197flux of DIC with a given PMC and an input of mineral
water at 68 ± 2 m depth was built using the AQUASIM code
(Reichert 1994 ). The flow of the mineral water ( 138 md^31 −)
has been estimated by Assayag et al. ( 2008 ) from^18 O data,
revisited here at 120 md^31 −. Observed results are consistent
with a flux of 75 .mmolm d−−^21 of DIC (as mentioned above
Sect. 11.5.3.1) with a settling POC with 75 PMC at the bot-
tom sediment-water interface (as mentioned above Sect.
11.5.3.1) and an input of 2600 molday−^1 of 0 PMC DIC by
the mineral water (Fig. 11.2). Therefore, the ratio of bio-
genic DIC to “volcanic” carbon is 1500 /( 2600 = 06 .). The
disagreement between values previously obtained either by
(^13) C(» 2 ) and (^14) C(» 05 .) is related: (i) to the different
depths of biogenic C inputs (OM deposition onto the bot-
tom and subsequent remineralization) and volcanic C (ca.
68 m depth); (ii) to the recycling of dead^14 C in the lake,
leading to a relative low PMC for the organic matter pro-
duced into the lake. Such a large^14 C impoverishment of the
lake autochthonous organic production compare to aerial
allochthonous sources (tree leaves) may be inferred from
sedimentary archives since at least 1000 years ago (Albéric
et al. 2013 ).
The same model allows the determination of conservative
species such as alkalinity to 500 eqd−^1. Estimation of reac-
tive species is more difficult. However, CO 2 rich mineral
waters present low concentrations of both NH 4 + and soluble
phosphates (relatively to monimolimnion concentrations). A
tentative value for Fe^2 + derived from the model is about
200 eqday−^1.
The model shows also a significant difference of δ^13 C
between 88 and 68 m depth (about 1 ‰) in agreement with
the experimental data (Fig. 11.11).
11.5.3.3 The Chemocline and the Redox
Boundary
The large gradients observed in this layer correspond to the
low dispersion coefficient estimated at 0 004.m^21 d−.
Redox boundary occurs at mixolimnion-mesolimnion
interface. Fe(II), NH 4 + and CH 4 are completely oxidized in
this zone. The almost total methane oxidization is confirmed
by the drastic decrease of δ^13 C of DIC at about 60 m depth
with respect to values at higher depths. However, δ^13 CDIC
profile is very sensitive to the relative depths of the oxidation
zone and of the top of the low dispersion layer. The best
results are obtained with a complete oxidization of CH 4
between 59.5 and 61.5 m depth and a top of the low disper-
sion coefficient zone at 60 m in agreement with Lopes et al.
results ( 2011 ) (Fig. 11.12).
11.5.3.4 Fresh Water Input at ca. 53 m Depth
Sublacustrine inputs were inferred in ancient papers (e.g.
Perreau 1948 ), but the first attempt to calculate an hydrological
budget was proposed by Meybeck et al. ( 1975 ), based on tri-
tium data, rainfall budget on the catchment area and only four
outflow values (pygmy current meter method). Due to incoher-
ence between^3 H model (which gave 25 Ls−^1 ) and observed
outflow data (from 65 to 360 Ls−^1 ), the flow of sublacustrine
input was not clarified. They just inferred, from chemical com-
position of the monimolimnion, that missing input might cor-
respond to mineral water feeding the bottom water. Martin
( 1985 ) proposed a box model of the lake from tritium data. The
apparent water deficit was estimated to 40 Ls−^1 , corresponding
to a mineral water inflow located in the monimolimnion.
Camus et al. ( 1993 ) barely revisited this box-model and
deduced a sublacustrine input of about 35 Ls−^1 also located in
Fig. 11.11 Model and
data for^14 C
measurements in the
monimolimnion
The different curves
correspond to different
C inputs in mmold−^1
in order to check the
sensibility of the model
for this parameter
11 Carbon Cycle in a Meromictic Crater Lake: Lake Pavin, France
http://www.ebook3000.com