Carbon Cycle in a Meromictic Crater Lake: Lake Pavin, France

Abstract

Lake Pavin is a meromictic maar lake located in the French Massif Central. Its maximum depth is 92 m and its mean diameter is 750 m (0.445 km²). Waters below about 60 m are never mixed with overlying waters and are permanently anoxic. DIC (Dissolved Inorganic Carbon) and CH₄ concentrations in the deep monimolimnion are 14 and 4 mmol/L respectively. DIC has two origins: biogenic and volcanic.

Data on C species, ¹³C and ¹⁴C isotopes as well as auxiliary data were used to build a carbon mass balance in the lake. Main features of the carbon cycle are:

• 1. Between 70 and 92 m depth, DIC and methane fluxes coming from the sediment are estimated to 1.5 and $1.0\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$ respectively.

• 2. At about 65–70 m depth, a PM¹⁴C (percent of modern carbon) minimum suggests the inflow of a mineral water which brought $\mathrm{D}\mathrm{I}\mathrm{C} = 2.6 \pm 0.2\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$.

• 3. As suggested by a minimum of δ¹³C of DIC and very low CH₄ concentration at ca. 60 m depth, methane is almost quantitatively transformed mainly into DIC at the redoxcline. The methane escape toward the atmosphere is very low and negligible in the carbon mass balance.

• 4. The δ¹³C DIC, DIC vs. O₂ plots and some thermal and dissolved oxygen anomalies suggest a fresh water inflow at about 53 m depth; the corresponding input of DIC is estimated to $1.6 \pm 0.2\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$.

• 5. Between 20 and 50 m depth, mineralization of organic matter produced in the photic zone occurs partially. From a δ¹³C versus 1/DIC plot, a production of DIC of $0.3 \pm 0.2\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$ can be derived.

• 6. The maximum of photosynthesis is located within the metalimnion (10–15 m depth). A mean value for the corresponding DIC uptake of $3.5 \pm 0.5\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$ is assumed. During seasonal stratification of the mixolimnion (about April to December), water deeper than 10–12 m are isolated from the atmosphere and superficial inputs as shown by the low PM¹⁴C value of 45 % at 10 m depth.

• 7. In the superficial layer, an input of $1.3 \pm 0.2\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$ of DIC is brought by several small brooks, and an output of $3.0 \pm 0.2\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$ is calculated for the discharge.

• 8. CO₂ exchanges with atmosphere are highly variable seasonally; the lake is strongly under saturated in summer and oversaturated in November. An annual balance is derived from modeling the global lake functioning, leading to a CO₂ escape of $4 \pm 1\ \mathrm{kmol}\ {\mathrm{d}}^{-1}$.

Budgets for DIC and alkalinity (alk) are proposed. For this latter, we need to distinguish two kinds of dissolved species:

• “conservative” ions (e.g. ${\mathrm{Na}}^{+}$, ${\mathrm{K}}^{+}$, ${\mathrm{Ca}}^{2+}$, ${\mathrm{Mg}}^{2+}$, ${\mathrm{Cl}}^{-}$) brought by streams and other water venues

• “reactive” species (e.g. ${\mathrm{Fe}}^{2+}$, ${{\mathrm{NH}}_4}^{+}$, ${\mathrm{H}}_2{{\mathrm{PO}}_4}^{-}$), essentially produced within the sediment and consumed or precipitated at ca. 60 m depth i.e. at the redox interface.

CO₂ behaviour is deduced from both DIC and Alk budgets. The dissolved concentration of CO₂ in the monimolimnion, associated with those of CH₄ and N₂, allow assessing the gas outburst natural hazard, which is very low considering the actual conditions.