Radium isotope disequilibrium to delineate coastal mixing rates and submarine groundwater discharge

In the natural radioactive series, there are four radium isotopes: ²²⁶Ra (t1/2 = 1620 years);

228Ra (t1/2 = 5.75 years); ²²³Ra (t1/2 = 11.3 days); ²²⁴Ra (t1/2 = 3.66 days). Each isotope is produced from the decay of a thorium parent: ²³⁰Th (t1/2 = 7.54×0⁴ years); ²³²Th (t1/2 = 1.40×10¹⁰ years);

227Th (t1/2 = 18.7 days); ²²⁸Th (t1/2 = 1.91 years), respectively.

Because thorium remains tightly bound to particles while radium daughters are mobilized into the marine environment, sediments provide a continuous source of Ra isotopes to seawater, at rates set by their respective decay constants. Measurements of the Th isotope activities in the sediments and the distribution coefficient of Ra between the sediments and water provide a means of quantifying the potential input of each isotope to the ocean.

Two short-lived radium isotopes ²²³Ra and ²²⁴Ra can be used as tracers to measure the rate of exchange of coastal waters [9]. Shore–perpendicular profiles of ²²³Ra and ²²⁴Ra in surface waters along the coast may be modeled to yield eddy diffusion coefficients. Coupling the exchange rate with offshore concentration gradients, the offshore fluxes of dissolved materials are estimated. For systems in steady state, the offshore fluxes must be balanced by new inputs from rivers, groundwater, sewers or other sources. Also, it was observed that barium and ²²⁶Ra contents can be powerful indicators of groundwater input in marine systems, since they have high relative concentrations in the fluids and low reactivity in the coastal ocean. An estimate of the ²²⁶Ra offshore flux is made applying the eddy diffusion coefficients to the ²²⁶Ra offshore gradient. Complementary data of ²²⁶Ra in subsurface fluids provides a mean of calculate the fluid flux necessary to support the ²²⁶Ra concentrations found in the marine environment.

Moore [9] used the distribution of the short-lived Ra isotopes to estimate exchange rates in the coastal ocean. The change in concentration or activity (A) with time (t) as a function of distance offshore (x)

for a radioactive tracer with decay constant (λ) may be expressed as a balance of advection, If net advection can be neglected, this reduces to:

(5)

where K_h is the dispersion coefficient.

The criterion for setting ω = 0 is based on the offshore distribution of conservative tracers such as

226Ra and ²²⁸Ra. These long-lived isotopes decay little during the residence time of coastal waters. A constant offshore concentration gradient of these tracers provides evidence that dispersion dominates offshore or onshore advection [9]. In the case where ω = 0, the boundary conditions of eq.5 are as follows:

A =A_iat x= 0 A→0as x→ ∞

If K_h is constant and the system is steady state, exchange is dominated by dispersion rather than advection and if the system is steady state:

ln A _x= ln A₀– x λ

K_{h (7)}

In this case the slope, m = λ

K_{h (8)}

An alternate way to use the short-lived Ra isotopes is to utilize the ²²⁴Ra/²²³Ra AR to estimate the ages of shelf waters [18]. This method is based on the assumptions that the ²²⁴Ra/²²³Ra AR is initialized to a constant value near shore and only changes by decay as the water is isolated from the radium source.

The ²²⁴Ra/²²³Ra AR decreases with an apparent half life of 5.4 days, as follows:

3.3. Experimental

3.3.1. Application of the ²²²Rn approach to estimate SGD in Ubatuba coastal waters

We describe here an application of excess ²²²Rn to estimate SGD in a series of small embayments of Ubatuba, São Paulo State, Brazil.

Seawater samples were collected at several stations of Ubatuba embayments in top-to-bottom vertical profiles. Temperature and salinity profiles were obtained at the same stations using a 2.00” Micro CTD, from Falmouth Scientific Inc. In each station, seawater samples were collected at 1 – 2 m depth

dA dt = K_h ∂ ²A

∂ x ² –

λΑ

intervals using a peristaltic pump to purge the sampling tubes and then drawn into 4 L evacuated glass bottles. Seawater was purged for 5 minutes from the hose at each depth prior to filling the sampling bottles, and they were immediately sealed to prevent radon losses. ²²²Rn was extracted and counted using a modified emanation technique described by Cable et al. [5]. Once extracted, the radon gas was collected in a liquid nitrogen cold trap and transferred from the trap to an alpha scintillation cell. After radon stripping and transfer into alpha scintillation cells, samples were stored for 3 hours to allow

222Rn daughters, ²¹⁸Po and ²¹⁴Po to equilibrate and counting was performed using a portable radon monitor RDA–200, Scintrex [19]. After the initial radon analysis, the samples were sealed and stored for at least five days for ²²²Rn ingrowth and then flushed again in order to determine the ²²⁶Ra activity.

