92
tion rates in sediments (Koike and SS!lrensen 1988;
Seitzinger 1988). We will limit our discussion to
the most widespread and accurate assays, including
the most recent developments.
The acetylene inhibition technique has been one
of the most widely used and relies on the inhibition
by acetylene of nitrous oxide reduction, the final
step in the denitrification pathway (SS!lrensen 1978).
The accumulation of nitrous oxide is monitored at
a high sensitivity, which allows incubation times of
few hours. Unfortunately, acetylene also inhibits nitrification, which leads to an underestimation of
rates in sediments where this process is the dominating source of NO; . Additionally, inhibition can
be incomplete at low NO; concentrations or in the
presence of H 2 S (Koike and SS!lrensen 1988; Binnerup et al. 1992).
Denitrification rates may be determined directly
from benthic N2 fluxes (Kaplan et al. 1979). This
requires minimal manipulation of the sediment, and
is possible both in situ with benthic flux chambers
(Devol 1991; Devol and Christensen 1993) and in
core incubations (Lamontagne and Valiela 1995).
The main disadvantage is that the high background
concentration of N2 in water calls for high analytical precision and relatively long incubations,
which may lead to O2 depletion and thereby change
the nitrification activity and the diffusion distance
for bottom water NO; (Rysgaard et al. 1993; Lamontagne and Valiela 1995; Blackburn 1996). N2
fluxes are also sensitive to temperature-dependent
solubility changes (Lamontagne and Valiela 1995).
Alternatively, cores can be depleted for background
N2 before the flux determination (Seitzinger et al.
1980). This, however, requires either an equilibration time of approximately 10 days to deplete pore
water N2, or separate determination of the nonbiogenic N2 flux in anoxic cores, with possible artifacts
due to ceasure of bioirrigation (Nowicki 1994).
A third approach involves the use of the stable
isotope nitrogen-15 C 5 N) (pioneered by Goering
and Pamatmat 1970). Rates of both denitrification,
NO; reduction to ammonia, and nitrification can
be obtained by parallel incubation of cores in a continuous flow system with 15NO; or 15NH: added
to the inflowing water when the isotopic composition of effluent nitrogen compounds is monitored
(Nishio et al. 1982, 1983). In such assays, natural
NO; (>99% 14NO;) and O 2 concentrations and
the associated chemical gradients can be mainBo Thamdrup and Donald E. Canfield
tained. A significant improvement was achieved
with the isotope pairing technique, which is based
on 15NO; addition, separate determination of
15N14N and 15N15N fluxes, and calculation of the
14N14N flux assuming random isotope pairing
(Nielsen 1992). The assay includes the nitrificationcoupled denitrification. It can be applied in both
closed and continuous flow modes (Rysgaard et al.
1993; Risgaard-Petersen and Rysgaard 1995) as
well as in situ in benthic flux chambers (Nielsen
and Glud 1996). It is robust and, in the no-flow
mode, rapid and easy to apply. The use of the
method is restricted to sediments with a relatively
thin NO; zone, where a homogeneous mixing of
the 15NO; with the indigenous NO; pool can be
rapidly achieved. Mathematical simulations have
shown that the basic assumptions hold to a good
approximation (Middelburg et al. 1996a), and it has
been confirmed experimentally that the addition of
labeled NO; to the system does not significantly
affect the turnover of the indigenous NO; pool
(Nielsen and Glud 1996). The rate from the assay
can be operationally separated into denitrification
based on nitrified and water-column NO; , but the
reality behind this separation has recently been debated (Middelburg et al. 1996a; Nielsen et al.
1996). This separation was introduced to illustrate
the regulatory controls on denitrification rates, but
for mass balances of the N cycle, only total denitrification rates are needed, and the separation itself
may be somewhat confusing.
Manganese and Iron Reduction
Oxidized manganese and iron, Mn(III, IV) and
Fe(III), are found in sediments in a wide variety of
solid phases from free hydr(oxides) to silicate minerals with an associated large variation in reactivity
(Lovley and Phillips 1986; Burdige et al. 1992;
Canfield et al. 1992). Also, most Mn(II) and Fe(II)
formed in sediments is either adsorbed or precipitated and only a small fraction is found in the pore
water (SS!lrensen 1982; Canfield et al. 1993b). The
deposition flux of reactive Mn and Fe in sediments
is low compared with the organic C flux and recycling is therefore necessary for these elements to
be of importance in C oxidation. Due to the predominant association with solids, such recycling
depends strongly on sediment reworking by infauna
tion rates in sediments (Koike and SS!lrensen 1988;
Seitzinger 1988). We will limit our discussion to
the most widespread and accurate assays, including
the most recent developments.
