![Total inorganic nitrogen (TIN; μmol/L) versus ²²⁸Ra (disintegrations per minute [dpm]/L) in lake groundwater samples.](https://www.researchgate.net/publication/331549919/figure/fig4/AS:11431281250878275@1718047376764/Total-inorganic-nitrogen-TIN-mmol-L-versus-Ra-disintegrations-per-minute-dpm-L_Q320.jpg)
A preview of this full-text is provided by Wiley.
Content available from Journal of Geophysical Research: Oceans
This content is subject to copyright. Terms and conditions apply.
Transport of Radium and Nutrients Through Eastern
South African Beaches
Willard S. Moore
1
, Marc S. Humphries
2
, Claudia R. Benitez‐Nelson
1
, Letitia Pillay
2
,
and Caldin Higgs
2
1
School of the Earth, Ocean and Environment, University of South Carolina, Columbia, SC, USA,
2
Molecular Sciences
Institute, School of Chemistry, University of the Witwatersrand, Johannesburg, South Africa
Abstract Submarine groundwater discharge (SGD) plays a critical role in coastal and ocean
biogeochemistry. Elucidating spatially and temporally heterogeneous SGD fluxes is difficult. Here we use
radium isotopes to explore the external sources and mixing regime along the eastern coast of South Africa.
We demonstrate that the long‐lived radium isotope compositions are controlled by low inputs of low‐and
high‐salinity terrestrial groundwater. While activities of
228
Ra and
226
Ra in beach porewaters are similar to
coastal waters,
224
Ra is enriched by inputs of
228
Th from coastal seawater. Porewater ages, based on the
production of
224
Ra from
228
Th, range from 0.3 to 2.3 days, indicating rapid flushing of the beach system.
Unlike radium, however, nutrients follow a more complex pathway. We hypothesize that high total
dissolved nitrogen (TDN) and phosphorus concentrations in beach porewaters (TDN ranges from 1 to
>700 μM) and the coastal ocean (TDN ranges from 1 to >40 μM) are derived from a source not enriched in
radium. We speculate that this source is terrestrial water flowing below the dune barrier at depths exceeding
our beach sampling depths. This water likely flows upward through breaches in the confining layer into
the beach or enters the ocean directly through paleochannels. The presence of high nutrient concentrations
in the coastal ocean unaccompanied by high
228
Ra activities leads to the hypothesis of this additional
nutrient source. These combined inputs may be of considerable importance to the coastal ecology of
southeastern Africa, an oligotrophic ecosystem dominated by the nutrient‐poor Agulhas Current.
Plain Language Summary Waves and tides constantly move water through beaches. This may
appear to be a simple circulation of seawater through the beach, but it is more complex. Inputs of
groundwater modify the composition of the beach porewater. This mixture is further changed by bacteria
that oxidize organic material within the beach and release nutrients, dissolved nitrogen and phosphorus.
Thus, the water that returns to the ocean is quite different chemically from the incoming seawater. Here we
use naturally occurring radium isotopes to trace these processes. We estimate that the beach porewater is
flushed every 1–2 days. This does not allow much time for the bacteria in the beach to strongly modify the
chemical composition within the porewater. However, in some cases we measured high nutrient
concentrations in beach porewaters and in the adjacent ocean. These sites were often coincident with
obvious seepage of water emerging from the beach and by algae growth on rocks. We speculate that these
high nutrient concentrations result from the movement of deeper groundwater into the beach or directly
offshore through ancient river channels. These nutrient inputs are important for maintaining the biological
productivity of the area in a region where nutrient concentrations in the ocean are low.
1. Introduction
As waves run up on beaches, a significant fraction of the seawater infiltrates the sand and returns to the
ocean as submarine groundwater discharge (SGD). While tides have longer wavelengths, the tidal cycle
causes a similar infiltration and return flow (McLachlan et al., 1985). Combined with terrestrial inputs of
both low‐salinity groundwater (LSGW) and high‐salinity groundwater (HSGW), these processes may signif-
icantly alter the biogeochemical reactions of coastal sediments and their porewaters and thus the composi-
tion and the magnitude of nutrient and trace metal release into the coastal ocean (Figure 1; Anschutz et al.,
2009 and references therein). Numerous studies have shown that SGD plays a critical role in both coastal and
open ocean biogeochemistry (e.g., Knee & Paytan, 2011; Moore, 2003a, 2010; Paytan et al., 2006; Taniguchi
et al., 2002), contributing as much or more than freshwater riverine discharge (e.g., Kwon et al., 2014; Moore
et al., 2008). Elucidating the magnitude and composition of SGD fluxes is difficult as they are spatially and
©2019. American Geophysical Union.
All Rights Reserved.
RESEARCH ARTICLE
10.1029/2018JC014772
Key Points:
•
224
Ra activities are highly enriched
relative to
228
Ra activities in South
African beach porewaters
•Nutrient‐rich beach porewaters are
rapidly flushed from eastern South
African beaches
•Paleochannels and a rocky
semiconfining layer are likely major
pathways of episodic nutrient input
into the coastal ocean
Supporting Information:
•Supporting Information S1
Correspondence to:
W. S. Moore,
moore@geol.sc.edu
Citation:
Moore, W. S., Humphries, M. S.,
Benitez‐Nelson, C. R., Pillay, L., &
Higgs, C. (2019). Transport of radium
and nutrients through eastern South
African beaches. Journal of Geophysical
Research: Oceans,124, 2010–2027.
https://doi.org/10.1029/2018JC014772
Received 20 NOV 2018
Accepted 21 FEB 2019
Accepted article online 5 MAR 2019
Published online 23 MAR 2019
MOORE ET AL. 2010
temporally heterogeneous. This is particularly true in the Southern Hemisphere off the coast of Africa (Knee
& Paytan, 2011; Petermann et al., 2018).
There are three main components of SGD: meteoric (fresh) water, circulated seawater, and connate ground-
water, that is, older water, which may contain salts dissolved from the host aquifer (Burnett et al., 1996). In
coastal regions, most SGD studies typically focus on fresh SGD or salty SGD that results from the infiltration
and exchange of seawater (Knee & Paytan, 2011). While salinity is often used to determine if there is a fresh
groundwater component in beach porewaters, it does not adequately discern other sources of salty SGD.
Here we use radioactive disequilibrium within the
232
Th decay series to investigate SGD and mixing within
coastal ocean sandy beaches. As long‐lived
232
Th decays, two radium daughters (
228
Ra, half‐life = 5.7 years,
and
224
Ra, half‐life = 3.66 days) and one thorium daughter (
228
Th, half‐life = 1.9 years) are produced
(Figure 1). Because thorium is a ubiquitous component of most rocks, small concentrations exist in almost
all sand. However, this thorium primarily resides in the crystal lattice of minerals, meaning the daughter
products are not typically released as seawater circulates over these coarse particles. Because of its slow rate
of regeneration after being washed from the sand, dissolved
228
Ra activities in excess of that produced from
mineral‐bound
232
Th decay require an external source. This source could be either fresh or salty SGD origi-
nating farther inland (Rama & Moore, 1996) and results in “excess”
228
Ra in coastal ocean waters that can be
used as a tracer to provide an integrated estimate of SGD over larger spatial and temporal scales (Figure 1).
The presence of dissolved
228
Ra in seawater leads to the continual production of
228
Th in the nearshore
ocean. This
228
Th is scavenged from seawater by biodetritus and other particles on a time scale of days to
weeks (Nozaka et al., 1987). As seawater infiltrates the beach, both dissolved
228
Th and
228
Th that has been
scavenged by particles may become enriched in the beach sand. On the time scale of its mean life, 2.5 years,
228
Th is expected to reach a steady state such that additions of
228
Th from the ocean are balanced by loss via
radioactive decay. Because
228
Ra is slowly regenerated from
232
Th, very little is expected to be released from
beach sand during seawater infiltration, which occurs on daily time scales. In contrast, the shorter half‐life of
224
Ra results in it readily being regenerated in beach sands from
228
Th decay. This process leads to an in situ
source of
224
Ra in excess of that originating from seawater and groundwater infiltration and leads to high
224
Ra/
228
Ra activity ratios (ARs) in beach porewater. Bokuniewicz et al. (2015) and Tamborski et al.
(2017) demonstrated that the residence time of beach porewater may be obtained by quantifying (1) the
amount of
224
Ra originating from seawater and additions from SGD and (2) the amount generated by
228
Th decay in the beach sand and released back into the beach porewater.
While there have been numerous studies of SGD worldwide, there have been few direct studies of SGD along
the African coast (Petermann et al., 2018). Such information is particularly important along the southeastern
coast of Southern Africa, the location of the UNESCO World Heritage Site, iSimangaliso Wetland Park. The
park includes 280 km of coastline, where adjacent coastal waters contain over 100 identified species of hard
and soft corals, sponges, and many other invertebrates and more than 1,200 species of fish (Connell & Porter,
Figure 1. Conceptual model of water flow through beaches. Inputs include seawater (SW), low‐salinity groundwater
(LSGW), and high‐salinity groundwater (HSGW). Decay of
228
Ra in the coastal waters produces
228
Th, which may
enter the beach along with the SW. Decay of
228
Th in the sediments produces
224
Ra, some of which becomes mobile upon
water intrusion. Biogeochemical reactions within the beach alter the composition of the porewaters within the mixing
zone. These processes produce submarine groundwater discharge (SGD) with a composition quite different from the
original end‐members.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2011
2013). Here we measured
228
Ra,
228
Th, and
224
Ra activities and inorganic nutrient concentrations in beach
porewaters and terrestrial groundwaters within iSimangaliso Wetland Park, South Africa, in 2012, 2013, and
2014. We compare these results with other studies, such as Patos Lagoon, Brazil; the Wadden Sea, Germany;
the South and East China Seas; Smithtown Bay, NY; and the southeastern coast of the United States. Our
results suggest that although some of these sites (e.g., Patos Lagoon) are comparable with regard to their
hydrology, they differ fundamentally in the relative sources and inputs of fresh and salty SGD to the coastal
ocean, confirming the critical importance of continued investigations of these tracers along different
sandy coasts.
