Carbon cycling in the epilimnion of Lake Kivu (East Africa): surface net autotrophy and emission of CO2 to the atmosphere sustained by geogenic inputs

Carbon cycling in the epilimnion of Lake Kivu (East Africa): surface net autotrophy and emission of CO

to the atmosphere sustained by geogenic inputs

A.V. Borges

_{, Bouillon S.}

_{, Morana C.}

_{, P. Servais}

_{, J.-P. Descy}

_{, F. Darchambeau}

1

_{Université de Liège (BE)}

_{KULeuven (BE),}

_{Université Libre de Bruxelles (BE),}

_{Université de Namur (BE), * alberto.borges@ulg.ac.be}

the seasonal variations (March 2007 - mid rainy season, September 2007 - late dry season, June 2008 - early dry season, and April 2009 - late rainy season). The partial pressure of CO₂(pCO₂) in surface waters of the main basin of Lake Kivu showed modest spatial variations (coefficient of variation between 3% and 6%), and modest seasonal variations with an amplitude of 163 ppm (between 579±23 ppm on average in March 2007 and 742±28 ppm on average in September 2007). The most prominent spatial feature of the pCO₂distribution was the very high pCO₂ values in Kabuno Bay (a small sub-basin of Lake Kivu) ranging between 13158 ppm and 14213 ppm (between 18 and 26 times higher than in the main basin). Surface waters of the main basin of Lake Kivu were a net source of CO₂to the atmosphere at an average rate of 5.9 mmol m-2_d-1_{which is lower than the global average reported for freshwater} (16.0 mmol m-2_d-1_{), saline (81 to 105 mmol m}-2_d-1_{) and volcanic lakes (458 to 51,183 mmol m}-2_d-1_{). In Kabuno Bay,} the CO₂emission to the atmosphere was on average 274.8 mmol m-2_d-1_{(~46 times higher than in main basin).} The deepening of the mixed layer and entrainment of deeper waters to the surface mixed layer was main driver of the seasonal variations of CH₄as shown by the positive correlation of pCO₂and CH_4,of pCO₂and the mixed layer depth, and of pCO2and δ13CDIC. Indeed, deeper waters have higher pCO2and CH4, and dissolved inorganic carbon (DIC) is more13_{C-depleted than that in surface waters.}

atmosphere epilimnion TA fluxes (mmol m-2_d-1₎ March 2007 rivers Ruzizi vertical input closing term 1.3 40.5 49.9 10.7 (FCaCO3= 5.4) atmosphere epilimnion TA fluxes (mmol m-2_d-1₎ September 2007 rivers Ruzizi vertical input closing term 1.5 44.7 54.0 20.8 (FCaCO3= 10.4) atmosphere epilimnion TA fluxes (mmol m-2_d-1₎ June 2008 rivers Ruzizi vertical input closing term 0.6 46.0 55.4 10.0 (FCaCO3= 5.0) atmosphere epilimnion TA fluxes (mmol m-2_d-1₎ April 2009 rivers Ruzizi vertical input closing term 1.3 41.3 44.0 4.0 (FCaCO3= 2.0) atmosphere epilimnion DIC fluxes (mmol m-2_d-1₎

March 2007 rivers Ruzizi vertical input closing term 1.3 36.9 69.6 29.6 (FPOC= 24.2) atmosphere epilimnion DIC fluxes (mmol m-2_d-1₎

September 2007 rivers Ruzizi vertical input closing term 1.5 41.5 74.4 26.6 (FPOC= 16.2) atmosphere epilimnion DIC fluxes (mmol m-2_d-1₎

June 2008 rivers Ruzizi vertical input closing term 0.6 42.6 74.5 26.7 (FPOC= 21.7) atmosphere epilimnion DIC fluxes (mmol m-2_d-1₎

April 2009 rivers Ruzizi vertical input closing term 1.5 38.3 60.9 18.4 (FPOC= 16.4) 4.4 CO2efflux 7.8 CO2efflux 5.8 CO2efflux 5.7 CO2efflux (FPOC= 21.7) (FPOC= 21.5) (FPOC= 23.2) (FPOC= 22.9) 65.7 71.3 75.0 69.4 March 2007 584 563 597 594 552 599 605 581 590 534 537 14213 592 587 586 Kabuno September 2007 11213 760 765 781 736 767 759 746 778 730 704 701 704 718 739 Kabuno June 2008 13158 597 602 626 635 640 612 637 614 627 629 612 640 605 597 Kabuno April 2009 673 627 667 695 706 652 637 618 13393 616 581 583 605 628 617 711 Kabuno Ze Chl-a PNPP BP BR PNPP-BR PER

(m) (mg m-3_{) (mmol m}-2 _d-1_{) (mmol m}-2 _d-1_{) (mmol m}-2 _d-1_{) (mmol m}-2 _d-1_{) (%)} March 2007 15/03/2007 Kibuye 18 38.3 27.0 23.2 35.7 -8.7 54 17/03/2007 Kibuye 20 48.4 42.5 25.8 39.9 2.6 33 23/03/2007 Ishungu 17 36.1 49.7 40.5 49.6 0.1 32 September 2007 09/09/2007 Kibuye 19 56.4 42.9 35.9 47.9 -5.0 42 12/09/2007 Kibuye 18 55.1 45.9 34.0 45.9 0.1 36 04/09/2007 Ishungu 20 48.2 n.d. 16.0 29.9 n.d. n.d. June 2008 23/06/2008 Kibuye 24 42.8 46.0 7.7 21.6 24.4 3 11/07/2008 Kibuye 20 37.8 42.0 11.1 24.3 17.7 17 03/07/2008 Ishungu 19 28.1 40.7 3.9 13.6 27.1 -4 April 2009 04/05/2009 Kibuye 21 22.9 14.2 49.8 61.2 -47.0 82 21/04/2009 Ishungu 24 39.3 24.5 43.5 58.9 -34.4 68 pCO2(ppm) 20 30 40 50 60 70 500 550 600 650 700 750 800 850 March 2007 June 2008 September 2007 April 2009

Mixed layer depth (m)

p C O2 ( p p m ) 2.5 3.0 3.5 4.0 4.5 500 550 600 650 700 750 800 850 δ δδ δ13_C DIC (‰) p C O2 (p p m ) 0 25 50 75 100 125 150 500 550 600 650 700 750 800 850 CH4 (nM) p C O2 (p p m )

In order to test if vertical mixing was the only driver of seasonal variations of pCO2, we computed DIC whole-lake mass balance of bulk concentrations and of stable isotopes. The budgets show that the input of DIC from depth and export by the Ruzizi (the only outflow of the lake) were the main players in the DIC budget. Export of CaCO₃ was significant. There was a net loss of carbon from surface waters that can only ascribed to export of POC to depth, hence, we show that the epilimnion of Lake Kivu was net autotrophic. This is due to the modest river inputs of organic carbon owing to the small ratio of catchment area to lake surface area (2.15), and implies that the CO₂ emission to the atmosphere must be sustained by DIC inputs from depth of geogenic origin from deep geothermal springs.

Based on metabolic rate measurements and mass balance

considerations, we show that bacterial respiration is not solely sustained by particulate primary production, and is substantially sustained by dissolved primary production. This leads to moderate to high levels of Percent of extracellular release (PER).

Table :Photic depth (Ze), chlorophyll-a concentration (Chl-a), particulate net primary production (PNPP), bacterial production integrated over Ze (BP), bacterial respiration (BR), and percent of extracellular release (PER, %), at two stations in the main basin of Lake Kivu [Kibuye, Ishungu], in March 2007, September 2007, June 2008, and April 2009