JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 103, NO. D5, PAGES 5679-5694, MARCH 20, 1998
A dry deposition
parameterization
for sulfuroxides
in a chemistry and general circulation model
Laurens Ganzeveld, Jos Lelieveld, and Geert-Jan Roelofs
Institutefor Marine and AtmosphericResearchUtrecht (IMAU), Utrecht, Netherlands
Abstract. A dry depositionscheme,originallydevelopedto calculatethe deposition
velocitiesfor the trace gases03, NO2, NO, and HNO3 in the chemistryand general
circulationEuropean Centre Hamburg Model (ECHAM), is extendedto sulfur dioxide
(SO2)andsulfate(SO42-).
In orderto reducesomeof theshortcomings
of theprevious
model versiona local surfaceroughnessand a more realisticleaf area index (LAI),
derived from a high-resolutionecosystemdatabaseare introduced.The current model
calculatesthe depositionvelocitiesfrom the aerodynamicresistance,a quasi-laminary
boundarylayer resistanceand a surfaceresistanceof the surfacecover, e.g., snow/ice,bare
soil, vegetation,wetted surfaces,and ocean. The SO2 depositionvelocity over vegetated
surfacesis calculatedas a function of the vegetationactivity,the canopywetness,turbulent
transportthroughthe canopyto the soil, and uptake by the soil. The soil resistanceis
explicitlycalculatedfrom the relative humidity and the soilp H, derived from a highresolutionglobal soilp H database.The snow/iceresistanceof SO2 is a function of
temperature.The SO2 depositionvelocityover the oceansis controlledby turbulence.The
sulfatedepositionvelocityis calculatedconsideringdiffusion,impaction,and
sedimentation.Over sea surfacesthe effect of bubble bursting,causingthe breakdownof
the quasi-laminaryboundarylayer, scavengingof the sulfate aerosolby sea spray,and
aerosolgrowth due to high local relative humiditiesare considered.An integrated sulfate
depositionvelocityis calculated,applyinga unimodal masssize distributionover land and
a bimodal masssize distributionover sea. The calculatedsulfate depositionvelocityis
about an order of magnitudelarger than that basedon a monodisperseaerosol,which is
often applied in chemistry-transportmodels. Incorporation of the new dry deposition
schemein the ECHAM model yieldssignificantrelative differences(up to -50%) in mass
flux densitiesand surfacelayer concentrationscomparedto those calculatedwith a simple,
constantdry depositionscheme.
1.
modelsis illustratedby resultsof previousstudiesof the global
troposphericsulfur budget. Roughly 50% of the emitted gas-
Introduction
Atmosphericsulfur has been studiedextensivelyin relation
to deleterioushealth effectsand the decline of ecosystemsdue
to anthropogenicemissions.The initial interestfocusedon the
impact of sulfur oxideson plant nutrition [Chamberlain,1980].
In the 1970sthe acidificationof ecosystemsthrough wet and
dry depositionof SO2 and sulfatewas widely investigated.In
the 1980sthe interest decreasedmainly becauseof successful
abatement strategiesof sulfur emissionsin western Europe,
Canada,and the United States[Erismanand Baldocchi,1994].
However, in recent years there has been a revival of the interest for atmosphericsulfur due to the stronglyincreasingemissionsin developingcountries,particularlyin Asia [Arndtet al.,
1997; Intergovernmental
Panel on Climate Change (IPCC),
1995] and the role of sulfateaerosolsin climatechange[Charlson et al., 1992;IPCC, 1995].
In this study we focus on the representationof the dry
depositionprocessof sulfur oxidesin a chemistryand general
circulationmodel (GCM). The ultimate aim of developing
such a coupled chemistryand GCM is to study interactions
betweenatmosphericchemistryand climate. The significance
of a realistic sulfur dry deposition representation in global
eoussulfur(-100 Tg S yr-•), mostlyin the form of sulfur
dioxide(SO2) and dimethylsulfide(DMS), is oxidizedto sulfate, while -10% is removedby wet deposition.The remaining
40% is removedthroughdry deposition.The sulfateformed is
mainly removed through wet deposition(-80%), while the
remaining 20% is removed through dry deposition [Langner
and Rodhe, 1991; Chin et al., 1996; Feichter et al., 1996; Kasi-
bhatla et al., 1997]. The importanceof dry deposition,in particular of SO2, underscoresthe need for an accurate description of this process.Langnerand Rodhe [1991] and Feichteret
al. [1996]usedconstantSO2andSO42-deposition
velocities
over the ocean, land, and snow/ice.Chin et al. [1996] and
Kasibhatlaet al. [1997] calculatethe SO2 depositionvelocity
using a parameterizationof the aerodynamicand surfaceresistancesfollowingWesely[1989].Kasibhatlaet al. [1997] pre-
scribea sulfatedeposition
velocityof 0.2cms- • whereasChin
et al. [1996]explicitlycalculatea SO•- surfaceresistance
fol-
Copyright1998 by the American GeophysicalUnion.
Paper number 97JD03077.
0148-0227/98/97JD-03077509.00
5679
lowing Weselyet al. [1985] and Hicks et al. [1989a]. However,
the validity of usingthe Wesely[1989] parameterizationin a
global-scalemodel may be questionedsince it was originally
developedfor regional scale air quality models. The surface
resistances
refer to 11 land usetypesand 5 seasonalcategories
representativeof the North American continent, not for surface coverssuchastropicalforestsand savannas[Wesely,1989].
5680
GANZEVELD
ET AL.: A DRY DEPOSITION
PARAMETERIZATION
FOR SULFUR
OXIDES
Moreover,the SO42-surfaceresistance
parameterization
by concentrations
of CH4, CO, NOy (NO, NO2, HNO3,HNO4,
Wesely[1985] and Hicks [1989a]is basedon observations
of NO3, and N205), OH, and 03 and hasbeen extendedwith a
sulfatedepositionovervegetations,
not for oceans,bare soil, sulfurcyclemodel [Feichteret al., 1996;Lelieveldet al., 1997].
and snow/ice.
In a previousstudywe presentedthe incorporationof a dry
depositionschemeinto the chemistrygeneralcirculationEuropeanCentre HamburgModel (ECHAM) [Ganzeveld
and
Lelieveld,1995]. The schemecalculatesdepositionvelocities
accordingto the "big-leaf"concept[e.g.,Hicks et al., 1987]
from the turbulenttransferand vegetationactivitycomputed
by the GCM supplemented
with representative
uptakerates
for soil, water, and snow/iceon a global scale.The further
developmentof this scheme,as presentedhere, aimsto improvethe descriptionof tracegasexchanges
betweenthe atmosphereand surface,consistent
with temporaland spatial
dependencies
of the model physicsand chemistry.We will
showthat the new schemeyieldssignificantrelativechangesin
mass flux densities and concentrationsin the atmospheric
boundarylayer comparedto the simplerapproaches
usedin
mostpreviousstudies.
2.
ECHAM Model and Chemistry Scheme
Oxidation of SO2 and DMS by OH and NO 3 and in-cloud
sulfateformation are included explicitly.Emissionsof DMS,
SO2[Spiroetal., 1992],NOx, CO, andCH4 are considered,
and
wet depositioncalculationsuse the ECHAM4 parameterization schemes
for large-scaleandconvective
clouds[Roelofs
and
LelieveM,1997]. The latter schemedistinguishes
betweenincloudand below-cloudscavenging
of aerosolparticlesand soluble gasesand the releaseof thesespeciesto the atmosphere
by cloud and precipitationevaporation.
3.
Surface Roughnessand Leaf Area Index
The removal of trace speciesby dry depositionis representedby the massflux densitycalculatedby
Fc - C•Vd
whereFc isthedownward
fluxof thespecies
(molecules
m-2
s-1), c• is its concentration
(molecules
m-3) at a reference
heightz, andl/• isthedeposition
velocity
(m s-1) at heightz.
The generalcirculationmodelECHAM hasbeendeveloped The time-integratedmassflux densityis hereafterreferredto
from the numericalweather predictionmodelsof the Euro- as "deposition."The dry depositionvelocityis relatedto specificcharacteristicsof surfacesand the atmosphericconditions
pean Center for Medium Range Weather Forecasts over these surfaces:
(ECMWF) [Roeckner
et al., 1996].In this studywe usedthe
T30 horizontal resolutionof ECHAM4 correspondingwith a
1
grid sizeof about3.75ø and a time stepof 30 min. The model
[/d=Ra
+ Rb+ Rsurf
has 19 verticallayersin a hybridrr-p coordinatesystem.Vorticity,divergence,
temperature,surfacepressure,
humidity,and whereR a is the aerodynamicresistanceto turbulenttransfer
boundarylayer;
cloudwater are prognosticvariables.The modelcontainspa- from a referenceheightz to a quasi-laminary
rameterizationsof radiation, cloud formation and precipita- transportthroughthislayeris largelycontrolledby molecular
by R•,, the quasi-laminary
boundarylayer
tion, convection, and horizontal and vertical diffusion. The diffusionexpressed
(QBR); Rsurfisthe combinedresistance
of all transseasonalcycleof sea surfacetemperatureis prescribedas a resistance
boundarycondition.Land surfaceprocesses
are described
by a fer pathwayswhichplay a role in the uptakeof tracegasesby
five-layerheat conductivitysoil model and by a hydrological the surface.For a more detaileddescriptionof the calculation
we refer to GanzeveMand LelieveM[1995].
model to determineevaporationand runoff. Over land each of the resistances
grid cell is subdivided
into four subcellsto distinguish
between ConcerningR a, the major differencefrom earlier work is the
for momentum
snowcoverage,bare soil, water in the skin reservoir(water useof a local(or vegetation)surfaceroughness
In the previousmodelversionwe usedthe ECHAM
storedwithin the canopyand on bare soil), and vegetation. (Z0mloc).
surfaceroughness
[DKRZ, 1992].The problemwith
Permanentice coveroverland is prescribedby a glaciermask. large-scale
Over land the roughnesslength is geographically
prescribed, thiswasthat extremelyhigh HNO 3 depositionvelocitieswere
while over ice-free sea it is calculated following Charnock calculatedover mountainousregionsbecauseof a smallR a
sincethe ECHAMz0m canbe aslargeas20 m in theseregions.
[1955][Deutsches
Klimarechenzentrum,
Modelbetreuungsgruppe
of the introductionof a localsurfaceroughness
(DKRZ), 1992, and referencestherein].Transportof water A disadvantage
with the GCM
vaporand trace gasesis describedby a semi-Lagrangian
ad- in the calculationof Ra is the lossof consistency
vection scheme. The main differences between ECHAM4
and
sincethe frictionvelocityand the stabilitycorrectionterm are
theversionusedin the previousstudy(ECHAM3), relevantfor still calculatedfrom the large-scalesurfaceroughness[Ganthis study,are the representationof the vegetationcoverand zeveldand Lelieveld,1995]. In order to removethis inconsisverticaltransport.In ECHAM4 the vegetationsurfacerough- tencywith the ECHAM frictionvelocityit is assumedthat the
nesslengthz0v, leaf area index (LAI), vegetationratio, and surfacedragin eachgrid squareconsistsof draggeneratedby
elements,
forest ratio are assignedto each grid cell on the basis of the orographyanddraggeneratedbylocalroughness
representation
of the major ecosystem
complexes
by Olsonet accordingto the drag partition theory of H. Schlichting[see
al. [1983][Claussen
et al., 1994].This classification
discerns
43 Claussen,1995]. A local drag coefficientfrom the reference
and a stabilitycorrection
ecosystems
andtheir characteristics
on a 0.5ø x 0.5øresolution heightz, the localsurfaceroughness,
[Olsonet al., 1983].Verticaltransportis calculatedon the basis term are calculated for the fraction coveredwith vegetation,
of the turbulent kinetic energy closure accountingfor the
transportof generatedturbulencethroughthe actionof turbulent diffusion,in contrastto the conventionaleddy diffusivity model usedin ECHAM3 [Brinkopand Roeckner,1993].