2. Materials and Methods
2.1. Site Description
Lake St. Lucia is a large (350 km
2
) shallow estuarine lake system that is situated on an extensive sand‐
covered coastal plain within iSimangaliso Wetland Park (Figure 2). The lake stretches ~70 km along the east
coast of South Africa and is separated from the ocean by a 100‐to 140‐m‐high vegetated dune barrier. The
only contemporary surface link to the ocean is via a sinuous channel in the south of the study area known
as the Narrows, although the estuary mouth is prone to prolonged periods of closure, as it was during our
investigations. The lake is fed by five rivers, with the Mkhuze River in the north and the Mfolozi River in
the south supplying the largest quantities of freshwater. The rivers are seasonal, flowing primarily during
the wet summer. With an average water depth of ~1 m, the lake basins are susceptible to extreme changes
in ecological state as a result of evaporation and variable river inflow.
Several other large and ecologically important water bodies occur in the area, namely, Lake Sibaya and Kosi
Bay. Unlike Lake St. Lucia, the Sibaya and Kosi systems maintain relatively deep basins (>20 m) and lack
significant surface inputs, suggesting they are almost exclusively groundwater fed. Lake Sibaya is an isolated
freshwater lake, while Kosi Bay maintains a permanent connection to the ocean via an estuary mouth.
Together, these systems form part of iSimangaliso Wetland Park, which spans 3,320 km
2
and was named
a UNESCO World Heritage Site in 1999.
The Maputaland coastal plain hosts an extensive unconsolidated primary aquifer (Meyer et al., 2002).
Groundwater flow from this aquifer plays an important role in maintaining lake levels. Consolidated
Cretaceous siltstones, consisting of the St. Lucia, Mzinene, and Makatini Formations, underlie much of
Maputaland to form the basement of the unconfined aquifer (Figure 3; Maud, 1980). The majority of the sedi-
mentary succession above the Cretaceous basement can all be treated as potential aquifer units that include
calcarenites of the Uloa Formation and loosely consolidated sands and clays of the Port Durnford Formation
(Hobday & Orme, 1974; Miller, 2001). The Port Durnford Formation is overlain by fluvial and eolian sands of
middle to upper Pleistocene and Holocene age. These sands are predominantly fine grained and largely
unconsolidated (Meyer et al., 2002). The Holocene cover sands exceed a thickness of 70 m in places.
The large coastal lakes in the area occupy river valleys that were incised into the underlying Cretaceous bed-
rock during the Last Glacial Maximum regression, ~18–20 kyr BP (Wright et al., 2000). The continental shelf
in this region is characterized by several incised submarine canyons of varying size, which extend from
Leven Point in the south to Bhanga Nek in the north (Figure 3). These paleovalleys represent previous drai-
nage systems that discharged onto the exposed continental shelf and were subsequently drowned during the
postglacial sea‐level transgression. Rising sea levels and coastal barrier accretion during the Holocene ulti-
mately sealed these outlets, prompting the development of backbarrier lakes. Ephemeral flow via the com-
bined Mfolozi and St. Lucia estuary mouth is the only major fluvial outlet to the ocean south of Kosi Bay.
This outlet was closed at the times of sampling.
Water levels in Lake St. Lucia and Kosi Bay are perched approximately 1 m above sea level, while Lake
Sibaya maintains a head of approximately 20 m. Between 2006 and 2016 the Maputaland region experienced
a prolonged drought, one of the longest on record. This led to hypersaline conditions in Lake St. Lucia, with
large areas of the lake basin drying out.
Expeditions to Lake St. Lucia in January and August 2012 and November 2013 collected surface and ground-
water samples from the lakes and adjacent beaches along the eastern coast (Figure 2). In October 2014, we
revisited Lake St. Lucia and extended our sampling northward along the coast to include Lake Sibaya and
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2012
Figure 2. (a) Study area, (b) larger‐scale map of iSimangaliso Wetland Park showing northern sampling sites collected in 2013 and 2014, and (c) detailed map of
Lake St. Lucia and the coastal barrier with 2012–2014 sampling.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2013
Kosi Bay. Additional samples were collected along two short (~300 m) transects in 2012, one extending from
the shoreline at Lake St. Lucia into the lakes (Station 3) and the second extending from the St. Lucia estuary
mouth toward the ocean (Station 5). Transect sampling at the estuary mouth was repeated in 2014.
Groundwater and porewater samples (typically 10 L) were collected from established monitoring wells
(Stations 19 and 20) and from PushPoint samplers inserted 0.5–2 m into the ground (Moore et al., 2006).
The penetration of the PushPoint into beach sands was occasionally limited by an underling rock layer
that frequently outcropped at the shore. This layer could act as a confining layer; however, the outcrops
suggested that it is often fractured. In one case (Station 5 in 2014) we first dug a hole in the beach and
Figure 3. Regional geological map of the Maputaland coastal plain (from data of Watkeys et al., 1993) showing the location of major paleochannels and former
drainage outlets associated with prominent coastal lakes (Benallack et al., 2016; Green, 2009; Wright et al., 2000).
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2014
inserted the PushPoint to a depth of 3.1 m. Surface water samples (20–50 L) were collected in sample‐rinsed
cubitainers or large carboys. Approximately 8 sites were targeted in 2012 (n= 18), 22 sites that overlapped
with 2012 sampling were targeted in November 2013, and 39 sites (n= 70) that overlapped with both
2012 and 2013 were targeted in October 2014 (Figure 2). In October 2014, we targeted stations that had active
seepage as identified visually. Salinity and temperature were measured in the field using a portable multi-
parameter meter calibrated against appropriate conductivity standards.
2.2. Radium and Nutrient Analyses
Samples for radium isotopes were allowed to stand until particles could settle in the few samples that con-
tained high particulate matter concentrations; the clear water was subsequently decanted and passed
through a column of Mn fiber to remove radium (Moore, 1976). Samples were returned to the University
of South Carolina where
224
Ra and
228
Th were measured on a radium delayed coincidence counter
(RaDeCC) system (Moore & Arnold, 1996). Almost all samples yielded measurable activities of both isotopes;
however, we do not report
224
Ra results for samples greater than 15 days old (four
224
Ra half‐lives; Moore,
2008). After the RaDeCC measurements were complete, the samples were leached with a mixture of hydro-
chloric acid and hydroxylamine hydrochloride and filtered. Dissolved Ba was added followed by sulfate to
precipitate BaSO
4
, which carries radium. The precipitate was measured in a well germanium detector for
226
Ra and
228
Ra (Moore, 1984).
Total and dissolved inorganic nutrient samples (15–40 mL) were collected into high‐density polyethylene
centrifuge tubes after prerinsing with filtered sample. Samples were either collected directly or filtered with
an inline 0.45‐μm glass fiber filter (GF/F) in the field prior to being stored refrigerated and then frozen.
Samples were returned to the University of South Carolina for measurement. Analyses for total inorganic
nitrogen (TIN; NO
3
,NO
2
, and NH
4
) were made using a Lachat QuickChem Flow Injection Analysis
+8000 series following established sampling protocols (Lachat QuikChem methods). Lower limits of detec-
tion were 0.05 μM. Phosphate or soluble reactive phosphorus concentrations were determined colorimetri-
cally (Koroleff, 1983) with a detection limit 0.07 μM.
2.3. Leaching Experiments and Determination of K
d
224
Ra may be leached from beach sands during seawater infiltration. In order to determine the
224
Ra activity
released during seawater infiltration, four sand samples from 1‐to 2‐m depth were collected from the bea-
ches on the ocean side of Lake St. Lucia within iSimangaliso Wetland Park. Dried aliquots were sealed in
plastic vials, and the total activity of
228
Th intrinsic to the sand was determined using gamma spectrometry
(Moore, 1984).
228
Th activities ranged from 0.2 to 11 disintegrations per minute (dpm)/g. Based on these ana-
lyses, about 30 g of the different sand samples was leached with 500 ml of seawater that had been passed
through Mn fiber to remove radium (Moore, 1976). The samples were mixed well several times over a period
of 2 hr, and the sediment was allowed to settle between mixing events. The seawater was decanted and fil-
tered through a 0.7‐μm GF/F and then passed through a column of Mn fiber to adsorb released Ra. The Mn
fiber was washed with distilled water, partially dried and measured in a RaDeCC system (Moore & Arnold,
1996). Very low activities, near the detection limit of
224
Ra, were measured. To obtain more precise measure-
ments, the leaching experiment was repeated on fresh beach sands (N= 3) using about 150 g of sand and
1.5 L of seawater following the procedure outlined above.
Some of the
224
Ra produced from
228
Th decay on the surface of the beach sand will remain adsorbed and not
released to beach porewaters. Experiments were conducted using these sands to measure the partition (dis-
tribution) coefficient, K
d
, following the procedure described in Beck and Cochran (2013). Dried sand sam-
ples (~25 g each) were mixed with seawater spiked with known activities of
228
Th in equilibrium with
224
Ra (about 40 g containing 1.7 dpm/g at pH = 8) for 1 hr. The mixture was filtered and the recovered liquid
volume measured. The procedure of Moore and Cai (2014) was used to separate
224
Ra based on precipitating
MnO
2
by mixing KMnO
4
and MnCl
2
solutions. The resulting MnO
2
precipitate was mixed, filtered onto a
0.7‐μm GF/F, and washed with distilled and deionized water. The GF/F was counted directly on the
RaDeCC system after being inserted into a column containing a small wad of moist acrylic fiber.
To calculate K
d
,wefirst determined the activity of
224
Ra that remained adsorbed to the sand (q
e
):
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2015
qe¼AL−Ai
ðÞVL
Ms
;(1)
where A
L
is the measured activity in the liquid after equilibration (dpm/g), A
i
is the initial activity in the
liquid, V
L
is the volume of liquid, and M
S
is the mass of sand. From this we obtain an apparent distribution
coefficient, K*
d:
K*
d¼qe
AL
:(2)
Here K*
dis related to K
d
by the following equation (Beck & Cochran, 2013):
Kd¼K*
dϕ
ρs1−ϕðÞ
;(3)
where ϕis the porosity and ρ
s
is the dry density of the sand. Note the units cancel out, so K
d
is dimensionless.
