efficiency
(E). An E value of 1 indicates a perfect fit of the data, and E <
0 indicates that taking the average of all measured P sorption
values in the isotherm would give a better prediction than the
model (Bolster and Hornberger, 2007).
Flow and sampling data were synchronized with Flow Link
software (Teledyne Isco Inc., Lincoln, NE) when downloaded
directly from the automatic samplers. Because flow rate mea-
surements were taken every minute, the discrete runoff volume
produced at any given minute can be determined by:
Discrete runoff volume = flow rate * 1 [2]
where discrete runoff volume is expressed in liters and flow
rate in L min-1. Discrete runoff volume was calculated at every
minute for each flow event. erefore, the total runoff volume
produced for a given time period could be determined by
the sum of all discrete runoff volumes over that time period.
Weighted average flow rate (L min-1) was calculated as:
total runoff volume
Weighted average flow rate
total runoff time
= [3]
where total runoff volume and time are in units of liters and
minutes, respectively. Phosphorus loading to the structure
between each sampling point was calculated by integrating P
concentrations with respect to flow volume. e sum of all P
loads for each sampling point interval represents the total P
load for an event. is value is used to calculate flow-weighted
P concentrations (mg L-1):
P load
Flow-weighted P concentration
total flow volume
= [4]
where P load and total volume are in units of milligrams and
liters, respectively. After P loads were determined for inflow
and outflow (treated) water, the P removal (mg) could be cal-
culated as a mass balance:
P removed = inlet P load – outflow P load [5]
where inlet and outflow P load are expressed as milligrams.
Retention time (in minutes) of the structure at different
flow rates was also estimated as described in Penn and
McGrath (2011):
total structure pore space
Retention time
flow rate at outlet
= [6]
Journal of Environmental Quality • Volume 41 • X–X 2012
where total structure pore space and flow rate at outlet are in
units of liters and L min-1, respectively. Total pore space (574 L)
was calculated based the total mass of material (2712 kg), bulk
density (1.8 g cm3), and porosity (38%).
Prediction of Field Results Using an Empirical Model
A series of empirical flow-through equations developed by Penn
and McGrath (2011) was used to compare field results of the
P removal structure with the predicted amount of P removed.
Although details of the general use of these empirical equa-
tions appear in a companion paper (Stoner et al., 2012), we
provide a brief description here. e following equations were
originally developed by Penn and McGrath (2011) to predict
the amount of discrete P removal (% P removal) with P loading
to sieved slag (x in mg P kg-1) using an exponential equation:
Discrete P removal (%) = bemx [7]
where b is the Y intercept and m is the slope coefficient for this
relationship. Because this is an exponential decay equation, m
is always negative. e following equations (significant at P <
0.01; R2 = 0.68 and 0.48 for Eq. [8] and [9], respectively) are
used to estimate the b and m parameters for Eq. [7] as a func-
tion of RT and inflow P concentration (Penn and McGrath,
log-m = (0.08506RT) - (0.07416Cin) - 2.53493 [8]
log b = (0.06541RT) - (0.00864Cin) + 1.60631 [9]
where Cin is the inflow P concentration (mg L-1). As described
in greater detail in Stoner et al. (2012) and Penn and McGrath
(2011), these equations were developed from a series of labora-
tory flow-through cell experiments in which a known mass of
slag was exposed to a flowing P solution at five different RTs
and five different inflow P concentrations. When parameters
m and b are inserted into Eq. [7], the result is a predicted P
removal curve specific to the inflow P concentration and RT
conditions that were input into Eq. [8] and [9]. Integration
of the predicted P removal curve (Eq. [7]) yields a prediction
of cumulative P removal (%) at any given level of P added (x;
mg kg-1):
0( )d
Cumulative P removed
be x
=ò [10]
Phosphorus removal approaches zero (1%) as described by the
equation for the predicted P removal curve (Eq. [7]) when the P
inflow concentration ≈ P outflow concentration (i.e., the point
at which the PSM is “spent”). Insertion of 1% for cumulative P
removed into Eq. [10] and subsequent rearrangement to solve
for x results in an estimate of the maximum amount of P that
can be delivered to the P removal structure before the PSM is
spent. Such a rearrangement results in the following equation:
Maximum P added b
=- [11]
Insertion of the maximum amount of P that can be added to
the P removal structure as determined from Eq. [11] into Eq.
[7] results in the total amount of P predicted to be removed by
the PSM under the conditions (i.e., RT and inflow P concen-
tration) used for the flow-through equations (Eq. [8] and [9])
used to produce the predicted P removal curve.
Results and Discussion
Phosphorus Removal Structure: Flow
Results from the dye test indicated that when a flow rate of
57.1 L min-1 was applied to the structure, the average RT was
9.3 min as estimated by CXTFIT (R2 = 0.97 between measured
and predicted dye outflow concentrations). is RT is similar
to the calculated value of 10 min estimated by Eq. [6].
During the 5-mo period in which all runoff was monitored,
there were 54 total runoff events. Twenty of the events were
rainfall, and 34 were due to irrigation of nearby golf course
greens (Table 1). Over that time period, the rainfall totaled
24.6 cm; the largest rainfall event was 4 cm on 8 Sept. 2010.
e P removal structure was able to treat all water delivered to
it, as evidenced by the fact that no water crested the overflow
weir, which was continuously monitored with an ultrasonic
probe. During the largest rainfall event, the maximum flow
rate through the structure was 506 L min-1.
As expected, rainfall events produced higher flow rates
through the structure than irrigation events from nearby golf
greens, which translated into a lower average RT for the rain-
fall runoff events (Table 1). All runoff samples were analyzed
for total dissolved P, and several random irrigation and storm
runoff samples were analyzed for dissolved reactive P (i.e.,
orthophosphate). Because the entire area immediately drain-
ing into the structure was well covered with grass, there was no
sediment in the samples, and thus >90% of the total dissolved
P was orthophosphate. e overall flow-weighted average total
dissolved P concentration in runoff delivered to the P removal
structure (0.50 mg L-1) is comparable to other studies, includ-
ing those conducted on agricultural land. Harmel et al. (2004)
showed that several agricultural subwatersheds consisting of
cultivated crops or pasture that received 0 to 358 kg P ha-1 yr-1
Table 1. Summary of the suburban phosphorus removal structure performance over the rst 5 mo of operation.