ECHAM4 is coupled to a chemistryschemedevelopedby
Roelofsand Lelieveld[1995, 1997].The schemecalculatesthe
bare soil, and snowof each grid square.The local drag coefficient is used to calculate a local friction velocitywhere we
assumethat the wind speedat the referenceheight(=30 m) is
unaffectedby localshearstress;that is, the referenceheightis
positionedabovethe blendingheightl•,. Also, the ECHAM
stabilityis not entirelyconsistentsincethe stabilitycorrection
GANZEVELD
term is a function
of the ECHAM
ET AL.: A DRY
Richardson
DEPOSITION
PARAMETERIZATION
number which is
calculatedfrom the wind and temperature gradient for the
large-scalesurfaceroughness.However, this inconsistency
has
been ignored since the sensibleheat flux, which reflects the
temperaturegradient,is mainly determinedby surfacecover
and not by surfaceroughness[Claussen,1995]. Hence the Richardsonnumber and the stability correction term will not
changesignificantlybecauseof a changein surfaceroughness
assumingthat the wind speedat the referenceheight is unaffected by the local surfaceroughness.
The localsurfaceroughness
for vegetationandbare soil and
snow is derived from assignedvalues of ZOr
n for 13 surface
types[Henderson-Sellers
et al., 1986] on the basisof the update
of the original Davenport classificationof roughnessby 14qeringa [1991] [Claussenet al., 1994]. The Olson et al. [1983]
ecosystemdatabaseis reduced to these 13 surface types to
estimatethe local surfaceroughnessof the ecosystems.
The
local surface roughnessfor each ECHAM grid cell is calculated from the 0.5ø x 0.5ølocal surfaceroughnessby averaging
the ecosystem
neutral drag coefficients[Claussenet al., 1994].
A major differencewith the previousmodel version,calculating the canopyresistancefrom the leaf resistanceand the
LAI [Ganzeveldand Lelieveld, 1995], is the use of a more
realisticLAI. The state of the canopyin ECHAM3 was not
expressed
by a variableLAI, but it had a constantvalue of 4 for
all vegetation types, independent of time and location; the
model merely resolveda seasonallydependentvegetationarea
fraction. Comparisonof the calculatedmonthly averagediurnal cycleof the depositionvelocityof 03 with the observations
indicatesthat this approachyieldsrealistic03 dry deposition
velocitiesover mostlocationsfor different seasons[Ganzeveld
andLelieveld,1995].However,for somelocations,e.g.,tropical
forests,the poor representationof the local amount of biomass,expressedby the LAI, causeddiscrepancies
betweenthe
observedand calculated03 depositionvelocities.Additional
explanationsfor suchdifferencescan be related to the parameterizationof the stomataluptake [Sellers,1986],whichdid not
accuratelysimulatethe very efficient stomataluptake by forests,and misrepresentationof the vegetationfraction at individual grid cells. In ECHAM4 a more realisticLAI is derived
for each grid cell from the arithmetic mean of a summer and
winter LAI for each Olsonet al. [1983] ecosystemcategory.
Thesewinter and summerLAI valuesare determinedby allocatingOlsonet al. [1983] ecosystemcategoriesto H. Lieth and
G. Esser'svegetationtypesand their assignedLAIs [Claussen
et al., 1994]. In the near future an annual cycle in the LAI
based on a biogeochemicalmodel will be incorporated in
ECHAM4 (M. Claussen,personal communication, 1996).
However, for this study we have tentatively incorporated a
parameterizationof the annualcyclein the LAI usinga clipped
FOR
SULFUR
OXIDES
5681
not yield a proportional increase in uptake becauseof the
extinctionof radiationand turbulencewithin the canopy.However, the canopyresistancein ECHAM4 is calculatedusinga
relationshipwhichaccountsfor the extinctionof the photosynthetically active radiation (PAR) within the canopy [Sellers,
1985, 1986]. The impact of turbulence extinction has been
ignored sincethe quasi-laminaryboundarylayer resistanceof
the leavesis usuallysmallerthan the stomatalresistance[Sellers, 1985].
Improvement of the canopy descriptionhas consequences
for soil uptake processessincethe LAI determinesthe partitioning of deposition between the soil and vegetation. In
sparselyvegetatedlocationsthe dry depositionvelocity is not
only controlledby the vegetationactivitybut alsoby transport
through the canopyto the underlyingsoil and the destruction
rate at the soil surface.For trace gaseswhich are removedvery
efficientlyby the soil,e.g.,SO2,to a wet soilwith relativelyhigh
p H (see section4), turbulent transferto the soil surfacecan
becomethe controllingfactor. To accountfor this process,we
have incorporateda parameterizationof the aerodynamicresistancewithin the canopyas a functionof canopyheight,LAI,
the friction velocityu *, and an empiricalcoefficient[Erisman
and Van Pul, 1994].The resultsof this parameterizationare in
reasonableagreementwith those of Wesely[1989] [Erisman
and Van Pul, 1994]. The canopy height is not defined in
ECHAM4. However, the grid average canopy height can be
estimatedfrom the ECHAM4 forest ratio cF, which is specified from the forestdescriptionby Olsonet al. [1983] [Claussen
et al., 1994].It is requiredthat cF _<vegetationratio (c •), and
correctionsare applied for different ecosystemcategoriesto
accountfor the canopyspacingof different forest types. For
example, for broad-leaved evergreen, cv is 0.95 whereas for
savannacv is 0.4 [Claussenet al., 1994]. For consistency
with
ECHAM4
we have used this forest ratio to determine
an av-
eragecanopyheightfor eachgrid cell by multiplyingcv with an
assumedforest canopyheight of 20 m.
4.
Sulfur
Dioxide
Surface
Resistances
Table 1 liststhe valuesadoptedfor the cuticleresistancercut,
the mesophyllresistancer .... and the water resistancerwat and
the parametersused for the calculationof the snow/iceresistancer .... /iceand the soil resistancer•o• for SO2.The cuticle
resistancercut of SO2 by far exceedsthe canopystomatalresistance[Baldocchi,1993,and referencestherein],expressed
by
a largercut,whereasthe mesophyllresistancerm½
• is negligibly
low. A distinction
has been made between
the water resistance
rwat and the wet skin resistanceof SO2,rw•, sinceSO2 deposition to wetted surfacesover land is controlled by both the
aerodynamic and the surface resistancein contrast to SO2
sine function to account for differences between the winter and
depositionto the oceanswhich is solelycontrolledby turbulent
summerLAIs. The length of the growth seasonis a function of transfer and diffusion.The surfaceresistanceof oceansis very
latitude, being 3 monthsat latitudeshigher than 60øNand S, 6 small because of the high SO2 solubility associatedwith an
monthsfor a latitude of -45øN and S, and up to 12 monthsfor oceanwaterpH of-8 [Spedcling,
1972]. Recent regional-scale
latitudes below 30øN and S. Plates la and b show the simulated
depositionmodels distinguishbetween foliage wetnesscaused
Januaryand July LAIs for a 0.5ø x 0.5ø horizontal grid reso- by rain and by dew to account for their different chemical
lution. A maximumLAI of -10 occursin the tropicalforestsof compositions[Wesely,1989].In our schemethe effect of foliage
SouthAmerica, centralAfrica, and southeastAsia throughout wetnesson SO2 dry depositiondue to rain or dew hasnot been
the year whereasthere is a distinctannual cyclein the LAI of treated separatelysincefoliagewetnessin ECHAM4 is the net
nonconiferousforestsat higher latitudes,with maximum dif- result of both processes.The model resultsindicate that the
ferencesbetween the January and July LAIs of -6. Linear monthlyaveragewet skinfractionexceeds0.25 only for a small
upscalingof the leaf resistanceto the canopy scaleusing the area, suggesting
that SO2uptake throughthis mechanismdoes
LAI breaks down for an LAI >-3-4
since an LAI increase will
not contributesignificantlyto dry depositionon a global scale.
5682
GANZEVELD
ET AL.'
A DRY
DEPOSITION
PARAMETERIZATION
Table 1. Soil, Cuticle, Mesophyll,Water/Wet Skin, and
SULFUR
OXIDES
lated from the ECHAM4 surfacetemperature,applyinga relationshipderivedfrom observations
of SO2 uptake by snow
[MuhlbaierDasch and Cadle, 1986], yielding a SO2 snow/ice
Snow/Ice
Resistance
(s m-1) for SO2
Resistances
FOR
Values
deposition
velocity
(for anassumed
Rail + Rb of 125s m-l),
/'soil
f(soilpH, rh (2 m))
whichincreasesexponentiallyfrom a minimumvalue of --•0.01
rcut
105
Fme
s
0
cms-1 for a temperature
of -20øCto a valueof ---0.25cms-1
rwat/ws
rsnow/ice
0/100
max(10, f(r(surface)))
The abbreviations
are definedasfollows:
rso,•,
soilresistance;
rh,
relative humidity;rcut,cuticle resistance;r .... mesophyllresistance;
rwat/ws
, water andwet skinresistance;
rsnow/ice
, snow/iceresistance;
and
T, temperature.
However, local and transient uptake processesmight be affected by foliage wetness.Also, the SO2 snow/iceresistance
dependson the chemicalcompositionof the snow/icesurface
(i.e., oxidantconcentrations
andionicstrengths)andthe physical characteristics
(liquid-likecharacteristics
andsurfacearea)
of the surfacelayer [Conklinet al., 1993].A generalaspectof
studiesthat considerthe uptake of SO2 and the chemicaland
physicalcharacteristics
of snow/ice[Cadleet al., 1985;Muhlbaier Dasch and Cadle, 1986; Granat and Johansson, 1983;
Sommerfeldand Lamb, 1986; Valdez et al., 1987; Clapsaddle
and Lamb, 1989;Conklinet al., 1993]is the strongdependence
of the snow/iceSO2 uptake on temperature,which manifests
itself particularlyin the temperaturerange -10ø-0øC [Mitra et
al., 1990]. In our model the SO2 snow/iceresistanceis calcu-
m for 0øC.SO2depositionvelocitiescalculatedwith a physicalchemical model which considersgaseousdiffusion into the
snowpack,air-waterpartitioning,and aqueous-phase
reactions
[Baleset al., 1987] also showsan exponentialincreasewith
temperature
from ---0.01cms-1 for -20øC, althoughthe in-
crease
isslightly
less(i.e.,a Vdso2
of '-'0.15cms-1 for0øC).
The soilresistanceof SO2is a functionof the soilpH andthe
relative humidity [Baldocchi,1993, and referencestherein].
Biologicalactivitydoesnot greatlyaffect SO2 uptake as indicatedby a small decreaseof uptake of SO2 due to soil sterilization [Murphyand Sigmon,1989].A parameterizationof the
SO2 soil resistance, being a function of soil p H and the
ECHAM4 relativehumidityat 2 m altitude,hasbeenincorporated in the model. The soilp H is derived from a 0.5ø x 0.5ø
global soilp H database,shownin Plate 2, which discernsfive
different soilp H classesaveragingover the top 30 cm of the
soil [Batjes, 1995]. The distribution of soil p H reflects the
distributionof vegetation;the mostacidicsoilsoccurin densely
vegetatedregionswhereasalkalinesoilsare found in sparsely
or nonvegetatedregions.For each model cell the fraction of
the five soil p H classesis determined. The soil resistanceis
calculated from these fractions and the soil resistances for each
Aerosol dry deposition velocity and sulfate mass distr.
lOO
6o
5o
lO
40 .-•
30
'•'
0
o
20
o.1
10
O.Ol
o
O.Ol
o.1
1
lO
radius [um]
Vd for u=10, u*=0.5.
Aerosol mass size distributions: .......
mr. cont. SO4 ....
marine SO4
Figure1. Particledrydeposition
velocity(cm s- 1) as a functionof particleradius(pm) for a wind speedof
10m s-1 anda frictionvelocityof 0.5 m s- 1. Also shown are the observed mass size distributions of sulfate
(nanoequivalents
m-3) associated
withruralcontinental
andmarineaerosols
(databyMehlmann
[1986]and
adaptedfrom Warneck[1988]).
•
GANZEVELD
ET AL.:
A DRY
DEPOSITION
PARAMETERIZATION
FOR
SULFUR
OXIDES
5683
January LAI derivedfrom ecosystemdistr.acc. to Olson et al. [1983]
60N
30N
o.