3. Results
3.1. Surface Water, Groundwater, and Porewaters
All surface water, groundwater, and porewater results are given in Tables 1–3 for iSimangaliso Wetland
Park. Lake St. Lucia groundwater salinities ranged from <1 to >35; there were no annual or larger‐scale spa-
tial trends (e.g., declining salinity with increasing distance from St. Lucia Estuary mouth). Rather, small‐
scale spatial variability was driven at least in part by distance from the lake shore, with samples collected
closest to shore having higher salinities. Lake St. Lucia surface water salinities were less variable, averaging
18.7 ± 5.8. The enclosed freshwater lakes, Lake Bhangazi and Lake Sibaya, had surface and groundwaters
that were significantly fresher, <5. Beach porewaters and ocean surface samples were characterized by sali-
nities >30, with the exception of porewater samples collected across the St. Lucia Estuary barrier (Station 5)
and near the mouth of the Mflozoi River. In 2014, active seepage with a salinity of 7.5 was observed at Station
46 (on the beach adjacent to Lake Sibaya).
224
Ra activities within iSimangaliso Wetland Park ranged from below detection (BD) to 15 dpm/L
(excess
224
Ra in groundwater samples collected adjacent to the lakes); there were no annual or larger‐
scale spatial trends.
226
Ra and
228
Ra activities were highest in groundwater samples collected adjacent
to the lakes (0.56 ± 0.41 dpm/L [BD to 1.49 dpm/L] and 3.83 ± 3.40 dpm/L [0.06 to 10.5 dpm/L],
respectively) and were significantly lower (p< 0.001) in lake surface waters (0.27 ± 0.14 dpm/L [0.06
to 0.59 dpm/L] and 0.96 ± 0.58 dpm/L [0.09 to 2.61 dpm/L], respectively). Ocean surface and beach
porewater
226
Ra activities were lowest and did not differ from one another (~ 0.09 ± 0.08 dpm/L).
While beach porewater
228
Ra activities were also significantly lower than those measured in ground-
water samples collected adjacent to the lakes (0.77 ± 1.38 dpm/L [0.08 to 6.74 dpm/L]), they were still
significantly higher than those measured in ocean surface waters (0.14 ± 0.15 dpm/L; p< 0.001). Excess
224
Ra was highest in groundwater samples collected adjacent to the lakes (4.89 ± 4.01 dpm/L [BD to
15.11 dpm/L]) and beach porewaters (3.42 ± 2.05 dpm/L [0.10 to 9.87 dpm/L]), followed by lake surface
waters (1.08 ± 0.77 dpm/L [0.04 to 3.69 dpm/L]) and ocean surface waters (0.33 ± 0.22 dpm/L). While
228
Ra versus
224
Ra activities were well correlated in groundwater samples collected adjacent to the lakes
(R
2
= 0.84, Slope = 1.06), and consistent with the ingrowth of
224
Ra from
228
Ra decay via
228
Th, there
were no similar correlations in lake and ocean surface waters or in beach porewaters. Although more
variable, radium activities were also well correlated with salinity in groundwater samples collected adja-
cent to the lakes (R
2
= 0.56).
Relative to the radium activities, nutrient concentrations were generally high and more variable. TIN
and phosphate concentrations ranged from BD to 1,400 μM and from BD to 21.6 μM, respectively
(Tables 1–3). Lake surface and groundwater and beach porewater had similar TIN concentrations and
ranges (111 ± 270 μM [0.7 to 1,151 μM], 116 ± 236 μM [0.1 to 1,404 μM], and 91 ± 135 μM [1.4 to
784 μM], respectively) and were generally higher than those measured in ocean surface waters
(42 ± 112 μM [0.6 to 519 μM]). In contrast, phosphate concentrations in groundwater samples collected
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2016
adjacent to the lakes were more variable and significantly higher than those measured in the other sample
types (2.63 ± 4.64 μM [BD to 21.6 μM] versus 0.80 ± 0.36 μM [BD to 1.6 μM], p< 0.001). There were
significant correlations between TIN concentrations and
228
Ra (
224
Ra) activities in the groundwater
samples collected adjacent to the lakes (Figure 4), but not in lake surface waters or beach porewaters. We
will later discuss the irregular TIN concentrations and
228
Ra activities in ocean surface waters. There were
no significant correlations between radium activities and phosphate concentrations in any of the
sample types.
3.2. Leaching Experiments
The partition coefficients (K
d
) determined from our leaching experiments are given in Table 4. Using
these K
d
values, we then calculated the amount of
224
Ra that remained adsorbed to the sand and thus
the total leachable
224
Ra (Table 5). The activity of total leachable
224
Ra ranged from 0.01 to 0.04 dpm/g
and averaged 0.024 ± 0.010 dpm/g. We can convert
224
Ra activities to a per‐liter basis by multiplying
by the dry sand density (2.7 g/ml), 1 −ϕ(0.6), and 1,000 mL/L to obtain an average
228
Th activity of
39 ± 16 dpm/L in the beach sand that produces leachable
224
Ra. This does not include
228
Th locked in
the mineral grains.
4. Discussion
4.1. Sources of Radium
The
228
Ra activities in the beach porewaters must result from external sources, as a very small fraction of the
total radium associated with the sand could be leached. Once the leachable
228
Ra is washed from the sand,
months are required to regenerate significant activities. On the other hand,
224
Ra can regenerate quickly, on
a time scale of hours; thus,
228
Th on the surface of the beach sand or in particles associated with the sand
provides a continuous source of
224
Ra.
The most striking aspects of the radium data are the beach porewater data from the ocean side of St. Lucia,
where extremely high activities of
224
Ra were measured relative to
228
Ra (Figure 5). In contrast, beach pore-
waters from other locations around the world have considerably lower activities of
224
Ra compared with
228
Ra, with typical ARs closer to 2:1. The extremely high
224
Ra/
228
Ra AR along the eastern coast of South
Africa makes this location unique among sites studied.
Table 1
Samples Collected in 2012 From iSimangaliso Wetland Park, South Africa
Station Location Sample ID/type Long. (E) Lat. (S) Depth (m) Salinity
226
Ra
228
Ra ex
224
Ra
228
Th TIN PO
4
5a St. Lucia Estuary Beach PW 32.422 28.384 −1.5 41.2 0.60 6.74 9.87 0.12 141 0.16
5b St. Lucia Estuary Beach PW 32.423 28.384 −1.5 4.5 0.06 0.53 1.19 0.01 60.1 0.22
5c St. Lucia Estuary Beach PW 32.425 28.384 −1.5 1.0 0.01 0.08 0.10 0.00 206 0.10
5d St. Lucia Estuary Beach PW 32.426 28.384 −1.5 35.0 0.14 1.13 3.61 0.05 32.1 0.16
3a Lake St. Lucia Lake Shore GW 32.486 28.221 −0.5 9.0 0.28 1.47 BD BD 26.8 BD
3b Lake St. Lucia Lake Shore GW 32.483 28.222 −0.5 38.8 0.96 4.72 3.57 0.07 110 0.15
3c Lake St. Lucia Lake Shore GW 32.486 28.221 −0.5 57.6 0.51 4.82 7.90 0.11 74.5 2.56
3d Lake St. Lucia Lake Shore GW 32.484 28.221 −0.5 22.2 0.93 5.70 7.02 0.13 70.8 0.63
3e Lake St. Lucia Lake Surface 32.483 28.222 0 8.1 0.19 0.96 0.59 0.03 1151 1.38
5e St. Lucia Estuary Lake Surface 32.423 28.383 0 8.0 0.32 1.22 BD BD 41.8 0.07
5e St. Lucia Estuary Lake Surface 32.421 28.384 0 13.2 0.43 1.63 0.98 0.03 32.5 0.25
6 Lake Bhangazi Lake Surface 32.538 28.136 0 1.0 0.08 0.14 BD BD 230 0.10
4 Lake St. Lucia Ocean Surface 32.424 28.393 0 35.7 0.13 0.14 BD BD 37.3 0.15
5g Lake St. Lucia Ocean Surface 32.423 28.383 0 34.8 0.08 0.12 0.32 0.01 47.6 0.14
7 Lake St. Lucia Ocean Surface 32.558 28.121 0 36.7 0.08 0.11 BD BD 36.1 0.18
13 Lake St. Lucia Ocean Surface 32.487 28.276 0 34.8 0.09 0.08 0.19 0.01 18.7 1.27
19 Lake St. Lucia Permanent Well 32.489 28.193 >10 0.0 0.16 0.20 BD BD 39.9 23.1
20 Lake St. Lucia Permanent Well 32.487 28.220 >10 0.0 0.06 0.08 BD BD 33.6 0.02
Note. TIN = total inorganic nitrogen; PW = porewater; GW = groundwater. Radionuclide activities are dpm L
−1
, nutrient concentrations are μM. Lower case
letters denote samples collected along a transect. BD is below detection.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2017
Why do these beach porewaters have such differences in
224
Ra and
228
Ra? One primary difference is that the
activity of
228
Ra dissolved in most St. Lucia beach porewater is 4 to 40 times lower than the activities mea-
sured in other beaches. Numerous studies have shown that the high
228
Ra activities measured in these other
upper porewaters are due to the supply of
228
Ra from external sources, namely, groundwater (Moore, 2003b,
2010; Rama & Moore, 1996; Tan et al., 2018; Windom et al., 2006). We therefore argue that the low
228
Ra
activities measured in St. Lucia must be due to differences in the groundwater components that flow into this
coastal system.
Moore (2003b) developed a three‐end‐member mixing model that used radium isotopes to estimate the frac-
tions of (1) ocean water, (2) water from a shallow aquifer, and (3) water from a deep aquifer that were present
in samples from the coastal ocean off western Florida (see Figure 1). Inherent in applying this model is the
assumption that the composition of the beach porewater is determined by mixtures of these three compo-
nents. For the St. Lucia ecosystem, we have modified this model to estimate the various components present
in beach porewater: (1) LSGW, (2) HSGW, and (3) ocean surface water. Using the following equations, we
constructed a mass balance model for water, salt, and
228
Ra activities.
