The young scientist school and conference CITES-2009, Krasnoyarsk, 2009 V. .M M . . Stepanenko Stepanenko, , E E. .E E. . Machul Machul’ ’skaya skaya V Moscow State University Moscow State University THE DEVELOPMENT THE DEVELOPMENT OF LAKE MODELS OF LAKE MODELS FOR W EATHER AND FOR W EATHER AND CLI MATE STUDI ES CLI MATE STUDI ES The w ork is supported by RFBR grants NN 0 9 -0 5 -0 0 3 7 9 -a and 0 7 -0 5 -0 0 2 0 0 -a
Outline Outline • The overview of lake models • Lake Model Intercomparison Project (LakeMIP) project: goals and design • The Lake Sparkling intercomparison results • Observation evidences for the role of lakes in methane budget of the climate system • The methane model for thermokarst lakes • Initial results of model’s evaluation • Other issues in land surface hydrology modeling
Num erical w ater reservoir m odels for Num erical w ater reservoir m odels for coupled lake – – atm osphere studies atm osphere studies coupled lake 1 ) 3 1 ) 3 - - dim ensional dim ensional (~ oceanic) (~ oceanic) 2 ) 2 2 - - dim ensional dim ensional 2 ) • vertically averaged (Shlychkov, 2008) • averaged in one lateral direction ( CE-QUAL x.x model) 3 ) 1 1 - - dim ensional dim ensional 3 ) • single-colum n ( GOTM model (Burchard et al. ), LAKE model (Stepanenko & Lykosov, 2005); • laterally averaged m odels (Vasiliev et al. , 2007) – applicable in many applications ½ - 4 ) ½ - dim ensional dim ensional – the vertical profiles of temperature, 4 ) salinity etc. are parameterized ( FLake model, Mironov et al. , 2008) – high computational efficiency → application in operational weather forecast 5 ) 0 5 ) 0 – – dim ensional dim ensional (mixed models, e.g. Goyette model)
Model FLake FLake Model ( Mironov, Golosov, Kirillin Mironov, Golosov, Kirillin et al. ) ( et al. ) Theoretical basis : the concept of self-similarity of temperature profile . Advantages: • computationally efficient (orders of magnitude less than k- ε ) • capable of representing surface temperature and heat fluxes with accuracy of k- ε models Shortcom ings: • the bottom temperature • does not capture the details of real temperature profile ( e.g. Brunt-V ä is ä la frequency) • the maximal lake depth is limited to 50-60 m
ε model (LAKE) - ε One- -dimensional dimensional k k- model (LAKE) One Heat equation ∂ ∂ ∂ ∂ r r T T 1 S 1 U E a S ( ) ∫ = − + ⋅ k u n Tdl ∂ ∂ ∂ ρ ∂ T E s t z z c z A Г p H,LE A Mom entum equations ∂ ∂ ∂ u u Snow = + − ⋅ α − + 2 2 k fv g tg C u u v , ∂ ∂ ∂ M x veg t z z Ice ∂ ∂ ∂ v v = − − ⋅ α − + 2 2 k fu g tg C v u v ∂ ∂ ∂ Water M y veg t z z K- ε turbulence closure 2 E = , k C ε M e Soil ∂ ∂ ∂ E k E = ν + + + − ε M , P B ∂ ∂ σ ∂ t z z E ∂ε ∂ ∂ε ε k ( ) = ν + + + − ε M c P c B c ε ε ε ∂ ∂ σ ∂ 1 3 2 t z z E ε
Workshop "Parameterization of Workshop "Parameterization of Lakes in Numerical Weather Lakes in Numerical Weather Prediction and Climate modeling“ “ , , Prediction and Climate modeling St - - Petersburg, September, 2008 Petersburg, September, 2008 St Topics • lake parameterizations for weather and climate models; • the global databases on lakes; • data assimilation issues; • … Topics • modeling the lake level changes (paleoclimate tasks); • CO 2 and CH 4 emissions by lakes; • lake ecosystems under the climate changes; • …
LakeMI P project LakeMI P project (Lake Model Intercomparison Project) (Lake Model Intercomparison Project) http: / / www.unige.ch/ climate/ lakemip/ index.html http: / / www.unige.ch/ climate/ lakemip/ index.html V. Stepanenko, S. Goyette, A. Martynov, M. Perroud, V. Stepanenko, S. Goyette, A. Martynov, M. Perroud, X. Fang, D. Mironov, K. J ö ö hnk X. Fang, D. Mironov, K. J hnk Objectives: • Assessment of the range of applicability of existing one- dimensional model formulations, i.e. their capabilities and limitations in reproducing lake-atmosphere interactions, as well as internal lake thermodynamics. • Evaluation of the interaction mechanisms between lakes and the atmosphere in weather and climate models for weather prediction and climate projections. I m plem entation phases: 1) LakeMIP1, the intercomparison of one-dimensional models, using observations on a number of lakes representing a wide range of climate and lake mixing regimes. 2) LakeMIP2, will aim at studying the impacts of lakes on regional-scale weather and climate using coupled lake- atmosphere m odels.
