4.13

RecordofMineral Aerosolsand TheirRole in the Earth SystemK.E. Kohfeld

SimonFraserUniversity, Burnaby, BC, Canada

and

I. Tegen

Leibniz Institute forTropospheric Research, Leipzig, Germany

4.13.1 INTRODUCTION

1 4.13.1.1 What is Dust? 1 4.13.1.2 Why is Dust Important? 2 4.13.1.3 Dust Cycle Processes: Entrainment, Transport, and Deposition 3

4.13.2 GEOLOGIC RECORDS OF DUST: PALEOENVIRONMENTS AND TOOLS

5 4.13.2.1 Paleoenvironments 5 4.13.2.2 Dust Deposition 7 4.13.2.3 Dust Sources and Emissions 7

4.13.2.3.1 Geochemical fingerprinting

7 4.13.2.3.2 Grain size indicators 8

4.13.3 GEOLOGIC VARIABILITY OF THE DUST CYCLE

9 4.13.3.1 Dust over Long Time Periods 9 4.13.3.2 The Distribution of Dust over the Last Glacial Period 10 4.13.3.3 Glacial–Interglacial Cycles 11

4.13.4 MODERN RECORDS OF DUST

11 4.13.4.1 Source Areas 11 4.13.4.2 Transport 13 4.13.4.3 Deposition 16

4.13.5 MODERN VARIABILITY OF THE DUST CYCLE

16

4.13.6 FILLING IN THE GAPS: THE ROLE OF GLOBAL DUST CYCLE MODELS

17

4.13.7 DUST FEEDBACKS WITHIN THE EARTH SYSTEM

18 4.13.7.1 Radiative Feedbacks 18 4.13.7.2 Biogeochemical Feedbacks 19

4.13.8 FUTURE OUTLOOK

20

ACKNOWLEDGMENTS

21

REFERENCES

21

Change begets changeyThe mine which Time hasslowly dug beneath familiar objects is sprung in aninstant; and what was rock before, becomes but sandand dust.

Charles Dickens

All we are is dust in the wind, dude.Ted, Bill and Ted’s Excellent Adventure

1

4.13.1 INTRODUCTION

4.13.1.1 What is Dust?

Soil dust is a type of mineral aerosol, andcomprises one component of the total atmos-pheric aerosol loading. Atmospheric aerosols

Figure 1 Satellite image of a Saharan dust storm. A long line of Saharan dust swept across Mali, Mauritania,and Western Sahara and out over the Canary Islands on March 3, 2004 when the Moderate ResolutionImaging Spectroradiometer (MODIS) acquired these images during the morning (a) and afternoon (b)passovers. As the day progressed, the dust grew thicker and the storm extended farther west. Image courtesy of

the NASA Visible Earth Observatory (Visible Earth, 2006).

2 Record of Mineral Aerosols and Their Role in the Earth System

play an important role in the climate system,affecting the energy balance of the atmosphereby controlling the amount of radiative energythat enters and exits, through the scatteringand absorption of radiation. They can influenceimportant atmospheric processes such as cloudformation, which in turn affect regional energybalances as well as the chemistry (e.g., pH) andquantity of precipitation. The degree to whichatmospheric aerosols, both anthropogenic andnatural, affect climate remains an importantquestion in earth system science, and has beenaddressed in Chapter 4.04. For the purpose ofthis review, we will focus strictly on soil dustand its role within the earth system.

Atmospheric soil dust is defined here as par-ticles that are formed in soils and emitted intothe atmosphere during strong surface windevents. Soil dust particles are a combination ofirregular-shaped mineral grains such as quartz,feldspars, and calcite, clay minerals rich inelements such as aluminum, and oxides andhydroxides rich in elements such as iron. Thecomposition of dust can change depending onthe source region. A compilation by Claquinet al. (1999) indicates that quartz is the mostabundant mineral in the silt size fraction of thesoil (particles in the size range of 2–50 mm

diameter), and illite is the most abundant claymineral. The chemical properties of dust canchange during atmospheric transport when dustsurfaces become coated with soluble sulfate andnitrate species. Soil particles that are generallyo60–100mm in diameter can be picked up bywind, and far-traveled dust is typically around2mm in diameter (Seinfeld and Pandis, 1997).Atmospheric soil dust particles are visible in theatmosphere as a haze. High dust aerosol loadsoriginate from dust storms that can reduceatmospheric visibility to o1 km and are evenvisible from space (Figure 1).

4.13.1.2 Why is Dust Important?

Atmospheric dust can have substantial regio-nal impacts on agriculture and human health.Dust storms affect agriculture through ex-tensive erosion of the soil surface (Nordstromand Hotta, 2004). Dust can also affect thequality of the air that we breathe. Minuteparticles of soil dust with diameters smallerthan 10 mm (PM10) in the atmosphere havebeen shown to be damaging to human res-piratory health (Chapter 9.07). Furthermore,certain toxic metals, such as mercury and

Introduction 3

arsenic, can be associated with the transport ofaeolian dust (Gill et al., 2002; Cannon et al.,2003; Givelet et al., 2004).

Although there have been efforts to quantifythe interactions between atmospheric dust andregional climate (e.g., Seinfeld et al., 2004;House et al., 2006), the overall radiative impactof soil dust on the climate cycle is still not wellquantified. In fact, it is not known if soil dustaerosols on the whole will have a net warmingor cooling effect (Houghton et al., 2001). Dustcan either increase or decrease planetaryalbedo, depending on both the single scatter-ing albedo of the dust particle as well as thealbedo of the underlying surface. For example,dust over optically ‘‘dark’’ regions, such asoceans or forests, can have a cooling effect (see,e.g., Tegen et al., 1996). Alternatively, dust canincrease the absorption of solar radiation overhigh reflective regions such as ice sheets, snow,or even desert surfaces. The role of dust ascondensation nuclei in cloud formation andprecipitation is unknown but potentially im-portant (e.g., Rosenfeld et al., 2001; Levinet al., 2005). The radiative forcing of high at-mospheric dust concentrations in the geologicpast have been proposed as a means of coolingthe tropics (Claquin et al., 2003) and the mid-continental United States (Roberts et al., 2003),but also as a means of accelerating deglaciation(Overpeck et al., 1996).

Dust can also impact the earth system viabiogeochemical feedbacks (see Section4.13.7.2). Soil dust contains micronutrients(such as iron, zinc, and cadmium) and macro-nutrients (such as phosphate and silica) all ofwhich can fertilize terrestrial and marine eco-systems. Iron in particular has been suggestedto limit phytoplankton growth in ocean re-gions, which are otherwise rich in macronutri-ents, such as nitrate (Martin and Fitzwater,1988). As such, changes in dust inputs to theocean could affect marine productivity as wellas the sequestration of carbon in the deep sea(Martin, 1990).

Assessing the magnitude of these regionalimpacts, radiative forcing, and biogeochemicalfeedbacks in the earth system requires an abil-ity to quantify the amounts of dust produced,both in the past and in the present. Predictingfuture changes in the dust cycle and its inter-action with climate are even more uncertain.Estimating the importance of dust depends onour ability to estimate (1) the natural processescontrolling dust entrainment, transport, anddeposition; (2) changes in these natural proc-esses due to global and regional climate change;and (3) changes in dust production due tohuman impacts on land use. Our ability tosimulate these processes requires correct

parameterization of these processes as well asdata to constrain them.

The purpose of the chapter is to review thecurrent knowledge about the record of mineralaerosols. This chapter will first review the mainprocesses of the dust cycle—entrainment,transport, and deposition (Section 4.13.1.3).The tools and records used to measure changesin entrainment, transport, and deposition ofdust in the geologic past will be discussed inSection 4.13.2, followed by an elaboration ofhow dust has varied through geologic history(Section 4.13.3). We follow with parallel dis-cussions of how soil dust processes are meas-ured for the modern climate (Section 4.13.4),and what these measurements have revealedwith regard to recent variability in the dust cy-cle (Section 4.13.5). We proceed by explaininghow data are used to inform models of the dustcycle (Section 4.13.6), and summarize whatis currently known about the radiative andbiogeochemical feedbacks of atmospheric dustwithin the earth system (Section 4.13.7). Weconclude with a future outlook of importantquestions that still need to be addressed in thefield of dust research (Section 4.13.8).

4.13.1.3 Dust Cycle Processes: Entrainment,Transport, and Deposition

Most atmospheric dust has its origin in aridto semi-arid regions of the Earth and is emittedfrom regions that are sparsely vegetated (i.e.,t15% vegetation cover). Today, most dust isproduced in deserts of northern Africa(Sahara, Sahel) and the Middle Eastern regions(Figure 2), contributing up to 70% to the mod-ern global annual dust emission fluxes (Jickellset al., 2005). Other sources of importance in-clude the desert regions of Asia and Australia.

In regions where the soil is exposed and thesoils contain fine and loose particles, windsare able to lift soil particles into the air. Thedependence of the dust flux F on the surfacewind friction velocity u� can be written for agiven particle size as

Fpu3� 1þ u�tru�

� �1� u2�

tr

u2�

� �� �for u�4u�tr

with the friction velocity and u�tr the thresholdfriction velocity (Marticorena and Bergametti,1995). Surface wind friction velocities are com-puted from surface wind speeds and surfaceroughness data, assuming neutral stabilityconditions, as u� ¼ uðzÞk=Inðz=z0Þ; where u isthe surface wind speed at reference height z, kis the von Karman constant, and z0 is the

TOMS absorbing aerosol index, 1981−19920.0 0.06 0.10 0.16 0.26 0.4 0.7 1.0 1.7 2.7

Figure 2 TOMS aerosol index values averaged from 1981 to 1992. Largest source areas are located in aridtropical–subtropical regions of Africa and the Middle East. Green lines outline regions that are considered

dust sources. Orange arrows indicate the general direction of dust transport from these regions.

4 Record of Mineral Aerosols and Their Role in the Earth System

aerodynamic roughness length. The minimumwind friction velocity u�tr required to lift soilparticles is called the ‘‘threshold wind velocity’’and depends on a combination of soil texture,moisture, and the surface roughness. Silt-sized(2–50mm diameter) or fine sand (50–250mmdiameter) particles are the easiest to lift and re-quire the lowest surface wind speeds to becomeairborne, while larger particles are heavier andhave higher threshold wind velocities. Thesmallest, clay-sized particles (diameter o2mm)have a greater surface area-to-volume ratiowhen compared with silt-sized particles. Theseclay-sized particles tend to adhere to each otherby cohesive forces (van der Waals forces andelectrostatic charges (Iversen and White, 1982),and thus cannot be lifted directly from theground as individual particles. Soils containingmore than 20% clay-sized particles are con-sidered to be only weak emitters of dust aero-sols (Gillette, 1979; Marticorena andBergametti, 1995). Soil particles with diametersof 2mm or less, which can be transported bywinds over large distances, are usually dis-located from soil aggegates at the soil surfacethrough the impact of saltating sand grains(e.g., Lu and Shao, 1999).