Table 2
Samples Collected in 2013 From iSimangaliso Wetland Park, South Africa
Station Location Sample ID/type Long. (E) Lat. (S) Depth (m) Salinity
226
Ra
228
Ra ex
224
Ra
228
Th TIN PO
4
5d St. Lucia Estuary Beach PW 32.426 28.383 −1.2 36.0 0.07 0.25 3.41 0.03 1.42 0.51
7 Lake St. Lucia Beach PW 32.558 28.121 −0.5 34.4 0.08 0.46 3.78 0.02 5.13 0.76
11 Lake St. Lucia Beach PW 32.462 28.320 −0.6 35.8 0.08 0.19 4.88 0.01 4.07 0.35
12 Lake St. Lucia Beach PW 32.448 28.339 −1.0 35.7 0.09 0.33 3.38 0.02 2.22 0.91
13 Lake St. Lucia Beach PW 32.487 28.276 −1 35.9 0.09 0.22 3.10 0.03 10.3 0.97
13 Lake St. Lucia Beach PW 32.487 28.276 −1.4 35.4 0.08 0.27 3.73 0.02 22.9 ND
14 Lake St. Lucia Beach PW 32.594 27.931 −1.8 35.0 0.08 0.31 4.58 0.04 10.6 0.98
15 Lake St. Lucia Beach PW 32.587 27.968 −1.6 34.7 0.07 0.56 4.97 0.02 3.46 0.95
16 Lake St. Lucia Beach PW 32.575 28.017 −1.4 35.1 0.08 0.61 5.37 0.03 10.8 0.66
17 Lake St. Lucia Beach PW 32.563 28.076 −0.4 31.4 0.07 0.32 5.98 0.02 19.7 0.73
18 Lake St. Lucia Beach PW 32.554 28.150 −0.4 35.2 0.10 0.56 5.65 0.03 9.26 0.82
38 Sodwana Bay Beach PW 32.678 27.539 −0.8 36.0 0.09 0.18 3.38 0.04 2.22 0.73
1 Lake St. Lucia Lake Shore GW 32.456 28.194 −0.8 0.7 0.17 0.10 <0.1 0.01 9.13 0.22
2 Lake St. Lucia Lake Shore GW 32.463 28.157 −1.8 0.2 0.87 1.56 3.41 0.21 3.65 0.23
3a Lake St. Lucia Lake Shore GW 32.486 28.221 −1.1 0.1 0.71 0.50 0.82 0.03 BD BD
3c Lake St. Lucia Lake Shore GW 32.487 28.220 −8 0.4 0.04 0.06 0.15 0.01 25.8 21.55
4 Lake St. Lucia Lake Shore GW 32.424 28.393 −1.9 35.6 0.32 2.32 4.99 0.61 15.8 0.82
5g St. Lucia Estuary Lake Shore GW 32.423 28.383 −1.9 35.6 0.40 4.46 5.82 0.13 30.2 0.4
6 Lake Bhangazi Lake Shore GW 32.538 28.136 −0.4 0.1 0.04 0.06 0.71 0.00 11.4 0.66
1 Lake St. Lucia Lake Surface 32.456 28.194 0.5 16.3 0.29 1.34 1.05 0.04 1.00 0.25
2 Lake St. Lucia Lake Surface 32.463 28.157 0.5 16.4 0.27 1.10 0.99 0.03 2.05 0.54
3 Lake St. Lucia Lake Surface 32.486 28.221 0.5 16.4 0.33 1.37 1.08 0.02 0.72 0.17
4 Lake St. Lucia Lake Surface 32.424 28.393 0.5 30.5 0.15 0.26 1.07 0.05 2.60 0.13
5 St. Lucia Estuary Lake Surface 32.423 28.383 0.5 32.0 0.44 1.21 2.19 0.12 2.68 0.76
6 Lake St. Lucia Lake Surface 32.538 28.136 0.5 0.4 0.08 0.13 0.30 0.01 17.0 1.25
8 Lake St. Lucia Lake Surface 32.434 28.109 0.2 17.6 0.18 0.91 1.04 0.06 5.25 0.87
9 Lake St. Lucia Lake Surface 32.418 28.198 0.2 16.8 0.23 0.94 0.95 0.05 2.72 0.26
10 Lake St. Lucia Lake Surface 32.429 28.261 0.5 22.0 0.51 1.51 1.27 0.06 12.7 0.31
5g St. Lucia Estuary Ocean Surface 32.426 28.383 0.5 32.8 0.11 0.18 0.62 0.01 2.38 0.87
11 Lake St. Lucia Ocean Surface 32.462 28.320 0.5 37.9 0.06 0.05 0.23 0.01 1.35 0.87
12 Lake St. Lucia Ocean Surface 32.448 28.339 0.5 36.0 0.09 0.10 0.70 0.01 1.99 0.95
13 Lake St. Lucia Ocean Surface 32.487 28.276 0.5 35.8 0.13 0.11 0.46 0.02 1.25 0.94
14 Lake St. Lucia Ocean Surface 32.594 27.931 0.5 34.2 0.07 0.09 0.40 0.01 1.99 1.08
16 Lake St. Lucia Ocean Surface 32.575 28.017 0.5 34.6 0.08 0.06 0.75 0.01 1.18 0.77
18 Lake St. Lucia Ocean Surface 32.554 28.150 0.5 34.8 0.06 0.06 0.35 0.00 2.96 1.02
38 Sodwana Ocean Surface 32.691 27.542 0.5 35.6 0.07 BD 0.03 0.00 0.55 0.58
39 Sodwana Ocean Surface 32.678 27.539 0.5 33.4 0.09 0.08 0.42 0.01 BD 0.76
40 Sodwana Ocean Surface 32.708 27.548 0.5 35.6 0.10 0.09 0.03 0.01 2.54 0.74
41 Sodwana Ocean Surface 32.731 27.558 0.5 35.7 0.06 0.05 0.02 0.00 2.67 0.71
Note. TIN = total inorganic nitrogen; PW = porewater; GW = groundwater. Radionuclide activities are dpm L
−1
, nutrient concentrations are μM. Lower case
letters denote samples collected along a transect. BD is below detection.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2018
Table 3
Samples Collected in 2014 From iSimangaliso Wetland Park, South Africa
Station Location Sample ID/type Long. (E) Lat. (S) Depth (m) Salinity
226
Ra
228
Ra ex
224
Ra
228
Th TIN PO
4
46 Lake Sibaya Active seepage 32.725 27.397 0 7.5 0.13 0.18 1.37 0.03 179 1.09
5a St. Lucia Estuary Beach PW 32.422 28.385 −1.8 34.9 0.25 2.28 BD BD 174 0.81
5b St. Lucia Estuary Beach PW 32.422 28.384 −1.9 35.2 0.07 0.67 0.87 0.03 BD 0.88
5c St. Lucia estuary Beach PW 32.425 28.383 −0.5 6.8 0.08 0.93 1.60 0.05 BD BD
5d St. Lucia Estuary Beach PW 32.426 28.383 −1.7 35.4 0.06 0.20 2.03 0.05 BD BD
5e St. Lucia Estuary Beach PW 32.426 28.383 −3.1 15.8 0.10 1.38 2.45 0.06 784 1.16
7 Lake St. Lucia Beach PW 32.558 28.121 −1.2 35.8 0.08 0.27 2.57 0.19 116 1.09
13 Lake St. Lucia Beach PW 32.487 28.276 −0.85 36.1 0.04 0.09 0.94 0.10 38.5 0.89
14 Lake St. Lucia Beach PW 32.595 27.931 −0.6 36.1 0.11 0.33 BD BD 71.5 0.85
14 Lake St. Lucia Beach PW 32.594 27.931 −1.7 36.4 0.08 0.30 BD BD 300 1.25
15 Lake St. Lucia Beach PW 32.587 27.972 −0.6 37.0 0.05 0.52 1.17 0.23 70.6 1.14
16 Lake St. Lucia Beach PW 32.576 28.014 −0.75 36.2 0.09 0.23 1.55 0.19 78.6 1.12
17 Lake St. Lucia Beach PW 32.562 28.080 −0.85 35.9 0.08 0.47 1.67 0.20 139 1.14
22 Lake St. Lucia Beach PW 32.592 27.880 −0.7 36.2 0.09 0.65 BD BD 210 0.86
22 Lake St. Lucia Beach PW 32.592 27.881 −1.5 35.2 0.10 1.11 BD BD 53.9 0.87
23 Lake St. Lucia Beach PW 32.591 27.909 −0.8 35.9 0.07 0.43 BD BD 76.6 0.80
24 Lake St. Lucia Beach PW 32.590 27.954 −0.9 36.4 0.08 0.68 0.83 0.24 165 1.25
25 Lake St. Lucia Beach PW 32.568 28.054 −0.8 35.9 0.11 0.48 2.75 0.22 174 1.12
38 Sodwana Beach PW 32.680 27.540 −1.5 35.0 0.06 0.21 4.24 0.10 59.7 1.33
42 Lake Sibaya Beach PW 32.745 27.345 −1.2 34.8 0.07 0.96 5.59 0.11 85.6 1.20
43 Lake Sibaya Beach PW 32.737 27.364 −1.9 34.3 0.07 0.37 5.16 0.08 35.3 1.20
44 Lake Sibaya Beach PW 32.732 27.384 −1.2 34.8 0.07 0.17 2.52 0.05 26.1 1.04
45 Lake Sibaya Beach PW 32.725 27.397 −1.2 34.6 0.07 0.13 3.43 0.05 82.0 1.12
46 Lake Sibaya Beach PW 32.713 27.431 −1.2 34.4 0.06 0.10 3.18 0.04 30.0 0.83
48 Kosi Bay Beach PW 32.882 26.897 −1.3 35.1 0.10 6.39 6.97 0.23 151 0.97
48 Kosi Bay Beach PW 32.881 26.893 −1.9 34.8 0.07 0.30 5.28 0.09 52.3 0.81
52 Kosi Bay Beach PW 32.864 27.007 −1.2 35.2 0.08 0.09 1.47 0.04 29.0 1.09
3 Lake St. Lucia Lake Shore GW 32.486 28.221 −0.5 22.9 0.87 6.37 6.77 0.35 131 0.86
5 St. Lucia Estuary Lake Shore GW 32.421 28.384 −1.5 35.9 1.11 10.38 13.68 0.37 172 1.09
27 Lake St. Lucia Lake Shore GW 32.404 28.297 −1.1 8.5 1.49 5.27 3.35 0.38 41.1 0.14
28 Lake St. Lucia Lake Shore GW 32.409 28.345 −0.5 14.5 0.20 3.94 5.75 0.22 4.5 0.67
29 Lake St. Lucia Lake Shore GW 32.409 28.391 −0.5 36.6 0.77 10.37 7.60 0.39 149 0.89
30 Lake St. Lucia Lake Shore GW 32.460 28.233 −1 55.9 0.79 10.50 15.11 0.58 176 4.81
31 Lake St. Lucia Lake Shore GW 32.483 28.235 −0.8 32.9 0.50 5.67 7.66 0.36 88.1 1.20
33 Lake St. Lucia Lake Shore GW 32.564 27.976 −1 32.1 0.71 6.11 7.52 0.26 185 1.12
34 Lake St. Lucia Lake Shore GW 32.513 28.017 −1.5 4.0 0.04 0.45 0.30 0.04 24.9 9.72
37 Hluhluwe River Lake Shore GW 32.358 28.039 −1.5 28.3 1.35 8.08 8.02 0.20 1404 5.03
47 Lake Sibaya Lake Shore GW 32.700 27.418 −1 0.8 0.59 0.77 0.86 0.05 0.1 0.48
48 Kosi Bay Lake Shore GW 32.877 26.895 −0.7 1.2 0.01 BD BD BD 4.0 3.45
49 Kosi Bay Lake Shore GW 32.864 26.926 −1.9 17.8 0.13 0.45 0.42 0.02 21.2 ND
49 Kosi Bay Lake Shore GW 32.862 26.934 −1.9 21.4 0.31 1.10 4.64 0.09 29.0 0.67
50 Kosi Bay Lake Shore GW 32.860 26.946 −1.3 16.1 0.58 4.06 4.49 0.15 55.4 7.39
51 Kosi Bay Lake Shore GW 32.858 27.006 −0.3 3.5 0.31 0.28 1.67 0.04 24.4 1.02
3 Lake St. Lucia Lake Surface 32.487 28.221 0 17.0 0.35 1.45 3.69 0.04 21.4 0.83
5 St. Lucia Estuary Lake Surface 32.409 28.391 0 15.7 0.59 1.46 1.21 0.06 87.0 0.73
5 St. Lucia Estuary Lake surface 32.422 28.385 0 23.2 0.47 2.61 2.53 0.09 BD 0.76
5 St. Lucia Estuary Lake Surface 32.422 28.387 0 15.6 0.49 1.40 BD 0.09 105 0.86
26 Lake St. Lucia Lake Surface 32.412 28.281 0 15.4 0.46 1.47 0.53 0.10 25.1 0.72