EQ
30S
60S
180W
t50W
120W
90W
60W
30W
0
30E
60E
90E
120E
150E
180E
60E
90E
120E
150E
180E
July LAI
60N
"' '" "
x•
"•lllr
q'''
,½
3ON
EQ
3OS
60S
180W
150W
120W
0.0
90W
60W
25
30W
0
50E
5.0
75
100
Plate 1. (a) Januaryand (b) Julyleaf area indices(LAIs) on a 0.5ø x 0.5øhorizontalgrid resolution,derived
from the Olsonet al. [1983] ecosystemdatabase.
class, based on observationsof the SO2 soil resistance for
different soilp H and relative humidity by Payrissatand Beilke
[1975]. A correctionof the calculatedsoil resistance,derived
from the observations
by Payrissatand Beilke [1975], is applied
for a calculatedrelative humidity below 60% at 2 m. Implementationof this parameterizationyieldsdepositionvelocities
aslargeas0.8cms-• in desertareas,e.g.,theSahara,because
of a soilpH of 8. However, for very dry conditionsand chemicallyinert sandysoils,SO2uptakecanbe expectedto be small.
Observationsof SO2 uptake by grassand soil by Milne et al.
[1979] showedthat bare soil,which had a thin coveringof red
quartz gravel, exhibitedvery little affinity for SO2. Sensitivity
study indicatesthat a relative humidity of 40% (2 m) is a
reasonablethresholdfor arid regionswith sandysoils[Olsonet
al., 1983]. Despite the lack of experimentalsupportto apply
this thresholdfor all soilsthe rsoilof SO2 for arid regions is
calculated from the soil resistancefor a relative humidity of
40%, assuminga linear increase with a decreasingrelative
humidity to a maximum value of rsoi•arbitrarily chosento be
1000 s m -•.
The soil resistance is corrected for a surface
temperature <-2øC [Wesely,1989;Erismanand Wyers,1992].
5.
Sulfate Dry Deposition
Dry depositionof sulfateover vegetationis a functionof the
canopystructure[Bache,1979;Slinn, 1982; Wesely,1983] since
the depositionvelocityis controlledby the turbulent transferto
and throughthe canopyto the different receptor surfaces,e.g.,
the leafs, branches, tree trunks, and soil. However, we have
incorporated the straightforwardparameterization of sulfate
depositionvelocitiesfrom the stability and friction velocity by
Wesely[1985] sincethe degreeof detail of the canopydescrip-
5684
GANZEVELD
ET AL.: A DRY
DEPOSITION
PARAMETERIZATION
FOR
SULFUR
OXIDES
..
60N
•,•.
!
50N
EQ
30S
60S
-
180W
150W
120W
90W
60W
50W
;..:,....;....:...::,
......:..:% ,.,....;.:
pH <=5.5
0
. ..
50E
60E
90E
120E
150E
18(
..
5.5<pH<=7.3 7.3<pH<=8.5
8.5<pH
4<pH<=8.5
Plate 2. Global distributionof soilpH on a 0.5ø x 0.5øhorizontalgrid resolutiondistinguishing
five soilpH
classesof the top 30 cm of the soil:p H -< 5.5, 5.5 < p H _<7.3, 7.3 < p H -< 8.5, 8.5 < pH, 4 < pH _<8.5,
indicatedby the litmuscolors,exceptfor the fifth undefinedclass[Batjes,1995].
tion required for the mechanisticapproachby Bache [1979]
and Slinn [1982] is beyondthe capabilitiesof our GCM. For
other
surfaces the sulfate
surface
resistances
are assumed
to
this nss sulfate in sea-saltparticles is found in the 1-5 /am
range [Sieveringet al., 1992]. We have calculatedthe sulfate
deposition velocity for a mean mass radius of 0.37 /am
havea minimum
valueof 1 sm- l, sothatthesulfatedeposition (Vaso4(0.37))
andbyintegration
of themasssizedistribution
velocity is controlled by the aerodynamictransfer and diffu- (Vasoq(msd))
of theruralcontinental
andmarineaerosols
for
sion, impaction, and gravitationalsettling, dependent on the
particle radius.Figure 1 showsthe particle depositionvelocity
at a referenceheightof 30 m asa functionof particleradiusfor
different meteorologicalconditions,to studythe sensitivityof
Vaso4for the assumed
masssizedistribution.
We alsocalculatedthe marineVasoq(msd)
applying
the rural continental
a windspeedandfrictionvelocityof 10 and0.5m s-q, respec- masssizedistribution.Sincewe only considernsssulfatein this
tively.The depositionvelocityis calculatedusingthe parame- study,the actualintegrated
SO42-deposition
velocitydepends
terizationby Slinn [1976]for particledry depositionto smooth on the fraction of nsssulfatein the coarseparticle mode aswell
surfaces,
whichwe appliedto calculate
VdSO4
overbaresoil as the accumulation mode. However, this information is not
and snow/ice-covered
areas. Particles
in the accumulation
mode(0.1/am < r < 1/am) are removedwith a relativelysmall
directly available from the model since the applied marine
masssize distributionrepresentsboth the nsssulfate and sea-
deposition
velocityof---0.01cm s-• becauseof the diffusion saltsulfate.
Therefore
thecalculated
Vasoq(msd)
for therural
limitation whereas depositionvelocities of particles in the
continental and marine sulfate should be interpreted as the
coarse
particlerange(r > 1 /am)areup to 10cms-• because lowerand upperlimit, respectively,
of the integratedSO24
of the gravitationalsettling.Therefore it is necessaryto calculate the depositionvelocity from the sulfate masssize distri-
depositionvelocityas a functionof the fraction of nsssulfatein
the coarseparticle mode.
bution and not from a mean mass radius. Over land most of the
The depositionvelocityof marine sulfateis calculatedin this
sulfate is in the accumulation mode with a mean mass radius of
study using a parameterizationby Hummelshojet al. [1992].
0.37/am [Whitby,1978].Figure 1 alsoshowsobservedmasssize This is a modificationof the parameterizationfor particle dry
distributions
of sulfate associated with rural continental
and
deposition to "smooth" natural waters by Slinn and Slinn
marine aerosols(data by Mehlmann[1986] and adaptedfrom [1980], which accountsfor the effect of whitecapson the dry
Warneck[1988]). Integrationof the rural continentalmasssize deposition velocity through the destruction of the quasidistributionover the whole sizerange showsthat ---85% of the laminarylayer andwashout by spraydrops[Hummelshojet al.,
sulfate mass is in the accumulation mode. This is close to the
1992]. This is an obviousand much discussedphysicalexpla95% calculatedby Whitby[1978]on the basisof an analysisof nation of the possibleenhancementof the particle dry depocontinental sulfate observations at different locations. The masition velocity compared to that derived by Slinn and Slinn
rine sulfateexhibitsa bimodal size distributionwith a signifi- [1980], but, unfortunately,experimentalevidenceto support
cant fraction, ---35%, in the coarseparticle range. The major this is lacking.The effect of particle growthfor a large relative
sourceof sulfatein the coarseparticlerange is sea salt [War- humidity has been accountedfor by assumingthat the water
neck,1988]whereasa relativelysmallfraction originatesfrom vapor pressurein the quasi-laminarylayer is at saturationlevel.
the heterogeneousconversionof SO2 to non-sea-salt(nss) This is expressedby a maximum relative humidity which is
sulfate. Observationsin severalclean and anthropogenically limited over salt water to ---98% as a consequenceof Raoult's
influencedmarine regionsindicatethat ---30% of the nsssul- law [145Iliares,1982]. The wet particle radius, which is calcufate occursin the coarsemode [Andreae,1995].About 75% of lated accordingto Fitzgerald[1975], is about twice the dry
GANZEVELD
ET AL.: A DRY
DEPOSITION
PARAMETERIZATION
FOR
SULFUR
OXIDES
5685
VdS04 (msd) and VdS04 (0.37 um) vs wind speed
Marine
case
10
0.1
0.01
0
5
10
15
20
25
30
windspeed (30 m) [m s-1]
•
int. Vd over sea .......
Vd (r=0.37 um) over sea ....
Vd over sea for continental distrib.
Figure2. Particledrydeposition
velocity(cms-•) overseaasa functionof windspeed(m s-•), calculated
from a mean massradiusof 0.37 •m and by integration over the masssize distributionof sulfate associated
with rural continentaland marine aerosols(see Figure 1).
with vegparticleradius,with a consequentincreaseof the dry deposi- strongdifferencesin depositionvelocities,associated
tion velocity by an order of magnitude. Figure 2 showsthe etation activity,chemistry,and meteorology,are expected.The
velocitiesaccountfor the rural
integrated
marineSO42-deposition
velocityfor differentmass calculatedSO42-deposition
sizedistributions
andthemarineSO42-deposition
velocityfor continentaland marine sulfatemasssize distributionsas prea mean mass radius of 0.37 •m, calculated as a function of
sentedin Figure 1.
windspeed.
The Vaso4(0.37)
overtheoceanisaboutanorder
ofmagnitude
smaller
thantheVaso4(msd)
overthewholewind 6.1. Impact of the Introduction of Local Surface
Roughness and LAI
speed
range.Thecalculated
marineVaso4(msd)
forthemarine
mass size distribution is a factor of 5 larger than the
Our previousmodelversion,whichusedthe ECHAM3 sur-
Vaso4(msd)
for the ruralcontinental
aerosol
because
of the face roughnessand a constantLAI of 4, has been shown to
larger dry depositionvelocityof the sulfatemassin the coarse reproduceobserveddepositionvelocitiesover severalsurfaces
particlerange.TheVaso4(msd)
for theruralcontinental
mass [Ganzeveldand Lelieveld,1995].However, somediscrepancies
size size distribution, which is not shown here, is more than an
remainedwhich could partly be explainedby the limited rep-
orderofmagnitude
largercompared
to theVas04(0.37).
These resentationof the surfacecover and roughness.Observed diresults
confirmtheimportance
of calculating
theSO42-depo- urnal ozonedepositionvelocitiesover tropicalforestsare -2.5
the previous
modelversion
sition velocityon the basisof a masssize distributionrather cms-• [Fanet al., 1990]whereas
than a mean massradius sincethe latter method significantly calculated
a maximum
Vao•of 1.2cms-•. However,
thecurunderestimates
the dry removal.Moreover,the partitioningof rentmodelcalculates
a maximum
January
Vao3overthetropthe sulfatebetweenthe accumulationand coarseparticlemode ical forests of South America of -2.4 cm s-•. This increase can
largelybe explainedby the increasein LAI from 4 to 10. Our
mustbe accountedfor explicitly.
previousstudyindicateda large overestimationof the ozone
depositionvelocity during July in Alaska with a calculated
6.
Results
maximum,
average,
andminimumVao3of 1.0,0.7,and0.4cm
Thenewmodelversioncalculates
ozonedepWe present the effects of the model improvementsfrom s- • respectively.
incorporatingthe local surfaceroughnessand a more realistic ositionvelocitieswhich agreewell with observationsover tunLAI. We also showa comparisonbetweensimulatedand ob- dra [Jacobet al., 1992;Ritteret al., 1992].Calculatedmaximum,
serveddiurnallyandannually
varyingSO2andSO42-deposi- average,
andminimum
Vao3are0.45,0.15,and0.05cms-•,
tion velocities. Further, we show the global distribution of
respectively,compared to the observedmaximum, average,
V•s02andV•s04for the monthsJanuaryandJulyfor which andminimum
Vao3of 0.35,0.2,and0.1cms-•, respectively.
5686
GANZEVELD
ET AL.' A DRY DEPOSITION
PARAMETERIZATION
FOR SULFUR OXIDES
The previousmodelversionyieldedHNO3 depositionveloci-
ties in mountainous
regionsup to 10 cm s-• becauseof the
high ECHAM surfaceroughness
in theseregions(---20m).
The introductionof the localsurfaceroughness
yieldsa Janu-
aryaverage
V•n4NO3
of"-1cms-• compared
to5 cms-• inthe
previousmodel version.Despite the lack of observationsto
validate the model HNO 3 depositionvelocitiesover theselo-
++
+
cations we feel that the current model calculates more realistic
HNO 3 depositionvelocities.
6.2. DiurnalandAnnualCycles
of VdSO2SO4:
Comparison With Observations
It shouldbe emphasizedthat evaluationof the model by
comparingit with experimentalresultsis difficult becauseof
the large differencein the spatialresolutionof measuredand
model-deriveddepositionvelocities.The currently applied
ECHAM modelversion(T30) hasa spatialresolutionof 3.75ø
x 3.75ø (300-400 km at midlatitudes),
whilemeasureddepositionvelocitiesare site specific,with typicalspatialscalesof
severalkilometers(assuminghomogeneous
terrain). There-
foretheevaluation
of V•tso2/so4
mainlyfocuses
onqualitative
comparisons
of diurnalandseasonal
cycles,
whichare notvery
sensitiveto the specificsurfaceparametersbut which are to a
large extent controlledby turbulenceand irradiance(e.g.,
throughstomataluptake).