Rainfall runo events Irrigation runo events All runo events
Number of runo events 20 34 54
Maximum ow rate, L min-1506 47 506
Weighted average ow rate, L min-130.3 11.5 29.8
Weighted average retention time, min 18.9 50 19.3
Maximum runo P concentration, mg L-11.61 0.97 1.61
Flow-weighted runo P concentration, mg L-10.59 0.44 0.50
Total P input to structure, mg kg-192.1 10.7 102.8
Total P removed by structure, mg kg-119.3 6.6 25.9
produced average dissolved P concentrations of 0.09
to 2.29 mg L-1. Among 35 agricultural catchments
monitored over 4 yr in Ireland, runoff-dissolved
P concentrations ranged from 0.01 to 0.70 mg L-1
(Daly et al., 2002). A golf course in Texas produced
an average dissolved P concentration of 0.13 mg L-1
over 5 yr (King et al., 2007).
Figure 2 shows hydrographs and corresponding
inflow total dissolved P concentrations for typical
runoff events from rainfall and irrigation. Not only
did rainfall runoff events produce higher dissolved P
concentrations than irrigation runoff events (Table
1), but rainfall runoff events also tended to produce
increasing P concentrations with flow rate into the
P removal structure. is suggests that hydrologi-
cal connectivity increased among certain portions of
the watershed as soils became saturated with mois-
ture and runoff increased, allowing runoff from these
“variable source” areas (Sharpley et al., 2008) in the
watershed to reach the outlet, which is the ditch P
removal structure. Similarly, Pionke et al. (1999)
found that dissolved P concentrations delivered from
an agricultural watershed increased with flow rate. In
our case, we speculate that high-P soils contribute P
to the structure only during large events when they
become “connected” and such runoff is able to reach
the outlet. Because the irrigation events that occurred
throughout the monitoring period were from the
same location, runoff produced from such events
typically displayed relatively steady runoff P concen-
trations delivered to the structure between 0.3 and
0.5 mg L-1 (Fig. 2).
Phosphorus Removal Structure:
Phosphorus Removal
e sum of total dissolved P delivered to the structure over
the 5-mo period was 0.282 kg or 0.0047 kg ha-1; 88% of
this P delivery occurred during rainfall induced runoff events
(Table 1). Among all dissolved P transported in runoff to the
P removal structure, 75% of this was delivered during the six
largest rainfall events. Various authors have suggested that
large rainfall events export the majority of P from watersheds
(Sharpley et al., 2008; Udawatta et al., 2004; Pionke et al.,
1999; Pionke et al., 1997). For example, Pionke et al. (1997)
found that 70% of annual dissolved P loads were exported by
the seven largest storms.
During the 5 mo of monitoring, the P removal structure
sorbed 25.9 mg P kg-1 slag, which was 25.2% of the total dis-
solved P delivered to it (Table 1). Of the 25.9 mg P kg-1 sorbed,
approximately 75 and 25% occurred during rainfall and irriga-
tion runoff events, respectively. Phosphorus transported during
irrigation runoff events was more efficiently removed by the
structure compared with rainfall runoff events (i.e., 62 versus
21% P removal for irrigation and rainfall events, respectively)
(Table 1). e difference in P removal efficiency among rain-
fall and irrigation events is likely due to the fact that rainfall
runoff events resulted in higher P concentrations and flow
rates. Higher structure flow rates during rainfall runoff events
translated into a RT that was more than two times less than
irrigation events (Table 1). Regarding the impact of flow rate
and RT on P removal by the structure, P removal on an event
basis was negatively correlated to the weighted average event
flow rate (Fig. 3). Similarly, in a previous study (McDowell et
al., 2008) involving slag placed in subsurface drainage pipes, it
was noted that larger events resulted in less contact time with
the slag and lesser differences in dissolved P concentrations
relative to control drains.
Although the weighted average RT for all rainfall runoff
events was 18.9 min, the RT for the six largest rainfall events
that delivered 75% of the P to the P removal structure was only
8.9 min. In addition, 54% of all the P removed by the structure
(14.1 mg kg-1) occurred over these six largest rainfall events.
Predicting Lifetime and Performance of the Structure
A predicted P removal curve estimated by the equations devel-
oped in Penn and McGrath (2011) for the electric arc furnace
steel slag is shown in Fig. 4. is curve (Eq. [7]) describes the
effect of P loading to the PSM on discrete P removal. is curve
was produced by estimating its Y intercept (b) and its slope
coefficient (m) with Eq. [8] and [9] in which RT and P inflow
concentration are used as inputs. For the RT of the runoff in
the P removal structure, we used 8.9 min (i.e., the RT for the
Fig. 2. Typical hydrograph and corresponding inow total dissolved phosphorus
(P) concentrations to the ditch P removal structure from a rainfall-induced (a) and
irrigation-induced (b) runo event. The 3.73-cm rainfall/runo event shown in (a)
occurred on 17 Aug. 2010, and the irrigation/runo event occurred on 3 Aug. 2010.
Journal of Environmental Quality • Volume 41 • X–X 2012
six largest rainfall events that delivered 75% of the P to the P
removal structure), whereas the average flow-weighted P inflow
concentration was set at 0.74 mg L-1. e predicted P removal
curve can be used to estimate the potential “lifetime” of the P
removal structure. When discrete P removal approaches nearly
zero (i.e., 1%), then the slag is effectively “spent” and needs to
be replaced with fresh PSM because the P inflow concentra-
tion will nearly equal the outflow concentration. e structure
“lifetime” can be predicted using an estimate of P loading to
the structure per unit time and the predicted maximum P load-
ing to the P removal structure at the point in which the PSM
is “spent” (Eq. [11]). Using predicted values of the Y inter-
cept (b) and the slope coefficient (m) from the flow-through
equations (see above), a maximum cumulative loading of the
P removal structure amounting to 345 mg kg-1 was calculated
using Eq. [11]. Based on the current P loading rate of the P
removal structure (i.e., 20.5 mg kg-1 mo-1), this would cor-
respond to a potential lifetime of 16.8 mo. e measured P
removal curve that was fitted to the field data of the
actual discrete P removal and P loading of the P removal
structure is shown in Fig. 4. Using the fitted values of
the Y intercept (b) and the slope coefficient (m), a maxi-
mum cumulative loading of 316 mg kg-1 was estimated
with Eq. [11], which corresponds to a structure lifetime
of 15.4 mo. us, the lifetime prediction of 16.8 mo
differs by a factor of only 1.09 of the projected lifetime
using current structure performance data. In practice,
one may be inclined to remove the slag material before
P saturation if environmental thresholds such as total
maximum daily loads are exceeded. is estimate of filter
lifetime does not take into account processes of sorbed
P on the slag changing forms and allowing for more P
sorption sites to become available, as described in Drizo
et al. (2008). Apparently, such a factor did not have a sig-
nificant impact on predicting filter lifetime due to near
agreement (16.8 vs. 15.4 mo). However, a slow P sorp-
tion mechanism as described by Drizo et al. (2008) that
was active would result in an underestimation of filter
lifetime by the predicted P removal curve. e steel slag used
in this study differed from that of Drizo et al. (2008) in that it
was sieved to exclude fine particles.
e predicted P removal curve shown in Fig. 4 can be inte-
grated to estimate the cumulative amount of P that the struc-
ture will remove as a function of P added (Eq. [10]). Figure 5
shows the predicted cumulative amount of P removed by the P
removal structure as a function of P loading. For comparison,
the measured values of the cumulative amount of P removed
from runoff as a function of P loading of the P removal structure
are shown. e predicted cumulative P removal compared with
the measured values showed that the flow-through equations
used to produce the predicted P removal curve overestimated
P removal. For example, after 5 mo and a total P input of 103
mg kg-1 to the P removal structure, the integrated predicted P
removal curve estimated 79 mg kg-1 of P sorption, whereas the
actual measured P sorption was 25.9 mg kg-1 (Table 1).