30 Lake St. Lucia Lake Surface 32.460 28.233 0 17.9 0.29 1.34 0.66 0.05 513 1.31
32 Lake St. Lucia Lake Surface 32.481 27.888 0 20.7 0.15 0.64 0.76 0.03 264 1.60
33 Lake St. Lucia Lake Surface 32.564 27.976 0 20.6 0.18 0.86 1.05 0.06 158 1.40
35 Lake St. Lucia Lake Surface 32.430 28.004 0 20.9 0.14 0.84 0.61 0.04 122 1.05
36 Naylazi River Lake Surface 32.386 28.105 0 21.7 0.13 0.75 0.82 0.04 66.7 0.76
37 Hluhluwe River Lake Surface 32.358 28.039 0 21.4 0.30 1.15 0.56 0.06 173 1.04
47 Lake Sibaya Lake Surface 32.700 27.418 0 0.8 0.20 0.11 BD 0.01 25.0 0.25
48 Kosi Bay Lake Surface 32.877 26.895 0 34.1 0.06 0.09 0.35 0.01 11.7 0.81
48 Kosi Bay Lake Surface 32.868 26.895 0 25.9 0.14 0.39 1.47 0.04 3.6 0.88
49 Kosi Bay Lake Surface 32.864 26.926 0 22.8 0.22 0.66 1.76 0.03 49.3 0.72
50 Kosi Bay Lake Surface 32.853 26.942 0 15.2 0.25 0.62 0.60 0.02 97.8 0.70
53 Kosi Bay Lake Surface 32.835 26.981 0 3.6 0.23 0.15 0.04 0.01 BD 0.32
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2019
fLþfHþfO¼1:00;(4)
SLfLþSHfHþSOfO¼SM;(5)
228RaLfLþ228 RaHfHþ228RaOfO¼228RaM:(6)
In equation (4), f
L
is the fraction of LSGW, f
H
is the fraction of HSGW, and f
O
is the fraction of ocean surface
water end‐members. In equations (5) and (6), these end‐members are multiplied by the corresponding sali-
nity and
228
Ra activity (dpm/L) end‐members to predict the salinity and
228
Ra activity measured in the sam-
ple, S
M
and
228
Ra
M
.
These equations are solved for the fractions of each end‐member below (Moore, 2003b) in equations (7)–(9).
fH
228RaM−228 RaL
228RaO−228 RaL
−SM−SL
SO−SL
228RaH−228 RaL
228RaO−228 RaL
−SH−SL
SO−SL
;(7)
fO¼SM−SL−fHSH−SL
ðÞ
SO−SL
;(8)
fL¼1:00−fH−fO:(9)
The groundwater end‐members are based on samples collected along the Lake St. Lucia shoreline during
2012–2014. These samples were clearly delineated into low salinity (2013 Stations 1–3, average sali-
nity = 0.35) and high salinity (2012 Station 3; 2013 Stations 4 and 5; and
2014 Stations 5, 29, 30). We assume these end‐members represent ground-
water reaching the beach. For the ocean surface water end‐member, we
used sample data with salinities >35.5 measured adjacent to Lake St.
Lucia in 2013 at Stations 11–13 and in 2014 at Station 14 as these were
characterized by a salinity usually measured in offshore waters and
matched the highest salinity measured in the beach porewater; the other
ocean samples had lower salinities, indicating some dilution. We did not
include the high salinity measured at Station 7 in 2012 as this was influ-
enced by visible seepage that had significantly higher nutrient concentra-
tions than the other sites (see below).
Figure 6 gives the results of the salinity‐
228
Ra mixing model for St. Lucia.
The fraction of LSGW in the beach porewater is very low, averaging
4 ± 3%, with one exception; the Station 4 sample from 2014 collected at
the St. Lucia estuarine mouth had a 69% LSGW component. The fraction
of HSGW averaged 0.0 ± 0.1%, with the highest fraction being 3%. This
explains in part the very high
224
Ra/
228
Ra ARs measured in the beach
Table 3 (continued)
Station Location Sample ID/type Long. (E) Lat. (S) Depth (m) Salinity
226
Ra
228
Ra ex
224
Ra
228
Th TIN PO
4
4 Lake St. Lucia Ocean Surface 28.393 32.424 0.5 11.6 0.33 0.63 BD 0.13 ND 0.30
5 St. Lucia Estuary Ocean Surface 32.426 28.384 0.5 34.4 0.08 0.12 BD BD 38.7 0.99
5 St. Lucia Estuary Ocean Surface 32.426 28.383 0 35.2 0.06 0.08 0.29 0.01 BD BD
13 Lake St. Lucia Ocean Surface 32.487 28.276 0 35.3 0.07 0.08 0.31 0.02 24.8 0.88
14 Lake St. Lucia Ocean Surface 32.595 27.931 0 36.1 0.07 0.12 BD BD 7.3 0.88
42 Lake Sibaya Ocean Surface 32.745 27.345 0 34.8 0.07 0.08 0.28 0.01 22.9 0.86
52 Kosi Bay Ocean Surface 32.864 27.007 0 34.5 0.05 0.06 0.26 0.01 13.5 1.12
21 Mfolozi River River Surface 32.422 28.397 0.5 9.3 0.31 0.58 BD BD 519 0.45
Note. TIN = total inorganic nitrogen; PW = porewater; GW = groundwater. Radionuclide activities are dpm L
−1
, nutrient concentrations are μM; a‐e denote
samples collected along a transect. BD is below detection.
Figure 4. Total inorganic nitrogen (TIN; μmol/L) versus
228
Ra (disintegra-
tions per minute [dpm]/L) in lake groundwater samples.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2020
porewater (Figure 5), since
228
Ra is primarily derived from the HSGW
component. This observation is in direct contrast to that measured in
lake groundwaters where the
224
Ra/
228
Ra AR is ~1 and at
secular equilibrium.
Given the very low fraction of LSGW and HSGW that reaches the St. Lucia
beaches, it is perhaps surprising that the
224
Ra activities are similar to
those measured in other beaches (Figure 5). This suggests an additional
source of
224
Ra to the beach sands that is unrelated to groundwater. Our
leaching experiments indicate that the fraction of total
228
Th that pro-
duces leachable
224
Ra is less than 10%, ranging from 3–9% of the total
228
Th in four of the samples and only 0.3% in the sample enriched in heavy
minerals. It is likely that most of the bulk
228
Th measured is due to small concentrations of
228
Th‐enriched
minerals that were more abundant in the high
228
Th activity sample. The thorium in these minerals must be
present in the crystal lattice, where the daughter is protected from efficient leaching. The
228
Th that does
produce leachable
224
Ra must therefore derive from another source, namely, dissolved and particulate
228
Th in the coastal seawater. Because the coastal seawater samples were not filtered, our measurements
include both dissolved and some particulate
228
Th. We do not know how effective the Mn fiber is in trapping
particulate matter, but clearly some is retained. The average
228
Th recovered from the ocean samples is
0.009 ± 0.005 dpm/L. As the coastal seawater circulates through the beach due to tides and waves, some
of the dissolved fraction must adsorb to particle surfaces and some of the particulate fraction will be retained,
similar to a sand filter. Because the
228
Th mean life is 2.5 years, considerable activity could be produced in
the beach sand by the continued circulation of coastal waters. For example, 1 m
3
of seawater infiltrating into
the beach each tidal cycle for 2.5 years, the mean life of
228
Th, would result in an inventory of 18 dpm.
Although the
228
Th may initially be concentrated near the surface, the inventory is likely mixed downward
during the 2.5‐year mean life by storms and other turbulent processes. Without detailed knowledge of such
factors as beach slope, swash period and length, tide range, and depth of infiltration, we cannot estimate the
total volume of water that circulates through the beach during a tidal period.
Due to its short half‐life, the
224
Ra released from fine‐grained particles must occur relatively quickly. We can
determine the in situ production rate of
224
Ra that may desorb into porewaters using our leaching experi-
ments. Because some of the
224
Ra is adsorbed to the sand after production, we must modify the leachable
value using the K
d
determined earlier to calculate the
224
Ra available to the circulating porewater
(Tamborski et al., 2017):
P′¼P
1þKd
;(10)
where Pis the average leachable
224
Ra (39 ± 16 dpm/L). Using an average K
d
for St. Lucia (1.37), we calcu-
late an exchangeable
224
Ra activity derived from in situ
228
Th decay, P′, of 16 ± 7 dpm/L. This value is at the
Table 4
Results of K
d
Determinations With Spiked
224
Ra for Sands From St
Lucia Beaches
Sample Porosity q
e
K*
dK
d
Station 7 0.62 0.70 2.15 1.37
Station 13 0.64 0.70 2.19 1.51
Station 14 0.63 0.59 1.51 1.00
Station 5 0.64 0.71 2.31 1.59
Note. For station locations see Figure 2.