Table 2 showsthe calculatedmonthlyaveragediurnaldry
depositionvelocities,representedby the average,maximum,
andminimumV•tso2/so4
for severalgridcells,andavailable
observations
for SO2andSO24
- undercomparable
conditions
(season,
surface,andsurfacecover).Insteadof presenting
the
average,
maximum,
andminimum
Vdso•./so4
pergrid,weshow
thosecalculatedover the (four) subgridfractionsin line with
the surfacecoverpertainingto the observations.
For evaluation of depositionvelocitiesover vegetations
we haveconsidered both the vegetationdepositionvelocityand the wet skin
depositionvelocity,despitethe fact that the wet skin fraction
alsorepresentsthe wet bare soil fraction.The grid cellsare
selectedon the basisof the distributionof major ecosystems
according
to Olsonetal. [1983].TheSO2andSO42deposition
velocitiesoverthe AtlanticOcean,northof the United Kingdom, in February,have not been measureddirectlybut are
derivedbyPrahmetal. [1975]usingan air trajectorymodeland
aerosol measurements
00000
OC•O
000
0
at the Faroes Islands and the British
Isles.Comparison
ofthecalculated
andderived
average
V•ts%
showsthat the model simulatesthe very efficientremovalof
SO2 in winter over the North Atlantic Ocean with a maximum
SO2deposition
velocityof 3.4 cm s-•. On the otherhand,the
sulfate depositionvelocityseemsto be overestimatedin this
location.Calculating
the SO24
- deposition
velocityfrom the
ruralcontinental
masssizedistribution
yieldsa smalleraverage
SO42-deposition
velocityof --•0.7cms-•, whichis,however,
stilltoo largeby almosta factorof 2. Thereisgoodagreement
between
theobserved
andcalculated
SO42deposition
velocity
over snow/ice.For the comparisonwith the observedammonium sulfatedepositionvelocitiesover snowby Ibrahimet al.
[1983],theaverage
V•tso4
of thetwoexperiments
forthesubmicrometerparticleswith a diameterof 0.7/•m are presented.
Theaverage
SO42drydeposition
velocity
oversnowissignificantly smaller than the constantvalue of 0.2 cm s-• used
previously
by LangnerandRodhe[1991],Feichteret al. [1996]
and Kasibhatlaet al. [1997] over snow/ice-covered
surfaces.
The SO2depositionvelocityover snow/icein Arizona is overestimatedby about a factor of 2. This is likely due to differ-
o
¸
.,..•
<
GANZEVELD
ET AL.'
A DRY
DEPOSITION
PARAMETERIZATION
encesin temperature,which largely controlsthe deposition
velocity through the surfaceresistanceas a result of the difference
between
the local elevation
of the measurement
FOR
SULFUR
OXIDES
0.s
I
5687
I
I
I
site
and the mean surfacealtitude of the ECHAM model grid cell,
2200versus700 m, respectively.There is generallygoodagreement betweenthe calculatedand observedvegetationdeposi-
0.6
tionvelocities,
exceptfor the35øN,85øWgridwherethemodel
underestimates
the May averageSO42-deposition
velocity. z 0.4
However,this underestimationcanbe explainedby the differ-
ence
between
thetypical
surface
roughness
ofdeciduous
forest
and the derived local surfaceroughnessin the model, -1 m
[Dolman, 1986] and 0.1 m, respectively.For the comparison
betweenthe observedand calculatedSO2 depositionvelocities
over deciduousforest in March and April the grid square at
45øNand90øWhasbeenselected
sincethe LAI of thegridcell
resembles
that of the location
I
I
I
I
I
I
I
I
I
I
2
3
4
5
6
7
8
9
10
11
2
of the observations.
Figure 3a showsthe simulatedannual cycle of the monthly
averageSO2 depositionvelocity for 35øN, 85øW with an LAI
and local surfaceroughnessrepresentativeof deciduoustrees.
Theannualcycleof Vaso2hasalsobeeninferredbyMattand
Meyers [1993] using the dry depositioninferential measure-
O8
ment(DDIM)technique.
Theinferred
annual
cycle
of Vaso2
indicates
a minimum
valueof -0.2 cm s-• duringwinter,
increasing
to a maximumvalueof 0.6 cm s-1 in May, then
decreasing
againin summerto a valueof •--0.4cm s-• in
Augustand September,and further decreasingin October to a
•
-
valueof 0.2cms-• because
of leaffall. Our modelreproduces
this cyclereasonablywell except during winter for which our
O2
modelcalculates
a Vaso2
up to -0.5 cms-• because
of a
smallercalculatedSO2 soil resistancerelative to the constant
valueof 300sm-• in DDIM. Figure3b shows
the annualcycle
of Vaso:overGreatBritain,50øN,0øW,and the observed
I
I
I
I
I
I
I
I
I
I
2
3
4
5
6
7
8
9
10
11
12
summerand winter mean SO2 depositionvelocityof 1.0 and
0.8cms-•, respectively,
overa sitecovered
withroughpasture,
shortgrass,and ploughedsoil [Nicholsonand Davies,1988].
\
•2
/
\
The lack of a distinctannualcyclein the observationsindicates
/
\
that otherprocesses
besidesstomataluptakecontributeto the
/
SO: depositionprocess.For many observationsthe fetch contained sparsevegetationor exposedsoil, and direct uptake by
the soil with a p H of 8 could have been a major influenceon
o
Vaso:
[Nicholson
andDavies,
1988].
Observations
byDavies
and Mitchell [1983] at the samesite 2 years earlier yielded an
annualmeanSO2deposition
velocitywhichis --•0.2cm s•-•
smaller
compared
totheannual
meanVaso2
of0.9cms-• by
04
Nicholson
andDavies[1988].Thiswasexplained
bythediffer-
02
encesin the densityof the surfacecover[Nicholsonand Davies,
00
1988].Themodelreproduces
a smallannualcyclein Vdso:but
underpredictsthe SO2 depositionvelocities.This can be explained by the relatively large calculated aerodynamicresistance comparedto the observedaerodynamicresistance.The
model calculatesa relatively small surface resistanceof 25 s
I
I
I
I
I
I
I
I
I
I
2
3
4
5
6
7
8
9
10
11
Figure
Annual
cycle
ofcalculated
monthly
mean
Vdso,
(cms- 1•o
for
threegrid
cells
(solidlines)
andthe
estimated
annual cyclefrom an inferential model (Figure 3a) and ob-
m-1, whichis in goodagreement
with the observed
surface servedannual cycles(Figures 3b and 3c) over comparable
resistance.During winter the calculatedsurfaceresistanceis
largely controlledby soil uptake associatedwith a soilp H of
•8
and an LAI
of 0.5 whereas
the summer
surface resistance
is controlledby both the soil uptake and vegetation.The impact of canopywetnesson the SO2 depositionvelocityis more
clearly shownin Figure 3c, which presentsthe simulatedannual cycle of the monthly mean SO2 depositionvelocity for
45øN,0ø. The Januaryaveragewet skin fraction for this location is 0.6, muchhigherthan the Juneaveragewet skinfraction
of 0.1. Also shownare the SO2 depositionvelocitiesobserved
by Vermetten
et al. [1991],the wintertimeaveragebeingashigh
surface cover (dash-dottedlines): (a) deciduousforest in
southeasternUnited States,35øN, 85øW, (b) grass/baresoil,
England,50øN,0ø, and (c) coniferousforest,westernEurope,
45øN, 0o.
as1.3cms-1 witha maximum
of 2.0cms-1 overDouglasFir
in the Netherlandsduringthe autumn and winter whereasthe
Vdso:during
thesummer
wasmuchsmaller,
i.e.,0.2cms-1.
We havenot selectedthe grid cell of the latitude and longitude
of the measurementsite in the Netherlandssincethis grid has
5688
GANZEVELD
ET AL.:
A DRY
DEPOSITION
PARAMETERIZATION
FOR
SULFUR
OXIDES
donuory overoge Vd S02 [cm s-t]
•;
..... -:.•...•::•:.•
........'
•" ,•
..........................
.-•::::::>•............................
':•:::.-.:-'.".--:'•:':,,•:?"
..
'*':'•
.............
::½•:
•
.•
".•'"•
::':L.:
•:::':•;½•'•"
•-:--':';'•½'
•'L•• •';•%•" • •
"....:•::,::'
•..x:
:::::::::::::::::::::::::::
............
:
....... :::.: ;.•::i;".:;:':::'::
================================================
..... . :::;•:;:*•.,,"
.:. .•.' ....
*.-½,'½'"• •.'
.•:::::::.-,;;•
r::-½?'
............
-...............................
:•.::.½:•:•:...4.:.•::
.............
::•:•:'½:':':::"
':"ß.....
'-:
.... :;:-:-:-":
•'•:.•
:•7-"'•:";
-TM
-;'
:-:::.:;,
:-:-.:•-.
'"•".:.::3•'--•-•
":•
•::..:•....:.
;. ":"-:-;-;.':
•::::•::?:•:•:½•.;;:;:;-:•:•;:•:---:•
':"'":,"-•;½:.
'½......-.•.-•
-*'"
-";':..
..;
............
::::::::::::::::::::::::::::::::::::::::::"'........
'
":;::::n•½½'•.'-•::•::•
::-:-•.•..•:'.:..:::;.":-::?:•E::
.....
•..::':•:.::•::
........
-:• %..•$:::;:;;::t
.... -'.;:a-;:-z,•'"
•:L
•'.:•.:%;::'"'""--':"
'"'•:•::
3.-.:.
':'•:-,;•
......
."'%-'..
::}•'".••.•:•.;':•:?;:.
3os;;'.
...........
:.:-:::::--------:::::
.'.'.'.'.•..2•;;}'•;•:-.-:
---.:';.:'•:'
•:'"''";:•;::::':::??•
:•:'?:L."L-::/.(•-').•:
.......
.:.:-.---.-;::.:::/',-.
:"•:.•.½•)-'"½'%•;
.......
....:::.
.................................
•::.:..•::•:•......::....:...
-• --•:..-•.••!:.--.-:.:...::.
..............
. ...::::::-.....::.-.-.
....................
.•....•!.C.•.•.........::::•.:
.... ..•
::::::::-:---....:..:-::.....-•
'•::
...:::::•
."•;;::..•.:;:..---:'
:-:--,j½.::
. . .........
..
:.::.: ......
......
:..............
. ............
........
%•:::.-"::
;:..•-•;•:-.-•:.:
.½.
•,•.'}.'
(.::•-:;:;•.::::.
-..................................
..::::::.::....:
:.::..;.:'::::
. ........................
:::..:
....:-:
...................
..-- •
.-.-....•::::.....-.-.
...........
-..:::.•
• :-- :::.
.:,8•::
.•:::::::.•::•:....
..................
..,.....'-----::::;:::::,:...'::.:'
.•::::::::::::•::•
............
•;'
':/•'"•;:•½:;;:'":'"'½;;;;;:::•:::•.•
::-?:;::•.:.
.... '......
;;,:.
•:::'•.,•:,;F:;.
': ..........
i•'•'•-'•:•••
::½-:-'•-:½,,
•:•
.................
::,,,;:
%
.....
-':-:•:•:4•;•s
::• %•{;L"?x:;z•5•3{•
-:::::::::::::::::::::::::::::::::::::::::
;•.•;•
................
::::::::::::::::::::::::,,:;•:::•::•:.•::::,::..:,,•,•::•:•,...'""•-½•
•,:.
.....
*•
':•"*&•
::;
::•:•½•;•;•:•:•::'"•:'":'•::•;•
':'•,,•a?;?
...................
6os:.:;.
..:
:..:,a;:;
.......................
::::::::::::::::::::::::::::::::::::'"•---•-•.
.......
.:•:-':
•::•:•:•:•:-:-:::::'-':-'"'"•':':"•-:•
'.-•:':":':':":':::':;
.................
;;%:•:•'-2
.•
180W
150W
isolines
120W
are:
0.025
•
.
. .'
90W
' .''"':'"•e.::--.