At the point of P saturation when the PSM is “spent,” the
integrated predicted P removal curve estimated a cumulative
removal of 101 mg P kg-1, or 28% of the total P added to
the structure. is estimated value was obtained from the pre-
dicted P removal curve (Fig. 4), which was produced using Eq.
[7–11] with an input of 8.9 min RT and 0.74 mg L-1 inflow
(i.e., the conditions of the six largest rainfall events that deliv-
ered 75% of the P). Specifically, flow-through Eq. [8] and [9]
predicted the P removal curve parameters (b and m) for Eq.
[7]; the resulting predicted design curve (Fig. 4) was integrated
(Eq. [10]) (Fig. 5), which produced an estimate of maximum P
removal under the conditions of the design curve (i.e., inflow P
concentration and RT).
Apparently, the empirical flow-through equations were
able to predict that P would be removed from runoff by the
P removal structure as the P loading increased, but not to
the correct degree in which it was occurring. is is likely
due to the fact that the equations were unable to accurately
predict the Y intercept (b) of the design curve (via Eq. [9])
(Fig. 4). e maximum amount of P projected to be removed
by the structure (i.e., 0.065 g kg-1 determined from integra-
tion of the curve fitted to measured field data in Fig. 4) is low
Fig. 3. Phosphorus (P) removal eciency presented per event as impacted by
the ow rate of runo water passing through the ditch P removal structure.
*Signicant at the 0.05 probability level.
Fig. 4. Discrete phosphorus (P) removal as a function of cumula-
tive P added to the ditch runo P removal structure. Predicted P
removal (dashed line) estimated based on average retention time
and P concentration of the six largest rainfall events that delivered
75% of runo P load (average weighted retention time, 8.9 min; total
dissolved P concentration, 0.74 mg L−1) using Eq. [7–10]. Measured
discrete P removal (open circles and solid line) calculated on a per-
event basis. Error bars indicate a 95% condence interval for the
predicted P removal curve. *Signicant at the 0.05 probability level.
compared with other studies that have investigated the use
of electric arc furnace steel slag for P sorption (Drizo et al.,
2006; Drizo et al., 2002). For example, Drizo et al. (2002)
achieved 1.35 to 2.35 g P removed kg-1; however, their study
used a much higher RT (~8 h) compared with the RT of the
runoff in the P removal structure in our study. In addition,
the large particle size fraction used in our study (i.e., 6.35–11
mm) compared with previous studies (Kostura et al., 2005;
Drizo et al., 2002) is not nearly as sorptive compared with the
finer slag fraction (Stoner et al., 2012). However, the benefit
of the large size fraction is higher hydraulic conductivity of
the structure, which reduces the “footprint” or area of the P
removal structure and allows more water to be treated com-
pared with a finer-sized fraction.
Equations [8] and [9], which were used to estimate the Y
intercept (b) and the slope coefficient (m) of the predicted
design curve in Fig. 4, were developed using slag with the same
size fraction collected from the same steel mill as slag used in
the P removal structure but was collected at a different time
(about 8 mo apart). In other words, cumulative P removal
predictions from equations developed by Penn and McGrath
(2011) are specific to their particular slag material, and any
variation in slag properties would likely result in deviation
from the predictions. is could explain why integration of
the predicted P removal curve with sample-specific parameters
indicated in Eq. [8] and [9] from Penn and McGrath (2011)
overpredicted cumulative P removal as compared with mea-
sured values (Fig. 5). For example, the slag placed in the P
removal structure contained less alkalinity and less total Ca and
Fe, and had a lower pH compared with the slag used to develop
the flow through equations of Penn and McGrath (2011). Slag
pH and alkalinity are integral to Ca phosphate precipitation
(Bowden et al., 2009; Kostura et al., 2005). e role of Ca
and Fe in P sorption by industrial by-products has been well
documented (Penn et al., 2011; Leader et al., 2008). Lesser
amounts of Ca and Fe would result in less Ca phosphate pre-
cipitation and P binding by Fe oxy/hydroxide minerals. e
Langmuir K value was also much less for the slag sample used
in this study compared with that used for development of flow-
through equations (i.e., 0.00126 vs. 2.43 L mg-1, respectively,
from Penn and McGrath, 2011).
Other Water Quality Parameters
Average pH of inflow and outflow treated water was 7.7 and
9.2 (SE, 0.04 and 0.08, respectively). e increase in pH of the
treated water was expected due to the elevated pH of the PSM
tested in the laboratory (i.e., 9.4) (Table 2). However, alkalinity
of the treated water was similar to inflow water; average inflow
and outflow alkalinity was 77 and 81 mg CaCO3 L-1 (SE, 21 and
23, respectively). A minimum alkalinity of 20 mg L-1 is required
for ecosystems, and an alkalinity up to 400 mg L-1 has no impact
on human health (USEPA, 1986).
For all inflow and treated water, Zn, Cu, Cr, and Mn con-
centrations were all below detection limits (i.e., 0.01 mg L-1).
Average B concentrations were similar among inflow and out-
flow treated waters (i.e., 0.14 and 0.15 mg L-1; SE, 0.003 and
0.005, respectively). However, these B concentrations are not
considered hazardous to aquatic life or B-sensitive agricultural
crops (USEPA, 1986).