Within the deserts, certain regions represent‘‘hot spots’’ of dust production, emitting higheramounts of dust compared with other desertareas. These regions containing large amounts

of fine and loose soil particles that can formdust aerosols are referred to as preferential dustsources (Gillette, 1999; Ginoux et al., 2001;Prospero et al., 2002; Tegen et al., 2002; Tegen,2003). Typically, dried paleolake beds such asthe Bodele depression in Chad or the LakeEyre basin in Australia, playa sediments, oralluvial fans can be ‘‘hot spot’’ dust sources(Pye, 1987). For example, the Bodele depres-sion contains the most active dust sourceworldwide (Prospero et al., 2002). There, dustemission occurs from massive diatomite sedim-ents that formed 6,000 years ago when thebasin was covered by the paleolake Mega-Chad(Washington et al., 2006).

Estimating global annual dust emissions iscomplicated because we do not have observat-ional datasets that are extensive and detailedenough to quantify dust emissions on a globalscale (see Cakmur et al., 2006 for a study ofusing various datasets to constrain emissions).Currently, emissions are generally calculatedusing global dust models, but uncertaintiesare introduced from model treatments of land-surface characteristics, the parameterizationscontrolling the mobilization of dust, and thecontrolling meteorological factors. As a result,current estimates of dust emissions span a largerange, between 1,000 and 2,500Tg (1012 g)of dust emitted globally each year (Zenderet al., 2004).

Geologic Records of Dust: Paleoenvironments and Tools 5

Our knowledge of the atmospheric burden, orthe amount of dust that remains in the atmos-phere at any one time, is even less concise thanour knowledge of emissions, and varies by afactor of 4. The values range from 8 to 36Tg(Zender et al., 2004), with uncertainty resultingfrom observational uncertainties as well asa difficulty constraining what controls theremoval of dust from the atmosphere, or dep-osition. The removal of dust from the atmos-phere occurs through two processes known as‘‘wet’’ or ‘‘dry’’ deposition. During dry deposi-tion, airborne particles fall to the ground bygravitational settling (where the settling velocitydepends on the square of the particle diameter),and mixing to the surface by turbulent air mo-tions. As a consequence of the gravitationalsettling, large particles (with diameters greaterthan B5 mm) are removed from the air ratherclose to their source of entrainment, but finerparticles remain suspended in the atmosphereand can be transported long distances. As a re-sult, areas of deposition that are remote fromdust source areas (such as the polar ice caps) aredominated by fine particles with sizes on theorder of 1–2mm. During wet deposition, dust isscavenged by precipitation droplets and rainedout of the atmosphere. Highest rates of dustdeposition are found directly downwind of thelargest dust sources such as the Sahara and Sa-hel desert regions of North Africa. These largedust fluxes are not surprising because the largestparticles are removed from the air near thesource. The fine dust from North Africa can betransported across the tropical Atlantic andreach as far as the Amazon basin. Other areasof high deposition are associated with thedeserts of central Asia. When transportedwithin higher atmospheric layers at several kilo-meters altitude, soil particles from Asian duststorms occurring in spring can be transported asfar as North America and Greenland, and insome cases, Europe (Grousset et al., 2003).Less-extensive dust sources in Australia, south-ern Africa, and Patagonia have a morelocal influence on dust deposition (Duce et al.,1991). Regions that are far from any majorsources of dust, such as Greenland andAntarctica (Figure 2), are characterized by ex-tremely low dust deposition (Figure 3a).

4.13.2 GEOLOGIC RECORDS OF DUST:PALEOENVIRONMENTS ANDTOOLS

4.13.2.1 Paleoenvironments

The factors controlling the dust cycle—arid-ity, vegetation cover, winds, and precipitation—are all related to climatological factors. Thus, it

is not surprising that dust emissions, transport,and deposition have changed with climate ongeological timescales, and will change overfuture timescales as well. In fact, the geologicrecord of dust provides a rich history of climatecycles, as archived on the land (through terres-trial loess deposits and lake sediments), in theocean (via marine sediments), and in remotepolar areas (from ice cores).

The geologic record of dust is by nature arecord of ‘‘deposition,’’ stored in several dif-ferent paleoenvironments. Geological depositsof dust are found on all parts of the globe, atthe bottom of lakes and also as sedimentarysequences on land called ‘‘loess’’ (Pye, 1987,1995). Loess deposits are present on every con-tinent and are composed of wind-blown, silt-sized dust particles (2–50 mm) that are generallyfound very near to dust source regions becauseof their relatively large sizes. Although some‘‘desert loess’’ is found in subtropical aridregions downwind of arid sources (e.g., Smal-ley and Krinsley, 1978), much loess is foundnear glaciogenic source regions, where materi-als have been ground up by glacial activity andthen transported by rivers to areas that are re-latively arid and free of vegetation (Smalleyand Smalley, 1983). Largest thicknesses of loessbegan to accumulate on the continents over thelast 2.8Myr, likely associated with the climatechanges associated with glacial–interglacialcycles (Derbyshire, 2003).

Aeolian materials are also found in lake andmarine sediments (for full review see Kohfeldand Harrison, 2001). In general, the aeolianmaterials archived in marine sediments arelocated further from dust source regions andso, with rare exception, are composed ofsmaller grain size material, and in lesser abun-dances compared with lake sediments. Theadvantage of lake and marine sedimentaryarchives is that the records tend to be morecontinuous than terrestrial records, and there-fore can provide an uninterrupted record ofchanges in dust deposition through geologictime. Furthermore, these sediments providemore direct information about the inputs tolacustrine and marine ecosystems, whichmay be important when considering biogeo-chemical feedbacks of dust. There are, how-ever, some disadvantages to these records. Forexample, the extraction of aeolian signals issometimes more complicated because of theinfluence of detritus from rivers and ice rafting,as well as sediment winnowing or resuspensionprocesses (Rea, 1994). Sediment records canalso be affected by mixing due to bioturbation,and this process can make it difficult to deter-mine abrupt changes in dust depositionwhen sedimentation rates are low (see, e.g.,

LGM

0 0.010.02 0.05 0.1 0.2 0.5 1 2 5 10 20 50 100 200

(g m−2 yr−1)

Late Holocene

(a)

(b)

(c)

0 0.1 0.2 0.5 1 2 5 10 20 50

Ratio

LGM/Late Holocene

6 Record of Mineral Aerosols and Their Role in the Earth System

Geologic Records of Dust: Paleoenvironments and Tools 7

discussion from Ruddiman, 1997). Finally, un-like loess deposits, marine and lacustrine se-diments are not 100% aeolian, butff aregenerally dominated by biogenic materials suchas carbonate and opal. One standard procedurefor isolating the terrigenous fluxes is by deter-mining the differences between the total fluxesand the biogenic components. This differencingprocedure produces large uncertainties, in par-ticular when fluxes of terrigenous materials arevery low. One promising means of gettingaround the dilution effects is to use trace ele-ments, such as 232Th, which are enriched incontinental crust material and therefore canserve as more direct geochemical tracers of dust(e.g., Anderson et al., 2006). In spite of theseissues, ocean sediments make up the majorityof the Quaternary archive of long-distancetransported dust, and can be useful for provid-ing a first-order indication of depositionchanges in mineral soil dust over longer time-scales.

Glaciers and ice sheets also serve as collec-tors of atmospheric dust that is transportedand deposited directly onto the ice and/or snowsurface. Dust either settles or is rained out andis directly incorporated into the ice and is dis-turbed by fewer postdepositional problemscompared with other depositional regimes.With the exception of some tropical glaciers,the aeolian material that has been retrievedfrom ice cores tends to be long distances fromdust sources.

4.13.2.2 Dust Deposition

The mass accumulation rates (MAR) of dustare determined in the following manner:

MAR ¼ LSR�DBD� f

where LSR is the linear sedimentation rate(m yr�1), DBD the bulk density of the sediment(gm�3), and f the fraction of the sedimentthat is considered aeolian. Some variant onthis equation is necessary for measuring amass accumulation rate no matter what thepaleoenvironment (see Kohfeld and Harrison,2001).

In ice cores, information regarding the frac-tion of wet versus dry deposition has sometimesbeen inferred by combining both simple andcomplex precipitation models with additional

Figure 3 Mass accumulation rates of dust for the (a) latthe ratio between them. Squares represent estimates frommodern marine sediment traps. Data are derived from thwith additional data from Delmonte et al. (2002, 2004a, b

information about precipitation rates andsnowfall (Alley et al., 1995; Werner et al.,2003). In general, however, changes in the massaccumulation rates of dust provide us with acombined picture of the overall changes in theintensity of the dust cycle, integrating aspectsof source, transport, and deposition.

4.13.2.3 Dust Sources and Emissions

While changes in deposition rates do not givea direct measure of dust emissions and trans-port, the deposited archive can provide us withsome clues through factors such as grain size,mineralogy, and isotopic characterization. Claymineralogy and isotopic characterization hasfrequently been used as a means of determiningthe general location of dust source areas. Thechanges in grain size of the deposited dust canreflect a combination of distance from source,wind strength as well as deposition regime.

4.13.2.3.1 Geochemical fingerprinting

Dust that enters the atmosphere carries withit a distinct geochemical fingerprint, a signaturedefined by its clay mineralogy, elemental com-position, and isotopic signature. One assump-tion is that the chemical characteristics ofsuspended and deposited dust can be used asa means of tracing its origin to the original,parent soil from which it was formed (forreview see Grousset and Biscaye, 2005). Prove-nance is determined by comparing the chemicalsignatures of dust with the signature of poten-tial source rocks. Source rocks can include bothprimary sources, which are parent rock mate-rials, or secondary sources, which are a mixtureof particles that have already been transportedby aeolian or fluvial processes. This methodof identifying source regions requires extensivesampling of potential source areas of dust fromaround the world, and requires that the distinctgeochemical signature of the airborne dust isnot altered during transport, perhaps viachemical alteration in clouds, or as a result offractionation during wet deposition.

The clay mineralogy is thought to reflect theparent rock material, drainage, climate, and thedegree of physical or chemical weathering. Forexample, kaolinite is formed primarily throughchemical weathering from feldspars and micas,at warm temperatures, and therefore at lower

e Holocene period, (b) the Last Glacial period, and (c)sediment records. Squares in (a) represent values frome DIRTMAP database (Kohfeld and Harrison, 2001),), Anderson et al. (2006) and Mahowald et al. (2006).

8 Record of Mineral Aerosols and Their Role in the Earth System

latitudes. Alternatively, chlorite is found insoils from cold, high-latitude regions wherephysical-weathering processes dominate. As aresult, the ratio between these two clay miner-als in dust provides a first-order indicator ofthe location of a dust source region (Biscaye,1965; Biscaye et al., 1997).

In the mid-continental United States, a com-bination of clay mineralogy and grain size pat-terns has been used successfully to trace thesource regions for loess deposition in Coloradoand Iowa (Aleinikoff et al., 1999; Muhs et al.,1999; Muhs and Bettis, 2000; Muhs and Bene-dict, 2006). Loess is typically associated withareas close to rivers that drained the LaurentideIce Sheet, but it is also distributed widely overthe central Great Plains of Nebraska, Kansas,and Colorado. Geochemical analyses (e.g., con-centrations of K2O, Rb, TiO2, and Zr) of thePeorian Loess in western Iowa indicate that thisaeolian material was likely derived directly fromthe Missouri River Valley during some periods,and also from regions west of the MissouriRiver in eastern Nebraska during the latterstages of deposition (Muhs and Bettis, 2000).Analyses of the trace elements Ti, Nb, Zr, Ce,and Y on silts found in the soils in the ColoradoFront Range (Muhs and Benedict, 2006) indi-cate that some portions of these soils are notderived from the parent rocks just below, butrather are aeolian in nature and must have beentransported from a more western basin duringthe early-to-mid Holocene Period (Muhs andBenedict, 2006).