"•:-.•.- ---.----•
60W
O. 1
0.2
30W
0.4
':':':
....
0
0.6
,
-
$OE
0.8
....
::::?
/:•::::•:•½•::":"
'•:;;"i'::,;:'";.•,•;:-:;;:;;:.•;:•
;'•'
'-•:---'
....
60E
•'.':.•::•::
90E
120E
150E
180E
1.5
July overogeVd S02 [cm s-1]
....
.......................
....
'.,....;:
.............
.......... .........L"i'.
.......
'...........
:".:..-•;;.;-:
'-'
'...-.-.:-•:::;.,.,.....::.2'
•':...:..'%.
'"'".::;:•::--'--',;'"'
','• -::-•-;.'•
-....•-•.
".',,,,•..~
..:::
• • ß ""..c:'
• -"•'-::"-'--::•:*•½::
...::•.•:::•'•:•:•:•'•*•:'%...•:•*:•:,-........
•......:.:..•-'•:'::'::'::•:;::•;:i•::•i::i•::::::•::::::-•::-::..If•
': .t'----r
.....
=%.
;•.:•
,A:.'..T•.
•':::::::::'
"":.•:.':::'"-"•
.............
•::":::*•::::::::ii::::::::•:•;•'•"":'•:---'•;
......."'-'•-:'$--.
½'.,,i11•,...•":::-.•
::•'. ':':"
.:•..;;7•:;i:i8::::::•73•.'":.',-•--•:•?'--•-';•
:;=:
:=F•½J:}.
".'"-•. '• •'-'"'="•!</•*%•:
•'=:==,-•"'•'::<•'•-•-,:.•c:=:=•.t
...•; ;*"
'""•.:.... -: L\x :•:;•:•
.....•:•'""•:::•
.........
•::::::::::':'•'":'
......'"'":'""11•'
';"•
....... i•:::".
............
::..•::'
...... ,,,,,/•:
.........
•*:":....
'•":'.:•:'"':-'"""':
...... ::::::::::::::::::::::
..................
-,½='=•*E,
•;:;a :a•:•
.......
[•. --x,...½:=
.........
•.••••.••"•:,•.:•,:.•.•`•.:•*•:"•.•:::•.;:•.•:..•:•:•:........•..:.•.
...........
-.::.>:....•:::::::.::,.:::•:-:::•;•,•.•.•;•:•.,,::,;,•.,,.,,.,•::,
........................
.•.•.::.•.•.•.•.,•.•::::..,•.:::.•..•.`•.::.:•..•``
:::::•..
.....
:•,.::.:::-•--•
.........
:.•....: '"•:
.• •½:--'
'-•--•'•t•.:.•.
:•- '--:-:-•
'•'•:.t3•'
.4:' ":/:;•i•?- :•X.'::'" .:"::•::•::•:•:-:-'-':...:::•:•::'..::.:•t:..*x'•:•
...•:•:::'":.:..::.'•
•r...-:-.::•½.----::-:.
,• ...........
•.- -:-,:-'-""-*:=*-=
=:=•:=
•
--:::::::::::::::::::::::::
.......•: • ..... -'-•%•%
;.-:.: ::::.:.:
::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
-%':
'....:•'
:!•: :::::::::::::::::::::::::
•::::B
s:::===========================================
..............
======================================
.........
•'-,,.
'::•. '-::-.,
.:-,:-:•;:(
........
•ali
'4•,
•. --•!:,:•*?
ii: .:.i½
•...............
**½:!::ii
:•:-'
"
30N ":"-
......
';'
:..-•?
:..
....
. "
I•
::.
.........
• '•:t•:•
•:'"'
.............
:•:"--:'•:':::•:'
....i,:::;:;•;::•::•:•"
•,•::•:.•:::•;':•:•:::::'.,-'""-':-•,,-'.'.'
....
.--:-:"-•-;--'-'•.
- '"'"'•':•••:.:..
•
::'<.•-
...:...-:--':'.
'. ....
.' :.::;::::::
t "':::-:-:':
.:.":'•::::.
- • ..
_--•.•...•.. :?:>:•--':'.::•:::!'...::•.-..:
-,:.•:?.
•;:
..................................
;;:..::•.?::
•"• ....................
'"":'
•t•:::..
.,:..-'::-'
.........
:...
' ' ::o¾..'.'.::'.:
:.........
.....-._•,.•
..........
.•,•
.....
............
-•. .....:'.
• "'..........:-.:'""'"'::::::.'..i!!?•::•::•::-:•::::.---.::<-:::?•::.
::' • :.
.:{:-:•,::
:•,.
:*:?:"".
"?•::::::::...'.t
........
•:•:;:;:;:.::;:•
.....
:.::.::.:"':•
:'"'::X'"'":.::.:::!::'.:::"'.;:.
..... ..,•/•'t:-..
.;½"
'::•
..,•=:
....•=:,'
½•:•
O.•
..t
....
•.'':::
.....
'•'•::.
""::'""
'-:".'::':-"'.,
---'
.......
.::::::•:•:
X?:---:::-•!::;::.-"":':'""•:;-:'
'"tl•.
........
:::-'"
..-"-:•:
"::................
:½
'•'•'•• :..':;
. , ' ø'
. ..,===========================
===
":
,:.
.........• •:
.:,q.
....
3os
• ':'':::"•
60S
•:::
••' "':'
:•
.... •::•:;:;•
.......
"':::?t•::'•:::'"'•::::::'"":"•
"/*:•::'
"'•:
.......%.-.•½":...--:.:.,::....
(...
ß :.-"
:",,:
.......
""*•;;'--'?.':
'•::•::.!½....::',:.-'../:..i•',•'"::-:•:•'
............
;.............
"':;•.'
:::.:
...........
•............
.....................
..........................
................
..........::.:.
.......................
.:....'::•:
:.:•...•:::....,
!•'"•-•
••••"•-,
......z._.
r
....... ......•_. _ .,.,J
180W
150W
isolines (]re:
120W
0.025
--:'
.....
,.--:::½.
.........
,.....
..............
....
.:...
.:.:..-::..:...'.;:?•
-. ,...;•..:..
-.. ..............
-..:•.:--•.-•:
.?::•::....:!::;•:•
,'•.',,
-:---:'",•
..
ti,'-::•:................
:::'.,
; ::•::
:::•
::ii
'?,
.•. :....?•;ii
..........................
,.....•
:'i..,'.•" .•F•'::::
":"
'::'"
:'½;:
•'•' ....
'":'":-'-•C•
,::\
•"'•"•"
.'
• ":"
.......--•-J•'
90W
0.1
......
60W
0.2
-,....
......:::::':':':'.:'•:.:•:,:.•.,._
•":'?::::...??'"'"'";:
",
..............
30W
0.4
0.6
•
0
-
30E
0.8
.....•.
•_
60E
-
''.;.':
"',
½;::.-.-.-;:::-:-::--:
:i:..!i?'"
.......
?!
............
":-•'.iiii;'
'!!'
'-
............................
90E
- .........................................................
-:- .........:.,:::
, .
....
. -•:......
;'.•.-,•-•---:
120E
150E
180E
1.5
Figure4. Monthlyaverage
SO2 deposition
velocity(cms-1): (a) Januaryand(b), July.
a winter LAI of only 0.25 comparedto the LAI of the Douglas
Fir of 10 [Vermettenet al., 1991].The selectedgrid squarehas
6.3.
Global Deposition and Concentration Calculations
6.3.1. Sulfur dioxide. Figures4a and 4b showthe January
a JanuaryaverageSO2 soilresistance
of _+75s m-• and an
globalV•so2distribution.
Distinctspatial
in-canopy
transferresistance
of _+125s m-•, sothatthe dep- andJulyaverage
gradientsoccurover the continents,mostlyrelated to variable
ositionvelocityis largelycontrolledby stomataluptake and the
whereasover sea the spatial distribuwet cuticle.The model doesnot simulatethe large deposition surfacecharacteristics,
tion
of
V•so•
reflects
the windspeeddistribution.
In July,
velocities during the winter and autumn, which can be exrelatively
large
values
of
V•so•
occur
in
dense
vegetated
areas,
plainedby the smallerLAI of --•3.5and the assumedwet skin
e.g.,
the
temperate
forests
in
the
northern
hemisphere
(NH).
resistance
of 100 s m-• comparedto the observed
canopy
resistance
of _+25s m-• for wet conditions
[Vermetten
et al., In Januarythe depositionvelocityin northwesternEurope is
1991]. However, the model seemsto reproducethe impact of stillrelativelylargebecauseof the canopywetnessthroughdew
the wet winter on the annual cycle of the SO2 deposition formationand precipitationinterception.Over the desertsof
velocity,characteristicof the northwesternEuropean climate, Africa andthe Middle Eastthe depositionvelocityisvery small
with wintertime depositionvelocitiescomparableto those in for a relative humidity as low as 10%. Large areaswith deposition velocities <0.025 cm s-• in winter occur over snow/icesummerdespitethe small vegetationactivityin the winter.
GANZEVELD
ET AL.: A DRY DEPOSITION
PARAMETERIZATION
FOR SULFUR OXIDES
5689
Rel. diff. [%] in SO2 conc. betweennew ond "constantVd" scheme, Jonuory
• •- ..•....
.....
...........................
•j ,.•:•:.•:•
• •....•::
........
•:.:.
:..::.:...:•:•:•::::•
.........
...::::•:::
...................................
..•.•,..•..,:•.
6ON
.•
•.-._:•.,
::::::::::::::::::::::::::::::::::::::::::::::::
............
•(•
.......
.•.:.•::.•=;j::.---:•;?::½•;.::•.•.•.:•:7.:;5;;;::•:•:::::
•:•..•>
"
'• t •:::.:::•:.:::•:::•.•:::•;•:•::::•:•::•::•::•::::::•:::::•..•:•:•::•::•:::•:•----:;"
Y'
. :.'- -"
"'•:':":'2•
• 4 %?• '"':5:'•::•:•:•t:J•:;::':::•;•
........
•t•/'"'.'.•-''
. • •:•:.•
...................
<::•?•::::•:::•:•.•.•:::;;;•;.•.
•ttt.
'.•,: ..........
ß
.......................
;::•
............
:•-"-•..
ß -..,
:?,:..
:•.;:.•;k•
• .....
::: :t--.•:•:.:•
..:.:.-:::::;...
5ON
EQ
.' ...•
::• -•*,
;.•. '•';•:'
•½•:.•:•:"•'?:•::•.
?"•:'•.,.•::•-'•:'•4•::.•--.....•:•:•
•
', - t;,,.•.-:•;:.•<::::.:............. •,
.......:::.:
....
...:
...................
:-•-'•::':::::.'•'."
•':•'""•:•:•:•::•½•,•-?;:'"•'
•...•..•
•...'Z7'-.-.::..-•-;;?
:•:. :.
,.:•.•'-:•:::•::::,.
............
%..
.........
:::::::::::::::::::::::::,•.•
ß ,. ;---.•::::::
..... :•:•:..•
::....
=================================================
•:: .........
:' ....-.
.... •'";-:::::":•
.................
•.•;;
..............
•:;::F"':::':.]..
':•....:•,;:':/• :.......
-?:':'":'::'-'
'•'•'•':':..":>:•...-'".•.:
' '--'"&
't•..
......
ß
.
.
':
•"•2:%
::.'•:----:---:
::•---:•--'?•:::
•:•:•:
::"-•..
.•
....•
...
::::::::::::::::::::::::::
....v-•'.•:-''
'•..•:.:;:•;:?t.
:::-:::.-•?i::•:•tSt•g•'..";;:
"•':
,4'•-.
::::::::::::::::::::::::
•:::•::•.....•':'t
::.::::.•)" .......
-"•
30S :.
:•;?•
•;•:•
'"•:•:
"•....•
'••..-•;':':"...
•:::::•'"'::•:-•::•:'•:½•v':•:.
•''::
::•'"'•A'•'•'•':•:'•:•
•'..-•"':.':-:":";:•
.......
'::::::•:.•;•;;•::&•
•.':::::.
•'•':";-•"•...----.-•'•
•,5:-;;:::•;;:•::•½.•
.... ,•:
•
'::':•:•:::--.;•:•:.•::•:•:•:::•::•-=-------•:..•';:
• '.....
•:•v/•..•- •:•::•:::•
.•
• ....• .........
