Recent activities on aerosols in TM5 Achim Strunk Twan van Noije, Michiel van Weele AeroCom-2 contribution by KNMI Preliminary source sink analysis Online.

Recent activities on aerosols in TM5

Achim StrunkTwan van Noije, Michiel van Weele

• AeroCom-2 contribution by KNMI

• Preliminary source sink analysis

• Online dust module

• EC-Earth/TM5 decadal runs

TM-Meeting Wageningen 2011-11-29

AeroCom-2: Overview

• Aerosol Comparisons between Observations and Models• “The AEROCOM-project is an open international initiative of

scientists interested in the advancement of the understanding of the global aerosol and its impact on climate.”

• AeroCom-1: accomplished before IPCC-AR4• AeroCom-2: Ongoing, next stimulus is IPCC-AR5; contribution by

many groups/models (CAM*, ECHAM*, FMI-SALSA, GeosChem, GISS*,

GLOMAP, GMI, GOCART, HadGEM, LSCE, OsloCTM, SPRINTARS, TM5, …) to various experiments (hindcast, direct/indirect forcing, …)

• TM5 contribution to AeroCom-1 by Maarten & Frank• TM5 contribution to AeroCom-2 by Twan & Achim

– HCA-IPCC (years 2000-2009), including target year 2006

– pre-industrial simulation (1850)

– TM5 chem V3 with a few new features

– AR5 (RCP4.5) emissions with natural/biogenic/soil/oceanic from MACCity


AeroCom-2: Diagnostics

• 0d (@ stations): – hourly means of full aerosol tracer information (2006)– daily means of full aerosol tracer information (1850, HCA)

• 1d (@ stations), vertical profiles: – daily means of of full aerosol tracer information, dry aerosol extinction &

absorption PM10 (2006)

• 2d:– daily means of optical properties (1850, HCA):

• AOD @ {550,440,870}nm, fractions for {SO4,NO3,BC,OA,Dust,SS,H2O}• fine-mode AOD, absorption AOD, asymmetry parameter, dry aerosol

extinction / absorption PM10, ambient aerosol extinction

– daily means of {emissions, surface concentrations, wet/dry deposition, load, nitrogen budget} (1850, HCA)

– daily means of nitrogen budget (1850, 2006)– daily means of SW & VIS radiation fields (1850, 2006) Twan

• 3d: – monthly means of extinction, absorption, asymmetry param., etc. – monthly means of full aerosol tracer information, etc. (1850, 2006)


AeroCom-2

Participating in the following studies:• radiative forcing (Gunnar Myhre)• all-aerosol-tracer (Graham Mann)• organic aerosols (Kostas Tsigaridis)• snow & ice radiative forcing (Mark Flanner)• hindcast evaluation (Michael Schultz)• … • no active participation from KNMI foreseen

Most striking result by TM5:

Very low global mean AODs!


AOD


Revision 2

• Bugs discovered: – SO2 emissions high by a factor of 2 (SO4 correct)

– SeaSalt emissions low by a factor of 2 (mass)

– NH3 emissions 20% low

re-run, re-submit to AeroCom database

before after


AOD against observations

bias: -0.111


AOD fine mode against observations

bias: -0.093


Absorption AOD against observations


Possible reasons

1. Sources for aerosols too low (emissions)

2. Sinks for aerosols too high (deposition, sedimentation) lifetime!

3. Ageing and coagulation too fast in M7 (too few coarse particles)

4. Hygroscopic growth underestimated

5. Optics wrong

6. Combinations of 1. - 5.

7. …?


Possible reasons





5. Optics wrong


7. …?


Possible reasons





5. Optics wrong


7. …?


Check of the optics code

• Originally developed by Michael Kahnert (SMHI)• Lookup-tables for {extinction, single-scattering-albedo, asymmetry-parameter}

via– effective refractive index for each log-normal mode in M7

– count mean radius of mode (cmr)

– wavelength (wl)

• Update by Maarten/Joost replacing (cmr & wl) by (cmr/wl) and production of a new lookup-table


Lookup-Table Revision

• Extinction coefficients – as a function of mode-radius

– here: for 3 wavelengths and charact. refractive indices of Dust and SO4

– dashed lines: V1; solid lines: V2



• Originally developed by Michael Kahnert (SMHI)• Lookup-tables for {extinction, single-scattering-albedo, asymmetry-parameter}

via– effective refractive index for each log-normal mode in M7

– count mean radius of mode (cmr)

– wavelength (wl)

• Update by Maarten Krol replacing (cmr & wl) by (cmr/wl) and production of a new lookup-table