In places where clay mineralogy is not suffi-cient for determining a unique source, naturaland radiogenic isotope tracers can be used totrack the different parent lithologies of the sourceareas (for a recent review, see Grousset and Bis-caye, 2005). The isotopic ratios of Sr (87Sr/86Sr),Nd (143Nd/144Nd), and Pb (206Pb/ 207Pb) havebeen used extensively to detect provenance in themid-continental United States (Aleinikoff et al.,1999), in the North Pacific Ocean (Pettke et al.,2000), off Africa (Grousset et al., 1998), inAntarctica (Basile et al., 1997; Delmonte et al.,2004a, b), and Greenland (Biscaye et al., 1997;Svensson et al., 2000; Bory et al., 2002, 2003a, b).These radiogenic isotopes are particularly ef-fective because of the range of isotopic valuesfound between relatively young, volcanic rocks,radiogenically older, crustal-derived materials,and parent soils that comprise a mixture of thetwo. Used together, the isotope signatures of Sr,Nd, and Pb can help one to discriminate betweendifferent aeolian sources.

The use of isotopic fingerprinting has beenparticularly effective in identifying the sourcesof dust deposited at the polar ice caps. In

Antarctica, the Sr and Nd isotope signatures ofdust have demonstrated that southern SouthAmerica, New Zealand, and regions of theAntarctic Dry Valleys might all serve as sourcesto the Vostok and EPICA Dome C ice cores(Delmonte et al., 2004a). However, the authorsuse the small size of the latter two source areasplus the glacial circulation to reason that south-ern South America remains the most likelysource of dust. In Greenland, several studieshave used clay mineralogy, Sr, Nd, and Pb iso-topes to pinpoint Asian deserts as the source ofdust both during cold stadial and warm inters-tadial periods based on the isotopic character-ization of the dust (Biscaye et al., 1997;Svensson et al., 2000).

4.13.2.3.2 Grain size indicators

The grain size distribution of deposited dustis frequently used in paleoclimate studies as anindicator of distance of deposited dust fromthe dust source, changes in wind strength thatcarries dust particles, or some combination ofthe two. Distance-related sorting of dustparticles is simply illustrated by the grain sizedistributions of deposited dust from differentpaleoenvironments. Near-source sedimentarydeposits such as loess are relatively coarse-grained with diameters of 20–60 mm (see, e.g.,Muhs and Bettis, 2003). In contrast, dustdiameters in marine sediments located at a fur-ther distance from source areas range from 0 to8 mm (Rea, 1994). Aeolian deposits from theGreenland ice cores, that are located largedistances from source areas, have grainsize distributions ranging from 0.4 to 2.0 mm(Steffensen, 1997). Modal grain sizes rangefrom 1.7 to 2.3 mm in diameter in Antarcticcores, with maximum sizes smaller than 5 mm(Delmonte et al., 2004b).

The relationship between the mean grain sizeof particles and carrying capacity of wind as-sociated with wind strength has been demon-strated using aeolian sediments collected insediment traps in the Indian Ocean (Clemens,1998), where strong positive correlations areobserved among median grain size, overlyingwind speed, and regional pressure gradients.Long-term trends in aeolian grain size in IndianOcean marine sediments show a decreasingtrend between 3.0Ma until today, whichClemens (1998) interprets to indicate a steadydecrease in the intensity of the winds associatedwith the southwest monsoon. At great dis-tances from source areas in Greenland andAntarctica, changes in the strength of the wind-carrying sediments have often been inferredfrom changes in sediment grain size. The mean

Geologic Variability of the Dust Cycle 9

grain size diameters from Greenland ice coreshave been shown to increase during cold, sta-dial periods (e.g., the Younger Dryas or LastGlacial Periods), in comparison with relativelywarmer periods such as the early-to-lateHolocene (Steffensen, 1997; Zielinski andMershon, 1997). These grain size increases arecorrelative with substantial increases in dustdeposition and suggest a possible enhancementof wind intensities during the colder climaticperiods.

Grain size distributions of dust in ice coresare not always simply related to changes inwind speed and can also result from changes inthe transit time of dust. Recent analyses fromthe EPICA Dome C core in Antarctic haveshown finer grain sizes at the last glacialperiod compared with data from the Holocene(Delmonte et al., 2002), in contrast to resultsfrom other east Antarctic cores (Thompson,1977; Petit et al., 1981; Briat et al., 1982; DeAngelis et al., 1984). Although the dust isotopicsignature suggests that the dust at these coresshare a common provenance, Delmonte et al.(2004a) propose that the finer-grained dust atthe EPICA Dome C site was transported viahigh-altitude air masses that resulted in alonger transit time than for dust at the otherAntarctic sites.

In actuality, the grain size distribution of at-mospheric dust particles can be affected byseveral additional factors, including the grainsize distribution of soils at the source area, thedeflation process, the nature of the deposition,and postdepositional changes. For example,higher surface wind speeds can cause saltatingsoil particles to dislocate smaller dust particlescompared with lower wind speeds. This occursbecause the high kinetic energy of saltatingparticles overcomes the binding energies of clayaggregates (as shown in experimental studies byAlfaro et al., 1998). Higher surface wind speedsin source areas thus can cause enhanced dislo-cation of both smaller and larger particlescompared with particle sizes dislocated at lowerwind speeds (Grini and Zender, 2004; Alfaroet al., 1998; Marticorena and Bergametti,1995). This effect should be taken into consid-eration for interpreting the particle size infor-mation in aeolian sediments.

The nature of the deposition process can alsostrongly influence the grain size distributionthat is recorded. In the atmosphere, the relativecontributions of wet versus dry deposition canalso affect the grain size distribution of the dustdeposited (e.g., Gillette et al., 1974; Johnson,1979). Studies in Australia have shown that thedust deposited by dry deposition has a uni-modal grain size distribution centered around12–13 mm (Hesse and McTainsh, 1999). In

contrast, indiscriminate scavenging of all par-ticle sizes occurs during rainfall events, whichresults in the observed bimodal distribution ofdeposited dust, with peaks centered at 3 and10.5 mm (Hesse and McTainsh, 1999). Oncedust has entered the ocean, hemipelagic con-tamination from river plumes, ice rafting, andturbidites, and bottom water currents can com-pletely alter the sediment grain size distribution(e.g., Rea, 1994). Thus, caution must be exer-cised in interpreting grain size distributions be-cause deciphering a single cause of changes ingrain sizes is not always straightforward.

4.13.3 GEOLOGIC VARIABILITY OFTHE DUST CYCLE

What changes have been observed overmillions of years; what do we know aboutthe regular ‘‘cycles’’ of dust; and what detailedinformation about the dust cycle can we re-construct about the last glacial period? Thegeologic record of dust has provided a wealthof information about changes in dust relative topast climates.

4.13.3.1 Dust over Long Time Periods

Terrestrial loess deposits and marine sedi-ments both provide important informationabout the changes in atmospheric dust activityover the last 12Myr (Pettke et al., 2000), asso-ciated with the uplift of the Tibetan Plateau.The longest aeolian sequences on land are fromthe Chinese Loess Plateau. These loess–paleo-sol sequences provide a relatively continuousrecord of changes in moist and arid conditionsassociated with fluctuations in the East Asianmonsoon over the last 2.6Myr (see, e.g., An,2000), and more recent evidence suggests thatthe underlying Red Clay sequence is alsoaeolian (Zheng et al., 1992; Ding et al., 1998;Guo et al., 2001), dating to as early as 8.35Ma(Qiang et al., 2001).

The onset of the Red Clay Sequence, whichindicated enhanced aridity over the Asian con-tinent, was also marked by a sudden increase inaeolian deposition in the North Pacific region(Zheng et al., 2004). A more recent phase ofthe uplift of the Tibetan Plateau (ca. 3.6Ma)effectively blocked Indian moisture sourcesfrom reaching East Asia and created condi-tions favorable for the development of theEast Asian monsoon, which modulates thesouthward flow of cold, dry air masses acrossthe East Asian continent in winter. As glacialcycles intensified, they enhanced conditions of

10 Record of Mineral Aerosols and Their Role in the Earth System

stronger winds and higher aridity that favordust emission during cold periods. Compari-sons between loess and marine sedimentrecords downwind of Asian deserts in theNorth Pacific Ocean suggest that dust deposi-tion was enhanced by an order of magnitudefollowing this later uplift event (Pettke et al.,2000; Zheng et al., 2004). The mineralogy andisotope chemistry of the deposited dust sug-gests that the source areas (basins north of theTibetan Plateau as well as the Gobi desert)have remained basically the same since theonset of aridification, but have experienceddifferent intensities of chemical weatheringfollowing the initiation of major NorthernHemisphere glacial cycles (Pettke et al., 2000).

The onset of Northern Hemisphere glaciat-ion has also affected the aridification andaeolian activity associated with Africa (De-Menocal, 2004). Marine sediment records ofdust from both the North Atlantic and ArabianSea also suggest a shift from a predominantlymonsoon-driven climate toward more arid con-ditions after 2.8Ma, modulated by the perio-dicity of Northern Hemisphere glaciations(DeMenocal, 1995). DeMenocal (2004) associ-ates this increased aridity with the intensifica-tion of Northern Hemisphere glaciations,which resulted in cooler North Atlantic sur-face ocean temperatures that inhibited the ex-pansion of the North African summermonsoon into the African continent, therebyincreasing Saharan and Sahelian aridity.

4.13.3.2 The Distribution of Dust overthe Last Glacial Period

Much information has been collected on dustdeposition during the last glacial period. Thesedata have been compiled using records from icecores, marine sediments, and terrestrial recordsas part of the Dust Indicators and Records ofTerrestrial and MArine Paleoenvironments(DIRTMAP) project (Kohfeld and Harrison,2001), which was established to archive dustinformation for the last 130,000 years. Themajority of this database is dedicated to the es-timation of mass accumulation rates of dust fordifferent time periods, for easy comparisonwith model simulations of the dust cycle (e.g.,Mahowald et al., 1999, 2006; Reader et al., 2000;Tegen et al., 2002; Werner et al., 2003).

In addition to including some new marinesediment sites in the Equatorial Pacific Ocean(Anderson et al., 2006) and new results fromrecent ice cores (Delmonte et al., 2002, 2004a),the revised version of DIRTMAP providedhere includes efforts to incorporate mass

accumulation rates of last glacial loess depositsdistributed around the world (Bettis et al., 2003;Chlachula, 2003; Eden and Hammond, 2003;Frechen et al., 2003; Hesse and McTainsh,2003; Kohfeld and Harrison, 2003; Muhs et al.,2003; Zarate, 2003). These data were recentlypublished in a special issue of Quaternary Sci-ence Reviews (Derbyshire, 2003) and then com-piled and included for a first time in comparisonwith model simulations of the last glacial period(Mahowald et al., 2006). One difficulty in com-paring loess accumulation rates with modelsimulations is that typical loess deposits show aparticle size mode in the range 20–60mm (Pye,1987), which is larger than the grain sizes sim-ulated by dust models of typically up to 10mm.To facilitate a more realistic comparison withtheir particular model simulation, Mahowaldet al. (2006) also estimated the fraction of loesssediments with grain size diameters smaller than10mm. However, here we present the bulk massaccumulation rates (Figures 3a and 3b).

