ICON Physics: General Overview

Martin Köhler and ICON team

ICON physics

• The standard Reynolds decomposition and averaging, leads to co-variances that need “closure” or “parametrization”.

• Radiation absorbed, scattered and emitted by molecules, aerosols and cloud droplets plays an important role in the atmosphere and needs parametrization.

• Cloud microphysical processes need “parametrization”.

• Parametrization schemes express the effect of sub-grid processes on resolved variables.• Model variables are U,V,T,q, (l,i,r,s,a)

What is parametrization and why is it needed

1 hour100 hours 0.01 hour

microscaleturbulence

• Diffusive transport in the atmosphere is dominated by turbulence.• Time scale of turbulence varies from seconds to half hour.• Length scale varies from mm for dissipative eddies to 100 m for transporting eddies.• The largest eddies are the most efficient ones for transport.

spectral gap

diurnal cycle

cyclones

data: 1957

Space and Time Scales

courtesy to Anton Beljaars

Space and time scales

Parametrized processes

courtesy to Anton Beljaars

dt

d

z

w

y

v

x

u

gwz

p

z

ww

y

wv

x

wu

t

w

vy

pfu

z

vw

y

vv

x

vu

t

v

ux

pfv

z

uw

y

uv

x

uu

t

u

1

1

1

1

2

2

2

Basic equations

mom.equ.’s

continuity

Reynolds decomposition

'.,'

',' ,'

pPp

wWwvVvuUu

o

Substitute, apply averaging operator, Boussinesq approximation (density in buoyancy terms only) and hydrostatic approximation (vertical acceleration << buoyancy).

Averaging (overbar) is over grid box,i.e. sub-grid turbulent motion is averaged out.

UUuUu 'Property of averaging operator:

1 ' '

1 ' '

o

o

U U U U P u wU V W fV

t x y z x z

V V V V P v wU V W fU

t x y z y z

Reynolds equations

Boundary layer approximation(horizontal scales >> vertical scales), e.g. :

High Reynolds number approximation (molecular diffusion << turbulent transports), e.g.:

z

wu

x

uu

''''

z

wuU

''2

Reynolds Stress

Shear production Turbulenttransport

Buoyancy

Mean flow TKE advection

Turbulent Kinetic Energy equation

2 2 2' 1/ 2( ' ' ' )E u v w local TKE:

Derive equation for E by combining equations of total velocity components and mean velocity components:

Dissipation

Storage

)'''(2/1 222 wvuE mean TKE:

Pressure correlation

' ' ' ' ' ' ' ' ' '

o

E E E EU V W

t x y z

U V g p wE w u w v w w

z z z z

Simple closures

Mass-flux method:

z

UKwu

''

K-diffusion method:

2

2

' 'u w UK K U

z z z z

Uuuz

UuMwu

upup

up

)(''

analogy tomolecular diffusion

mass flux (needs M closure)

entraining plume model

Process Authors Scheme Origin

Radiation

Mlawer et al. (1997)Barker et al. (2002)

RRTM (later with McICA & McSI) ECHAM6/IFS

Ritter and Geleyn (1992) δ two-stream GME/COSMO

Non-orographicgravity wave drag

Scinocca (2003)Orr, Bechtold et al. (2010)

wave dissipation at critical level IFS

Sub-grid scaleorographic drag

Lott and Miller (1997) blocking, GWD IFS

Cloud coverDoms and Schättler (2004) sub-grid diagnostic GME/COSMO

Köhler et al. (new development) diagnostic (later prognostic) PDF ICON

MicrophysicsDoms and Schättler (2004)

Seiffert (2010)prognostic: water vapor, cloud water, cloud ice, rain and snow

GME/COSMO

ConvectionBechthold et al. (2008) mass-flux shallow and deep IFS

Plant, Craig (2008) stochastic based on Kain-Fritsch LMU, Munich

Turbulent transfer

Raschendorfer (2001) prognostic TKE COSMO

Mironov, Mayuskava (new) prognostic TKE and scalar var. ECHAM6

Neggers, Köhler, Beljaars (2010) EDMF-DUALM IFS

Land

Heise and Schrodin (2002), Helmert, Mironov (2008, lake)

tiled TERRA + FLAKE + multi-layer snow

GME/COSMO

Raddatz, Knorr, Schnur JSBACH ECHAM6

Physics in ICON

ICON dynamics-physics cycling

Slow PhysicsSlow Physics

Non-Orographic Gravity Wave

Drag

Non-Orographic Gravity Wave

Drag

Sub-Grid-Scale Orographic DragSub-Grid-Scale

Orographic Drag

Land/Lake/Sea-Ice

Land/Lake/Sea-Ice

dtime

dt_gwd

dt_sso

dt_conv

dt_rad

dtime * iadv_rcf

dt_conv

Fast PhysicsFast Physics

OutputOutput

Tracer AdvectionTracer Advection

DynamicsDynamics

Turbulent DiffusionTurbulent Diffusion

MicrophysicsMicrophysics

Satur. Adjustment

Satur. Adjustment

Satur. Adjustment

Satur. Adjustment

RadiationRadiation

Cloud CoverCloud Cover

ConvectionConvection

„dt_output“

Tendencies

T-tendencies due to solar radiation scheme

[K/day]

