6. Benthic Respiration in Aquatic Sediments
or through physical processes (Aller 1990; Canfield
et al. 1993b; Aller et al. 1996).
Oxidized Mn and Fe in sediments can be reduced
through bacterial respiration coupled to the oxidation of organic matter or hydrogen (Lovley 1991).
Fermentative organisms may also utilize the metals
as electron sinks, though this process is a minor part
of their metabolism (Lovley 1991) and, therefore,
likely of little significance in C oxidation. Bacterial
Mn and Fe reduction competes with rapid abiotic
reactions with H 2 S and, for manganese, Fe2+ (Pyzik and Sommer 1981; Postma 1985; Yao and Millero 1993). Manganese reduction by sulfide may
also be bacterially catalyzed (Aller and Rude 1988;
Lovley and Phillips 1994). Humic substances have
been shown to shuttle electrons from bacteria to Fe
oxides during the oxidation of organic matter in
some pure cultures (Lovley et al. 1996). Whether
such interactions play a role in sedimentary iron
reduction is not yet known.
Dissimilatory Mn and Fe reductions are the most
rarely quantified pathways of benthic respiration.
They may contribute significantly to C oxidation in
some continental margin and freshwater wetland
sediments (Aller 1990; Canfield et al. 1993a; Roden
and Wetzel 1996; Thamdrup and Canfield 1996),
whereas in some cases Fe and, in particular, Mn
reductions are mainly coupled to inorganic reoxidation (Canfield et al. 1993a; Aller 1994; Thamdrup et al. 1994a). For a recent review see Thamdrup (2000).
Due to the co-occurrence of respiratory reduction
and reactions with reduced inorganic species, similar problems are met in the quantification of Mn
and Fe respiration as for aerobic respiration, with
additional complications associated with the analysis of oxidized and reduced Mn and Fe pools. The
use of radiolabeled Fe(III) oxide as a tracer of Fe
reduction is precluded by rapid electron exchange
that causes the formation of labeled Fe(II) without
net Fe reduction (Roden and Lovley 1993). Iron
reduction rates have been estimated by monitoring
Fe(III) consumption or Fe(II) production by means
of wet chemical Fe extractions in anoxic incubations of homogenized sediment from discrete depth
intervals (S0rensen 1982; Lovley and Phillips
1986; Roden and Wetzel 1996). This approach has
some potential with freshwater sediments with low
rates of SO~- reduction and, hence, of H 2 S production. There are, however, several complications
93
associated with the use of changes in Fe pools for
measuring Fe reduction. Long incubation times
(approximately 10 days) are mostly necessary to
obtain sufficient changes in the Fe(II) or Fe (III)
pools due to high background levels. Also, it is not
clear that iron extractions retrieve all reacting
Fe(III) or produced Fe(II). Particularly, latticebound Fe in clay minerals, which is not targeted by
conventional extractions, may be redox active
(Wallmann et al. 1993; Kostka et al. 1996). The
extraction efficiency toward Fe reduced during
incubation may be tested by measuring the Fe
reduction/2:C02 production ratios in sediment
known to be dominated by Fe reduction (Roden and
Wetzel 1996). As a further problem, oxidation or
reduction artifacts may be associated with the extractions (Phillips and Lovley 1987; Jensen and
Thamdrup 1993; Petersen 1993). Mossbauer spectroscopy has been used as an alternative to, and a
calibration of, wet extractions, but this technology
is not generally accessible (Wallmann et al. 1993).
If SO~ - reduction is of significance, interpretation
of Fe(III) consumption is difficult as it is not clear
to what extend the H 2 S formed reacts with Fe(III).
Independence of Fe reduction on H 2 S oxidation has
been demonstrated by addition of molybdate to inhibit SO~ - reduction, or by direct determination of
SO~ - reduction rates with radiotracer (S0rensen
1982; Canfield 1989; Roden and Wetzel 1996).
However, inhibition of SO~ - reduction with molybdate does not exclude the reaction of Fe (III) with
reduced S already present in the sediment. Finally,
Fe reduction may also be obscured by a rapid reoxidation of Fe(II) by Mn oxides. To exclude this reaction, Ferrozine, a strong Fe 2 + chelator has been
added to the incubated sediment (Canfield et al.
1993b).
Manganese reduction rates have been estimated
from the accumulation of dissolved Mn 2 +, either
by assuming an adsorption coefficient or by incubating sediment in large volumes of anoxic water
in order to minimize adsorption or precipitation
(Aller and Mackin 1989; Aller 1990, 1994). Reduction of Mn by reduced Fe or S was, however,
not excluded in these experiments. Furthermore,
adsorption and precipitation of Mn 2 + is poorly understood, and rates based on Mn 2 + may seriously
underestimate actual rates of reduction (Aller and
Rude 1988; Canfield et al. 1993b; Aller 1994).
