394
Walter et ai.
between ocean basins and changes in the chemical
composition of settling particles, become predominant and limit the use ofthe 231PaI 230Th ratio
xs
xs
as an indicator of paleoproductivity.
Theory
Both 231Pa (half-life: 32.500 years and 230Th (halflife: 75.400 years) are natural radionuclides, which
are continuously produced in seawater by in-situ
decay of their dissolved progenitors 23SU and 234U
(their only significant source in the ocean). Since
the residence time of uranium in the ocean (several hundred thousand years; Cochran 1992; Chen
et a!. 1986) is long compared to the mixing time of
the oceans (approx. thousand years, Broecker and
Peng 1982), the distribution of U throughout the
oceans is very homogenous. As a result, both 231Pa
and 23°Th are produced in seawater at a constant
rate with an initial 231PaI 230Th activity ratio of
0.093. In contrast to U, 231Pa and 23 0 Th are particle reactive and are scavenged to the sea floor on
time scales of about 50-200 years and 5-40 years,
respectively (Nozaki et a!. 1981; Anderson et al.
1983a,b; Nozaki and Nakanishi 1985; Nozaki and
Yamada 1987; Huh and Beasley 1987; Rutgers van
der Loeff and Berger 1993; Guo et a!. 1995). Because these timescales are short compared to their
half-lives, radioactive decay in the water column
can be neglected. Consequently, if2 3 0Th and 231Pa
are scavenged from the water column without
fractionation, we should expect an ocean-wide
,,231PaI,,23°Th ratio in newly formed sediment equal
to the production ratio of 0.093.
Variable ,,231Pa/,,230Th ratios in surface
sediments throughout the oceans indicate that both
radionuclides follow different pathways of removal
from seawater. Because of its short residence time
in the water column, horizontal transport of230Th
before scavenging is minimal over most of the
oceans. As a result, the flux of230Th to the seafloor
is nearly identical to its rate of production in the
water column (Anderson et al. 1983a,b; Bacon et
al. 1985; Taguchi eta!. 1989; Yu 1994). In contrast,
the longer oceanic residence time for dissolved
231Pa allows a large scale diffusive transport of this
nuclide over basin-wide distances prior to scavenging, resulting in its preferential removal in high particle flux regions (Anderson eta!. 1983a,b; Yang et
al. 1986; Bacon 1988; Lao et a!. 1992b; 1993; Yu
et al. 1996). Since 231Pa is often found at ocean
margins (DeMaster 1979; Anderson et al. 1983a,b;
Shimmield et al. 1986) this preferential removal is
known as "boundary scavenging".
Boundary Scavenging
In principle, three requirements have to be met for
effective boundary scavenging of a particle reactive element. First, there must be intensified scavenging at the ocean margins or other areas of the
oceans to produce horizontal gradients in the dissolved radionuclide concentration. These gradients
will then maintain a net diffusive transport from low
particle flux regions (with low scavenging rates) to
high particle flux regions (with high scavenging
rates). Second, the scavenging residence time of
the radionuclide must be longer than the timescale
required for diffusive lateral mixing within the ocean
basin, so that significant quantities of the dissolved
radionuclide generated in low flux regions can be
transported to high flux regions prior to being removed from the water column. Third, the mean
residence time of water in the ocean basin must be
long compared to the scavenging residence time of
the radionuclide and the time required for lateral
mixing within the basin.
In order to demonstrate how these three factors affect the large-scale distributions of231Pa
and 230Th throughout the ocean, a simplified version of the scavenging model developed for the
Pacific Ocean by Bacon (1988) is used (model A,
in Fig. 1). In contrast to the original model, in our
illustration the influence of the chemical composition of particulate matter on the scavenging of231Pa
and 23 0 Th is not considered. The ocean (volume V)
is divided into two boxes, an interior box represented
with suffix i, with 80% of the total volume ofthe
ocean (0.8V), and a margin box with suffix m containing 20% (O.2V) ofthe total volume ofthe ocean
(Fig. 1 ). Each box has average radionuclide concentrations C; and C m , respectively. There is horizontal fluid exchange between boxes i and m at a rate
vox. The scavenging rates in each box are given by
rate coefficients k. The material balances for 23 0Th
in boxes i and m at steady state are respectively,
Walter et ai.
between ocean basins and changes in the chemical
composition of settling particles, become predominant and limit the use ofthe 231PaI 230Th ratio
xs
xs
as an indicator of paleoproductivity.