During the first 5 mo of operation, the P removal structure
trapped 25% of runoff dissolved P. is could be improved
by using the smaller particle size fraction of the slag, which is
much more sorptive than the large fraction used in this study
(Stoner et al., 2012). However, the smaller-sized fraction
would reduce the hydraulic conductivity, thereby reducing
the amount of water that can be treated during a large runoff
event. Alternatively, the filter dimensions could be adjusted to
allow for a higher RT. e flow-through equations presented
in Penn and McGrath (2011) predicted a lifetime of 16.8 mo,
which is similar to the projected lifetime of 15.4 mo based on
current measurements. However, the flow-through equations
overestimated current P removal (79 vs. 26 mg P kg-1) by the
P removal structure. Differences in P removal between pre-
dictions and measurements were likely a result of variability
in slag chemical properties among slag used in the P removal
structure and for development of flow-through equations. is
emphasizes the need to develop a universal” flow-through
Fig. 5. Cumulative phosphorus (P) removal by the ditch P removal
structure over a 5-mo period as measured and predicted (dashed
line) using a series of ow-through based equations (Eq. [7–10]).
Predicted P removed estimated by integration of the curve pre-
sented in Fig. 4 using Eq. [10]. Error bars indicate 95% condence
interval for the predicted P removal based on the standard error for
each model coecient.
Table 2. Chemical properties of the steel slag used in the suburban phosphorus removal structure.
Smax K pH Alkalinity Total‡ Water soluble
Ca Mg S Fe Al Ca Mg S Fe Al
mg kg–1 L mg–1 mg CaCO3 kg–1 — mg kg–1
† Smax is the maximum sorption capacity of the soil. Langmuir isotherm Smax and K values were estimated using Eq. [1].
‡ Determined by EPA3051 digestion method.
§ Values in parentheses indicate standard error.
Journal of Environmental Quality • Volume 41 • X–X 2012
model that takes into account chemical characterization of
sorption materials in addition to RT and P concentrations.
Because 75% of all P delivered to the structure occurred over
the six largest rainfall events, P removal structures should be
designed for handling these events to maximize P removal.
Compared with other best management practices, poultry
litter transport programs and limitation of fertilizer P applica-
tions only prevent soil P from increasing further. is technol-
ogy can help to prevent P losses to surface waters in the short
term. In addition, the structure provides an easily quantified P
removal that not only can be removed from the watershed but
also may be useful to nutrient trading programs that are anal-
ogous to current carbon credit exchange programs (USEPA,
2001). Such programs apply a monetary value to P discharged
or transported from a site or prevented from being transported.
e authors thank the associate editor for his considerable time and
effort invested in improving this manuscript.
Agrawal, S.G., K.G. King, J.F. Moore, P. Levinson, and J. McDonald. 2011.
Use of industrial byproducts to filter phosphorus and pesticides in golf
green drainage water. J. Environ. Qual. 40:1273–1280.
Baumann, T., S. Muller, and R. Niessner. 2002. Migration of dissolved heavy
metal compounds and PCP in the presence of colloids through a hetero-
geneous calcareous gravel and a homogeneous quartz sand: Pilot scale
experiments. Water Research 36:1213–1223.
Bolster, C.H., and G.M. Hornberger. 2007. On the use of linearized Lang-
muir equations. Soil Sci. Soc. Am. J. 71:1796–1806. doi:10.2136/
Bowden, L.I., A.P. Jarvis, P.L. Younger, and K.L. Johnson. 2009. Phosphorus
removal from waste waters using basic oxygen steel slag. Environ. Sci.
Technol. 43:2476–2481. doi:10.1021/es801626d
Cucarella, V., and G. Renman. 2009. Phosphorus sorption capacity of filter
materials used for on-site wastewater treatment determined in batch
experiments: A comparative study. J. Environ. Qual. 38:381–392.
Daly, K., P. Mills, B. Coulter, and M. McGarrigle. 2002. Modeling phospho-
rus concentrations in Irish rivers using land use, soil type, and soil phos-
phorus data. J. Environ. Qual. 31:590–599. doi:10.2134/jeq2002.0590
Drizo, A., J. Cummings, D. Weber, E. Twohig, G. Druschel, and B. Bourke.
2008. New evidence for rejuvenation of phosphorus retention capacity
in EAF steel slag. Environ. Sci. Technol. 42:6191–6197. doi:10.1021/
Drizo, A., C. Forget, R.P. Chapuis, and Y. Comeau. 2006. Phosphorus removal
by electric arc furnace steel slag and serpentininte. Water Res. 40:1547–
1554. doi:10.1016/j.watres.2006.02.001
Drizo, A., Y. Comeau, C. Forget, and R.P. Chapuis. 2002. Phosphorus satura-
tion potential: A parameter for estimating the longevity of constructed
wetland systems. Environ. Sci. Technol. 36:4642–4648. doi:10.1021/
Gallimore, L.E., N.T. Basta, D.E. Storm, M.E. Payton, R.H. Huhnke, and
M.D. Smolen. 1999. Water treatment residual to reduce nutrients in
surface runoff from agricultural land. J. Environ. Qual. 28:1474–1478.
Harmel, R.D., H.A. Torbert, B.E. Haggard, R. Haney, and M. Dozier. 2004.
Water quality impacts of converting to a poultry litter fertilization strat-
egy. J. Environ. Qual. 33:2229–2242. doi:10.2134/jeq2004.2229
King, K.W., J.C. Balogh, K.L. Hughes, and R.D. Harmel. 2007. Nutrient
load generated by storm event runoff from a golf course watershed. J.
Environ. Qual. 36:1021–1030. doi:10.2134/jeq2006.0387
Koiv, M., M. Liira, U. Mander, R. Motlep, C. Vohla, and K. Kirsimae. 2010.
Phosphorus removal using Ca-rich hydrated oil shale ash as filter mate-
rial: e effect of different phosphorus loadings and wastewater com-
positions. Water Res. 44:5232–5239. doi:10.1016/j.watres.2010.06.044
Kostura, B., H. Kulveitova, and J. Lesko. 2005. Blast furnace slags as sor-
bents of phosphate from water solutions. Water Res. 39:1795–1802.
Leader, J.W., E.J. Dunne, and K.R. Reddy. 2008. Phosphorus sorbing materi-
als: Sorption dynamics and physicochemical characteristics. J. Environ.
Qual. 37:174–181. doi:10.2134/jeq2007.0148
Lee, J., R. Horton, and D.B. Jaynes. 2002. e feasibility of shallow time
domain reflectometry probes to describe solute transport through un-
disturbed soil cores. Soil Sci. Soc. Am. J. 66:53–57. doi:10.2136/
Makris, K.C., and W.G. Harris. 2006. Time dependency and irreversibil-
ity of water desorption by drinking-water treatment residuals: Implica-
tions for sorption mechanisms. J. Colloid Interface Sci. 294:151–154.
McDowell, R.W., A.N. Sharpley, and W. Bourke. 2008. Treatment of drain-
age water with industrial by-products to prevent phosphorus loss from
tile-drained land. J. Environ. Qual. 37:1575–1582. doi:10.2134/
Penn, C.J., and R. Bryant. 2006. Application of phosphorus sorbing materials
to streamside cattle loafing areas. J. Soil Water Conserv. 61:303–310.