•.
•:.:.'•::.:-...-•
.
.
..................
60S
;:•
:.•.........
...•(..??,
.....•,•,•-•
....................
.....................................
'•'•..:.:
..................
•••'•"••
....,••
.,•• •:'•::'""'-•:•"::•::•
'••••••-.-•
•:::::::..<...-•.;:.===============================================
-•...••
ß.
///•••:
....
•...•,...............
•::::::•
...............................
•
........
'•••••=__••,....-:..•
• ..•'.....
:•.•'.......................................
?....
180W
15OW
isolines
ore:
120W
-25
90W
- 10
60W
10
30W
0
25
50
'..............
..
30E
1 O0
60E
90E
...•:....•
120E
150E
180E
150
Figure5. Relativedifferences
(percent)of the SO2surfacelayerconcentrations
betweenthe newscheme
andtheconstant
Vas02
scheme,
January.
Dashed
andsolid
lines
indicate
adecrease
andincrease,
respectively,
of theconcentrations
calculated
bythenewscheme
compared
to theconstant
Vas02scheme.
covered
surfaces.
TheJulyVaso2
is --•0.15
cms-• overthe
Arctic snow/icesurfaces.In the southernhemisphere(SH) the
6.3.2. Sulfate. Figure 6a and 6b showthe Januaryand
JulyaverageSO•- deposition
velocities
by applyingthe rural
January
andJulyVaso2
values
arerelatively
smallinAustralia, continentaland marine masssize distributionshownin Figure
1. Overland,Vaso4
isrelatively
smallwithminimum
deposifor example,
Julyaverage
Vaso2
overthetropical
forestin South
America
is tionvelocitiesof--•0.05cms- • overevensurfaces,
with valuesof --•0.2cm s-• whereasthe averageJanuaryand
ashighas0.8cms-•. A strikingfeatureis thelargedifference the snow/iceareas and over the deserts.Over vegetated surfor example,the tropical
betweenthe JanuaryandJulyaverageVaso2oversouthern faceswith a largesurfaceroughness,
SO•- deposition
velocities
areashighas0.25cms-•.
Africa and southof the tropicalforestin SouthAmerica,0.6 forests,
and0.05cms-•, respectively.
Overthe oceansthe SO2depo- Over
sea,Vaso4
ranges
between
---0.1
cms-t inthesubsidence
andup to 1.0cms-• overtheNorthAtlanticin Jansitionvelocity
isrelatively
high,i.e.,up to 2.0cms-• overthe regions
uary.Figure7 showsthe relativedifferences
betweenthe JanuaryaverageSO•- concentrations
in the modelsurfacelayer
s-• in subsidence
(lowwindspeed)regions.
The SO2 depositioncalculatedwith the new schemehas calculated
withthenewscheme
andtheconstant
Vaso•scheme
North Atlantic in January,and it has a minimumof --•0.3cm
is0.2cms-• overallsurfaces
[see
Feichter
etal.,1996]).
beencompared
to a "constant
Vaso2scheme
(Vaso2is0.6cm (Vaso•
calculated
s-• overlandwithoutsnow/icecover,0.1 cm s-• oversnow/ice Over land the SO•- surfacelayer concentrations
surfaces,
and 0.8 cm s-• oversea[seeFeichteret al., 1996]). with the new schemeare generallyhigherthan thoseof the
The relativedifferencesare calculatedas (new schememinus constant
Vaso•scheme,
withrelative
increases
of morethana
"constantVax" scheme)/(constant
Vax scheme).Overlandthe factor of 2. Over marine regionswith high wind speedsthe
calculated
withthenewscheme
areup to
SO2 depositioncalculatedwith the new schemeis generally SO•- concentrations
smaller
thantheconstant
Vaso2,
withrelativedecreases
upto
-50% lowerthan thosecalculated
with the constantVaso•
100% over snow-covered surfaces. The new scheme calculates
scheme.
SO2deposition
velocities
oversnow/ice
<0.01 cms-•. Over
seathe SO2depositioncalculatedwith the new schemeshows 6.4. Global BudgetDifferencesby the New Scheme
relativeincreases
compared
to theconstant
Vaso2scheme
up
Table 3 showsthe relative differencesin column SO2 and
to 75% in the regionswith high wind speeds.This can have
significant
consequences
for SO2concentrations
in the marine
boundarylayerwhich,in turn,affectsthe calculatedamountof
precursor
gasesavailablefor newparticleformation.Figure5
SO•- burdensin four approximately
equalareasof the globe
(00-30øand300-90øN andS), comparing
thenewschemeand
the constantVax schemefor Januaryand July(for the calculation of the relativedifferences,seesection6.3). The global
shows the relative differences between the January average
totals,calculatedwith the new scheme,are shownaswell. The
new schemeyieldslargerSO2burdensin all areasfor January
SO2concentrations
in the modelsurfacelayer(--•30m) byboth
schemes.
Althoughtheseare generallysmallerthanthe differencesin SO2depositionassociated
with negativefeedbacksin
the SO2 concentrationchangesthroughdry depositionand
chemistry,relative differencesbetweenthe new and constant
Vaso2
scheme
of --•25%occur
upto analtitude
of 5 km.
andJuly,with a relativeincreaseof--•20%for the 30øN-90øN
areawhichcontains75% of the globalSO2burdenin January
and 45% in July.This alsoreflectsthe globalrelativeincrease
of about20%. The Januaryand JulyglobalSO•- burdens
increaseby 9% and 15%, respectively,
becauseof the larger
5690
GANZEVELD
ET AL.' A DRY
DEPOSITION
PARAMETERIZATION
FOR SULFUR
OXIDES
JonuoryoverogeVd S04 [cm s-1]
...
.....
.
,: q:•.::-:::.........
•:•
:::
•::,...'•:'-"..:.'•-?:,,,;;;?;;;;;;;;;:=:";:?
,½'-
....
:? .....
:.•½•:..";
..... :':-
.,.;•;::•,•,
m• r'"':"*•:'
-....
................
....•-..-..•..
............
. ::, •:, .½•:..
' '"'"":'
........................
::"-::.:::.
'%.
' ..
½ • .......';. T'. ':..' ......
:.'"=.':72'-':
-- . -::..•
•'
•
• ....................
' .....
.
..
180W
I50W
12•W
90W
60W
0.0t
0.025
0.05
isolines ore:
........
.
.....
50W
O. 10
0
...
....
50E
0.20
..
60E
0.50
90E
120E
150E
180E
0,75
July overogeVd S04 [cm s-1]
• •
•?-'-:•
.......
.• •
•
....
-
•
.
'-m.. '•*
ß
:-• o
....
•
......<',z,•'½,•!
..... .
•'•.
•' ...•..
............
:..•...• •
• •..• .... .•. *
:•.• ::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::.
• •'-•.' .:,-•..•:--'.,
---.'.:,
--.-•
'•,•'•::?,(•;'•"::• •:.'"w•/...... •:•.;Z:•:•
:.--: .-'?;.. •:•:
'--r'.-"•
.%;%• ......
% *•.• •..-, .. :..::%.,.....
•........:•..._..;.,,,.•.
•,•;½: . ½........ :: .
. .;..:.%:%%7:4-•.•;--,•r-•.:.<.
....•.<
...:
....
•:. •...:
.........:....;•...•...:.:::::::::::::::::::::::
....:•
.::....,
-.... ß •
K. X
:../..2
.......
305
.
o
: -•:•':'
..........
-
½:-,;;
.•'•::-'•.• ';;;:%....
::::::::::::'"
.......
%•:
.........
.'" '::'.
:.....:.:.
..............
....:: •:•½•
•
•
......
..,•
•i
....
•::•'.
:" .'• ":::
. .. •
- ß..... ::::r-:. . •::..•...•½:•:;:;::
•'-::::--'•'-'•5::.%.........':•:;
, -:½
:';::::.::::•
":::
•'•:"•
&'
.
.......
ß
o-::•::...' "'- ==========================================================
................................
=:'
'::
.:..;:
......, •'":'-:::•
...... :-'-'-•--'
...................
'...... ß":':'""•':•-..
' .:':!
......
60S
•'•:':'•'"
'....................
:::':::"
•;"'
:............................
..• ......
..........
....
' .....................
:•.........-• ß
180W 150W 120W
isolines ore:
0.01
.......
.:.....?L?;;.;.;;;;;;::;;:z-•.-:'":
"•'•':
•.•?:•z.•
.;
•;•:•
....•......,
...................
•:.., ....
•:•:::•:• :•:•::•:•:•,•:•:•:•
.......
•
-' .-' .-:-::"......... '.----•,.,.½•:•-...........................................
.•::..'•.--
90W
60W
0.025
0.05
30W
0.10
0
0.20
........ ...
•:?::::.:<:
.::•,.......
-•
. ..:::,...::
':•;;::..::':;:;:;?•:'½•
• .::•......:
.:..
.... ..:...•.....• •' - •...••
•OE
0.50
60E
90E
;•'•:•r•:•
120E
• .....
........................
'
.......
150E
180E
0.75
[i•ure 6, Monthlyaverage
SO•- deposition
veloci•(cms-•) for (a) Janua•and(b) July.
SO2 burden. This increasein SO2 is more pronouncedthan
ioussurfacecovers.The introductionof a local surfacerough-
that due to the improvedSO42-dry deposition
formulation. ness and a more realistic LAI representationhas reduced
The new schemedoes not yield a significantchangein the
severalshortcomings
of the model. Calculationof the aerodynamic resistanceR a in the new model versionhas improved
crease of 7% in July in the 30øN-90øN area. These results significantly.
Further improvementwill rely on the ECHAM
indicatethat a realisticsimulationof the globalsulfateburden turbulentexchangerepresentationand the descriptionof ecorequires,in particular,a realisticSO2dry depositionrepresen- systemsand their characteristicsin global vegetationdatatation. In addition, regional differencesin boundarylayer bases.The amount of standingbiomassis more realistically
SO24
- concentrations
relatedto drydeposition
parameteriza-representedthroughthe LAI as a functionof vegetationcover
tions are significant.
and season.However,there appearsto be somemisrepresentation of the LAI at somelocations,for example,a too large
7.
Discussion
January LAI in the central United States, becauseof a misThe new dry deposition routine improves the model- representationof the vegetationcoverfor that locationor the
calculateddepositionvelocitiesat different locationswith vat- assignedLAI. Therefore it needs further validation based on
JanuaryglobalSO42-burdenwhereasthereis a relativein-
GANZEVELD
ET AL.:
A DRY
DEPOSITION
PARAMETERIZATION
FOR
SULFUR
OXIDES
5691
Rel. diff. [%] in S04 conc.betweennewand "constantVd" scheme,donuary
..........
••
-•
120W
90W
-50
-25
• .............
•.-';.•.....:•;Zf.&"
--•'---•............
%.-•JR•:t•:.
•. w-;•-•'" .•"..•-T.,.-•:g
%
• •'-
.........
30N
EQ
30S
-
180W
150W
isotines
are:
60W
30W
- 10
10
0
30E
25
50
60E
90E
120E
I50E
180E
1 O0
Figure7. Relativedifference
(percent)of SO42-surfacelayerconcentration
betweenthe newscheme
and
theconstant
V•so4scheme
(seeFigure5), January.
observations,especiallyconcerningthe seasonalcycle.We intend to use the normalized difference vegetation index
(NDVI) data by Kidwell[1990]for the validationand introduction of a realisticseasonallydependentLAI followingGao and
Wesely[1995],who haveusedsatellitedata for the modelingof
dry depositionon a regional scale.