• effective refractive index retrieved by mixing rules for internally mixed modes, crucial input:– refractive indices and densities of SO4/BC/POM/Dust/SeaSalt/H2O

SO4 SeaSalt BC Dust POM H2O

refractive index

OPAC* OPAC* OPAC* ECHAM-HAM ECHAM-HAMSegelstein (1981)

densities[kg/m3]

1841 2165 2000 2650 2000 1000

*) Hess et al, 1998



• Test(s): – update of densities, esp. carbon compounds, both in M7 and optics

– update of optical properties of aerosol components

• Results:– increase of max. 10% for global mean AODs (July 2007)

– change of AAOD not checked (BC)

• Higher sensitivity of AODs to mean particle radii, due to dependence of particle cross section and mass on rg

– redistribution of particle mass and numbers!


Possible reasons




4. Hygroscopic growth too low

5. Optics wrong


7. …?


Global mean burdens

• Comparison with “products” that capture observed AOD frequency distribution:– TM5 (year 2006)

– GOCART (year 2006)

– HadGEM2 (year 2006)

– OsloCTM2 (year 2006)

– AeroCom-1 (year 2000)

• Importance ranking (partial AOD):– SO4 (20-40%)

– SeaSalt (15-40%)

– Dust (15-30%)

– POM (10-30%)

– BC ( 0- 5%)

• Main problem: No unique model reference available…


Example: SO4 zonal means

sconc

emi wetdep

load


Example (cont.): SO4 characteristic times

• lower sink terms in GOCART / OsloCTM• higher chemical production also possible for GOCART / OsloCTM

[d] TM5 AeroCom1 GOCART OsloCTM

burden / emissions

120 110 162 252

burden / deposition

4.63 4.34 5.68 7.90


Summary

SO4: • Surface concentrations good (NH-midlat.), loads on the low side• Rather high sink terms (wetdep), esp. in the tropics

SeaSalt:• Despite doubling the emitted mass, the load is still low• Rather high sink terms (wetdep)

Dust:• Loads on the low side• Rather high sink term

POM:• Loads on the low side (SOA!)• Rather high sink terms (wetdep), esp. in the tropics and the NH

BC: • Surface concentrations good, loads on the low side• Rather high sink terms (wetdep), esp. in the tropics


Con(cl/f)usion

• Low loads cause low AODs, but presumably the masses can NOT explain the AODs being low by ~50%...

• More likely due to wrong size distributions, so are they correct – in M7 (water uptake / coagulation)??

– in the emissions (radii & masses specified)??

• Size distributions interact with removal processes (deposition, sedimentation).

• Very low fine mode AOD another hint on wrong particle size distribution.

• M7 nucleation/condensation is a very critical process, as seen by the effect of reducing SO2 emissions (H2SO4 precursor).


Outlook

• Principle check of removal processes– previous problems known (aan de Brugh, 2010)

– first analysis already done for wetdep

– adaptation of more sophisticated approaches planned (e.g., ECHAM-HAM; Croft et al., ACP, 2010; )

– especially: • wet removal of number and mass individually• usage of current mode radii instead of predefined ones• take into account cloud type for in-cloud scavenging• …

– …

• Check of emission routines and underlying assumptions desirable as well


Dust emissions: Introduction

• So far: module by Elisabetta Vignati (presented June 2008)– Offline production of daily emissions using ECMWF meteo fields

• Algorithm is based on Tegen et al. (JGR, 2002) – definition of possible dust uptake areas:

• calculation of the distribution of potential vegetation types (biomes)• non-forest biomes are considered as potential source

– usage of normalised difference vegetation index (NDVI) to derive global fraction of absorbed photosynthetically active radation (FPAR)

– “shrinking” of available uptake area in case of • grass dominated vegetation (FPAR!)• shrub dominated vegetation (FPAR!)• snow cover

– supressing dust uptake when soil moisture exceeds a certain threshold

– vertical particle flux from soil properties (size distribution, alpha) and surface wind stress u* (derived from 10m wind speed)


Dust emissions: Offline Implementation

1. Parameters not dependent on time and grid cell o Properties of soil size distribution based on soil population and texture

2. Parameters calculated on daily basiso Snow cover from snow deptho Areas of emissions using potential sources, vegetation, cultivation, lsm,

fpar (and moisture)