Penn, C.J., R.B. Bryant, M.A. Callahan, and J.M. McGrath. 2011. Use of
industrial byproducts to sorb and retain phosphorus. Commun. Soil Sci.
Plant Anal. 42:633–644. doi:10.1080/00103624.2011.550374
Penn, C.J., R.B. Bryant, P.A. Kleinman, and A. Allen. 2007. Sequestering
dissolved phosphorus from ditch drainage water. J. Soil Water Conserv.
Penn, C.J., and J.M. McGrath. 2011. Predicting phosphorus sorption onto steel
slag using a flow-through approach with application to a pilot scale sys-
tem. J. Water Resour. Prot. 3:235–244. doi:10.4236/jwarp.2011.34030
Penn, C.J., J.M. McGrath, and R.B. Bryant. 2010. Ditch drainage manage-
ment for water quality improvement. p. 151–173. In M.T. Moore (ed.)
Agricultural drainage ditches: Mitigation wetlands for the 21st century.
Research Signpost, Kerala, India.
Pionke, H.B., W.J. Gburek, R.R. Schnabel, A.N. Sharpley, and G.F. Elwing-
er. 1999. Seasonal flow, nutrient concentrations and loading patterns
in stream flow draining an agricultural hill-land watershed. J. Hydrol.
220:62–73. doi:10.1016/S0022-1694(99)00064-5
Pionke, H.B., W.J. Gburek, A.N. Sharpley, and J.A. Zollweg. 1997. Hydro-
logic and chemical controls on P loss from catchments. p. 225–242. In
H. Tunney, O.T. Carton, P.C. Brookes, and A.E. Johnston (ed.) Phos-
phorus loss from soil to water. CAB International Publishers, New York.
Rhoton, F.E., and J.M. Bigham. 2005. Phosphate adsorption by ferrihydrite-
amended soils. Soil Sci. Soc. Am. J. 34:890–896.
Sharpley, A.N., P.J.A. Kleinman, A.L. Heathwaite, W.J. Gburek, J.L. Weld,
and G.J. Folmar. 2008. Integrating contributing areas and index-
ing phosphorus loss from agricultural watersheds. J. Environ. Qual.
37:1488–1496. doi:10.2134/jeq2007.0381
Stoner, D., C.J. Penn, J.M. McGrath, and J.G. Warren. 2012. Phosphorus
removal with by-products in a flow-through setting. J. Environ. Qual.
(in press). doi:10.2134/jeq2011.0049
Toride, N., F.J. Leij, and M.. van Genuchten. 1999. e CXTFIT code
for estimating transport parameters from laboratory or field tracer ex-
periments, version 2.1. Research Rep. 137. USDA–ARS, U.S. Salinity
Laboratory, Riverside, CA.
Udawatta, R.P., P.P. Motavalli, and H.E. Garrett. 2004. Phosphorus loss and
runoff characteristics in three adjacent agricultural watersheds with clay-
pan soils. J. Environ. Qual. 33:1709–1719. doi:10.2134/jeq2004.1709
USEPA. 1986. Quality criteria for water. EPA Rep. 440/5-86-001. USEPA,
Office of Water Regulations and Standards, Washington, DC.
USEPA. 1997. Method 3051a: Micro-wave assisted acid dissolution of sedi-
ments, sludges, soils, and oils. 2nd ed. U.S. Gov. Print. Office, Wash-
ington, DC.
USEPA. 2001. Chesapeake Bay program nutrient trading fundamental prin-
ciples and guidlines. EPA 903-B-01-001. CBP/TRS 254/01. USEPA
Report Collection, Regional Center for Environmental Information,
USEPA Region 3, Philadelphia, PA.
Wei, X., R.C. Viadero, Jr., and S. Bhojappa. 2008. Phosphorus removal by
acid mine drainage sludge from secondary effluents of municipal waste-
water treatment plants. Water Res. 42:3275–3284. doi:10.1016/j.
... Slag had an elevated pH and high Ca concentration, both of which are required for the slag to be effective at dissolved P removal via precipitation of Ca phosphate [7,[15][16][17][18]. Although slag contained some trace metals, there was no environmental concern about release to treated water since EAF slag has been shown to possess negligible metal solubility [9,19] and even may remove trace metals [20]. Flow-through analysis ( Figure 2) revealed typical P removal for a slag sample mostly devoid of fines less than 1 mm in diameter [19,21,22]. ...
... Although slag contained some trace metals, there was no environmental concern about release to treated water since EAF slag has been shown to possess negligible metal solubility [9,19] and even may remove trace metals [20]. Flow-through analysis ( Figure 2) revealed typical P removal for a slag sample mostly devoid of fines less than 1 mm in diameter [19,21,22]. The higher removal of dissolved P at 2.5 mg L −1 , relative to 0.2 mg L −1 , by the EAF slag ( Figure 2) is typical for EAF slags and other PSMs dominated by Ca-P precipitation [7,23]. ...
... Thus, about 32 mg of dissolved P per kg of slag was removed from April to December 2018. The cumulative P removal by a relatively large size slag fraction in this study is comparable to both field and laboratory studies conducted on smaller EAF slag [19,21,22]. Dissolved P loads removed by the blind inlet are shown on a per-event basis in Table 4. ...
Full-text available
Blind inlets are implemented to promote obstruction-free surface drainage of field depressions as an alternative to tile risers for the removal of sediment and particulate phosphorus (P) through an aggregate bed. However, conventional limestone used in blind inlets does not remove dissolved P, which is a stronger eutrophication agent than particulate P. Steel slag has been suggested as an alternative to limestone in blind inlets for removing dissolved P. The objectives of this study were to construct a blind inlet with steel slag and evaluate its ability to remove dissolved P, nitrogen (N), and herbicides. A blind inlet was constructed with steel slag in late 2015; data from only 2018 are reported due to inflow sampling issues. The blind inlet removed at least 45% of the dissolved P load and was still effective after three years. The dissolved P removal efficiency was greater with higher inflow P concentrations. More than 70% of glyphosate and its metabolite, and dicamba were removed. Total N was removed in the form of organic N and ammonium, although N cycling processes within the blind inlet appeared to produce nitrate. Higher dissolved atrazine and organic carbon loads were measured in outflow than inflow, likely due to the deposition of sediment-bound particulate forms not measured in inflow, which then solubilized with time. At a cost similar to local aggregate, steel slag in blind inlets represents a simple update for improving dissolved P removal.