The different records of glacial–interglacialchanges in mineral dust deposition agree inshowing increased dust fluxes during glacialperiods. Ice core records, which are the longestdistance from source regions, have demon-strated a 2–20-fold increase in depositionduring the Last Glacial Maximum (LGM)compared to today (Figure 3c). Various pat-terns of deposition have been observed in ma-rine sediments, but the overall change has beenabout two- to fivefold increase at the LGM.Land records have also demonstrated largechanges in deposition rates over the last periodof deglaciation. Highest rates of depositionduring the last glacial period are recordedfrom Peoria Loess in the mid-continentalUnited States, where deposition rates were ashigh as 17,500 gm�2 yr�1 during the last glacialperiod (Bettis et al., 2003), and B3,500–11,500 gm�2 yr�1 on the deglaciation in centralNebraska (Roberts et al., 2003). For compar-ison, the largest accumulation rates from theChinese Loess Plateau were on the order of1,000 gm�2 yr�1 (Sun et al., 2000; Kohfeld andHarrison, 2003). Comprehensive estimates ofchanges in mass accumulation rates of dusthave demonstrated approximately a threefoldincrease over the Chinese Loess Plateau be-tween the last glacial period and late Holocenetimes (Sun et al., 2000; Kohfeld and Harrison,2003), a pattern that is repeated over multipleglacial–interglacial cycles (Sun and An, 2005).

The LGM has also been a major focus fordust modeling because both aeolian depositionrates and boundary conditions (e.g., insolationforcing, ice sheet extent, and atmospheric tracegas concentrations) are known with some con-fidence. Incorporating expanded dust source

Modern Records of Dust 11

areas resulting from less extensive vegetationcover during glacial climate conditions in-creased the modeled atmospheric dust loadsat high latitudes by factors of 1.4 up to 2.5,consistently with the high dust fluxes observedin polar ice cores (Mahowald et al., 1999, 2006;Werner et al., 2003). The potentially large effectof dust aerosol supplied by glacial outwash hasbeen difficult to incorporate into global mod-eling schemes, largely because of the relativelysmall, subgrid-scale nature of glaciogenicprocesses. Using an inverse modeling scheme,Mahowald et al. (2006) infer that these glac-iogenic sources increase glacial dust emissionsby almost 60% over simulations that do notaccount for these sources. Still, potential in-creases in the extent of dry lake bed areasduring the last glacial period, caused by theweakened hydrological cycle, have not yet beenfactored into global dust models, and thesemay also enhance dust emissions during glacialclimates.

4.13.3.3 Glacial–Interglacial Cycles

The geologic record of mineral aerosols alsodemonstrates some intriguing temporal rela-tionships between dust and atmospheric CO2,as recorded in air bubbles of polar ice cores(Petit et al., 1999). Close comparison betweenthe dust concentration and atmospheric CO2

concentrations in the Vostok ice core revealthat dust concentrations do not even begin toincrease until atmospheric CO2 values reach acritically low level of B220 ppm. In otherwords, atmospheric CO2 levels seem to dropby B50 ppm below interglacial levels beforedust concentrations even began to increase.Further examination of marine sediments dem-onstrates that this relationship may not be lim-ited to the Vostok ice core, and may possibly begeneralized to the entire ocean (Figure 4).

This relationship between CO2 and dust isrepeated over multiple glacial–interglacialcycles and is likely to represent an importantrecurring feedback between climate, CO2, anddust. Model simulations of the last glacial pe-riod using the ECHAM3 climate model havedemonstrated a significant increase in theareal extent of dust sources. This was a resultof the response in plant evapotranspiration(and consequently vegetation cover) to changesin atmospheric CO2 on glacial–interglacialtimescales (Harrison and Prentice, 2003). Sen-sitivity studies suggested that about one-half ofthe increase in glacial dust emissions were thusdue to the glacial reduction in atmosphericCO2 (Mahowald et al., 1999). The remainingchanges in dust emissions were due to changes

in vegetation cover in response to changed cli-mate variables such as temperature andvegetation, as well as changes in wind speedand soil moisture that directly impact on dustemission fluxes. Recent simulations with theclimate model CCM3 have included the effectsof lower carbon dioxide on vegetation at thelast glacial period, and resulted in a 35% in-crease in source areas. In simulations withoutthe effects of lowered CO2, including only theinfluence of glacial climate on vegetation,source area extent is only 15% greater thantoday (Mahowald et al., 2006). These simula-tions suggest a mechanistic response betweenclimate, atmospheric CO2, and dust (see, e.g.,Ridgwell and Watson, 2002).

4.13.4 MODERN RECORDS OF DUST

While geologic data provide some interestinginsights into past changes in the dust cycle, amore extensive range of data describing differ-ent aspects of the modern dust cycle are avail-able and can be used to constrain recent changesin the global dust cycle, and provide more de-tailed information about changes in dust emis-sion, source area, transport, and deposition. Forexample, various in situ and remote-sensing ap-plications allow quantification of several aspectsof the atmospheric dust cycle.

4.13.4.1 Source Areas

Remote sensing with satellite instrumentsprovides information on vegetation cover andsurface properties that are a prerequisite foridentification of dust source areas. Maps ofsurface conditions that are derived from satel-lites are used extensively in locating areas thatare likely to act as preferential dust source ar-eas, and these maps are frequently used as inputparameters for modeling purposes (see Section4.13.6). However, there are very few direct andspatially extensive measures of dust emissions.

Changes in visibility in response to atmos-pheric dust storms are qualitative indicators ofchanges in atmospheric dustiness close to sourceregions. A dust storm is defined as an event thatreduces atmospheric visibility to o1 km due tothe presence of dust. Several studies have useddust storm frequencies to describe wind erosiontrends in specific regions (e.g., Changery, 1983;Middleton, 1984, 1986a, b; Wheaton andChakravarti, 1990; Goudie and Middleton,1992). Only one study (Engelstaedter et al.,2003) has examined the global distribution ofdust storms, using a climatological average ofannual average dust storm frequency,

20

8

6

4

2

0

25

20

15

10

5

0

0.016

0.012

0.008

0.004

0.000

15

10

5

10

8

6

4

2

01.8

1.4

1.0

0.6

0.2

280

260

240

220

200

1800 50

Age (ka)

100 150 200

0 50

North Pacific

Tropical Atlantic

g m

−2 y

r−1g m

−2 yr −1g m

−2 yr −1g

m−2

yr−1

Dus

t con

c. (

ppm

)

Dust flux (g m

−2 yr −1)

CO

2 (p

pm)

Indian Ocean

Australia

Vostok Dust

Vostok CO2

Age (ka)

100 150 200

12 Record of Mineral Aerosols and Their Role in the Earth System

Modern Records of Dust 13

approximately covering the time period 1970–1990 (Figure 5). That study demonstrated astatistical relationship between annual averagedust storm frequency and land-surface condi-tions, in spite of the potential impact of atmos-pheric transport on visibility measurements.This relationship was also utilized later byTegen et al. (2004a) to reassess the contributionof agricultural sources to the global dust emis-sions. However, this climatology does not allowfor the examination of temporal changes in duststorm frequencies. Furthermore, in some casesthe amount of data is too limited to producestatistically significant results, for example, forregional comparisons or when trying to estab-lish relationships between dust storm frequencyand degree of cultivation or soil protection(Tegen et al., 2004a, b). Nevertheless, the cur-rently available meteorological station datahave the potential to provide a global time se-ries of changes in dust storm frequencies thatcan be compared with model simulations overthe past 50–100 years (Figure 5).

4.13.4.2 Transport

Modern satellite data retrievals provide amore direct means of measuring changes in dusttransport than individual in situ measurementsof dust concentration or deposition fluxes.However, quantification of dust optical thick-nesses from such satellite retrievals requires apriori assumptions on dust optical properties,surface albedo, and the presence of cirrus clouds.In situ ground-based measurements offer the op-portunity for long-term measurements that aredifficult to achieve for satellite instruments.

Over the oceans, aerosol optical thicknessproducts are derived from different satelliteinstrument retrievals, including Meteosat,AVHRR, SeaWifs, POLDER, MODIS, andMISR (for an overview see, e.g., Yu et al.,2006). Generally, satellite measurements oflight at specific wavelength that has been re-flected by airborne particles and scatteredback to space are used to quantify aerosolproperties in cloud-free areas. This retrieval is

Figure 4 Changes in dust inputs to the ocean over thecores, as compared with dust deposition and changes incore. Shaded regions indicate time periods when atmosand coincide with periods of highest dust inputs to thefluxes are from North Pacific cores H3571 (pink, 34.91N(purple, 32.31N, 157.81E, 2612m; Kawahata et al., 199921.01W, 3082m; Tiedemann et al., 1994), and ODP 6Indian Ocean core RC27-61 (16.631N, 59.861E, 1806mSO36-61 (yellow, 30.61S, 161.41E, 1340m; Hesse, 199Hesse, 1994), LH3166/NGC97 (35.51S, 161.01E, 3166m(brownish green, 40.31S, 168.31E, 910m; Hesse, 1994).

the Advancement of Science f

impeded by light reflected from the Earth’ssurface, which can cause considerable uncer-tainties for aerosol retrievals above soils withhigh and variable surface reflectances. Overoceans with low surface albedo, aerosol op-tical thicknesses can be accurately retrievedfrom remote sensing. The estimate of dustoptical thickness from these retrievals is com-plicated by the presence of other aerosols likesea-salt particles or anthropogenic sulfatesand carbonaceous particles. Over land, quali-tative distributions of aerosols absorbing atUV wavelengths are observed by the TotalOzone Mapping Spectrometer (TOMS) satellite(Hermann et al., 1997) and more recently, OzoneMapping Instruments (OMI, aboard the Aurasatellite). Another instrument allowing for theretrieval of aerosols over some land surfaces isthe Multiangle Imaging Spectro Radiometer(MISR) Instrument aboard the Terra satellite(Kahn et al., 2005), which provides additionalinformation by multiangular retrievals.

In addition to views from the air, the opticalthickness of dust in the atmosphere can beestimated from the Earth’s surface, using Sunphotometers. The Aerosol Robotic Network(AERONET: Holben et al., 1998) includesseveral hundred sites where sun and sky radi-ances are measured by intercalibrated sun pho-tometers at four wavelengths (Figure 6). Theretrieval algorithm inverts measurements of aer-osol optical thickness and diffuse sky radiationand derives aerosol particle size distribution inthe range from 0.05 to 15mm and the complexrefractive index at the four wavelengths (Dubovikand King, 2000). Several sites are located in ordownwind of desert areas, and information fromthese sites can be used to determine the temporalvariation of dust optical thickness and opticalproperties. AERONET retrievals are a usefulground validation tool for satellite retrievals andmodel estimates of aerosol optical thickness.