Table 5
Results of Leaching Sand From St Lucia Beach With Seawater (SW)
Station
Sand
(g)
Sea water
(ml)
Leached
224
Ra
(dpm) K
d
Adsorbed
224
Ra
(dpm)
Leachable
224
Ra
(dpm/g)
Bulk
228Th
(dpm/g)
Leachable
fraction
Depth
(m)
7‐L1 29.9 500 0.25 1.37 0.34 0.020 0.45 0.044 −0.7
7‐L2 151 1,500 1.58 1.37 2.16 0.025 0.45 0.055 −0.7
13‐L1 29.8 500 0.17 1.51 0.25 0.014 0.35 0.040 −0.8
14‐L1 30 500 0.14 1.00 0.14 0.009 0.27 0.034 −0.6
14‐L2 150 1,500 1.18 1.00 1.18 0.016 0.27 0.058 −0.6
5‐L1 30.1 500 0.46 1.59 0.73 0.040 0.43 0.092 −0.5
5‐L2 151 1,500 1.96 1.59 3.12 0.034 0.43 0.078 −0.5
13HM‐L1 30.1 500 0.44 1.30 0.57 0.033 11.16 0.003 −0.1
Note For station locations see Figure 2. Station 13HM refers to a surface sample enriched in heavy minerals.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2021
upper end of values for Smithtown, NY, beach reported by Tamborski et al. (2017), whose results ranged
from 11 to 16 dpm/L for salinities >26.
Since the
224
Ra decay constant is much greater than the
228
Th decay constant, we can write
ARa ¼P01−e−λt
;(11)
where A
Ra
is the measured
224
Ra activity in 1 L of porewater, P′is the
228
Th activity in 1 L of sediment that
produces exchangeable
224
Ra (16 dpm /L), λis the decay constant for
224
Ra (0.19/day), and tis the time the
water has been in contact with the sediment. Other sources of
224
Ra to the porewaters must also be
accounted for, namely, seawater and groundwater. Here we modify the equation of Tamborski et al.
(2017) to include the HSGW and LSGW although they make a minor contribution:
ARa ¼P01−e−λt
þfSW ASW e−λt
þfLSGW ALSGW e−λt
þfHSGW AHSGW e−λt
(12)
where A
SW
,A
LSGW
,A
HSGW
are the initial
224
Ra activities in seawater, LSGW and HSGW, and f is the fraction
of seawater, LSGW or HSGW as determined from the mixing model. This approach could underestimate the
external supply of
224
Ra because of decay; however, the low fractions of groundwater derived from the mix-
ing model indicate that this correction is minor.
Figure 5. The activities of
224
Ra and
228
Ra measured in (a) South African beach porewaters compared with similar measurements in other beaches. Samples from
the estuary mouth are designated EM. The 2:1 line is shown for reference. Data from (b) Patos Lagoon from Windom et al. (2006); from (c) East and South
China Seas from Tan et al. (2018); from (d) Wadden Sea from Moore et al. (2011); from (e) North Carolina, South Carolina, and Georgia beaches from Peterson et al.
(2016) and Table S1; and from (f) Smithtown Bay from Tamborski et al. (2017).
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2022
Solving this equation for t(equation (13)) allows the calculation of the
beach porewater age based on their
224
Ra activities (Table 6):
t¼
−ln ARa−P0
−P0þfSWA0
SWþfLSGW A0
LSGWþfHSGW A0
HSGW
hi
λ:(13)
The ages range from 0.3 to 2.3 days, the same age range determined by
Tamborski et al. (2017) for their high‐salinity barrier beach samples from
<2‐m depth (0.14–2.6 days). While consistent with previous work, the
time scale of less than 2 days to completely flush the upper 2 m of beach
porewater is surprisingly rapid. Nevertheless, it is difficult to make this
period of time longer. Porewater age would increase if the activity of
228
Th in a liter of sediment that produces leachable
224
Ra was over esti-
mated, since this is the primary source of
224
Ra to the beach porewater.
However, four different beach samples gave similar
228
Th activities
(Table 5).
4.2. Sources of Nutrients
Groundwater near all of the lakes sampled was characterized by a strong
positive correlation between TIN concentrations and
228
Ra activities
(R
2
= 0.61, excluding Station 37 located in the muddy marsh near the river
mouth of the Hluhluwe River; Figure 4). There are few studies of ground-
water composition in this region, making source identification difficult.
However, radium activities are consistent with a primary groundwater
aquifer dominated by calcium and carbonate that flows from the coastal
dune complex on the eastern side of Lake St. Lucia (Bjorkenes et al.,
2006). Given the remote location and groundwater source, nutrients are
likely derived from organic matter decomposition, increasing as water
percolates through the dune sands and into the wetland (Kelbe
et al., 2013).
Lake surface waters showed no trends between nutrient concentrations
and
228
Ra activities. Rather, lake waters contained high TIN concentra-
tions (average = 116 μM) but significantly lower phosphate and radium
activities (p< 0.05, Tables 1–3). These results are consistent with previous
work showing that groundwater only provides about 6–7% of the total lake
volume of Lake St. Lucia on an annual basis during nondrought condi-
tions (Kelbe et al., 2013). Rather, the major sources of the high TIN con-
centrations measured in lake surface waters are due to riverine input
and sheet flow. These inputs are modified by incomplete mixing and bio-
geochemical reactions that occur within the variable bathymetry of lake
ecosystems. Lake St. Lucia, the major lake sampled in this study, is fed
by a combination of land surface runoff and the Mkhuze, Mfolozi,
Nyalazi, and Hluhluwe Rivers (Figure 2c). TIN concentrations measured
at stations in close proximity to river inflow ranged from 66.7‐to
264‐μM TIN and 0.76‐to 1.6‐μM phosphate and averaged
0.85 ± 0.26 dpm
228
Ra/L. These measured riverine nutrient concentra-
tions are consistent with average lake surface water concentrations, which
averaged 115‐μM TIN, 0.70‐μM phosphate, and 0.96 dpm
228
Ra/L, and
previous studies (Connell & Porter, 2013). Although these watersheds
are lightly populated, sugar cane is a major crop. Runoff containing ferti-
lizer from these cultivated fields may explain some of the very high
TIN concentrations.
Figure 6. Results of the three‐end‐member mixing model based on salinity
and
228
Ra for the 2013 and 2014 beach porewater samples from St. Lucia.
Black dots are 2013; red dots are 2014. With a few exceptions, the fractions of
high‐and low‐salinity groundwater (GW) are minor.
Table 6
Ages Calculated for 1–2 m Deep Porewaters From St Lucia Beaches
Salinity OW LSGW HSGW
224
Ra (dpm/L) Age (day)
2013
35.8 0.97 0.01 0.03 4.88 1.6
35.7 0.92 0.01 0.07 3.38 1.2
36.0 0.95 0.00 0.05 3.41 1.3
35.9 0.96 0.00 0.04 3.10 1.1
35.4 0.94 0.02 0.05 3.73 1.4
35.0 0.91 0.03 0.06 4.58 1.8
35.1 0.82 0.02 0.15 5.37 2.2
34.7 0.83 0.03 0.13 5.65 2.3
31.4 0.82 0.13 0.05 4.97 2.0
2014
34.99 0.98 0.00 0.02 4.24 1.4
34.34 0.94 0.02 0.04 5.16 2.0
35.80 0.99 0.00 0.03 2.57 0.9
34.80 0.98 0.01 0.01 2.52 0.9
34.55 0.98 0.02 0.00 3.43 1.3
34.41 0.98 0.02 0.00 3.18 1.2
35.23 1.00 0.00 0.00 1.47 0.5
34.76 0.96 0.01 0.03 5.28 2.1
35.39 0.99 0.00 0.02 2.03 0.7
36.40 0.96 0.00 0.08 0.83 0.3
35.85 0.97 0.00 0.05 1.67 0.6
36.07 1.02 0.00 0.00 0.94 0.3
35.94 0.97 0.00 0.06 2.75 1.0
36.20 1.01 0.00 0.02 1.55 0.5
36.98 0.99 0.00 0.06 1.17 0.4
Note. OW = ocean surface water; LSFW = low‐salinity groundwater;
HSGW = high‐salinity groundwater. The table also gives the fractions
of ocean water (OW), low salinity groundwater (LSGW) and high salinity
groundwater (HSGW) present in the interstitial water based on the mix-
ing model.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2023
Explaining the nutrient distributions in the beach porewaters and off-
shore waters is more problematic. The three‐end‐member mixing model
that successfully explained the salinity and
228
Ra distributions in the
beach porewaters fails to explain the nutrient distributions because the
end‐members we selected for
228
Ra and salinity are clearly not representa-
tive of nutrient inputs to the beach. For example, the TIN concentration in
the LSGW end‐member is 4.6 μM, the HSGW is 10.2 μM, and the ocean
water is 3.8 μM. However, the concentration of TIN in the beach ground-
water is as high as 784 μM, and in the ocean it reaches 47 μM. This obser-
vation requires another source of TIN. This can either be an internal
source where regeneration of N occurs within the beach sediments or
an external source that we did not consider with the
228
Ra‐salinity mixing
model. This external source could be groundwater occurring deeper than
our 1‐to 2‐m penetration on the beach that selectively enriched some of
the beach stations and/or that discharged directly into the ocean at some
of the stations.
The high TIN signal in the beach porewaters is highly variable spatially
and temporally. For example, at Station 5, the estuary mouth, the concen-
tration at nearby sampling sites ranged from 32 to 206 μM in 2012 and
from 174 to 784 in 2014; the only sample collected at this station in 2013 was 1.4 μM. The ocean surface water
at Station 5 followed this general trend measuring 48 μM in 2012, 2.4 in 2013, and 39 in 2014. The ages of the
beach porewaters were all ≤2 days. It is difficult to imagine that regeneration of N within the upper 2 m of
the beach could produce such extremely variable TIN concentrations when it is flushed so quickly. We think
it is more likely that if regeneration of Nis a significant source, it is occurring at deeper depths and reaching
the surficial sands heterogeneously in time and space. Additionally, groundwater flow beneath the beach
from distal sources could emerge through preferential flow paths, that is, fractures in the rocky confining
layer under the beach, and enrich the porewaters in TIN. The deepest porewater sample collected on the
beach (3.1 m, 2014) contained the highest TIN of any beach porewater sample, 784 μM, had a salinity of
15.8, but was only slightly enriched in
228
Ra, 1.4 dpm/L (Table 3).
