Earth system model of INM RAS Volodin E.M., Galin V.Ya., Diansly N.A., Gusev A.V., Smyshlyaev S.P., Yakovlev N.G. Institute of Numerical Mathematics RAS

Model includes 2 main blocks: atmosphere and ocean. Atmosphere: spatial resolution is 5х4, 2х1.5 or 1.25х1 degrees in longitude and latitude. In vertical 21 levels (uppermost level at 30 km) or 39 levels (uppermost level at 90 km), or 80 levels (uppermost level at 90 km). Time step is 3-12 minutes. Equations are solved by finite difference method. Ocean: spatial resolution is 1x0.5 degrees in longitude and latitude and 40 levels in vertical. Time step is 2 hours. Equations are solved by finite difference method at spherical coordinates with shifted poles. Model includes sea ice dynamics and thermodynamics block. Exchange between atmosphere and ocean occurs at each time step of ocean model, without flux adjustment.

The model can include also: 1. Chemistry block that can take into account 74 spices and about 150 reactions of chlorine, bromine, nitrogenous, oxigenous, hydrogenous, sulfuric cycles. 2. Carbon cycle (calculation of carbon of plants, soil, ocean and atmosphere) 3. Methane cycle (emission from wetlands, atmospheric concentration). 4. Dynamical vegetation. 5. Parameterization of electric phenomena (flashes, ionospheric potential)

The model code is written as 3 independent tasks: Atmospheric dynamics, oceanic dynamics and atmospheric chemistry, that exchanges data through hard disk. The model with atmospheric resolution 2x1.5L21 and oceanic resolution 1x0.5L40 that is used for CMIP5, runs for 6 years in 1 day at 56 cores of Intel Xeon 2.66 GHz.

Annual mean sea surface temperature error

Sea ice compactness in March (left) and September (right) in the model (top) and observations (bottom)

RMS of SST in El-Nino region for observations (top) and model (bottom)

Annual cycle of total ozone (model)

Annual cycle of total ozone (TOMS)

Model NPP (umol/(m2 s))

Estimation of observed NPP

OCEAN CARBON (10-3 MOL/M3) AT 3000 M. MODEL. OBSERVATIONS

Methane flux in the model (g/(m2 year))

Model flash climatology fl/(km2 year) and amount of flashes cloud - surface

Model diurnal cycle of ionospheric potential, KV

Observed ionospheric potential

DJF Near-surface air temperature minus for model (top) and NCEP (bottom)

JJA Near-surface air temperature minus for model (top) and NCEP (bottom)

CARBON EMISSION DUE TO FUEL BURNING (A1B) СО 2 CARBON EMISSION DUE TO LAND USE

The percent of absorbed carbon, year Model atmosphere land ocean F HadCM3LC 72(49) 5(30) 24(20) 1.47 IPSL-CM2C 47(40) 22(30) 32(30) 1.17 NCAR-CSM1 54(52) 25(26) 21(22) 1.04 MPI 54(46) 22(30) 24(24) 1.17 LLNL 41(36) 44(49) 15(15) 1.14 FRCGC 63(60) 10(10) 27(30) 1.05 UMD 64(55) 1(6) 35(39) 1.16 UVic (48) 17(28) 23(26) 1.23 CLIMBER 58(52) 22(27) 20(21) 1.12 BERN-CC 48(42) 26(32) 26(26) 1.14 INM 58(49) 20(31) 22(20) 1.19 AVER. 56±8(48±7) 19±11(27±11) 25±5(25±6) 1.17±0.11

Near-surface daily maximum temperature change in June-August induced by doubling of CO 2 in control run (top), run with decrease evaporation from plants (middle) and the difference (bottom)

LPJ DGVM coupled to INMCM. Results Change in boreal forest fractional cover from 1860 to 2100

DJFM temperature change induced by geoengineering with stratospheric aerosols

Relative change of annual mean precipitation

Air temperature change

Total ozone change

Change of NPP