215Mineralogical study was not able to detect Fe(III) miner-
als in the sediments, while they were still present in the deep-
est sediment trap at 86 m in the water column (Cosmidis
et al. 2014 ). This suggests complete Fe(III) reductive disso-
lution at the sediment-water interface, probably favored by
the high availability of fresh and reactive organic matter. A
complete reduction of Fe(III) minerals is also supported by
the δ^56 Fe values (near 0 ‰) of sediment porewaters (Busigny
et al. 2014 ), because partial Fe(III) reduction would produce
dissolved Fe(II) with markedly negative δ^56 Fe values, which
are not detected (Severmann et al. 2006 ; Crosby et al. 2005 ).
An important question to address in future studies is
whether Fe(III) is mostly reduced in the water column and/or
at the water-sediment interface, and in which proportion?
From the profiles obtained on Fe in SPM and turbidity (Fig.
12.3), it appears that a large part of the Fe(III) particles pre-
cipitated at the redox transition zone re-dissolves rapidly
when particles sink in the anoxic water column. This conclu-
sion is deduced from geochemical modeling (Fig. 12.7d).
However, the high concentration of Fe(II) and SRP at the
lake bottom requires benthic fluxes from the sediment, which
should be derived from the dissolution of an Fe-P-rich phase.
This phase could be either (1) a residual Fe(III)-P rich phase
formed at the redox transition zone and having escaped water
column reductive dissolution, (2) a detrital Fe(III)-phase
with adsorbed P (as evidenced in Fig. 12.5), (3) a metastable
Fe(II) phosphate, formed in the anoxic Fe-P-rich deep
waters, that transforms into vivianite during sediment early
diagenesis, or (4) a combination of these three cases. It is
worth noting that the flux of Fe(III)-P phase reaching the
lake bottom sediment can be increased occasionally for
instance if particles located at the redox transition zone are
flushed by diatom blooms, resulting in an increase of the par-
ticle sedimentation flux. The flux of Fe(III) can also be
increased during periods of lake mixing down to the redox
boundary, inducing a drastic precipitation of Fe(III)-P phase.
In this case, the steady-state regime used for our modeling is
no longer valid and these punctual disturbances may explain
that a larger fraction of the Fe(III) flux reaches the water-
sediment interface. In any case, a benthic flux of Fe(II) andTable 12.3 Description of the parameters and constants used for the biogeochemical model of Lac Pavin
- Notation of variables
c_Fe54: concentration of dissolved^54 Fe
c_Fe56: concentration of dissolved^56 Fe
x_Fe54: concentration of particulate^54 Fe
x_Fe56: concentration of particulate^56 Fe
c_P: concentration of dissolved PO 4 3− - List of reactions (kinetic equations)
In the mixolimnion
Iron oxidation V = v_ox_Fec_Fe/(c_Fe-c_M)
Phosphate adsorption V = v_ads_Pc_P
In the monimolimnion
Iron reduction V = v_red_Fex_Fe
Vivianite precipitation V = v_viv_Fe(c_Fe*c_P-Ks) - Kinetic constants and adjusted parameters
Monod parameter (c_M) 0.01 mol.m−3
Fe concentration in the subsurface spring 0.2 mol.m−3
PO 4 3− concentration in the subsurface spring 0.05 mol/m^3
Benthic Fe flux (from sediment) 9.2 × 10−4 mol.m−2.day−1
Benthic PO 4 3− flux (from sediment) 4.5 × 10−4 mol.m−2.day−1
Vivianite solubility product 0.12 mol^2 .m−6
Fe oxidation kinetic constant (v_ox_Fe) 1 × 10−3 day−1
PO 4 3− adsorption kinetic constant (v_ads_P) 3 × 10−4 m^3 .mol−1.day−1
Fe reduction kinetic constant (v_red_Fe) 0.2 × 10−4 m^3 .mol−1.day−1
Vivianite precipitation kinetic constant (v_viv_Fe) 2.7 × 10−4 m^6 .mol−2.day−1
Fe isotope fractionation factors (^56 Fe/^54 Fe ratio)
oxidation/precipitation 1.0009
reduction/dissolution 0.999
vivianite precipitation 0.9995
Fe isotope compositions (^56 Fe/^54 Fe ratio)
benthic Fe flux 0.2 ‰
subsurface ferruginous spring −1 ‰
12 Iron Wheel in Lac Pavin