Changes in oxidation state of solid-phase Mn have
or through physical processes (Aller 1990; Canfield
et al. 1993b; Aller et al. 1996).
Oxidized Mn and Fe in sediments can be reduced
through bacterial respiration coupled to the oxidation of organic matter or hydrogen (Lovley 1991).
Fermentative organisms may also utilize the metals
as electron sinks, though this process is a minor part
of their metabolism (Lovley 1991) and, therefore,
likely of little significance in C oxidation. Bacterial
Mn and Fe reduction competes with rapid abiotic
reactions with H 2 S and, for manganese, Fe2+ (Pyzik and Sommer 1981; Postma 1985; Yao and Millero 1993). Manganese reduction by sulfide may
also be bacterially catalyzed (Aller and Rude 1988;
Lovley and Phillips 1994). Humic substances have
been shown to shuttle electrons from bacteria to Fe
oxides during the oxidation of organic matter in
some pure cultures (Lovley et al. 1996). Whether
such interactions play a role in sedimentary iron
reduction is not yet known.
Dissimilatory Mn and Fe reductions are the most
rarely quantified pathways of benthic respiration.
They may contribute significantly to C oxidation in
some continental margin and freshwater wetland
sediments (Aller 1990; Canfield et al. 1993a; Roden
and Wetzel 1996; Thamdrup and Canfield 1996),
whereas in some cases Fe and, in particular, Mn
reductions are mainly coupled to inorganic reoxidation (Canfield et al. 1993a; Aller 1994; Thamdrup et al. 1994a). For a recent review see Thamdrup (2000).
Due to the co-occurrence of respiratory reduction
and reactions with reduced inorganic species, similar problems are met in the quantification of Mn
and Fe respiration as for aerobic respiration, with
additional complications associated with the analysis of oxidized and reduced Mn and Fe pools. The
use of radiolabeled Fe(III) oxide as a tracer of Fe
reduction is precluded by rapid electron exchange
that causes the formation of labeled Fe(II) without
net Fe reduction (Roden and Lovley 1993). Iron
reduction rates have been estimated by monitoring
Fe(III) consumption or Fe(II) production by means
of wet chemical Fe extractions in anoxic incubations of homogenized sediment from discrete depth
intervals (S0rensen 1982; Lovley and Phillips
1986; Roden and Wetzel 1996). This approach has
some potential with freshwater sediments with low
rates of SO~- reduction and, hence, of H 2 S production. There are, however, several complications
93
associated with the use of changes in Fe pools for
measuring Fe reduction. Long incubation times
(approximately 10 days) are mostly necessary to
obtain sufficient changes in the Fe(II) or Fe (III)
pools due to high background levels. Also, it is not
clear that iron extractions retrieve all reacting
Fe(III) or produced Fe(II). Particularly, latticebound Fe in clay minerals, which is not targeted by
conventional extractions, may be redox active
(Wallmann et al. 1993; Kostka et al. 1996). The
extraction efficiency toward Fe reduced during
incubation may be tested by measuring the Fe
reduction/2:C02 production ratios in sediment
known to be dominated by Fe reduction (Roden and
Wetzel 1996). As a further problem, oxidation or
reduction artifacts may be associated with the extractions (Phillips and Lovley 1987; Jensen and
Thamdrup 1993; Petersen 1993). Mossbauer spectroscopy has been used as an alternative to, and a
calibration of, wet extractions, but this technology
is not generally accessible (Wallmann et al. 1993).
If SO~ - reduction is of significance, interpretation
of Fe(III) consumption is difficult as it is not clear
to what extend the H 2 S formed reacts with Fe(III).
Independence of Fe reduction on H 2 S oxidation has
been demonstrated by addition of molybdate to inhibit SO~ - reduction, or by direct determination of
SO~ - reduction rates with radiotracer (S0rensen
1982; Canfield 1989; Roden and Wetzel 1996).
However, inhibition of SO~ - reduction with molybdate does not exclude the reaction of Fe (III) with
reduced S already present in the sediment. Finally,
Fe reduction may also be obscured by a rapid reoxidation of Fe(II) by Mn oxides. To exclude this reaction, Ferrozine, a strong Fe 2 + chelator has been
added to the incubated sediment (Canfield et al.
1993b).
Manganese reduction rates have been estimated
from the accumulation of dissolved Mn 2 +, either
by assuming an adsorption coefficient or by incubating sediment in large volumes of anoxic water
in order to minimize adsorption or precipitation
(Aller and Mackin 1989; Aller 1990, 1994). Reduction of Mn by reduced Fe or S was, however,
not excluded in these experiments. Furthermore,
adsorption and precipitation of Mn 2 + is poorly understood, and rates based on Mn 2 + may seriously
underestimate actual rates of reduction (Aller and
Rude 1988; Canfield et al. 1993b; Aller 1994).
Changes in oxidation state of solid-phase Mn have