A dry depositionformulation for a global-scalemodel requires an adequate representation of uptake processesby
sparselyor nonvegetatedsurfacessincedeserts,grass,and tundras cover large areas. Most continentshave a bare soil fraction exceeding0.25, up to almost 1, for example, in North
Africa and the Middle East. This impliesthat soil uptake plays
an important role in the depositionprocessover these continents. The current SO2 soil resistanceparameterization accountsfor the influencesof soilpH and the relative humidity
on the SO2 uptake. It is possiblethat other soil propertiesalso
affect the SO2 uptake, for example,the organicsoil composition, but quantitative information about additional important
parametersis not available.Nevertheless,Murphy and Sigmon
[1989] showedthat a layer of partially decomposedorganic
Table 3. RelativeDifferences
(Percent)of SO2 andSO42ColumnsOver Four ApproximatelyEqual Areas of the
Globe, Comparingthe Constant l/•x Schemeand the New
Dry Deposition Scheme
802, Tg S
900-30ø
30ø-0øN
0ø-30øS
30ø-90øS
Global
SO4
2-, Tg S
January
July
January
July
20 (0.49)
23 (0.09)
12 (0.04)
6 (0.03)
20 (0.65)
17 (0.11)
21 (0.05)
18 (0.06)
8 (0.03)
18 (0.24)
1 (0.19)
0 (0.23)
0 (0.20)
-1 (0.11)
0 (0.72)
7 (0.45)
7 (0.33)
1 (0.18)
-2 (0.06)
5 (1.02)
Positivevaluesindicate an increasein the budget calculatedby the
new schemecomparedto the constantl/dx scheme.Total columnSO2
and SO4burdensfor theseareasare indicatedbetweenparentheses.
matter coveringthe soil can influencethe SO2 uptake. This is
probablyonly relevant for forest soilsbut will not significantly
influencethe SO2depositionvelocitysincedepositionis mainly
controlled by the vegetation.
A large part of the wintertime northern hemisphereis covered with snow and ice for which we calculate an SO2 dry
depositionvelocitycloseto zero, exceptfor someregionswith
surface temperaturesbetween 263 and 273 K. Although differencesbetween the calculated l/•so• and the fixed value of
0.1 cm s-• usedin the constantI/aso• schemeare not very
_
large, differencesin total dcpositionare significantsincethe
SO2 concentrationsare relatively high as a result of large SO2
emissions,a low SO2 oxidation rate, and suppressedvertical
exchangein the stablystratificd boundarylayer. The ncgligible
low SO2 dry depositionvelocity over snow and ice in the NH
winter yields surfaceconcentrationsof SO2 which are significantly larger than observed surface concentrationsin this region [LelieveMet al., 1997]. One possibleexplanationof this
discrepancy,which is a commonfeature in global sulfur models, is that we underestimateSO2 deposition. However, the
SO2 snow/ice resistanceparameterization has been derived
from observations,while the resultsare consistentwith a large
number of additional observations.Other explanationsfor the
overestimationare the strong horizontal and vertical concentration gradientsat some locationswhich are not reproduced
by the model becauseof the coarsehorizontal grid resolution
and the possiblemisrepresentation
of verticalmixing.Another
possibleexplanationis the potential importance of an additional SO2oxidationpathway(s)whichis not consideredin the
model [LelieveMet al., 1997].
Validation of our schemesuggeststhat we calculaterealistic
SO42-deposition
velocities
overvegetationandsnow-covered
surfaces.As far as we know, there are no observationsavailable
over bare soil, but it may be expectedthat the SO42-dry
depositionvelocitiesover bare soil are comparable to those
over snowand ice (similar surfaceroughness),neglectingpos-
5692
GANZEVELD
ET AL.: A DRY
DEPOSITION
PARAMETERIZATION
sibleresuspension
of particles.The relativelylargestuncertain-
FOR
SULFUR
OXIDES
Influencesof AnthropogenicTrace Gas Emissions).We thank the
PlanckInstitutf/JrMeteorologie
tiesareassociated
withthecalculation
of SO•- drydeposition DeutschesKlimarechenzentrum/Max
in Hamburg,Germany,for the useof computerfacilitiesandsupport.
velocitiesover seasurfaces.In fact, removalprocessesover the
ocean are still not well understood [Williams, 1982; HummelshOjet al., 1992].A main problemis the lackof observations References
of particle dry depositionto the sea surface.The difference
Andreae, M. O., W. Elbert, and S. J. de Mora, Biogenicsulfuremisbetweenthe particle dry depositionvelocitycalculatedby the
sionsand aerosolsoverthe tropicalSouthAtlantic,3, Atmospheric
parameterizationby HummelshOjet al. [1992] and the paramdimethylsulfide,aerosols,and cloud condensation,
J. Geophys.Res.,
100, 11,335-11,355, 1995.
eterizationby Slinn and Slinn [1980],whichdoesnot consider
the effect of whitecapson the dry depositionprocess,is rela- Arndt, R. L., G. R. Carmichael, D. G. Streets, and N. Bhatti, Sulfur
dioxideemissionsand sectoralcontributionsto sulfurdepositionin
tivelysmallforwindspeeds
<---t0 m s-•. TheJanuaryaverage Asia,
Atmos.Environ.,in press,1997.
windspeedat 10m in ECHAM4 is <10 m s-• exceptfor a few Bache,D. H., Particulatetransportwithin plant canopies,2, Predic-
smallregionsover the North Atlantic and the PacificOcean.
tionsof depositionvelocities,
Atmos.Environ.,13, 1681-1687,1979.
However,differences
in calculated
SO•- deposition
velocities Baldocchi,D. D., Depositionof gaseoussulfurcompoundsto vegetabetween the two parameterizationsare very large, i.e., more
thant orderof magnitude,
for a windspeedof 30 m s- • over
the North Atlantic Ocean in January. It has been shown in
section
6.3thatSO42-surface
layerconcentrations
decrease
by
---50%,associated
with a factorof t0 increaseof the SOldepositionvelocity. Sensitivityanalysisalso showsthat the
transfer
of a small fraction
of marine
sulfate
to the coarse
tion, in SulfurNutritionandAssimilationandHigherPlants,editedby
L. J. De Kok et al., pp. 271-293, SPB Acad., The Hague, Netherlands, 1993.
Bales,R. C., M.P. Valdez, and G. A. Dawson,Gaseousdepositionto
snow,2, Physical-chemical
model for SO2 deposition,J. Geophys.
Res., 92, 9789-9799, 1987.
Barrie, L. A., and J. L. Walmsley,A studyof sulphurdioxidedeposition velocities to snow in northern Canada, Atmos. Environ., 12,
2321-2332, 1978.
particlemodecan significantlyenhancethe depositionvelocity Batjes,N.H., A globaldata set of soilpH properties,Tech.Pap. 27,
Int. Soil Ref. and Inf. Cent., Wageningen,Netherlands,1995.
due to sedimentation.Therefore the representationof the
sulfateaerosolsize distributionsneedsfurther studybasedon Bergin,M. H., J.-L. Jaffrezo,C. I. Davidson,J. E. Dibb, S. N. Pandis,
R. Hillamo, W. Maenhaut, H. D. Kuhns, and T. Makela, The conadditionalobservations.Future effortswill focuson the explicit
tributionsof snow,fog, and dry depositionto the summerflux of
calculation
of sulfate mass size distributions
as a function
of
anions and cations at Summit, Greenland, J. Geophys.Res., 100,
16,275-16,288, 1995.
the controllingprocesses,for example,nucleation,condensation, and coagulation(J. Wilson, personalcommunication, Brinkop,S., and E. Roeckner,Cloud-turbulenceinteractions:Sensitivityof a generalcirculationmodelto closureassumptions,
Rep.17,
t996), providinga basisfor further improvements.
8.
Conclusions
Our dry deposition
schemecalculates
SO2and SO•- re-
Max-Planck-Inst.ffir Meteorol., Hamburg, Germany, 1993.
Cadle, S. H., J. Muhlbaier Dasch, and P. A. Mulawa, Atmospheric
concentrations
and the depositionvelocityto snowof nitric acid,
sulfurdioxideand variousparticulatespecies,Atmos.Environ.,11,
1819-1827, 1985.
movalratesof overmanylocationsfor differentmeteorological Chamberlain,A. C., Dry depositionof sulfurdioxide,in Atmospheric
conditions, consistentwith surface characteristicsthat control
SulfurDeposition,edited by D. S. Shrineret al., pp. 185-197,Butterworth-Heinemann, Newton, Mass., 1980.
dry deposition.Dry depositionof SO2 over sea is to a large Charlson,R. J., S. E. Schwartz,J. M. Hales, R. D. Cess,J. A. Coakley,
extent controlled by the aerodynamicresistance,while over
J. E. Hansen,and D. J. Hofmann,Climateforcingby anthropogenic
aerosols,Science,255, 423-430, 1992.
land it is determinedmostlyby the surfaceresistance(i.e., rsoi•,
rveg
, rwsand rs,•ow/ic½).
The SO2 soil resistance
is explicitly
Charnock,M., Wind stresson a water surface,Q. J. R. Meteorol.Soc.,
81,639-640, 1955.
calculatedfrom the soilp H and relative humiditywhereasthe Chin, M., D. J. Jacob, G. M. Gardner, M. Foreman-Fowler, and P. A.
SO2 snow/iceresistanceis a functionof temperature.IncorpoSpiro,A globalthree-dimensional
model of troposphericsulfate,J.
ration of the schemein the chemistryand generalcirculation
Geophys.Res.,101, 18,667-18,690,1996.
model ECHAM yields significantchangesin SO2 massflux Clapsaddle,C., and D. Lamb, The sorptionbehaviorof SO2on ice at
densitiesand SO2 and SO42-concentrations
in the lower
troposphere and in their column burdens compared to a
previousschemeusingconstantdepositionvelocities [Feichter et al., 1996]. For example, the new schemecalculatesup
to -50% larger SO2 concentrationsnear the surface in the
continental NH in January whereas there is a relative decreasein SO2 concentrationsover the North Atlantic Ocean
of ---50% in January.The global SO2 burden calculatedwith
the new scheme is ---20% larger compared to the constant
Vaso2
scheme.
The consequent
SO]- burdenincrease
is
relativelylarge comparedto the SO•- increaseresulting
from the new SO42-dry depositionparameterization.
This
emphasizesthe need for a realistic representation of the
SO2 dry deposition process.In addition, significantdiffer-
encesin SO]- massfluxdensities
andsurfacelayerconcentrations
are calculated
between
the new and the constant
Vaso4scheme.
temperaturesbetween -30øC and -5øC, Geophys.Res.Lett., 16,
1173-1176, 1989.
Claussen, M., U. Lohmann, E. Roeckner, and U. Schulzweida, A
global data set of land-surfaceparameters,Rep. 135, Max-PlanckInst. f/JrMeteorol., Hamburg,Germany,1994.
Claussen,M., Estimationof regionalheat and moisturefluxesin homogeneousterrain with bluff roughnesselements,J. Hydrol.,166,
353-369, 1995.
Conklin, M. H., R. A. Sommerfeld, S. K. Laird, and J. E. Villinski,
Sulfurdioxidereactionson ice surfaces:Implicationsfor dry deposition to snow,Atmos. Environ. Part A, 27, 159-166, 1993.
Davies,T. D., and J. R. Mitchell, Dry depositionof sulphurdioxide
onto grassin rural easternEngland (with somecomparisons
with
other formsof sulphurdeposition),in Precipitation
Scavenging,
Dry
Depositionand Resuspension,
edited by H. R. Pruppacher,R. G.
Semonin,and W. G. N. Slinn, pp. 795-806, Elsevier,New York,
1983.
DeutschesKlimazechenzentrum,
Modellbetreuungsgruppe
(DKRZ),
The ECHAM3 AtmosphericGeneralCirculationModel, Tech.Rep.
6, Hamburg,Germany,1992.
Dolman,A. J., Estimatesof roughness
lengthandzeroplanedisplacement for a foliatedand non-foliatedoak canopy,Agric.For. Meteorol., 36, 241-248, 1986.
Acknowledgments. This work was performed within the
SINDIcATE project (Study of the Indirect and Direct Climate
Dovland, H., and A. Eliassen,Dry depositionon a snow surface,
Atmos. Environ., 10, 783-785, 1976.
GANZEVELD
ET AL.: A DRY
DEPOSITION
PARAMETERIZATION
Erisman,J. W., and D. D. Baldocchi,Modelling dry depositionof SO2,
Tellus, Ser. B, 46, 159-171, 1994.
Erisman, J. W., and A. Van Pul, Parameterization of surface resistance
for the quantificationof atmosphericdepositionof acidifyingpollutants and ozone, Atmos. Environ., 28, 2595-2607, 1994.
Erisman,J. W., and G. P. Wyers,Continuousmeasurementsof surface
exchangeof SO2andNH3: Implicationsfor their possibleinteraction
in the depositionprocess,Atmos. Environ. Part A, 27, 1937-1949,
1992.