Although not a direct measure of transport,measurements of the atmospheric concentra-tions of dust, using mesh sampling and high-volume dust samplers, can provide an ongoingrecord of changes in dust concentrations over a

last 200,000 years as measured in marine sedimentatmospheric CO2 levels measured in the Vostok icepheric CO2 concentrations dropped below 220 ppm,Vostok ice core and each ocean basin. Terrigenous, 179.71E, 3571m; Kawahata et al., 2000) and S2612); tropical Atlantic sites ODP 659 (light blue, 18.11N,63 (blue, 1.21 S, 11.91W, 3708m; Schneider, 2002);; Clemens and Prell, 1990), and Tasman Sea cores4), C1-86-6GC3 (green, 32.981S, 160.01E, 1540m;; light green, Kawahata et al., 1999, 2000), and E26.1Adapted by permission of American Association forrom Kohfeld et al. (2005).

0 1 2 3 4 5 6 7 8 9 10 20 30 40 50 60 70 80 90 100 110

Simulated surface layer "dust concentration"

Top: days yr−1 Bottom: µg dust m−3 air

Dust storm frequency for visibility <1 km

Figure 5 Climatology of annual dust storm frequency (days yr�1) averaged from B1971 to 1996 (Engel-staedter et al., 2003) provides one semiquantitative means of approximating dust emissions. Here they are

compared with output from a dust simulation of Tegen et al. (2002).

14 Record of Mineral Aerosols and Their Role in the Earth System

given site. Dust concentration information hasbeen collected at several stations, as part of anetwork of mostly remote island sites (Pros-pero, 1996). These unique and very valuablelong-term records have revealed such things asa long-term change in the transport of dustfrom Africa to the North American continent.Regional projects such as the InteragencyMonitoring of Protected Visual Environments(IMPROVE) project in the United States havealso attempted to measure changes in the at-mospheric particulates. The IMPROVE projectendeavores to quantify the soil dust contents of

the atmosphere by measuring the total particlemass concentrations (as PM10 and PM2.5,which refers to particle mass of the particleswith sizes of o10 and 2.5mm, respectively). Theprogram also approximates the total soil dustconcentration from the trace metal compositionof the collected samples. These analyses suggestthat as much as 40% of the atmospheric par-ticulates smaller than 2.5mm in diameter in thesouthwestern United States are comprised ofsoil dust (Malm et al., 2004).

In conclusion, in situ measurements allowquantification of dust properties, but have

−120 −60 0 60 120

−120 −60 0 60 120

−60

−30

0

30

60

−60

−30

0

30

60

ISH met stations

AERONET sun photometer

IMPROVE

U Miami dust conc. network

USGS sediment traps

Figure 6 Summary of data available atmospheric optical thickness, visibility, dust particle concentration, and sediment deposition. These include the Integrated SurfaceHourly meteorological station data available from the National Climatic Data Center (http://www4.ncdc.noaa.gov, June 2, 2006), AERONET sun photometer network(Holben et al., 1998); Visibility, PM10, and PM2.5 measurements from IMPROVE stations started before 1995 (Malm et al., 2004); a global network of atmospheric dust

concentration measurements (Prospero, 1996); and data from dust sediment traps (Reheis, 2003).

16 Record of Mineral Aerosols and Their Role in the Earth System

usually only sparse spatial and temporal cover-age. Reliable quantification of dust propertiescan only be achieved by combining several typesof measurements and modeling (Figure 6).

4.13.4.3 Deposition

There are several means by which moderndust deposition can be measured, includingland-based sediment traps (Reheis and Kihl,1995), marine sediment traps (e.g., Honjo et al.,2000), and snow samples in polar environmentswhere dust falls directly on snow and ice (e.g.,Bory et al., 2002, 2003a, b). Few measurementsdetermine both dry and wet deposition of dustparticles simultaneously, which reduces the use-fulness of such data (Mahowald et al., 2005).

One difficulty of using land-based traps isthe issue of calibrating the different trap archi-tectures to demonstrate that quantitativeestimates of dust fluxes are consistent. How-ever, within a particular region using self-consistent traps, these traps have been able toprovide examples of seasonal-to-interannualvariability in dust deposition (e.g., Derbyshireet al., 1998; Okin and Reheis, 2002). Thus,although difficult to interpret in the context ofa global database, land-based traps haveproved to be useful for understanding proc-esses controlling regional dust cycle dynamics.

Marine-based sediment traps not only pro-vide a means of measuring dust fluxes to theocean surface, but also have complications.First, the terrigenous component must be iso-lated from the total flux of material using thesame procedure as with marine sediments toremove the several biogenic minerals producedby organisms in the surface waters. As a result,sediment trap materials share the same uncer-tainties, especially at low fluxes. Furthermore,trace elements such as 232Th show the samepromise for reconstructing low dust fluxes(see Section 4.13.2.1).

A second issue is that dust that falls on theocean surface must then be exported to thedepth of the sediment trap. Phytoplankton andzooplankton activity can influence the aggregat-ion process of dust particulates, and thusbiogenic productivity can affect the seasonaltiming of when these minerals are exported tothe traps below (Bory and Newton, 2000; Boryet al., 2001). Furthermore, minute dust particlescan be horizontally advected or redistributed inthe water column, before reaching the sedimenttrap. Several modeling (Siegel and Deuser,1997) and observational (Gardner et al., 1997)studies have attempted to quantify the extent ofthis problem, by taking into account particlesinking rates and horizontal velocities. The

observational data from the Vema Channelsuggested that while the horizontal fluxes pastdifferent traps varied by a factor of 37, thequantity collected by the traps deployed at var-ious depths differed by only a factor of 1.4(Gardner et al., 1997). Alternatively, a model-ing study suggested that for traps moored atwater depths of 1,500m, 95% of the collectedmaterials fall within 600km of the trap. Thirtypercent of the material comes from within200km (Siegel and Armstrong, 2002). Thesedistances lie in the order of magnitude of asingle grid cell of a global transport model,which usually has a length of 200–500km in onedimension. Thus, as with all data, it is impor-tant to be aware of the potential limitationswhen interpreting these dust flux data.

A third means of measuring dust depositionis through quantification of the particulatesthat land on ice and snow, both in polar andalpine regions. Quantification of depositionfrom these regimes requires a precise estima-tion of snow accumulation rates. Furthermore,chemical analyses on mineral dust can be com-plicated in regions that experience seasonalmelt layers. Nevertheless, snow pit studies inlocations such as Greenland have collected suf-ficient quantities of snow to isolate seasonalchanges in Asian dust sources to the ice sheet(Bory et al., 2002, 2003a, b).

4.13.5 MODERN VARIABILITY OFTHE DUST CYCLE

Quantifying changes in modern variability ofdust emissions and transport requires detaileddatasets appropriate for addressing the changeswe wish to study. The previous section revie-wed several types of data for assessing the in-ter- and intra-annual variability of atmosphericdust. Currently, there exists no global datasetthat allows both a data and modeling analysisof systematic changes in dust emissions overthe past several decades.

Nevertheless, enough data have been col-lected to document changes in the processesassociated with the dust cycle that have variedover the last 100 years. Changes in dust emis-sions can be modified by anthropogenic influ-ences either directly by land use, or indirectlythrough changes in surface wind speed, thehydrological cycle, or vegetation changes as aconsequence of anthropogenic climate changes(Zender et al., 2004). Model results suggest thatthe emissions and atmospheric dust loadingsince preindustrial times may have altered con-siderably due to changes in climate and vegetat-ion cover (Mahowald and Luo, 2003) Theauthors of this study conclude that the dust

Filling in the Gaps: The Role of Global Dust Cycle Models 17

loads may have increased by up to 24% or de-creased up to 60% for different climate andland-use scenarios. The uncertain and sparseobservations available for validating thoseresults make it difficult to decide which of thescenarios is the most probable.

Some regions have experienced dramaticchanges in dust production, such as in the NorthAmerican dust bowl region (e.g., Goudie andMiddleton, 1992), the Sahel region of WestAfrica (Mbourou et al., 1997; Prospero andLamb, 2003), and Australia (McTainsh et al.,1998). Recent dust storm frequency analyses inChina have also shown that the 1950s–1960swere much dustier than today (Zhao et al.,2004). Thus, the modern records providevidence of interdecadal variability in dustproduction.

During the past decades, there have beenseveral documented changes in regional landsurface conditions such as vegetation cover,and trends in temperature and precipitationthat are superimposed on observed climatefluctuations such as the El Nino/Southern Os-cillation and the North Atlantic Oscillation(Houghton et al., 2001), which can be related tochanges in the atmospheric dust content. Forexample, Okin and Reheis (2002) use dust fluxdata to support a correlation between dustevents in the southwestern United States andthe previous year’s winter temperature anom-alies from the Equatorial Pacific Ocean. Theysuggest that reduced precipitation duringLa Nina years results in less vegetation coverand enhanced dust activity in the southwesternUnited States. However, their results alsosuggest that enhanced precipitation and floo-ding associated with the strongest El Ninoyears results in a renewal of reworked materialto closed basins, making it subsequently avail-able for transport.

Multiyear dust aerosol retrievals from theMeteosat satellite indicate a correlation of dusttransport toward the Mediterranean and theNorth Atlantic with the North Atlantic Oscil-lation (NAO) (Moulin et al., 1997). Recentmodel simulations also suggest that interannualvariability in winter dust production in the Sa-hel desert region is likely correlated to changesin the NAO, but such variability is not apparentin dust concentration records in Barbadoswhich integrate both emission and transportchanges (Ginoux et al., 2004). Testingthe significance of these correlations is prob-lematic due to the decadal-scale variability ofthe NAO and short timescale of most dustrecords. In contrast, Moulin and Chiapello(2004) and Chiapello et al. (2005) find observat-ional evidence of a connection between multi-year satellite retrievals from the TOMS and

Meteosat instruments, the dust concentrationrecord at Barbados, and the Sahel drought.Analysis of the potential connections betweenthese factors requires an integrated approachcombining the different observations and modelresults.

4.13.6 FILLING IN THE GAPS: THEROLE OF GLOBAL DUST CYCLEMODELS

Most measurements of dust properties pro-vide only patchy and discrete information(both in time and space). This patchiness com-plicates any global assessment of the spatialand temporal variability in the emissions, theatmospheric burden, transport, and depositionof dust. Global dust models allow us to fill inthe gaps. When validated against the existingdata, global models allow us to assess changesin source areas. They can account for theintegrated impact of wind speeds, transport,and deposition processes on large-scale spatialdeposition patterns. Finally, global dust mod-els can provide a first-order estimate onchanges in dust emissions and the atmosphericburden of dust for geologic, modern, andfuture time periods.