Figure 7.
228
Ra (disintegrations per minute [dpm]/L) activities versus total
inorganic nitrogen (TIN) concentrations (μM/L) in ocean surface samples.
Stations are identified by no visual seepage (open triangles) and active visual
seepage (closed circles). Samples collected at the estuary mouth of Lake St.
Lucia are identified by gray squares. The 13.8:1 TIN:
228
Ra relationship
observed in lake groundwaters (Figure 4) is shown for comparison.
Figure 8. (a) A site of active seepage on the beach; algae is often associated with active seepage. (b) A vigorous freshwater
seep on the beach at Sibaya. (c) Algae coating coastal rocks. (d) Algae growth extends to the dunes.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2024
The coast of eastern South Africa from Lake St. Lucia to Kosi Bay has a number of paleochannels and his-
torical drainage outlets that provide an additional pathway for terrestrial waters to intermittently flow into
the coastal zone (Figure 3). A majority of the stations located close in proximity to these sites have beach
porewaters in excess of 200‐μM TIN and ocean surface water nutrient concentrations that exceed 10‐μM
TIN. All of the ocean surface samples exceeding 10‐μM TDN were collected either at a site of active seepage
on the beach or at the estuary mouth (Figure 7). The remaining samples, where there was no obvious see-
page, followed the 13.8 N:1
228
Ra relationship observed in the lake groundwater (Figure 4). The highest
TIN concentrations in the open ocean occurred at the estuary mouth, which was closed to direct discharge
by a wide sandbar, and adjacent to areas where we observed significant seepage (Figures 8a and 8b). These
sites were further characterized by thick coatings of green algae belonging to the Ulva genera on rocks and
sand (Figure 8c and 8d). Many of these species, such as Ulva intestinalis, tolerate a wide range of salinities
and are often found in eutrophic settings (Kamer et al., 2004). Our results are consistent with previous stu-
dies that found high nutrient seepage from beach porewaters could support and sustain biological activity in
the nearshore (e.g., Hays & Ullman, 2007). The intermittent nature of deeper inputs to porewaters is clearly
observed at Station 13 (Mission Rocks). In 2013, the surface ocean TIN concentration was 1.3 μM (Table 2).
In 2012 and 2014, TIN concentrations exceeded 18.7 μM and active seepage was observed (Tables 1 and 3).
We should note that a number of porewaters sampled north of Cape Vidal (Figure 2), but not within known
paleochannels or relict drainage outlets, were also characterized by higher TIN concentrations and were
chosen due to visible seepage (Figures 8a and 8b) and/or beach algal growth (Figure 8c and 8d). While many
of these sites are separated from the nearby lakes by high dune barriers, these barriers are often less than
0.5 km wide. We therefore hypothesize that at least some of the lake water likely flows through the uncon-
solidated sand dune barrier to reach the coast. The exact mechanisms of this flow and the biogeochemical
reactions that occur during transit remain enigmatic, particularly given that Lake St. Lucia is brackish, while
Lake Sibaya and Kosi Bay are relatively fresh. These results, however, do suggest that the water sources in
this region are likely influenced by more than the mixing of salt and fresh water (e.g., changes in ionic
strength influences radium) but also by redox and microbial reactions that preferentially impact nutrient
concentration and composition (Beck et al., 2017; Charette & Sholkovitz, 2006; Kim & Swarzenski, 2010;
Knee & Paytan, 2011; Moore, 2010; Tamborski et al., 2017).
Based on these considerations, we hypothesize that a likely additional source of the high‐nutrient and low‐
radium‐activity beach porewaters and ocean waters (Figure 9) is inputs from Lake St. Lucia, Lake Sibaya,
and Kosi Bay that occur deeper than our 2‐m measurements on the beaches and emerge through fractures
in the confining layer. Such inputs may be altered in concentration by regeneration of N in the
deeper aquifer.
5. Conclusions
Our results suggest that the daily flushing of beach porewaters supplies significant nutrients to coastal
waters off South Africa. Nutrient inputs probably also occur through deeper paleochannels connecting
Figure 9. Revised conceptual model of groundwater flow along the coast of eastern South Africa. In addition to LSGW and
HSGW, porewaters are further enriched with a deeper groundwater component enriched in nutrients, but not
228
Ra,
derived from adjacent lakes. LSGW = low‐salinity groundwater; HSGW = high‐salinity groundwater; SW = seawater;
SGD = submarine groundwater discharge.
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2025
Lake St. Lucia, Lake Sibaya, and Kosi Bay with offshore waters. It is the presence of high nutrient concen-
trations in the coastal ocean unaccompanied by high
228
Ra activities that leads to the hypothesis of this addi-
tional nutrient source. These combined inputs may be of considerable importance to the coastal ecology of
southeastern Africa, an oligotrophic ecosystem dominated by the nutrient‐poor Agulhas Current
(Bustamante et al., 1995; Lutjeharms, 2007).
Most research along the eastern shores of South Africa has focused on the KwaZulu‐Natal Bight, just south
of our study area, where the main drivers of coastal ocean biological production have been attributed to
mesoscale eddy‐induced upwelling of cold, nutrient‐rich waters (Barlow et al., 2008; Meyer et al., 2002).
More recent work from stable isotope analyses, however, suggests that a significant source of nutrients is
likely derived from terrestrial sources as well (De Lecea et al., 2013, 2016). Our results of temporally and spa-
tially heterogeneous sources of terrestrially derived nutrients to the coastal zone support the stable isotope
results. Clearly, more research is required to validate our hypothesis that deep SGD, flowing beneath the
beach system, carries significant nutrients to coastal waters. Monitoring wells drilled into some of these
paleochannels would allow us to test this hypothesis.
References
Anschutz, P., Smith, T., Mouret, A., Deborde, J., Bujan, S., Poirier, D., & Lecroart, P. (2009). Tidal sands as biogeochemical reactors.
Estuarine, Coastal and Shelf Science,84(1), 84–90. https://doi.org/10.1016/j.ecss.2009.06.015
Barlow, R., Kyewalyanga, M., Sessions, H., van den Berg, M., & Morris, T. (2008). Phytoplankton pigments, functional types, and
absorption properties in the Delagoa and Natal Bights of the Agulhas ecosystem. Estuarine, Coastal and Shelf Science,80(2), 201–211.
https://doi.org/10.1016/J.ECSS.2008.07.022
Beck, A. J., & Cochran, M. A. (2013). Controls on solid‐solution partitioning of radium in saturated marine sands. Marine Chemistry,156,
38–48. https://doi.org/10.1016/j.marchem.2013.01.008
Beck, M., Reckhardt, A., Amelsberg, J., Bartholomä, A., Brumsack, H. J., Cypionka, H., et al. (2017). The drivers of biogeochemistry in
beach ecosystems: A cross‐shore transect from the dunes to the low‐water line. Marine Chemistry,190,35–50. https://doi.org/10.1016/j.
marchem.2017.01.001
Benallack, K., Green, A. N., Humphries, M. S., Cooper, J. A. G., Dladla, N. N., & Finch, J. M. (2016). The stratigraphic evolution of a
large back‐barrier lagoon system with a non‐migrating barrier. Marine Geology,379,64–77. https://doi.org/10.1016/j.
margeo.2016.05.001
Bjorkenes, M. S., Haldorsen, S., Mulder, J., Kelbe, B., & Ellery, F. (2006). Baseline groundwater quality in the coastal aquifer of St.
Lucia, South Africa. In Urban groundwater management and sustainability (pp. 233–240). Dordrecht: Springer. https://doi.org/
10.1007/1‐4020‐5175‐1_18
Bokuniewicz, H., Cochran, J. K., Garcia‐Orellana, J., Rodellas, V., Daniel, J. W., & Heilbr un, C. (2015). Intertidal percolation through beach
sands as a source of
224,223
Ra to Long Island Sound, New York and Connecticut, United States. Journal of Marine Research,73(5),
123–140. https://doi.org/10.1357/002224015816665570
Burnett, W. C., Cable, J. E., Corbett, D. R., & Chanton, J. P. (1996). Tracing groundwater flow into surface waters using natural
222
Rn. In
R. W. Buddemeier (Ed.), Groundwater discharge in the coastal zone (pp. 22–28). The Netherlands: LOICZ, Texel.
Bustamante, R. H., Branch, G. M., Eekhout, S., Robertson, B., Zoutendyk, P., Schleyer, M., et al. (1995). Gradients of intertidal primary
productivity around the coast of South Africa and their relationships with consumer biomass. Oecologia,102(2), 189–201. https://doi.
org/10.1007/BF00333251
Charette, M. A., & Sholkovitz, E. R. (2006). Trace element cycling in a subterranean estuary: Part 2. Geochemistry of the pore water.
Geochimica et Cosmochimica Acta,70(4), 811–826. https://doi.org/10.1016/j.gca.2005.10.019
Connell, A. D., & Porter, S. N. (2013). The marine environment. In R. Perissinotto, D. D. Stretch, & R. H. Taylor (Eds.), Ecology and con-
servation of estuarine ecosystems: Lake St. Lucia as a global model (pp. 63–76). Cambridge: Cambridge University Press. https://doi.org/
10.1017/CBO9781139095723.006
De Lecea, A. M., Cooper, R., & Smit, A. J. (2016). Identifying the drivers of the pelagic ecosystem of an oligotrophic bight (KwaZulu‐Natal,
South Africa) using stable isotopes (δ
13
C, δ
15
N) and C:N ratio analysis. Marine and Freshwater Research,67(11), 1750–1761. https://doi.
org/10.1071/MF15256
De Lecea, A. M., Fennessy, S. T., & Smit, A. J. (2013). Processes controlling the benthic food‐web of a mesotrophic bight (KawZuluNatal,
South Africa) revealed by stable isotope analysis. Marine Ecology Progress Series,484,97–114. https://doi.org/10.3354/MEPS10311
Green, A. N. (2009). Palaeo‐drainage, incised valley fills and transgressive systems tract sedimentation of the northern KwaZulu‐Natal
continental shelf, South Africa, SW Indian Ocean. Marine Geology,263(1‐4), 46–63. https://doi.org/10.1016/j.margeo.2009.03.017
Hays, R. L., & Ullman, W. J. (2007). Direct determination of total and fresh groundwater discharge and nutrient loads from a sandy
beachface at low tide (Cape Henlopen, Delaware). Limnology and Oceanography,52(1), 240–247. https://doi.org/10.4319/lo.2007.52.