The acetylene inhibition technique has been one
of the most widely used and relies on the inhibition
by acetylene of nitrous oxide reduction, the final
step in the denitrification pathway (SS!lrensen 1978).
The accumulation of nitrous oxide is monitored at
a high sensitivity, which allows incubation times of
few hours. Unfortunately, acetylene also inhibits nitrification, which leads to an underestimation of
rates in sediments where this process is the dominating source of NO; . Additionally, inhibition can
be incomplete at low NO; concentrations or in the
presence of H 2 S (Koike and SS!lrensen 1988; Binnerup et al. 1992).
Denitrification rates may be determined directly
from benthic N2 fluxes (Kaplan et al. 1979). This
requires minimal manipulation of the sediment, and
is possible both in situ with benthic flux chambers
(Devol 1991; Devol and Christensen 1993) and in
core incubations (Lamontagne and Valiela 1995).
The main disadvantage is that the high background
concentration of N2 in water calls for high analytical precision and relatively long incubations,
which may lead to O2 depletion and thereby change
the nitrification activity and the diffusion distance
for bottom water NO; (Rysgaard et al. 1993; Lamontagne and Valiela 1995; Blackburn 1996). N2
fluxes are also sensitive to temperature-dependent
solubility changes (Lamontagne and Valiela 1995).
Alternatively, cores can be depleted for background
N2 before the flux determination (Seitzinger et al.
1980). This, however, requires either an equilibration time of approximately 10 days to deplete pore
water N2, or separate determination of the nonbiogenic N2 flux in anoxic cores, with possible artifacts
due to ceasure of bioirrigation (Nowicki 1994).
A third approach involves the use of the stable
isotope nitrogen-15 C 5 N) (pioneered by Goering
and Pamatmat 1970). Rates of both denitrification,
NO; reduction to ammonia, and nitrification can
be obtained by parallel incubation of cores in a continuous flow system with 15NO; or 15NH: added
to the inflowing water when the isotopic composition of effluent nitrogen compounds is monitored
(Nishio et al. 1982, 1983). In such assays, natural
NO; (>99% 14NO;) and O 2 concentrations and
the associated chemical gradients can be mainBo Thamdrup and Donald E. Canfield
tained. A significant improvement was achieved
with the isotope pairing technique, which is based
on 15NO; addition, separate determination of
15N14N and 15N15N fluxes, and calculation of the
14N14N flux assuming random isotope pairing
(Nielsen 1992). The assay includes the nitrificationcoupled denitrification. It can be applied in both
closed and continuous flow modes (Rysgaard et al.
1993; Risgaard-Petersen and Rysgaard 1995) as
well as in situ in benthic flux chambers (Nielsen
and Glud 1996). It is robust and, in the no-flow
mode, rapid and easy to apply. The use of the
method is restricted to sediments with a relatively
thin NO; zone, where a homogeneous mixing of
the 15NO; with the indigenous NO; pool can be
rapidly achieved. Mathematical simulations have
shown that the basic assumptions hold to a good
approximation (Middelburg et al. 1996a), and it has
been confirmed experimentally that the addition of
labeled NO; to the system does not significantly
affect the turnover of the indigenous NO; pool
(Nielsen and Glud 1996). The rate from the assay
can be operationally separated into denitrification
based on nitrified and water-column NO; , but the
reality behind this separation has recently been debated (Middelburg et al. 1996a; Nielsen et al.
1996). This separation was introduced to illustrate
the regulatory controls on denitrification rates, but
for mass balances of the N cycle, only total denitrification rates are needed, and the separation itself
may be somewhat confusing.
Manganese and Iron Reduction
Oxidized manganese and iron, Mn(III, IV) and
Fe(III), are found in sediments in a wide variety of
solid phases from free hydr(oxides) to silicate minerals with an associated large variation in reactivity
(Lovley and Phillips 1986; Burdige et al. 1992;
Canfield et al. 1992). Also, most Mn(II) and Fe(II)
formed in sediments is either adsorbed or precipitated and only a small fraction is found in the pore
water (SS!lrensen 1982; Canfield et al. 1993b). The
deposition flux of reactive Mn and Fe in sediments
is low compared with the organic C flux and recycling is therefore necessary for these elements to
be of importance in C oxidation. Due to the predominant association with solids, such recycling
depends strongly on sediment reworking by infauna