To better understand processes that controldust emission and deposition, quantify atmos-pheric loads, and estimate dust effects, not onlyfor present-day but also for past climates orprojections of future changes, increasingly com-plex regional and global-scale models of thedust cycle have been developed in the recentyears (Zender et al., 2004). Dust emissions insuch models are determined by surface windspeeds, and surface properties such as rough-ness length, vegetation cover, soil moisture, andsoil texture. Recently, developed dust modelsexplicitly prescribe seasonal vegetation cover todelimit the extent of dust sources (Werner et al.,2003) as well as use topographic depressions indry, unvegetated areas as indicators of prefer-ential sources containing fine and loose sedi-ment particles that can be easily deflated understrong wind conditions (Ginoux et al., 2001).Surface roughness retrievals from the POLar-ization and Directionality of the Earth’sReflectances (POLDER) instrument (Laurentet al., 2005) or the European Remote Sensing(ERS) microwave scatterometer (Prigent et al.,2005) are used as input parameters in large-scale dust emission models, substantially impro-ving distribution and magnitudes of calculateddust emission fluxes compared with the as-sumption of a constant surface roughness inthe emission model. Other large-scale datasetsof soil properties like texture are difficult to

18 Record of Mineral Aerosols and Their Role in the Earth System

obtain, as the most active dust sources lie inremote areas that cannot easily be accessed.

Global models are able to simulate the dis-tribution of atmospheric dust, the seasonal andinterannual changes, and the magnitude of op-tical thickness reasonably well. Major remain-ing problems in computing global dustemissions from wind deflation include theunavailability of soil properties datasets atglobal scale, and insufficient model resolution,as peak gusts that are responsible for thestrongest dust events remain largely unresolvedin large-scale global models. Parameterizationsof subgrid-scale variability in surface winds(e.g., Cakmur et al., 2004; Grini et al., 2005)have provided some improvements in compu-tations of dust emissions. Regional models ofdust emission, transport, and deposition re-solve surface wind speed variability better thanglobal models, and are well suited forinvestigations of individual dust events andcomparisons with in situ observations madeduring field experiments. Regional dust modelsexist for key regions, including the Sahara andAsia (Sokolik et al., 2001). However, while thetopography, soil conditions, and small-scaleextreme wind events are simulated more real-istically in regional compared to global models,uncertainties in the parameterization of keyprocesses and input data on meteorologicalfields and soil properties still remain.

Projections of dust emissions taking into ac-count expected changes in vegetation and me-teorology in the next 100 years yield changesranging from a 60% decrease (Mahowald et al.,2006) to a 233% increase in dust (Woodwardet al., 2005). The large difference in those studiesis mostly related to the difference in predictedexpansion or shrinking of the desert regions,from which dust aerosol can be emitted, by thedifferent vegetation models used in the studies.Mahowald et al. (2006) used an equilibrium-vegetation model and predicted a maximumreduction in dust emissions, including a sub-stantial reduction of the size of the Saharandesert. Alternatively, Woodward et al. (2005)used a dynamic, coupled vegetation–climatemodel and predicted maximum increase in dustemission, largely as a result of desertification ofthe Amazon rainforest region.

Agricultural land use and other human dis-turbances such as off-road traffic constructioncan increase soil dust emission. The degree towhich anthropogenically disturbed soils add tothe regional and global dust loading is still un-clear. Comparison of model estimates withavailable observations from remote-sensingand meteorological data lead to estimates thatdisturbed soils contribute between o10% andup to 50% of the modern global dust load in the

atmosphere (Houghton et al., 2001; Tegen et al.,2004a, b; Mahowald et al., 2004). The ability ofcurrent dust cycle models to assess these changesrepresents progress in the field. At the sametime, the uncertainty in the modeled responsesdemonstrates the future challenges for assessingthe role of dust that lies ahead.

4.13.7 DUST FEEDBACKS WITHIN THEEARTH SYSTEM

4.13.7.1 Radiative Feedbacks

A major climate impact of dust aerosols is itsdirect radiative forcing (Houghton et al., 2001).Dust particles reflect part of the incoming solarradiation back to space, thus reducing theradiative fluxes reaching the surface. The signand magnitude of the impact of dust particleson solar radiation depends on dust opticalproperties, the albedo of the underlying surfaceand cloud cover (e.g., Liao and Seinfeld, 1998).Beneath dust clouds this effect can locally re-duce the direct solar fluxes more than100Wm�2 (e.g., Haywood et al., 2003a, b).For reference, the mean annual incident solarradiation is B342Wm�2 at the top of the at-mosphere. Incoming sunlight can also be ab-sorbed by dust particles, which can heat thedust layer and result in a redistributionof energy within the atmospheric column. Therelation between absorbed and reflected solarradiation depends on the size and mineral com-position of the dust particles (Sokolik et al.,2001). Recent in situ retrievals of dust opticalproperties from sun photometer measurementssuggest dust only to be weakly absorbing atsolar wavelengths (Dubovik et al., 2002).Owing to their large sizes, dust particles alsocontribute to the greenhouse effect by absorb-ing and emitting outgoing terrestrial radiation.This effect depends on the vertical distributionof dust, and is expected to be smaller than thesolar effect (Houghton et al., 2001).

Haywood et al. (2005) recently estimated alongwave radiative effect of up to þ 50Wm�2

in cloud-free regions of the Sahara. Based onour current knowledge of the optical propertiesand distribution of dust aerosol, its net ra-diative effect is expected to be negative in theglobal mean, such that an increase of atmos-pheric dust load leads to a reduction of energyinput into the earth system. This causes notonly reduced surface temperatures in the pres-ence of dust clouds, but also a weakening of thehydrological cycle due to reduced evaporationbeneath dust clouds (Miller et al., 2004). Thedirect radiative forcing by dust aerosol leads toa stabilizing of the dust-laden atmosphere andto a feedback on the dust production itself;

Dust Feedbacks within the Earth System 19

reduced surface winds reduce dust emissions ina global model by 10–20% if the dust radiativeeffect is taken into consideration (Perlwitzet al., 2001). While a positive feedback of dustforcing upon dust production and atmosphericdust load is imaginable due to reduced precip-itation enhancing the availability of dust par-ticles for deflation and reduced wet removalfrom the atmosphere, model results by, for ex-ample, Perlwitz et al. (2001) indicate that thenegative-feedback processes are dominant.

Both the regional and global radiative im-pacts of dust would be expected to change un-der different climate conditions when theoverall distribution and total atmospheric bur-den of dust are expected to be different. Earlystudies hypothesized that atmospheric dustover ice sheets with high-surface albedos inhigh-latitude regions could create localized hea-ting (Peltier and Marshall, 1995). Global modelsimulations suggested that dust over ice sheetscould result in a surface warming of as much as1–4 1C over Greenland and North America(Overpeck et al., 1996). The authors suggestedthat this heating could be sufficient to invokerapid melting at the edges of ice sheets andprovide a sufficient feedback to cause rapidclimate events that were propagated throughthe climate system. Later, investigations usingmore realistic simulations of dust suggestedthat the largest radiative impacts of dust werefound in the tropical regions, with an overallnegative top-of-atmosphere (TOA) forcing ofB2–3Wm�2 associated with the largest dustsources (Claquin et al., 2003). This forcing wasof the same order of magnitude as the radiativecooling effect associated with the glacial reduc-tion in atmospheric carbon dioxide.

In high latitudes, the TOA radiative forcingwas dominated by the negative forcing of nearly�20Wm�2 associated with albedo effects fromgreatly expanded Northern Hemisphere icesheets (Claquin et al., 2003). Only a slightly po-sitive forcing due to dust was found over high-latitude ice sheets, as atmospheric dust loadingswere very low (even though greatly enhancedcompared to interglacial time periods). One int-eresting point about these simulations is thatglobal model experiments have not been able tosufficiently capture the changes in dust associ-ated with glacial conditions. Simulated dust ac-cumulation rates have been underestimated inChina (Kohfeld and Harrison, 2003) and inparticular in the mid-continental United States,where the highest mass accumulation rates ofdust have been recorded (Roberts et al., 2003)on the deglaciation. Roberts et al. (2003)suggested that the atmospheric dust may be amissing feedback that is needed to account forthe difference between the simulated late glacial

warming in this region (e.g., Bartlein et al.,1998) and the observed vegetation and faunathat indicate that the region remained cool wellinto the deglaciation. Recent model simulationshave used an optimization approach to infer thepotential contributions of glaciogenic sourcesto atmospheric dust loadings at the LGM(Mahowald et al., 2006). The inclusion of theseglaciogenic sources results in a regional en-hancement of the TOA radiative forcing by1–5Wm�2 at the ice sheet edge in North Amer-ica (Mahowald et al., in press), but also a local-ized cooling of the same magnitude just to thesouth of the ice sheet in the region of these highdust accumulation rates. This suggests thatindeed regional radiative impacts of dust maybe large enough to impact the energy balance ofclimate and ice sheet ablation.

Very few estimates have been made of futurechanges in the radiative forcing of dust. The re-sults of future changes in globally averaged ra-diative forcing by dust estimated in theindependent studies by Woodward et al. (2005)and Mahowald et al. (in press) are actually quiteclose with an increase by þ 0.17 andþ 0.14Wm�2, respectively, but these resultsare obtained for entirely different reasons.Woodward et al. estimated a fivefold increasein TOA forcing by dust between 2000 and 2100,increasing from 0.04 to 0.21Wm�2. This in-crease in forcing is a consequence of increaseddust emissions and an originally positive dustforcing. Alternatively, Mahowald et al. (inpress) estimated a less-negative TOA dust for-cing for future climate conditions as a conse-quence of reduced dust emissions with anegative dust forcing estimate for modern cli-mate conditions. In each instance, the globallyaveraged impacts of dust are small. However,regional impacts could be substantial (e.g.,410Wm�2 change in dust forcing in theAmazon basin in the Woodward et al. (2005)estimate). Furthermore, the potential for indi-rect effects of dust on cloud formation and pre-cipitation may prove more significant. Inparticular, the role of dust aerosol particles inproviding nuclei for ice cloud formation may beimportant (DeMott et al., 2003). This is an areathat requires much future work because thevariability between different climate model sim-ulations is larger than the changes in futuredustiness themselves, and the indirect aerosoleffects require further investigation.

4.13.7.2 Biogeochemical Feedbacks

Dust has also been recognized as a potentialsource of nutrients to both terrestrial andmarine ecosystems. Several studies have

20 Record of Mineral Aerosols and Their Role in the Earth System

documented the aeolian input of dust to soils(Chadwick and Davis, 1990; Muhs et al., 1990;Tiessen et al., 1991; Swap et al., 1992; Chadwicket al., 1999; Vitousek et al., 1999; Reynolds et al.,2006). In regions such as Hawaii where soils areformed on fresh basalts, near-surface layers con-tain as much as 30% aeolian-derived quartz(Kurtz et al., 2001). Saharan dust has beensuggested as a source of potassium and calciumto topsoils in Ghana (Tiessen et al., 1991) as wellas a long-term source of trace elements to theAmazon basin (Swap et al., 1992). Recent workon the Colorado Plateau suggests that aeolianinputs provide as much as 40–80% of plant nu-trients (Reynolds et al., 2006). One implication isthat changes in dust emissions, transport, anddeposition can affect terrestrial ecosystems ongeologic timescales but also on modern ones. Forexample, Reheis et al. (2002) demonstrate thatcertain geochemical ‘‘hot spots,’’ such as OwensValley, California, have the potential to contrib-ute toxic metals to soils in the southwesternUnited States, and that the impacts of these con-tributions are still relatively unstudied.