Excess radon was determined as the difference between the total ²²²Rn in samples and the supported

222Rn, assumed to be equal to the ²²⁶Ra activity. These values were decay corrected back to the time of sampling in order to assess the in situ excess radon concentrations. Once the concentrations have been determined, ideally as a complete profile through the water column, the inventory was calculated by integrating the excess radon concentrations over water depth intervals. Bottom sediment grab samples were also obtained at each site in order to assess potential diffusive fluxes of ²²²Rn from sediments.

3.3.2. Use of Ra isotopes in Ubatuba to study coastal dynamics and groundwater input

Activity concentrations of ²²³Ra, ²²⁴Ra, ²²⁶Ra and ²²⁸Ra have been also measured in seawater, surface and groundwater samples collected in Ubatuba coastal area. All the samples studied were collected during tree sampling cruises performed in June 2000 (winter), January 2002 (summer) and November 2003 (summer), respectively.

Large volume seawater samples (196 L) were pumped from 5 m below the surface into plastic drums on the R/V Velliger II. The sample volume was recorded and the seawater was percolated through a column of manganese coated acrylic fiber (<1 L/min) to quantitatively remove radium from seawater [8]. Samples for salinity and nutrients were also collected in each station.

Additional samples were obtained from a set of monitoring wells and from seepage meters installed in Flamengo Bay in 2003. These samples were processed in the same manner as the surface samples.

The samples collected in June 2000 were sent by express mail to the University of South Carolina at Columbia, SC, for measurement. Samples from January 2002 and November 2003 were measured at the University of São Paulo Marine Laboratory in Ubatuba. Each Mn fiber sample was partially dried with a stream of air and was placed in a closed–loop air circulation system as described by Moore and Arnold [20]. Helium was circulated over the Mn fiber to sweep the ²¹⁹Rn and ²²⁰Rn generated by ²²³Ra and ²²⁴Ra decay through a 1L scintillation cell where alpha particles from the decay of Rn and daughters were recorded by a photomultiplier tube (PMT) attached to the scintillation cell. Signals from the PMT were routed to a delayed–coincidence system. The delayed–coincidence system utilizes the difference in decay constants of the short-lived Po daughters of ²¹⁹Rn and ²²⁰Rn to identify alpha particles derived from ²¹⁹Rn or ²²⁰Rn decay and hence to determine activities of ²²³Ra and ²²⁴Ra on the Mn fiber. The expected error of the short-lived Ra measurements is 10%.

After the ²²³Ra and ²²⁴Ra measurements were complete, the Mn fiber samples were aged for 2–6 weeks to allow initial excess ²²⁴Ra to equilibrate with ²²⁸Th adsorbed to the Mn fiber. The samples were measured again to determine ²²⁸Th and thus to correct for supported ²²⁴Ra.

Later, the Mn fibers were leached with HCl in a Soxhlet extraction apparatus to quantitatively remove the long-lived Ra isotopes. The Ra was coprecipitated with BaSO₄. The precipitant was aged for 2 weeks to allow ²²²Rn and its daughters to equilibrate with ²²⁶Ra. The ²²⁶Ra and ²²⁸Ra were measured by gamma spectrometry of a Ba(Ra)SO₄ precipitate in a WeGe well germanium detector, after 21 days from the precipitation. The detector is a 78 cm³ coaxial intrinsic germanium crystal with a 1 cm diameter and 4 cm deep well produced by Princeton Gamma Tech. The ²²⁶Ra activities were determined by taking the mean activity of three separate photo peaks of its daughter nuclides: ²¹⁴Pb at

295.2 keV and 351.9 keV, and ²¹⁴Bi at 609.3 keV. The ²²⁸Ra content of the samples was determined from the 911 keV and 968 keV gamma ray peaks of ²²⁸Ac. Both measurements were performed at the Radioisotope Geochemical Laboratory of the University of South Carolina. The expected error of the long-lived Ra measurements is 7%.

4. Results and Discussion

4.1. Application of the ²²²Rn approach to estimate SGD in Ubatuba embayments and its seasonal