Lake sites and models Lake sites and models Latitude Deep/ Example Average shallow depth, m Geneva ( Sw itzerland, France) Equatorial deep 153 shallow Am erican Great Lakes ( USA, Canada) Mid-latitude deep 19-147 shallow Sparkling Lake ( USA, W isconsin) non- 20 freezing Mid-latitude deep freezing shallow Arctic deep shallow Toolik ( USA, Alaska) 7 Kossenblatter ( Germ any) Very 2 shallow High- altitude • LAKE ( k- ε m odel) • FLake ( 1 / 2 – dim ensional) Models: • Sim strat ( k- ε m odel) • Hostetler • LAKEoneD ( k- ε m odel) • MI NLAKE9 6
Lake Sparkling intercom parison Lake Sparkling intercom parison Experim ental setup: Mean m onthly • 2002 – 2005 yr m eteorological tem perature profiles forcing • unified initial conditions, optical characteristics and other input parameters (e.g. bathymetry) Surface tem perature Δ T ≈ 2-3 Celsius R = 0.99
Methane em ission from therm okarst lakes Methane em ission from therm okarst lakes (Semiletov, 2005; K. Walter et al. et al. , 2007) , 2007) (Semiletov, 2005; K. Walter Therm okarst lakes are abundant in Northern Siberia ( 2 2 -4 8 % of land surface) , tending to expand in w arm ing clim ate Methane turbulent and ebullition flux from lake talik Unfrozen “hotspot” – wintertime methane source • 8 - 5 0 % of anthropogenic m ethane em issions up to 2 0 0 0 depending on the em ission scenario
INM RAS soil model with the Walter&Heimann methane model against BOREAS data 1994 1996 Zero methane fluxes when the top soil layers are not saturated by water. Methane model is highly sensitive to the soil moisture content, predicted by the soil model. To simulate realistic soil moisture the correct soil parameters should be set.
Methane m odel for therm okarst lakes Methane m odel for therm okarst lakes Bastviken et al., 2 002 I n w ater colum n [ ] [ ] ∂ ∂ ∂ CH CH = − 4 4 , k V ∂ ∂ ∂ CH oxid t z 4 z [ ] [ ] ∂ ∂ ∂ O O = − 2 2 k 2 V ∂ ∂ ∂ O oxid t z 2 z [ ] [ ] CH O ( ) = 4 2 V V T a [ ] [ ] + + oxid oxid ,max CH a O oxidation ( Arah & Stephen, 199 8 ) CH 4 O 2 4 2 I n soil ( talik) [ ] [ ] ∂ ∂ ∂ CH CH = + − − − 4 4 k P V V E W alter & Heim ann, ∂ ∂ ∂ CH , m oxid plant t z z 4 ) ( ) 1 9 9 6 , 2 0 0 1 ( ) ( ) ( ⋅ − = 0.1 T T P R f z f t Q m f T production 0 org 10 in ( ) ( ) [ [ ] ] [ ] f T ( ) - step function = − ebullition E k f CH CH CH e 4 4 4 max ≈ ≈ V 0 V 0 - oxidation plant-m ediated transport oxid plant
Bottom sedim ents tem perature Bottom sedim ents tem perature • Krasnoe Lake, ( near S.-Petersburg) • 1 9 6 9 – 1 9 7 9 • Sortavala station forcing Bottom tem perature Bottom sedim ents tem perature ( 3 m depth) Observations: Kusm enko, 1 9 7 6 . Soil heat conductance: Cote and Konrad m odel ( Sen Lu et al., 2 0 0 7 )
Methane ebullition flux Methane ebullition flux at Lake Shuchi at Lake Shuchi Sim ulations: Observations • atmospheric forcing – station (K. Walter et al. , 2007): Kamenskoe, 1990-2000; • Lake Shuchi, 2003-2004, step • 1 hour; depth = 3 m depth = 10 m 900 • different sections of the lake; 800 • background and point-source 2 *day) 700 fluxes. center 400 Ebullition flux, mg/(m 600 non-thermokarst thermokarst margin 500 400 300 Background 2 *day) 300 flux CH 4 flux, mg/(m 200 200 100 0 1110 1140 1170 1200 1230 1260 1290 1320 1350 1380 1410 1440 100 Time, days Modeling production term ( labile organic m atter) 0 is crucial for sim ulating 0 30 60 90 120 150 180 210 240 270 300 330 360 390 Time, days (28.04.2003 - 30.06.2004) correct fluxes
Further m ethane m odel Further m ethane m odel developm ent developm ent • NPP modeling in lakes • estimates of labile organic matter in permafrost • inclusion the CO 2 equation in water column • parameterization of gas fluxes through unfrozen hotspots
Other issues Other issues • the global database on depth and lake optics (probably, other parameters) for land hydrology is needed; • the rivers representation in GCMs might require more sophisticated models then simple runoff schemes; • coupling the hydrological model (lakes + rivers + soil hydrology) to land carbon cycle model
ANY QUESTI ONS? ANY QUESTI ONS?
Recommend
More recommend