Dust is also an important external source ofnutrients, particularly iron, to the ocean (Duceet al., 1991; Prospero, 1996; Jickells et al., 2005).Atmospheric dust has a relatively low iron con-tent (3.5%), and it is estimated that on averageo2% of that iron is soluble in seawater.However, because most of the fluvial and glaci-ally derived iron is trapped in coastal regions,aeolian dust becomes a necessary source of ironto relieve this limitation (for review see Jickellset al., 2005). Several laboratory (e.g., Martinand Fitzwater, 1988) and in situ experiments(for summary see de Baar et al., 2005) havedemonstrated the potential of Fe for enhancingphytoplankton productivity and carbon uptakein Fe-limited regions. Furthermore, observa-tions of natural dust-fertilization events (Bishopet al., 2002) have shown that Asian dust is asignificant source of iron to Pacific Oceanwaters today, and the sediment record showssignificant fluctuations in dust influxes onglacial–interglacial timescales. A long-standinghypothesis (Martin, 1990) holds that increasediron supply during glacial times invigoratedthe ocean’s biological pump, and thus drewdown atmospheric CO2. Studies invokingFe-enhancement of the biological pump focusedon the Southern Ocean because of the largeinventory of unused surface nutrients and thestrong connection between surface and CO2-richdeep waters. Several modeling and data studieshave examined the response of marine biota toiron fertilization from dust and suggested thatthis effect could account for between a 10 and40ppm reduction in atmospheric CO2 levelsduring the last glacial period (Archer et al.,

2000; Watson et al., 2000; Bopp et al., 2003;Kohfeld et al., 2005). In other words, the ironfertilization effects of dust could account for asubstantial portion of the glacial–interglacialchanges in atmospheric CO2, but it is notlikely to explain the entire 80–100ppm changeobserved in ice core records.

4.13.8 FUTURE OUTLOOK

Biogeochemical modeling studies have fo-cused mostly on the impact of iron fertilizationin the modern and glacial oceans, with little em-phasis to date on the potential biogeochemicalimpacts of future changes in dust emissions dueto climate and anthropogenic activity (Ridgwelland Kohfeld, in press). To date, the projectedchanges in dust emissions by 2100 range fromdecreases of 60% (Mahowald et al., 2006) toincreases of 233% (Woodward et al., 2005) inresponse to climate changes. An increase in dustemissions by 233%, while extreme, is of the sameorder of the glacial–interglacial changes in dustinputs to the ocean, and thus the projectedbiogeochemical impacts could be quite substan-tial. While the radiative impacts of these changeshave been studied, the biogeochemical feedbacksare only beginning to be considered in projec-tions of future climate change (e.g., Ridgwellet al., 2002; Parekh et al., 2006). Yet, how canwe improve our understanding of the dust cycleto understand even the sign of change in thefuture atmospheric dust burden?

Sophisticated models of the dust cycle arenow able to reproduce the characteristics ofglobal and regional dust distribution for mod-ern and glacial climates, the diverging results indifferent future estimates of atmospheric dustshow, however, that the connection betweendust emissions and climate are not yet suffi-ciently understood. The role of vegetationchanges and thus the changes in the extent ofdust sources remains a major uncertainty factorin dust–climate models.

One means of improving our understanding ofthe relationship between dust emissions, climate,and human activity is to expand our datacoverage in the time domain. For example,utilizing the full extent of global meteorologicalstation data (including visibility, humidity,precipitation, wind speed, and gustiness) may al-low us to quantify changes in global dust stormfrequencies over the past century. Alternatively,in the past 10 years, vast improvements havebeen made in global syntheses of geologic data ofdust deposition, but more work could be done.For example, sedimentary records of dust can benormalized using constant flux proxies to helpaccount for problems of sediment redistribution.

References 21

Use of geochemical sediment tracers such as232Th (e.g., Chase et al., 2001; Anderson et al.,2006) can also help eliminate the methodologicaluncertainties associated with isolating terrigenousmaterials from the dominant biogenic materialsfound in sediment cores, thus making recon-structions of changes in dust fluxes to the oceanmore reliable. Finally, an important goal of thesedatasets is to make them comparable to modeledoutput of the dust cycle. The DIRTMAP data-base (Kohfeld and Harrison, 2001) was one stepin this direction in that data from multiple pale-oenvironments were interpreted in terms of theirmass fluxes so that they could be compared withmodeled output. Another step in this directionhas been made by estimating the portion of ter-restrial loess deposits that can be directly com-pared with model simulations, for example, byisolating the grain sizes that are actually trans-ported by certain modeling schemes (Mahowaldet al., 2006). This effort has been one step in thedirection of understanding the role of near-source loess deposits in the global dust cycle(and its resultant impact on climate).

Another avenue for advancement in our abil-ity to simulate changes in dust sources and emis-sions lies in expanding our current observationalnetworks for measuring changes in dust emis-sions and transport. New observations of dustdistribution and properties from upcoming sat-ellite missions (as e.g., the lidar measurements ofvertical distribution of aerosol and cloud prop-erties with the Cloud-Aerosol Lidar and InfraredPathfinder Satellite Observations (CALIPSO)instrument and field experiments (such asAMMA: African Monsoon MultidisciplinaryAnalysis and SAMUM: SAharan Mineral dUstExperiMent, in the southern and northern Sa-hara, respectively, with field measurements dur-ing 2006) will greatly enhance our knowledge ondust properties and dust production and trans-port. Global proxies of dust emissions also holdgreat promise, as does increasing efforts to usemultiple datasets to constrain model simulations(e.g., Cakmur et al., 2006).

ACKNOWLEDGMENTS

We thank D. Muhs and N. Mahowald forgraciously providing the global compilation ofterrestrial loess mass accumulation rates andother unpublished data for the improvement ofthis work.

REFERENCES

Aleinikoff J. N., Muhs D., Sauer R. R., and Fanning C. M.(1999) Late Quaternary loess in northeastern Colorado.Part II: Pb isotopic evidence for the variability of loesssources. Geol. Soc. Am. Bull. 111, 1876–1883.

Alfaro S. C., Gaudichet A., Gomes L., and Maille M.(1998) Mineral aerosol production by wind erosion:aerosol particle sizes and binding energies. Geophys. Res.Lett. 25, 991–994.

Alley R. B., Finkel R. C., Nishizumi K., AnankakrishnanS., Shuman C. A., Mershon G., Zielinski G. A., andMayewski P. A. (1995) Changes in continental and sea-salt atmospheric loadings in central Greenland duringthe most recent deglaciation: model-based estimates.J. Glaciol. 41, 503–514.

An Z. S. (2000) The history and variability of East AsianPaleomonsoon climate. Quat. Sci. Rev. 19, 171–187.

Anderson R. F., Fleisher M. Q., and Lao Y. (2006) Gla-cial–interglacial variability in the delivery of dust to thecentral equatorial Pacific Ocean. Earth Planet. Sci. Lett.242, 406–414.

Archer D., Winguth A., Lea D., and Mahowald N. (2000)What caused the glacial/interglacial atmospheric pCO2

cycles? Rev. Geophys. 38, 159–189.Bartlein P. J., Anderson K. H., Anderson P. M., Edwards M.

E., Mock C. J., Thompson R. S., Webb R. S., Webb T.,III, and Whitlock C. (1998) Paleoclimate simulations forNorth America over the past 21,000 years: features of thesimulated climate and comparisons with paleoenvironmen-tal data. Quat. Sci. Rev. 17, 549–586.

Basile I., Grousset F. E., Revel M., Petit J. R., Biscaye P.E., and Barkov N. I. (1997) Patagonian origin of glacialdust deposited in East Antarctica (Vostok and Dome C)during glacial stages 2, 4 and 6. Earth Planet. Sci. Lett.146, 573–589.

Bettis E. A., III, Muhs D. R., Roberts H. M., and WintleA. G. (2003) Last Glacial loess in the conterminous USA.Quat. Sci. Rev. 22, 1907–1946.

Biscaye P. E. (1965) Mineralogy and sedimentation of re-cent deep-sea clay in the Atlantic Ocean and adjacent seasand oceans. Geol. Soc. Am. Bull. 76, 732–803.

Biscaye P. E., Grousset F. E., Revel M., Van der Gaast S.,Zielinski G. A., Vaars A., and Kukla G. (1997) Asianprovenance of glacial dust (Stage 2) in the GISP2 icecore, Summit, Greenland. J. Geophys. Res. 102, 26315–26886.

Bishop J. K. B., Davis R. E., and Sherman J. T. (2002) Ro-botic observations of dust storm enhancement of carbonbiomass in the North Pacific. Science 298, 817–821.

Bopp L., Kohfeld K. E., Le Quere C., and Aumont O.(2003) Dust impact on marine biota and atmosphericCO2 during glacial periods. Paleoceanography 18,doi:10.1029/2002PA000810.

Bory A. J.-M., Biscaye P. E., and Grousset P. E. (2003a)Two distinct seasonal Asian source regions for mineraldust deposited in Greenland (NorthGRIP). Geophys.Res. Lett. 30, 1167.

Bory A. J.-M., Biscaye P. E., Piotrowski A. M., and Stef-fensen J. P. (2003b) Regional variability of ice core dustcomposition and provenance in Greenland. Geochem.,Geophys., Geosyst. 4, 1107.

Bory A. J.-M., Biscaye P. E., Svensson A., and Grousset F.E. (2002) Seasonal variability in the origin of recentatmospheric mineral dust at NorthGRIP, Greenland.Earth Planet. Sci. Lett. 196, 123–134.

Bory A., Jeandel C., Leblond N., Vangriesheim A.,Khripounoff A., Beaufort L., Rabouille C., Nicolas E.,Tachikawa K., Etcheber H., and Buat-Menard P. (2001)Downward particle fluxes within different productivityregimes off the Mauritanian upwelling zone (EUMELIprogram). Deep-Sea Res. I 48, 2251–2282.

Bory A. J.-M. and Newton P. P. (2000) Transport of air-borne lithogenic material down through the water col-umn in two contrasting regions of the eastern subtropicalNorth Atlantic Ocean. Global Biogeochem. Cycles 14,297–315.

Briat M., Royer A., Petit J. R., and Lorius C. (1982) Lateglacial input of eolian continental dust in the Dome C ice

22 Record of Mineral Aerosols and Their Role in the Earth System

core: additional evidence from individual microparticleanalysis. Ann. Glaciaol. 3, 27–30.

Cakmur R., Miller R., Perlwitz J., Geogdzhayev I., GinouxP., Koch D., Kohfeld K., Tegen I., and Zender C. (2006)Constraining the global dust emission and load by min-imizing the difference between the model and observat-ions. J. Geophys. Res.—Atmos. 111, Art. No. D06207.

Cakmur R., Miller R., and Torres O. (2004) Incorporatingthe effect of small-scale circulations upon dust emissionin an atmospheric general circulation model. J. Geophys.Res.—Atmos. 109, Art. No. D07201.

Cannon W. F., Dean W. E., and Bullock J. H., Jr. (2003)Effects of Holocene climate change on mercury deposi-tion in Elk Lake, Minnesota: the importance of eoliantransport in the mercury cycle. Geology 31, 187–190.