3. Parameters calculated on 3 hourly basiso Surface stress from wind speed fluxeso Mapping from emission bins (24) to TM5 modes (2: insoluble

accumulation and coarse)

from presentation by Elisabetta


Dust emissions: Input Parameters

• Soil type and size distribution• Preferential dust source areas (paleolakes)• Fraction of absorbed Photosynthetically Active Radiation (FPAR)• Cultivation• Surface roughness length• Vegetation type• Snow depth• Surface wind speed• soil moisture

from Ina TegenECMWF fields



Index Vegetation type H/L veg

1 Crops, Mixed Farming L

2 Short Grass L

3 Evergreen Needleleaf Trees H

4 Deciduous Needleleaf Trees H

5 Deciduous Broadleaf Trees H

6 Evergreen Broadleaf Trees H

7 Tall Grass L

8 Desert -

9 Tundra L

10 Irrigated Crops L

11 Semidesert L

12 Ice Caps and Glaciers -

13 Bogs and Marshes L

14 Inland Water -

15 Ocean -

16 Evergreen Shrubs L

17 Deciduous Shrubs L

18 Mixed Forest/woodland H

19 Interrupted Forest H

20 Water and Land Mixtures L

Using Soil ph

Dust emissions: Vegetation types



Dust emissions: Input Parameters


Dust emissions: Results for 2000

Tegen versus AEROCOM/Ginoux

0

200

400

600

800

1000

1200

1400

1600

1800

OP

EN

OC

EA

N

CA

NA

DA

US

A

LA

TIN

AM

ER

ICA

AF

RIC

A

OE

CD

EU

RO

PE

EA

ST

ER

N E

UR

OP

E

CIS

(FO

RM

ER

US

S

MID

DL

E E

AS

T

IND

IA R

EG

ION

CH

INA

RE

GIO

N

EA

ST

AS

IA

OC

EA

NIA

JA

PA

N

ISL

AN

DS

OC

EA

N

AN

TA

RC

TIC

A

WO

RL

DTegen/TM5 year 2000

ginoux



However,

(revised) online version desirable/needed

for EC-Earth/TM5


Online dust emissions: What’s new?

• Adaptation to TM5-chem-V3 • Input parameter now in one single NetCDF file• Reformulation of the uplift area as a function of FPAR• Lowered hard constraint of LSM == 100 to LSM > 99• Removed uninitialised parameters, speed optimisation, OpenMP• Identical offline & online version available


Online dust emissions: Evaluation

AeroCom-1 median

GISS HadGEM SPRINTARS OsloCTM(year 2006)

MPIHAMv2(year 2006)

dust emissions [Tg/yr]

1126 1110 2972 2596 1420 792

Source: http://aerocom.met.no/


Online dust emissions: Evaluation

Tegen versus AEROCOM/Ginoux

0

200

400

600

800

1000

1200

1400

1600

1800

OP

EN

OC

EA

N

CA

NA

DA

US

A

LA

TIN

AM

ER

ICA

AF

RIC

A

OE

CD

EU

RO

PE

EA

ST

ER

N E

UR

OP

E

CIS

(FO

RM

ER

US

S

MID

DL

E E

AS

T

IND

IA R

EG

ION

CH

INA

RE

GIO

N

EA

ST

AS

IA

OC

EA

NIA

JA

PA

N

ISL

AN

DS

OC

EA

N

AN

TA

RC

TIC

A

WO

RL

D

Tegen/TM5 year 2000

ginoux


Outlook

• Implementation of soil moisture• Evaluation of load, AOD etc. compared to usage of Ginoux (median

particle size!)• Replacement of input fields by EC-Earth parameters


Decadal runs with EC-Earth/TM5

2000-2009/10 and 2025-2035• offline and online simulations with TM5• offline: small pre-processor developed to use EC-Earth CMIP5

output for TM5• online: CMIP5 restart files for EC-Earth/TM5• base case: ERA-Interim driven TM5 simulations 2000-2009• first results for ERA-I (full period) vs. EC-Earth (2000-2002)


EC-Earth/TM5: SO4


EC-Earth/TM5: Decadal runs

2000-2009/10 and 2025-2035, (2060-70)• offline and online simulations with TM5 and EC-Earth/TM5, resp.• offline: small pre-processor developed to use EC-Earth CMIP5

output for TM5• online: CMIP5 restart files for coupled EC-Earth/TM5• base case: ERA-Interim driven TM5 simulations 2000-2009• first results for ERA-I (full period) vs. EC-Earth (2000-2002)

Outlook• Evaluation of inter-annual variability, extreme events, etc. by

comparison with observations (current climate)• Assessment of climate impact on (aerosol) chemistry by comparing

the two (three) time slice experiments

Recent activities on aerosols in TM5 Achim Strunk Twan van Noije, Michiel van Weele AeroCom-2 contribution by KNMI Preliminary source sink analysis Online.

Documents

maccitytmmeeting wageningen

aodtmmeeting wageningen

aerosol tracer information

global aerosol

ambient aerosol

absorption aod

hourly means

monthly means of extinction