1.0240
Hobday, D. K., & Orme, A. R. (1974). The Port Durnford Formation. A major barrier lagoon complex along the Zululand coast.
Transactions of the Royal Society of South Africa,77, 141–149.
Kamer, K., Fong, P., Kennison, R., & Schiff, K. (2004). Nutrient limitation of the macroalgae Enteromorpha intestinalis collected along a
resource gradient in a highly eutrophic estuary. Estuaries,27(2), 201–208. https://doi.org/10.1007/BF02803377
Kelbe, B. E., Taylor, R. H., & Haldorsen, S. (2013). Groundwater hydrology. In R. Perissinotto, D. D. Stretch, & R. H. Taylor (Eds.), Ecology
and conservation of estuarine ecosystems: Lake St Lucia as a global model (pp. 151–168). Cambridge: Cambridge University Press.
https://doi.org/10.1017/CBO9781139095723.010
Kim, G., & Swarzenski, P. W. (2010). Submarine groundwater discharge (SGD) and associated nutrient fluxes to the coastal ocean. In
K.‐K. Liu, L. Atkinson, R. Quinones, & L. Talaue‐McManus (Eds.), Carbon and nutrient fluxes in continental margins: A global synthesis
(pp. 529–538). Berlin Heidelberg: Springer‐Verlag. https://doi.org/10.1007/978‐3‐540‐92735‐8_11
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2026
Acknowledgments
Caroline Fox, Ricky Taylor, Sbu Mfeka,
Sue van Rensburg, and Triton Divers
assisted with the collection of some
samples. Jessica Frankle and Calyn
Crawford assisted in the laboratory. The
iSimangaliso Wetland Park Authority
and Ezemvelo KZN Wildlife kindly
granted us permission to work within
the wetland park. This work was
supported by a South African National
Research Foundation grants 105724
and 104255 to M. S. H. and a National
Science Foundation grant EAR‐
1316250 to W. S. M. All data referenced
in this paper are published herein or in
earlier publications.
Knee, K. L., & Paytan, A. (2011). Submarine groundwater discharge: A source of nutrients, metals, and pollutants to the coastal ocean. In
E. Wolanski & D. S. McLusky (Eds.), Treatise on Estuarine and Coastal Science (pp. 205–233). Waltham: Acade mic Press. https://doi.org/
10.1016/B978‐0‐12‐374711‐2.00410‐1
Koroleff, F. (1983). Determination of phosphorus. In Methods of seawater analysis (2nd ed., pp. 125–139). Weinheim: Verlag.
Kwon, E. Y., Kim, G., Primeau, F., Moore, W. S., Cho, H. M., DeVries, T., et al. (2014). Global estimate of submarine groundwater discharge
based on an observationally constrained radium isotope model. Geophysical Research Letters,41, 8438–8444. https://doi.org/10.1002/
2014GL061574
Lutjeharms, J. R. E. (2007). Three decades of research on the greater Agulhas Current. Ocean Science,3(1), 129–147. https://doi.org/
10.5194/os‐3‐129‐2007
Maud, R. R. (1980). The climate and geology of Maputaland. In M. N. Bruton & K. H. Cooper (Eds.), Studies on the ecology of Maputaland
(pp. 1–7). Grahamstown: Rhodes University.
McLachlan, A., Eliot, I. G., & Clarke, D. J. (1985). Water filtration through reflective microtidal beaches and shallow sublittoral sands and
its implications for an inshore ecosystem in Western Australia. Estuarine, Coastal and Shelf Science,21(1), 91–104. https://doi.org/
10.1016/0272‐7714(85)90008‐3
Meyer, A. A., Lutjeharms, J. R. E., & de Villiers, S. (2002). The nutrient characteristics of the Natal Bight, South Africa. Journal of Marine
Systems,35(1‐2), 11–37. https://doi.org/10.1016/S0924‐7963(02)00043‐X
Miller, W.R. (2001). The bathymetry, sedimentology and seismic stratigraphy of Lake Sibaya–Northern KwaZulu‐Natal. Bulletin 131,
Council for Geoscience, Pretoria.
Moore, W. S. (1976). Sampling
228
Ra in the deep ocean. Deep‐Sea Research,23,647–651.
Moore, W. S. (1984). Radium isotope measurements using germanium detectors. Nuclear Instruments and Methods in Physics Research,
223(2‐3), 407–411. https://doi.org/10.1016/0167‐5087(84)90683‐5
Moore, W. S. (2003a). The role of submarine groundwater discharge in coastal biogeochemistry. Journal of Geochemical Exploration,88,
389–393. 2006
Moore, W. S. (2003b). Sources and fluxes of submarine groundwater discharge delineated by radium isotopes. Biogeochemistry,66(1/2),
75–93. https://doi.org/10.1023/B:BIOG.0000006065.77764.a0
Moore, W. S. (2008). Fifteen years experience in measuring
224
Ra and
223
Ra by delayed‐coincidence counting. Marine Chemistry,109(3‐4),
188–197. https://doi.org/10.1016/j.marchem.2007.06.015
Moore, W. S. (2010). The effect of submarine groundwater discharge on the ocean. Annual Review of Marine Science,2(1), 59–88. https://
doi.org/10.1146/annurev‐marine‐120308‐081019
Moore, W. S., & Arnold, R. (1996). Measurement of
223
Ra and
224
Ra in coastal waters using a delayed coincidence counter. Journal of
Geophysical Research,101(C1), 1321–1329. https://doi.org/10.1029/95JC03139
Moore, W. S., Beck, M., Reidel, T., van der Loeff, M., Dellwig, O., Shaw, T. J., & Brumsack, H.‐J. (2011). Radium‐based pore water fluxes of
silica, alkalinity, manganese, DOC, and uranium: A decade of studies in the German Wadden Sea. Geochimica et Cosmochimica Acta,
75(21), 6535–6555. https://doi.org/10.1016/j.gca.2011.08.037
Moore, W. S., Blanton, J. O., & Joye, S. B. (2006). Estimates of flushing times, submarine groundwater discharge, and nutrient fluxes to
Okatee Estuary, South Carolina. Journal of Geophysical Research,111, C09006. https://doi.org/10.1029/2005JC003041
Moore, W. S., & Cai, P. (2014). Calibration of RaDeCC systems for
223
Ra measurements. Marine Chemistry,156, 130–137.
Moore, W. S., Sarmiento, J. L., & Key, R. M. (2008). Submarine groundwater discharge revealed by
228
Ra distribution in the upper Atlantic
Ocean. Nature Geoscience,1(5), 309–311. https://doi.org/10.1038/ngeo183
Nozaka, Y., Yang, H.‐S., & Yamada, M. (1987). Scavenging of thorium in the ocean. Journal of Geophysical Research,92(C1), 772–778.
https://doi.org/10.1029/JC092iC01p00772
Paytan, A., Shellenbarger, G. G., Street, J. H., Gonneea, M. E., Davis, K., Young, M. B., & Moore, W. S. (2006). Submarine groundwater
discharge: An important source of new inorganic nitrogen to coral reef ecosystems. Limnology and Oceanography,51(1), 343–348.
https://doi.org/10.4319/lo.2006.51.1.0343
Petermann, E., Knöller, K., Rocha, C., Scholten, J., Stollberg, R., Weiß, H., & Schubert, M. (2018). Coupling end‐member mixing analysis
and isotope mass balancing (222‐Rn) for differentiation of fresh and recirculated submarine groundwater discharge into Knysna
Estuary, South Africa. Journal of Geophysical Research: Oceans,123, 952–970. https://doi.org/10.1002/2017JC013008
Peterson, R. N., Moore, W. S., Chappel, S. L., Viso, R. F., Libes, S. M., & Peterson, L. E. (2016). A new perspective on coastal hypoxia: The
role of saline groundwater. Marine Chemistry,179,1–11. https://doi.org/10.1016/j.marchem.2015.12.005
Rama, M., & Moore, W. S. (1996). Using the radium quartet for evaluating groundwater input and water exchange in salt marshes.
Geochimica Cosmochimica. Acta,60(23), 4645–4652. https://doi.org/10.1016/S0016‐7037(96)00289‐X
Tamborski, J. J., Cochran, J. K., & Bokuniewicz, H. J. (2017). Application of
224
Ra and
222
Rn for evaluating seawater residence times in a
tidal subterranean estuary. Marine Chemistry,189,32–45. https://doi.org/10.1016/j.marchem.2016.12.006
Tan, E., Wang, G., Moore, W. S., Li, Q., & Dai, M. (2018). Shelf‐scale submarine groundwater discharge in the northern South China Sea
and East China Sea and its geochemical impacts. Journal of Geophysical Research: Oceans,123, 2997–3013. https://doi.org/10.1029/
2017JC013405
Taniguchi, M., Burnett, W. C., Cable, J. E., & Turner, J. V. (2002). Investigation of submarine groundwater discharge. Hydrological
Processes,16(11), 2115–2129. https://doi.org/10.1002/hyp.1145
Watkeys, M. K., Mason, T. R., & Goodman, P. S. (1993). The rôle of geology in the development of Maputaland, South Africa. Journal of
African Earth Sciences,16(1‐2), 205–221. https://doi.org/10.1016/0899‐5362(93)90168‐P
Windom, H. L., Moore, W. S., Niencheski, F. H., & Jahnke, R. A. (2006). Submarine groundwater discharge: A large, previously unrecog-
nized source of dissolved iron to the South Atlantic Ocean. Marine Chemistry,102(3‐4), 252–266. https://doi.org/10.1016/j.marchem.200
6.06.016
Wright, C. I., Miller, W. R., & Cooper, J. A. G. (2000). The late Cenozoic evolution of coastal water bodies in northern KwaZulu‐Natal,
South Africa. Marine Geology,167(3‐4), 207–229. https://doi.org/10.1016/S0025‐3227(00)00032‐3
10.1029/2018JC014772
Journal of Geophysical Research: Oceans
MOORE ET AL. 2027