... Structures are engineered for achieving a desired cumulative P removal and lifetime, as a function of PSM properties and site characteristics such as annual flow volume, dissolved P concentration, and peak flow rate [30]. While many PSMs have been shown to be effective in P removal structures, one of the most popular PSMs is steel slag, due to its low cost and widespread availability [24,28,[31][32][33][34][35][36]. Electric arc furnace steel slag has proven effective at trapping dissolved P in surface runoff [25,27,33,36,37] and wastewater [20,[38][39][40][41], but little to no research has been conducted on the use of steel slag for treating tile drainage water. ...
... While many PSMs have been shown to be effective in P removal structures, one of the most popular PSMs is steel slag, due to its low cost and widespread availability [24,28,[31][32][33][34][35][36]. Electric arc furnace steel slag has proven effective at trapping dissolved P in surface runoff [25,27,33,36,37] and wastewater [20,[38][39][40][41], but little to no research has been conducted on the use of steel slag for treating tile drainage water. A major challenge for design and construction of P removal structures on tile drains is the fact that many tile-drained fields have little relief and therefore do not provide the hydraulic head necessary for achieving appreciable flow rates through a structure and minimizing its footprint for a top-down flow design. ...
... In fact, there was a significant (p < 0.05) and positive relationship between flow rate and inflow dissolved P concentration: inflow P concentration = 0.52 * flow rate − 0.056, where P concentration and flow rate are in units of mg L −1 and L s −1 , respectively. A similar trend was also evident in non-point drainage monitoring in several other studies [36,66,67]. Similarly, mean flow-weighted concentrations for precipitation-induced flow events were much greater than baseflow concentrations, which was only 0.17 mg L −1 ( Table 2). ...
Full-text available
Reducing dissolved phosphorus (P) losses from legacy P soils to surface waters is necessary for preventing algal blooms. Phosphorus removal structures containing steel slag have shown success in treating surface runoff for dissolved P, but little is known about treating subsurface (tile) drainage. A ditch-style and subsurface P removal structure were constructed using steel slag in a bottom-up flow design for treating tile drainage. Nearly 97% of P was delivered during precipitation-induced flow events (as opposed to baseflow) with inflow P concentrations increasing with flow rate. Structures handled flow rates approximately 12 L s−1, and the subsurface and ditch structures removed 19.2 (55%) and 0.9 kg (37%) of the cumulative dissolved P load, respectively. Both structures underperformed relative to laboratory flow-through experiments and exhibited signs of flow inhibition with time. Dissolved P removal decreased dramatically when treated water pH decreased <8.5. Although slag has proven successful for treating surface runoff, we hypothesize that underperformance in this case was due to tile drainage bicarbonate consumption of slag calcium through the precipitation of calcium carbonate, thereby filling pore space, decreasing flow and pH, and preventing calcium phosphate precipitation. We do not recommend non-treated steel slag for removing dissolved P from tile drainage unless slag is replaced every 4–6 months.
... It is important to note that while this data suggests that the fine sediment content may not directly influence the SRP retention, the particle size distribution does influence the hydraulic conductivity and thus the porewater residence time in the system (Tirpak et al., 2021). This in turn may affect the redox conditions and rate-controlled reactions including Ca phosphate mineral precipitation and organic matter degradation (Marvin et al., 2020;Penn et al., 2012;Stoner et al., 2012). ...
Full-text available
Soluble reactive phosphorus (SRP) in urban runoff can contribute to eutrophication of downstream water bodies. Bioretention systems, a popular low impact development (LID) practice, have the potential to reduce SRP loads in urban runoff. In this study, six field-scale bioretention systems located in southern Ontario, Canada, with different soil media compositions were monitored to evaluate their performance with respect to SRP retention and to identify the geochemical processes influencing SRP behaviour within the systems. The physical and chemical characteristics of the soil media were analysed to compare the systems as well as to assess whether the installed soil media compositions met the LID bioretention design guidelines. To examine how SRP evolves as stormwater infiltrates through the different media, porewater was collected from multiple depths within the bioretention systems. The geochemical processes controlling SRP retention-release were identified by chemical analysis of the porewater and selective solid phase extractions. The geochemical processes governing SRP behaviour differed between the bioretention systems with processes validated via one-dimensional reactive transport simulations conducted in PHREEQC. Results indicate: 1) Soil media characteristics are highly variable between systems despite the systems having the same overall engineering design specifications; 2) Porewater SRP does not evolve uniformly with depth indicating the importance of overall system design including redox conditions, infiltration rate, and vertical layering of different materials (e.g., surface cover layer and bioretention media); 3) SRP retention-release processes often occur at shallow depth (< 0.1 m) and more attention is needed to assess the impact of different surface cover layers used in field installations; 4) Hydraulic conductivity controls the residence time of water in the systems which in turn influences SRP retention-release; 5) SRP adsorption onto Al- and/or Fe-oxides is the main retention mechanism in systems with high hydraulic conductivity while Ca phosphate mineral precipitation may be important in systems with low hydraulic conductivity. The study findings provide insights needed to improve the field performance of bioretention systems in retaining SRP including the testing and application of soil media amendments for enhancing SRP retention.
... It appears that the use of gypsum is a valuable strategy for controlling P movement [58,59]; likewise, there is a large potential in the use of waste materials such as bauxite residuals, fly-ash, wood ash, and slag [52]. Pen et al. [60] observed that using steel slag as the P sorption material in the P removal structure resulted in the removal of 25% of all dissolved P from rainfall and irrigation events during the first five months of structure operation. The other study demonstrated that the maximum adsorption capacity of the fly-ash and bauxite residuals were 29 and 25 g kg −1 , respectively [61], and P removal varied based on the chemical properties of the by-products [62,63]. ...
Full-text available
The long-term application of manures in watersheds with dense animal production has increased soil phosphorus (P) concentration, exceeding plant and soil assimilative capacities. The P accumulated in soils that are heavily manured and contain excess extractable soil P concentrations is known as legacy P. Runoff and leaching can transport legacy P to ground water and surface water bodies, contributing to water quality impairment and environmental pollution, such as eutrophication. This review article analyzes and discusses current and innovative management practices for soil legacy P. Specifically, we address the use of biochar as an emerging novel technology that reduces P movement and bioavailability in legacy P soils. We illustrate that properties of biochar can be affected by pyrolysis temperature and by various activating chemical compounds and by-products. Our approach consists of engineering biochars, using an activation process on poultry litter feedstock before pyrolysis to enhance the binding or precipitation of legacy P. Finally, this review article describes previous examples of biochar activation and offers new approaches to the production of biochars with enhanced P sorption capabilities.