Fan, S.-M., S.C. Wofsy,P.S. Bakwin,D. J. Jacob,and D. R. Fitzjarrald, Atmosphere-biosphere
exchangeof CO2 and O3 in the central
Amazon forest,J. Geophys.Res.,95, 16,851-16,864, 1990.
Feichter, J., E. Kjellstr6m, H. Rodhe, F. Dentener, J. Lelieveld, and
G.-J. Roelofs,Simulationof the troposphericsulfurcyclein a global
climate model, Atmos. Environ., 30, 1693-1707, 1996.
Fitzgerald,J. W., Approximationformulasfor the equilibriumsize of
an aerosolparticle as a functionof its dry size and compositionand
the ambient relative humidity, J. Appl. Meteorol., 14, 1044-1049,
1975.
FOR
SULFUR
OXIDES
5693
to estimatedry depositionof SO2,Atmos.Environ.Part A, 27, 493501, 1993.
Mehlmann, A., Gr6ssenverteilungdes aerosolnitratsund seinebeziehung zur gasf6rmigensalpetersgure,Ph.D. thesis,Univ. of Mainz,
Mainz, Germany, 1986.
Milne, J. W., D. B. Roberts,and D. J. Williams,The dry depositionof
sulfur dioxide field measurements with a stirred chamber, Atmos.
Environ., 13, 373-379, 1979.
Mitra, S. K., S. Barth,andH. R. Pruppacher,A laboratorystudyon the
scavengingof SO2 by snow crystals,Atmos. Environ. Part A, 24,
2307-2312, 1990.
Muhlbaier Dasch, J., and S. H. Cadle, Dry depositionto snowin an
urban area, WaterAir Soil Pollut., 29, 297-308, 1986.
Murphy, C. E., Jr., and J. T. Sigmon,Dry depositionof sulfur and
nitrogenoxidegasesto forestvegetation,in Acidic Precipitation,vol.
3, Sources,Depositionand CanopyInteractions,editedby S. E. Lindberg, A. L. Page, and S. A. Norton, pp. 217-240, Springer-Verlag,
New York, 1989.
Nicholson,K. W., and T. D. Davies, The dry depositionof sulfur
Fowler, D., and M. H. Unsworth, Turbulent transfer of sulfur dioxide
dioxide at a rural site, Atmos. Environ., 22, 2885-2889, 1988.
Olson, J., J. A. Watts, and L. J. Allison, Carbon in live vegetation of
major world ecosystems,
Rep. ORNL-5862, Oak Ridge Natl. Lab.,
Oak Ridge, Tenn., 1983.
Padro,J., H. H. Neumann,and G. Den Hartog, Dry depositionvelocity
1995.
estimatesof SO2 from models and measurementsover a deciduous
forest in winter, Water Air Soil Pollut., 68, 325-339, 1993.
Gao, W., and M. L. Wesely, Modelling gaseousdry depositionover
regional scaleswith satellite observations,1, Model development, Payrissat,M., and S. Beilke, Laboratorymeasurementsof the uptake
Atmos. Environ., 29, 727-737, 1995.
of sulfur dioxide by different European soils,Atmos. Environ., 9,
211-217, 1975.
Granat, L., and C. Johansson,Dry depositionof SO2 and NOx in
winter, Atmos. Environ., 17, 191-192, 1983.
Prahm, L. P., U. Torp, and R. M. Stern, Deposition and transformaHenderson-Sellers, A., M. F. Wilson, G. Thomas, and R. E. Dickinson,
tion rates of sulfur dioxidesduring atmospherictransportover the
Atlantic, Tellus, 28, 355-372, 1975.
Currentgloballand-surfacedata setsfor usein climate-relatedstudies, Tech. Note NCAR/TN-272+STR, Natl. Cent. for Atmos. Res.,
Ritter, J. A., J. D. Barrick, G. W. Sachse,G. L. Gregory, M. A.
to a wheat crop, Q. J. R. Meteorol.Soc.,105, 767-783, 1979.
Ganzeveld,L., and J. Lelieveld,Dry depositionparameterizationin a
chemistrygeneral circulationmodel and its influenceon the distribution of reactivetrace gases,J. Geophys.Res., 100, 20,999-21,012,
Boulder, Colo., 1986.
Hicks,B. B., M. L. Wesely,R. L. Coulter,R. L. Hart, J. L. Durham, R.
Speer,andD. H. Stedman,An experimentalstudyof sulfurand NOx
fluxesover grassland,BoundaryLayer Meteorol.,34, 103-121, 1986.
Hicks, B. B., D. D. Baldocchi,T. P. Meyers,R. P. Hosker Jr., and D. R.
Matt, A preliminary multiple resistanceroutine for deriving dry
depositionvelocitiesfrom measuredquantities,WaterAir Soil Pol-
Woerner, C. E. Watson, G. F. Hill, and J. E. Collins Jr., Airborne
flux measurementsof trace speciesin an Arctic boundarylayer, J.
Geophys.Res.,97, 16,601-16,625,1992.
Roeckner,E., K. Arpe, L. Bengtsson,M. Christoph,M. Claussen,L.
Dtimenil, M. Esch, M. Giogetta, U. Schlese,U. Schulzweida,The
atmosphericgeneral circulationmodel ECHAM-4: Model description and simulation of the present-dayclimate, Rep. 218, Maxlut., 36, 311-330, 1987.
Planck-Inst.ffir Meteorol., Hamburg, Germany, 1996.
Hicks, B. B., D. R. Matt, and R. T. Millen, A micrometeorological Roelofs, G.-J., and J. Lelieveld, Distribution and budget of O3 in the
investigationof surfaceexchangeof 03, SO2, and NO2: A casestudy,
tropospherecalculatedwith a chemistry-general
circulationmodel,
BoundaryLayer Meteorol.,47, 321-336, 1989a.
J. Geophys.Res., 100, 20,983-20,998, 1995.
Hicks,B. B., D. R. Matt, R. T. McMillen, J. D. Womack, M. L. Wesely, Roelofs, G.-J., and J. Lelieveld, Model studyof the influenceof crossR. L. Hart, D. R. Cook, S. E. Lindberg, R. G. de Pena, and D. W.
tropopauseO3 transportson troposphericO3 levels,Tellus,Ser.B,
49, 38-55, 1997.
Thomson,A field investigationof sulfatefluxesto a deciduousforest,J. Geophys.Res.,94, 13,003-13,011,1989b.
Sellers,P. J., Canopyreflectance,photosynthesis
andtranspiration,Int.
J. Remote Sens., 6, 1335-1372, 1985.
Hummelshoj,P., N. O. Jensen,and S. E. Larsen, Particle dry deposition to a sea surface, in PrecipitationScavengingand Atmosphere- Sellers,P. J., Y. Mintz, Y. C. Sud, and A. Dalcher, A simplebiosphere
SurfaceExchange,vol. 2, edited by S. E. Schwartzand W. G. N.
model (SiB) for usewithin generalcirculationmodels,J.Atmos.Sci.,
43, 505-531, 1986.
Slinn,pp. 829-840, Hemisphere,Washington,D.C., 1992.
Ibrahim, M., L. A. Barrie, and F. Fanaki, An experimentaland theo- Sievering,H., J. Boatman,E. Gorman, Y. Kim, L. Anderson,G. Ennis,
M. Luria, and S. Pandis, Removal of sulfur from the marine boundretical investigationof the dry depositionof particlesto snow,pine
trees and artificial collectors,Atmos. Environ., 17, 781-788, 1983.
ary layer by ozone oxidationin sea-saltaerosols,Nature, 360, 571573, 1992.
Intergovernmental
Panelon ClimateChange(IPCC), ClimateChange:
The Scienceof ClimateChange,Contributionof WorkingGroup 1 to Slinn,S. A., and W. G. N. Slinn,Predictionsfor particledepositionon
natural waters, Atmos. Environ., 14, 1013-1016, 1980.
the SecondAssessment
Reportof the Intergovernmental
Panel on Climate Change,edited by J. T. Houghton et al., Cambridge Univ. Slinn, W. G. N., Dry depositionand resuspension
of aerosolpartiPress, New York, 1995.
cles--A new look at some old problems, in Atmosphere-Surface
Jacob,D. J., S.-M. Fan, S.C. Wofsy,P. A. Spiro, P.S. Bakwin, J. A.
Exchangeof Particulateand GaseousPollutants,edited by R. J. EnRitter, E. V. Browell, G. L. Gregory,D. R. Fitzjarrald, and K. E.
gelmannand G. A. Sehmel,pp. 1-40, U.S. Dep. of Energ. Tech. Inf.
Moore, Depositionof ozoneto tundra,J. Geophys.Res.,97, 16,473Cent., Oak Ridge, Tenn., 1976.
16,479, 1992.
Slinn,W. G. N., Predictionsfor particledepositionto vegetativecanKasibhatla,P., W. L. Chameides,and J. St. John, A three-dimensional
opies,Atmos. Environ.,16, 1785-1794, 1982.
globalmodel investigationof seasonalvariationsin the atmospheric Sommerfeld,R. A., and D. Lamb, Preliminarymeasurementsof SO2
burden of anthropogenicsulfate aerosols,J. Geophys.Res., 102,
absorbedon ice, Geophys.Res.Lett., 13, 349-351, 1986.
3737-3759, 1997.
Spedding,D. J., Sulfur dioxide absorptionby seawater, Atmos. Environ., 6, 583-586, 1972.
Kidwell,K. B. (Ed.), GlobalVegetation
Index User'sGuide,45 pp., U.S.
Dep. of Comm., Washington,D.C., 1990.
Spiro, P. A., D. J. Jacob,and J. A. Logan, Global inventoryof sulfur
Langner,J., and H. Rodhe, A globalthree-dimensionalmodel of the
emissionswith 1ø x 1ø resolution,J. Geophys.Res., 97, 6023-6036,
1992.
troposphericsulfurcycle,J. Atmos.Chem.,13, 225-263, 1991.
Lelieveld, J., G.-J. Roelofs, L. Ganzeveld, J. Feichter, and H. Rodhe,
Valdez, M.P., R. C. Bales,D. A. Stanley,and G. A. Dawson,Gaseous
Terrestrial sourcesand distributionof atmosphericsulfur, Philos.
depositionto snow,1, Experimentalstudyof SO2 and NO 2 deposiTrans. R. Soc. London, Set. B, 352, 149-158, 1997.
tion, J. Geophys.Res., 92, 9779-9787, 1987.
Matt, D. R., and T. P. Meyers,On the useof the inferentialtechnique Vermetten, A. W. M., P. Hofschreuder,J. H. Duyzer, H. S. M. A.
5694
GANZEVELD
ET AL.:
A DRY
DEPOSITION
PARAMETERIZATION
Diederen, F. C. Bosveld, and W. Bouten, Dry depositionof SO2
onto a standof Douglas Fir: The influenceof canopywetness,Rep.
R-508, Dep. of Air Qual., Agric. Univ. of Wageningen,Wageningen,
Netherlands, 1991.
Warneck, P., Chemistryof the Natural Atmosphere,Academic, San
Diego, Calif., 1988.
Wesely,M. L., Parameterizationof surfaceresistances
to gaseousdry
depositionin regional-scalenumericalmodels,Atmos.Environ.,23,
1293-1304, 1989.
Wesely, M. L., D. R. Cook, and R. L. Hart, Fluxes of gasesand
particlesabove a deciduousforest in wintertime,BoundaryLayer
FOR
SULFUR
OXIDES
Whitby, K. T., The physicalcharacteristicsof sulfur aerosols,Atmos.
Environ., 12, 135-159, 1978.
Wieringa,J., Updating the Davenportroughnessclassification,
J. Wind
Eng. Ind. Aerodyn.,41-44, 357-368, 1991.
Williams,R. M., A model of the dry depositionof particlesto natural
water surfaces,Atmos. Environ., 16, 1933-1938, 1982.
L. Ganzeveld, J. Lelieveld, and G.-J. Roelofs, Institute for Marine
and AtmosphericResearchUtrecht (IMAU), Princetonplein5, 3584
CC Utrecht,Netherlands.(e-mail:[email protected])
Meteorol., 27, 237-255, 1983.
Wesely, M. L., D. R. Cook, and R. L. Hart, Measurementsand parameterization of particulate sulfur dry depositionover grass,J.
Geophys.Res.,90, 2131-2143, 1985.
(ReceivedMay 29, 1997;revisedOctober14, 1997;
acceptedOctober21, 1997.)