Numerical modelling of methane emissions from thermokarst lakes - PowerPoint PPT Presentation

ENVIROMIS-2010, 7 July, Tomsk, Russia V.M.Stepanenko 1 , Е .E.Machulskaya 1,2 , and M.V.Glagolev 3,4 1. Moscow State University, Research Computing Center 2. Deutscher Wetterdienst 3. Moscow State University, Faculty of Soil Science 4. University of Yugra, Chanty-Mansiisk Numerical modelling of methane emissions from thermokarst lakes The work is supported by grants: RFBR 09-05-13562- офи _ ц , 09-05-00379- а , 10-05-00981- а , П№ 1394

Atmospheric methane and its sources Sources of methane in a climate system ¡ Mtones СН 4 /yr ¡ Animals (mostly ruminants), without termites ¡ 106 ¡ Termites ¡ 23 ¡ Rice paddies ¡ 69 ¡ Natural ¡wetlands, ¡excluding ¡tundra ¡ 113 ¡ Tundra ¡ 19 ¡ Oceans ¡ 14 ¡ Lakes ¡ 5 ¡ Methane hydrates ¡ 4 ¡ IPCC report, Volcanoes ¡ 1 ¡ 2007 Other natural sources ¡ 6 ¡ Burials of solid waste products ¡ 33 ¡ Coal industry ¡ 46 ¡ Gas industry ¡ 54 ¡ Biomass burning ¡ 40 ¡ Automobiles ¡ 1 ¡ TOTAL ¡ ~530 ¡

Emission of methane by thermokarst lakes • thermokarst lakes in Northern Siberia occupy 22-48% of the area • satellite images indicate expanding of thermokarst lakes area Unfreezing “hotspot” – the source of methane during wintertime • 8 - 50% of anthropogenic emissions in XXI century depending on IPCC scenario (K. Walter et al., 2006, Nature )

Implication to climate change and climate modeling • Positive feedback: Climate warming Thermokarst Increase of methane development, fluxes from expansion of lakes thermokarst lakes We need a modeling tool, a parameterization of thermokarst lakes’ emissions in climate models

Methane emission: bogs and lakes Mechanism of methane production • On bogs the substrate for methane production comes from surface NPP -> modeling approaches are well developed • In lakes methane is produced (i) from lake bottom NPP and (ii) from the old organics, that has been sequestered in permafrost and comes to positive temperature region while talik is deepening -> the need for new parameterization Implication to annual cycle • On bogs cold season emission is very low; • In lakes methane is produced in talik, that is under positive temperatures all year round (40-50% of annual emission happen in cold period)

Methane concentration in lake talik [ CH ] [ CH ] ∂ ∂ ∂ 4 4 k P E F = + − − CH , m t z z 4 ∂ ∂ ∂ (B. Walter & Heimann, 1996, 2001) Neglected : vegetation transport F P Ebullition: new ( ) [ E k f CH CH , [ ] ] = Δ Δ e step 4 4 [ CH ] [ CH ] [ CH ] Δ = − P 4 4 4 max old Production: P P P = + new old P T - calibrated P P exp z f T q ( ) ( ) 10 new ,0 = −α parameter new new ,0 new step 00

Methane production from old organics decomposition • happens only under positive temperatures • is exponentially dependent on temperature • is proportional to decomposable organics content T * P P C f T q * ( ) 10 P = - calibrated parameter old old ,0 old step 00 old ,0 V C C ∂ C ,max old old , = − t C Michaelis-Menthen equation ∂ α + C old for decomposition (1) C f t t ( , , , V ) = α old 0 C C ,max Analytical law z C t 0 , h C t = = for talik deepening (2) t t t Combining (1) and (2) yields ⎛ 2 ⎞ 2 ( 2 2 ) C C ,0 2 1 2 C h z − ( ) = + λ − + λ + γ − ⎜ ⎟ old C C C t t ⎝ ⎠

Methane transfer in the water body (Bastviken et al., 2002) • dissolved gases : CH CH [ ] [ ] ∂ ∂ o methane ∂ 4 4 k V , = − o oxygen CH oxid t z z 4 ∂ ∂ ∂ o carbon dioxide O O [ ] [ ] ∂ ∂ ∂ • processes: 2 2 k 2 V , = − O oxid o turbulent diffusion t z z 2 ∂ ∂ ∂ o methane oxidation [ CO ] [ CO ] ∂ ∂ ∂ 2 2 k 2 V = + CO oxid CH 4 + O 2 = CO 2 + 2H 2 O t z z 2 ∂ ∂ ∂