... These are landscape-scale filters containing media with a high affinity for P and placed in hydrologically active areas that produce flow with appreciable dissolved P concentrations (Figure 1). Various filter media, known as P sorption materials (PSMs), as well as different forms and applications of this BMP have been demonstrated in a variety of situations for treating non-point drainage (Erickson et al., 2018;Gonzalez et al., 2020;Groenenberg et al., 2013;Mendes & Renato, 2020;Penn et al., 2012Penn et al., , 2014Penn et al., , 2020Shedekar et al., 2020;Vandermoere et al., 2018). While diverse in appearance, P removal structures possess several core similarities (Penn & Bowen, 2017): (a) sufficient mass of PSM for removing an appreciable amount of dissolved P load for the site, (b) ability to conduct an appreciable portion of the peak flow rate while allowing water to flow through the PSM at a sufficient retention time (RT), and (c) ability to contain the PSM and prevent it from being flushed out so that it may be replaced or regenerated when necessary. ...
Full-text available
Phosphorus (P) removal structures are a new best management practice for filtering dissolved P in non‐point drainage from legacy P soils through use of P sorption materials (PSMs). Structures must be designed according to characteristics of the site (hydrology and constraints) and PSMs to be utilized, as well as user‐defined goals (P removal, lifetime, and flow rate), making it a cumbersome process. A freely available P Transport Reduction App (P‐TRAP) allows users to quickly produce a custom design or evaluate a hypothetical or existing structure. The software includes a library of P removal flow‐through curves for many different PSMs conducted under various conditions of inflow P concentration and retention time. Design output includes the necessary PSM mass and orientation, pipe requirement, and a table of annual P removal. The software enables conservationists and engineers to quickly compare cost and efficiency among possible designs.
... The second generation of P removal structures sought to correct this problem by compromising P sorption ability for hydraulic conductivity to permit a greater treatment volume. Most of the second generation structures were constructed using >7 mm diameter steel slag or flu-gas gypsum that were able to convey high flow rates, but due to the chemistry of the PSM were not able to sorb as much P per unit mass of PSM compared to PSMs of first generation structures [7,8]. Currently, the third generation of P removal structures are being developed, which aim to produce a "happy medium" between hydraulic conductivity and P sorption ability. ...
Full-text available
The purpose of this special issue is to explore current challenges and develop a better understanding of the processes that control dissolved phosphorus (P) removal among P removal structures [...]
... If properly sieved to exclude fines, slag can possess an appreciable hydraulic conductivity and high P retention capacity [18]. Several studies have shown that slag is effective for treating surface water runoff [19][20][21] even when used in a ditch-style structure [13]. Nevertheless, a recent study by Penn et al. [22] found that subsurface tile drainage water decreases the lifetime of EAF slag in P removal structures due to presence of bicarbonate. ...
Full-text available
Several structural, treatment, and management approaches exist to minimize phosphorus (P) transport from agricultural landscapes (e.g., cover cropping and conservation tillage). However, many of these practices are designed to minimize particulate P transport and are not as effective in controlling dissolved P (DP) losses. Phosphorus removal structures employ a P sorption material (PSM) to trap DP from flowing water. These structures have been successful in treating surface runoff by utilizing aluminum (Al)-treated steel slag, but subsurface tile drainage has never been tested with this material. The goal of this study was to evaluate the performance and economics of a ditch-style P removal structure using Al-treated steel slag for treating agricultural subsurface drainage discharge. The structure treated subsurface drainage water from a 4.5 ha agricultural field with elevated soil test P levels. Overall, the structure removed approximately 27% and 50% of all DP and total P (TP) entering the structure, respectively (i.e., 2.4 and 9.4 kg DP and TP removal). After an initial period of strong DP removal, the discrete DP removal became highly variable. Flow-through analysis of slag samples showed that the slag used to construct the structure was coarser and less sorptive compared to the slag samples collected prior to construction that were used to design and size the structure. Results of this study highlight the importance of correctly designing the P removal structures using representative PSMs.
Riverbed sediments are the interface layer in riverine ecosystems connecting the overlying medium of water and the vadose zone. The transport behavior of phosphorus (P), which has been recognized as the primary cause of freshwater eutrophication, in riverbed sediments remains unclear. Understanding the impact of riverbed sediments on P transport is a necessary prerequisite for the development of appropriate strategies to reduce potential groundwater pollution. In this study, riverbed sediments were collected from the upstream, midstream, and downstream sections of the Beiyun River, China, and packed into vertical soil columns to perform leaching experiments to quantify P transport characteristics. In addition, the impact mechanisms were further explored by conducting laboratory batch tests of P adsorption and desorption. The results demonstrated that approximately 80% of P can be adsorbed by riverbed sediments in soil column leaching experiment, and a tailing phenomenon was observed in its desorption. The hydraulic conductivity properties of riverbed sediments were evaluated by the advection-dispersion equation, showing a gradually decreasing adsorption capacity for P from upstream to downstream sections, which was supported by the results obtained from adsorption–desorption thermodynamic and kinetic batch tests. The estimated annual leaching masses of P increased from 60.72 g/(m² a) in the upstream section to 132.31 g/(m² a) in the downstream section. The role of riverbed sediments as a source or sink of P is possibly determined by their coarse sand particles content, and the mean equilibrium P concentration (EPC0). The competitive relationship between P and other forms of nutrients also has an important influence on its source-sink role. These findings suggest that the prevention of the potential P leaching is most needed in the downstream sections of Beiyun River, and corresponding control strategies should be developed to avoid groundwater pollution.
Full-text available
Best management practices (BMPs) are site-specific and their implementation, long-term management, and maintenance are important for successful reduction of phosphorus (P) loss into headwater streams. This paper reviews published research on managing P loss from agricultural cropping systems in the Midwestern United States and classified the available research based on BMPs and their efficacy in reducing P loss. This review paper also identifies the areas where additional research could provide insight for managing P losses. Our literature review shows that cover crops, reduced tillage, saturated buffers, and constructed wetlands are the most evaluated areas of current research. However, additional research is necessary on the site-specific area to measure the effectiveness of BMPs in managing P loss. The BMPs that serve as a sink of P need further evaluation in long-term field-scale trials. Studies evaluating adsorption and desorption mechanisms of P in surface and subsurface soils with materials or amendments that bind P in the soil are needed. The time required and pathways, where the flush of available P is lost or fixed in the soil matrix, need further investigation. Measured P loss from BMPs like bioreactors and saturated buffers supplemented with P adsorption materials or filters need to be simulated with models for their prediction and validation. Field evaluations of P index and critical source area concepts should be investigated for identifying problematic areas in the watersheds. Identification of overlapping areas of high P source and transport can help in strategic planning and layout, thereby resulting in reducing the cost of implementing BMPs at field and watershed scales.