Theory
Both 231Pa (half-life: 32.500 years and 230Th (halflife: 75.400 years) are natural radionuclides, which
are continuously produced in seawater by in-situ
decay of their dissolved progenitors 23SU and 234U
(their only significant source in the ocean). Since
the residence time of uranium in the ocean (several hundred thousand years; Cochran 1992; Chen
et a!. 1986) is long compared to the mixing time of
the oceans (approx. thousand years, Broecker and
Peng 1982), the distribution of U throughout the
oceans is very homogenous. As a result, both 231Pa
and 23°Th are produced in seawater at a constant
rate with an initial 231PaI 230Th activity ratio of
0.093. In contrast to U, 231Pa and 23 0 Th are particle reactive and are scavenged to the sea floor on
time scales of about 50-200 years and 5-40 years,
respectively (Nozaki et a!. 1981; Anderson et al.
1983a,b; Nozaki and Nakanishi 1985; Nozaki and
Yamada 1987; Huh and Beasley 1987; Rutgers van
der Loeff and Berger 1993; Guo et a!. 1995). Because these timescales are short compared to their
half-lives, radioactive decay in the water column
can be neglected. Consequently, if2 3 0Th and 231Pa
are scavenged from the water column without
fractionation, we should expect an ocean-wide
,,231PaI,,23°Th ratio in newly formed sediment equal
to the production ratio of 0.093.
Variable ,,231Pa/,,230Th ratios in surface
sediments throughout the oceans indicate that both
radionuclides follow different pathways of removal
from seawater. Because of its short residence time
in the water column, horizontal transport of230Th
before scavenging is minimal over most of the
oceans. As a result, the flux of230Th to the seafloor
is nearly identical to its rate of production in the
water column (Anderson et al. 1983a,b; Bacon et
al. 1985; Taguchi eta!. 1989; Yu 1994). In contrast,
the longer oceanic residence time for dissolved
231Pa allows a large scale diffusive transport of this
nuclide over basin-wide distances prior to scavenging, resulting in its preferential removal in high particle flux regions (Anderson eta!. 1983a,b; Yang et
al. 1986; Bacon 1988; Lao et a!. 1992b; 1993; Yu
et al. 1996). Since 231Pa is often found at ocean
margins (DeMaster 1979; Anderson et al. 1983a,b;
Shimmield et al. 1986) this preferential removal is
known as "boundary scavenging".
Boundary Scavenging
In principle, three requirements have to be met for
effective boundary scavenging of a particle reactive element. First, there must be intensified scavenging at the ocean margins or other areas of the
oceans to produce horizontal gradients in the dissolved radionuclide concentration. These gradients
will then maintain a net diffusive transport from low
particle flux regions (with low scavenging rates) to
high particle flux regions (with high scavenging
rates). Second, the scavenging residence time of
the radionuclide must be longer than the timescale
required for diffusive lateral mixing within the ocean
basin, so that significant quantities of the dissolved
radionuclide generated in low flux regions can be
transported to high flux regions prior to being removed from the water column. Third, the mean
residence time of water in the ocean basin must be
long compared to the scavenging residence time of
the radionuclide and the time required for lateral
mixing within the basin.
In order to demonstrate how these three factors affect the large-scale distributions of231Pa
and 230Th throughout the ocean, a simplified version of the scavenging model developed for the
Pacific Ocean by Bacon (1988) is used (model A,
in Fig. 1). In contrast to the original model, in our
illustration the influence of the chemical composition of particulate matter on the scavenging of231Pa
and 23 0 Th is not considered. The ocean (volume V)
is divided into two boxes, an interior box represented
with suffix i, with 80% of the total volume ofthe
ocean (0.8V), and a margin box with suffix m containing 20% (O.2V) ofthe total volume ofthe ocean
(Fig. 1 ). Each box has average radionuclide concentrations C; and C m , respectively. There is horizontal fluid exchange between boxes i and m at a rate
vox. The scavenging rates in each box are given by
rate coefficients k. The material balances for 23 0Th
in boxes i and m at steady state are respectively,