Jan. 2012

T-tendencies due to terrestrial radiation scheme

[K/day]

Jan. 2012

T-tendencies due to turbulence scheme

Jan. 2012

[K/day]

T-tendencies due to convection scheme

[K/day]

Jan. 2012

T-tendencies due to SSO+GWD schemes

[K/day]

Jan. 2012

T-tendencies due to microphysics / sat.adj. scheme

[K/day]

Jan. 2012

microphysics saturation adjustment

Jan. 2012

JSBACH Land Surface ModelSchnur, Knurr, Raddatz, MPI Hamburg

JSBACH is the land surface parametrization within the ECHAM physics in the MPI Earth System Model.

Physical processes: Energy and moisture balance at the surface (implicit coupling within vertical

diffusion scheme of atmosphere) 5-layer soil temperatures and hydrology Snow, glaciers Hydrologic discharge (coupling to ocean)

Bio-geochemical processes: Vegetation characteristics represented by Plant Functional Types Phenology Photosynthesis Carbon cycle Nutrient limitation (nitrogen and phosphorus cycles) Dynamic vegetation Land use change

EDMF-DUALM turbulence scheme in ICON

Goals:

turbulence option to ICON that is scientifically and operationally appealing

reference for default TKE scheme

reserach (HeRZ and HD(CP)2)

potential for climate

Martin Köhler and ICON team

DUALM concept: multiple updrafts with flexible area partitioning

N

iuiuiui

upup waw1

)(´´ A

preVOCA: VOCALS at Oct 2006 – Low Cloud

Daniel Klocke‘s Jülich 100m ICON LES run: qc+qi

GCSS process: GEWEX Cloud System Study (1994-2010)

Randall et al, 2003

extra slides

Maike Ahlgrimm: CALIPSO trade cumulus

Tiedtke DUALM

call tree EDMF

3 parcel updrafts (test, sub-cloud, cloud)

mass-flux closure

z0 calculation

exchange coefficients

call TERRA to get land fluxes

ocean cold skin, warm layer description

TOFD, drag from 5m-5km orography

EDMF solvers for qt/T, u/v, tracer (e.g. aerosol)

multiple diagnostics including T2m, gusts

JSBACH in ICONSchnur, Knurr, Raddatz, MPI Hamburg

New development of unified JSBACH code that works with the ICON and ECHAM6 (MPI-ESM1) models.

Has its own svn repository (https://svn.zmaw.de/svn/jsbach) and is pulled into the ICON code on svn checkout/update via svn:externals property

Self-contained model; ICON code itself only contains calls to JSBACH for initialization and surface updating at each time step (src/atm_phy_echam/mo_surface.f90)

Currently, only the physical processes have been implemented in the new JSBACH code; bio-geochemical process to be ported to new code in the coming months

New structures for memory and sub-surface types (tiles) that allow a more flexible handling of surface characteristics and processes: PFTs, bare soil, lakes, glaciers, wetlands, forest management, urban surfaces, etc.

ICON physics upgrades and tunings 2013 Aug-Dec

• Non-orographic gravity wave tuning • Marine surface latent heat flux in TKE scheme - rat_sea • Land surface physics

• Exponential roots• Moisture dependent heat conductivity

• Cloud cover scheme • Tiedtke/Bechtold convection parameters• Bechtold diurnal cycle upgrade • Horizontal diffusion • new TURBDIFF code

non-orographic gravity wave tuninglaunch amplitude x10-3Pa

IFS analysis

URAP observation July 1992(Kristina Fröhlich)

3.75, default

U bias

non-orographic gravity wave drag tuning

1.0

launch amplitude x10-3Pa

3.02.5

2.0

U bias

In TERRA plant roots are a sink constant to a depth dependent on vegetation type.

Now: the uptake of moisture is described exponentially as a function of depth.

The default setting soil level 1-4 are moister than the IFS soil and the levels below 5-8 are dryer after 10 days simulation in July. The new formulation exactly counter acts those IFS/ICON differences with 1-4 becoming dryer and 5-8 becoming moister. So more moisture is left lower down and more is taken out near the top of the soil.

ICON: exponential roots

ICON: moisture dependent soil heat conductivity

Moisture dependent formulation based on Johansen (1975) as described in Peters-Lidard et al (1998, JAS).

The impact is most prominent in the Sahara, which has virtually no soil moisture, because the previous constant formulation was tuned to moist soils. The cooling in the Sahara in the top most soil level and a warming in the lowest dynamic soil level after 24 hours at 00UTC is shown. This night-time near-surface cooling is a signal of a larger diurnal cycle resulting from a smaller ground heat flux..

default level 2 moisture dependency level 2

default level 7 moisture dependency level 7