Phosphorous is one of the important pollutants, which exacerbates algal bloom in aqueous ecosystems. Also, presence of phosphorous (P) in surface water lets to a significant increase of water treatment cost at treatment plants. Many researches have made efforts to remove phosphorous from aqueous solution. One of these efforts is using adsorption process onto the industrial by-products, such as steel slag (SS). So, in this study, we are going to do a systematic review about the mentioned issue. All references were 137 in number, from which duplicates were removed manually and n = 91 records remained. After reading their titles, abstracts and full-texts, 59 relevant full-text articles remained in this systematic review. The relevant publications were from 1997 to 2019. In the studies, type of SS in different applications for removal of phosphorous at vary of pH value were used. The Electric Arc Furnace (EAF) SS (36%) and filtration system (37%) had many applications. Moreover, most relevant study (58%) reached to the optimum condition at alkaline pH value. In final, this review revealed that SS is promising and efficient material for removal of phosphorous from water and wastewater solutions. In addition, future perspectives of the application of SS as an industrial by-product were discussed in this study.
Full-text available
Excessive soil phosphorus (P) concentrations among cattle loafing areas located in close proximity to surface waters represent great potential for P transport. This study assessed the ability of several P sorbing materials in reducing P losses from streamside cattle loafing areas. Simulated rainfall was applied at seven (time 1) and 28 (time 2) days after P sorbing material applications to runoff plots on cattle loafing areas located at Amish farms. Treatments consisted of alum, water treatment residuals, fly-ash, gypsum, and no amendment (control). Alum addition reduced time 1 runoff P concentrations the most followed by water treatment residuals - gypsum, then fly-ash. However, runoff P losses from P sorbing materials were not significantly different from the control at time 2. These results suggest that P sorbing materials alone provide only a temporary solution to P losses from cattle loafing areas and should be used with other best management practices.
Full-text available
The potential of six industrial by-products for use as phosphorus-sorbing materials (PSMs) in solutions was evaluated. These included two different acid mine drainage treatment residuals (AMDR1 and AMDR2), water treatment residual (WTR), fly ash, bauxite mining residual, and flue gas desulfurization product (FGD). Characterization of the by-products and their mechanisms for sorption and retention of inorganic phosphorus (P) from solution identified those PSMs that sorbed primarily by an iron and aluminum (Fe/Al) mechanism, those that sorbed primarily by a calcium and magnesium (Ca/Mg) mechanism, and those that sorbed by both mechanisms. Degree of P sorption and associated mechanisms were strongly influenced by the pH, buffer capacity, ionic strength, and common ion effects.
Full-text available
Application of animal manures in excessive amounts can result in surface runoff of nutrients and degradation of surface water. Best management practices that use chemical or by-products to sorb nutri-ents can reduce nutrient loss from agricultural land. The objective of this work was to determine the ability of water treatment residual (WTR) to reduce N and P runoff from land treated with poultry litter. Different WTR (ABJ or WISTER) were used in two experiments different locations. Three WTR treatments were applied to plots that received poultry litter at 6.72 Mg ha -I broadcast on bermudagrass [Cynodon dactylon (L.) Pers.] pasture. Treatments were broadcast (11.2 or 44.8 Mg ha 1), and a buffer strip (44.8 Mg -1) tothebottom 2.44 m of the plot. Experimental plots received simulated rainfall for 75 min at 6.35 cm h -1 within 24 h of litter and WTR application. Nitrogen, NH4, P, AI, and dissolved solids in surface runoff were determined. Mean dissolved P of 15.0 mg L -1 was reduced to 8.60 mg L -~ by the high broadcast and to 8.12 mg L -~ by the buffer strip ABJ treatments. Reductions in runoff P were attributed to amorphous AI in the WTR. Soluble NH~-N was reduced from 33.7 to 11.3 mg L ~ (high broadcast) and to 17.9 mg -~ (buffer s trip) by ABJ. W IS-TER did not, however, reduce soluble NH4-N or total N. Reduction in NH~-N was related to cation-exchange capacity of the WTR. Land application of WTR did not increase dissolved solids or AI in sur-face runoff.
Full-text available
Reducing phosphorus (P) loads from soils to surface waters is necessary for solving the problem of eutrophication. Many industrial by-products have been shown to sorb appreciable amounts of dissolved P from solution and it has been proposed to use P sorption materials (PSMs) such as steel slag in landscape scale "filters" for trapping dissolved P in runoff. The objective of this study was to model the effect of retention time (RT) and P concentration on P sorption by steel slag and a surface modified slag in a flow-through system. Sorption of P onto steel slag and rejuvenated-modified steel slag was measured using a traditional batch isotherm and a flow-through setting at several RTs and P concentrations. Flow-through data were used to produce a model that estimated P sorption based on RT and P concentration. The model was tested on a pilot-scale pond filter consisting of the same slag materials. For both the materials, flow-through tests indicated an increase in RT increased P removal efficiency but decreased the total amount of P removed at saturation. The Langmuir model developed from batch isotherms overestimated and underestimated P sorption in normal and rejuvenated slag respectively, relative to flow-through. Normal and rejuvenated slag removed 38 and 36% of P in the pilot-scale pond filter after 2 weeks of pumping. The Langmuir equation poorly predicted P sorption in the pond filter while the flow-through model produced reasonable estimates. Results suggest that flow-through methodology is necessary for estimating P sorption in the context of landscape P filters.
Full-text available
Phosphorus (P) losses to surface waters can result in eutrophication. Some industrial by-products have a strong affinity for dissolved P and may be useful in reducing nonpoint P pollution with landscape-scale runoff filters. Although appreciable research has been conducted on characterizing P sorption by industrial by-products via batch isotherms, less data are available on P sorption by these materials in a flow-through context integral to a landscape P filter. The objectives of this study were to evaluate several industrial by-products for P sorption in a flow-through setting, to determine material chemical properties that have the greatest impact on P sorption in a flow-through setting, and to explore how retention time (RT) and P concentration affect P removal. Twelve materials were characterized for chemical properties that typically influence P removal and subjected to flow-through P sorption experiments in which five different RTs and P concentrations were tested. The impact of RT and P concentrations on P removal varied based on material chemical properties, mainly as a function of oxalate-extractable aluminum (Al), iron (Fe), and water-soluble (WS) calcium (Ca). Statistical analysis showed that materials elevated in oxalate-extractable Al and Fe and WS Ca and that were highly buffered above pH 6 were able to remove the most P under flow-through conditions. Langmuir sorption maximum values from batch isotherms were poorly correlated with and overestimated P removal found under flow-through conditions. Within the conditions tested in this study, increases in RT and inflow P concentrations increased P removal among materials most likely to remove P via precipitation, whereas RT had little effect on materials likely to remove P via ligand exchange.