One-dimensional k- ε model (LAKE) Heat equation r r T T 1 S 1 U E a S ∂ ∂ ∂ ∂ ⎛ ⎞ k u n Tdl ( ) = − + ∫ ⋅ ⎜ ⎟ T E s t z z c z A ∂ ∂ ∂ ρ ∂ ⎝ ⎠ p Г H,LE A Momentum equations u u Snow ∂ ∂ ∂ 2 2 k fv g tg C u u v , = + − ⋅ α − + M x veg t z z ∂ ∂ ∂ Ice v v ∂ ∂ ∂ 2 2 k fu g tg C v u v = − − ⋅ α − + Water M y veg t z z ∂ ∂ ∂ K- ε turbulence closure 2 E k C , = M e ε Soil E k E ⎛ ⎞ ∂ ∂ ∂ M P B , = ν + + + − ε ⎜ ⎟ t z z ∂ ∂ σ ∂ ⎝ ⎠ E k ⎛ ⎞ ∂ε ∂ ∂ε ε M c P c B c ( ) = ν + + + − ε ⎜ ⎟ 1 3 2 ε ε ε t z z E ∂ ∂ σ ∂ ⎝ ⎠ ε

Validation: sediments temperature • Krasnoe Lake, (near S.-Petersburg) • 1969 – 1979 • Sortavala station forcing Bottom temperature Bottom sediments temperature (3 m depth) Observations: Kusmenko, 1976. Soil heat conductance: Cote and Konrad model (Sen Lu et al., 2007)

Case study: Lake Shuchi • Time series of atmospheric variables as input to lake model are extracted from ERA-Interim reanalysis 1000 Control experiment êî í òðî ëüí û é ¡ýêñï åðèì åí ò ýêñï åðèì åí ò ¡ï ðè ¡ï î ñòî ÿí í î ì ¡ 900 Experiment with constant Ebullition methane flux, mg/(m 2 *day) àòì î ñô åðí î ì ¡ä àâëåí èè atmospheric pressure 800 90% of bubbles 700 during wintertime 600 are intercepted by ice cover 500 and are released 400 when the later thaws 300 200 100 0 420 480 540 600 660 720 780 840 900 Time, days

Model calibration: Lake Shuchi The measure 2 2 2 ( w w ) ( s s ) F F F F F Δ ≡ − + − of model error a a m , a a m , • Calibrated 0,82 Δ F parameters 0,80 55,20 0,78 * P , P 53,80 51,80 old ,0 new ,0 0,76 49,80 47,80 10 , ì î ëü/(êã*ñ) 45,80 0,74 43,80 F Δ • has 41,80 0,72 39,80 single 37,80 35,80 0,70 33,80 minimum 31,80 0,68 29,80 * *10 27,80 0,66 25,80 23,80 P old,0 21,80 0,64 19,80 17,80 0,62 15,80 13,80 11,80 0,60 9,800 0,58 200 210 220 230 240 250 260 270 280 290 300 310 2 F 10 mg m / Δ ≅ 10 , ¡ì î ëü/(ì 3 *ñ) P new,0 *10 min

Model validation Observations : Lake Shuchi (K. Walter et al., 2006) hourly observations of ebullition and diffusion methane fluxes in different lake sections for 2003 – 2004 Annual methane A part of open- A part of ice- emission, mg/ water period covered period (m 2 *yr) ¡ emission , % ¡ emission, % ¡ Observations ¡ 22658 ¡ 54 ¡ 46 ¡ Model ¡ 22588 ¡ 54 ¡ 46 ¡ Open water period ¡ Ice-covered period ¡ 47 ¡ 6 ¡ A part of young methane in emissions (observations) , % ¡ 61 ¡ 32 ¡ A part of young methane in net generation (model) , % ¡

Remarks on lake methane model • The values of calibrated parameters depend on errors (lack of observations!) of input parameters: lake depth, water turbidity, atmospheric forcing, etc. • The model should be verified on a significant number of thermokarst lakes • The model does not consider thermokarst lake development (deepening, drainage, etc.)

Regional atmospheric model NH3D_MPI Land surface model of Atmospheric INM: 3D dynamics in 1. Soil (including permafrost) σ -coordinates, 2. Vegetation methane 3. Snow cover transport and 4. Walter and Heimann methane chemistry model for bogs 5. A set of models for oxic soils carbon cycling • horizontal spacing 1-10 km • 30 levels in vertical LAKE model with • time step 5-10 s methane block • parallel implementation using MPI

Research perspectives • Inverse modeling of atmospheric methane transport on a regional scale to (i) identify sources (ii) validate and calibrate land surface methane models (bogs and lakes) using measurements of atmospheric methane concentration • Incorporation of lake methane model in regional and global climate models to assess regional feedback between climate change and thermokarst lakes and its global significance