Chadwick A. O. and Davis J. O. (1990) Soil-forming in-tervals causeed by eolian sediment pulses in the Lahon-tan Basin, northwestern Nevada. Geology 18, 243–246.

Chadwick O. A., Derry L. A., Vitousek P. M., Huebert B.J., and Hedin L. O. (1999) Changing sources of nutrientsduring four million years of ecosystem development.Nature 397, 491–497.

Changery M. J. (1983) A Dust Climatology of the WesternUnited States. NUREG/CR-3211. US Nuclear Requla-tory Commission, Washington, DC, Contract No. NRC-01-78-015, 42pp.

Chase Z., Anderson R. F., and Fleisher M. Q. (2001) Evid-ence from authigenic uranium for increased productivityof the glacial Subantarctic Ocean. Paleoceanography16, 468–478.

Chiapello I., Moulin C., and Prospero J. (2005) Unders-tanding the long-term variability of African dust trans-port across the Atlantic as recorded in both Barbadossurface concentrations and large-scale Total Ozone Map-ping Spectrometer (TOMS) optical thickness. J. Geophys.Res.—Atmos. 110, Art. No. D18S10.

Chlachula J. (2003) The Siberian loess record and its signi-ficance for reconstruction of Pleistocene climate changein north-central Asia. Quat. Sci. Rev. 22, 1879–1906.

Claquin T., Roelandt C., Kohfeld K. E., Harrison S. P.,Tegen I., Prentice I. C., Balkanski Y., Bergametti G.,Hansson M., Mahowald N., Rodhe H., and Schulz M.(2003) Radiative forcing of climate by ice-age atmos-pheric dust. Clim. Dynam. 20, 193–202, doi:10.1007/s00382-002-0269-1.

Claquin T., Schulz M., and Balkanski Y. (1999) Modelingthe mineralogy of atmospheric dust sources. J. Geophys.Res. 104, 22243–22256.

Clemens S. C. (1998) Dust response to seasonal atmos-pheric forcing: proxy evaluation and calibration. Pale-oceanography 13, 471–490.

Clemens S. C. and Prell W. L. (1990) Late Pleistocenevariability of Arabian Sea summer monsoon winds andcontinental aridity: eolian records from the lithogeniccomponent of deep-sea sediments. Paleoceanography 5,109–145.

De Angelis M., Legrand M., Petit J. R., Barkov N. I., Ko-rotkevitch Y. S., and Kotlyakov V. M. (1984) Solubleand insoluble impurities along the 950m deep Vostok icecore (Antarctica)—climate implications. J. Atmos. Chem.1, 215–239.

de Baar H. J. W., Boyd P. W., Coale K. H., Landry M. R.,Tsuda A., Assmy P., Bakker D. C. E., Bozec Y., BarberR., Brzezinski M. A., Buesseler K. O., Boye0 M., CrootP. L., Gervais F., Gorbunov M. Y., Harrison P. J., His-cock M. R., Laan P., Lancelot C., Law C. S., LevasseurM., Marchetti A., Millero F. J., Nishioka J., Nojiri Y.,van Oijen T., Riebesell U., Rijkenberg M. J. A., Saito H.,Takeda S., Timmermans K. R., Veldhuis M. J. W., WaiteA. M., and Wong C.-S. (2005) Synthesis of iron fertili-zation experiments: from the Iron Age in the Age ofEnlightenment. J. Geophys. Res. 110, doi:10.1029/2004JC002601.

Delmonte B., Basile-Doelsch I., Petit J. R., Maggi V.,Revel-Rolland M., Michard A., Jagoutz E., and Grous-set F. (2004a) Comparing the Epica and Vostok dustrecords during the last 220,000 years: stratigraphical cor-relation and provenance in glacial periods. Earth Sci.Rev. 66, 63–87.

Delmonte B., Petit J. R., Basile-Doelsch I., Maggi V., andYa Lipenkov V. (2004b) Dust size evidence for oppositeregional atmospheric circulation changes over east Ant-arctica during the last climatic transition. Clim. Dynam.23, 427–438.

Delmonte B., Petit J. R., and Maggi V. (2002) Glacial toHolocene implications of the new 27000-year dust recordfrom the EPICA Dome C (East Antarctica) ice core.Clim. Dynam. 18, 647–660.

deMenocal P. B. (1995) Plio-Pleistocene African climate.Science 270, 53–59.

deMenocal P. B. (2004) African climate change and faunalevolution during the Pliocene–Pleistocene. Earth Planet.Sci. Lett. 220, 2–24.

DeMott P. J., Sassen K., Poellot M. R., Baumgardner D.,Rogers D. C., Brooks S. D., Prenni A. J., and Krei-denweis S. M. (2003) African dust aerosols as atmos-pheric ice nuclei. Geophys. Res. Lett. 30, 1732,doi:10.1029/2003GL017410.

Derbyshire E. (2003) Loess, and the Dust Indicators andRecords of Terrestrial and Marine Palaeoenvironments(DIRTMAP) database. Quat. Sci. Rev. 22, 1813–1819.

Derbyshire E. X. M. and Kemp R. A. (1998) Provenance,transport and characteristics of modern aeolian dust inwestern Gansu Province, China, and interpretation of theQuaternary loess record. J. Arid Environ. 39, 497–516.

Ding Z. L., Sun J. M., Liu T. S., Zhu R. X., Yang S. L., andGuo B. (1998) Wind-blown origin of the Pliocene redclay formation in the central Loess Plateau, China. EarthPlanet. Sci. Lett. 161, 135–143.

Dubovik O., Holben B., Eck T., Smirnov A., Kaufman Y.,King M., Tanre D., and Slutsker I. (2002) Variabilityof absorption and optical properties of key aerosoltypes observed in worldwide locations. J. Atmos. Sci. 59,590–608.

Dubovik O. and King M. (2000) A flexible inversion algor-ithm for retrieval of aerosol optical properties fromSun and sky radiance measurements. J. Geophys. Res.—Atmos. 105, 20673–20696.

Duce R. A., Liss P. S., Merrill J. T., Atlas E. L., Buat-Menard P., Hicks B. B., Miller J. M., Prospero J. M.,Arimoto R., Church T. M., Ellis W., Galloway J. N.,Hansen L., Jickells T. D., Knap A. H., Reinhardt K. H.,Schneider B., Soudine A., Tokos J. J., Tsunogai S., Wol-last R., and Zhou M. (1991) The atmospheric input oftrace species to the world ocean. Global Bio-chem. Cycles 5, 193–259.

Eden D. N. and Hammond A. P. (2003) Dust accumulationin the New Zealand region since the last glacial maxi-mum. Quat. Sci. Rev. 22, 2037–2052.

Engelstaedter S., Kohfeld K. E., Tegen I., and Harrison S.P. (2003) Controls of dust emissions by vegetationand topographic depressions: an evaluation using duststorm frequency data. Geophys. Res. Lett. 30, 1294,doi:10.1029/2002GL016471.

Frechen M., Oches E. A., and Kohfeld K. E. (2003) Loessin Europe—mass accumulation rates during the LastGlacial Period. Quat. Sci. Rev. 22, 1835–1857.

Gardner W. D., Biscaye P. E., and Richardson M. J. (1997)A sediment trap experiment in the Vema Channel toevaluate the effect of horizontal particle fluxes on meas-ured vertical fluxes. J. Marine Res. 55, 995–1028.

Gill T. E., Gillette D. A., Niemeyer T., and Winn R. T.(2002) Elemental geochemistry of wind-erodible playa se-diments, Owens Lake, California. Nuclear Instruments andMethods in Physics Research Section B—Beam Interactionswith Materials and Atoms 189, 209–213.

References 23

Gillette D. A. (1979) Environmental factors affectingdust emission by wind erosion. In Saharan Dust (ed.C. Morales). Wiley, New York, pp. 71–94.

Gillette D. A. (1999) A qualitative geophysical explanationfor ‘‘hot spot’’ dust emitting source regions. Contrib.Atmos. Phys. 72, 67–77.

Gillette D. A., Bifford I. H., Jr., and Fryrear, D. W. (1974)The influence of wind velocity on the size distributionof aerosols generated by the wind erosion of soils.J. Geophys. Res. 79, 4068–4075.

Ginoux P., Chin M., Tegen I., Prospero J., Holben B.,Dubovik O., and Lin S.-J. (2001) Global simulation ofdust in the troposphere: model description and assess-ment. J. Geophys. Res.—Atmos. 106, 20255–20273.

Ginoux P., Prospero J. M., Torres O., and Chin M. (2004)Long-term simulation of global dust distribution with theGOCART model: correlation with North Atlantic Oscil-lation. Environ. Model. Software 19, 113–128.

Givelet N., Roos-Barraclough F., Goodsite M. E., Chebur-kin A. K., and Shotyk W. (2004) Atmospheric mercuryaccumulation rates between 5900 and 800 calibratedyears BP in the high Artic of Canada recorded by peathummocks. Environ. Sci. Technol. 38, 4964–4972.

Goudie A. S. and Middleton N. J. (1992) The changingfrequency of dust storms through time. Clim. Change 20,197–225.

Grini A., Myhre G., Zender C., and Isaksen I. (2005)Model simulations of dust sources and transport in theglobal atmosphere: effects of soil erodibility and windspeed variability. J. Geophys. Res.—Atmos. 110, Art. No.D02205.

Grini A. and Zender C. S. (2004) Roles of saltation, sand-blasting, and wind speed variability on mineral dust aer-osol size distribution during the Puerto Rican DustExperiment (PRIDE). J. Geophys. Res.—Atmos. 109,Art. No. D07202.

Grousset F. and Biscaye P. E. (2005) Tracing dust sourcesand transport patterns using Sr, Nd, and Pb isotopes.Chem. Geol. 222, 149–167.

Grousset F. E., Ginoux P., Bory A., and Biscaye P. E.(2003) Case study of a Chinese dust plume reaching theFrench Alps. Geophys. Res. Lett. 30, 1277.

Grousset F. E., Parra M., Bory A., Martinez P., BertrandP., Shimmield G., and Ellam R. M. (1998) Saharan windregimes traced by Sr–Nd isotopic composition of sub-tropical Atlantic sediments: last Glacial Maximum vs.today. Quat. Sci. Rev. 17, 395–409.

Guo B., Zhu R. X., Roberts A. P., and Florindo F. (2001)Lack of correlation between paleoprecipitation andmagnetic susceptibility of Chinese loess/paleosol se-quences. Geophys. Res. Lett. 28, 4259–4262.

Harrison S. P. and Prentice A. I. (2003) Climate and CO2

controls on global vegetation distribution at the last gla-cial maximum: analysis based on palaeovegetation data,biome modelling and palaeoclimate simulations. GlobalChange Biol. 9, 983–1004.

Haywood J., Allan R., Culverwell I., Slingo T., Milton S.,Edwards J., and Clerbaux N. (2005) Can desert dust ex-plain the outgoing longwave radiation anomaly over theSahara during July 2003? J. Geophys. Res.—Atmos. 110,D05105.

Haywood J., Francis P., Dubovik O., Glew M., and Hol-ben B. (2003a) Comparison of aerosol size distribu-tions, radiative properties, and optical depthsdetermined by aircraft observations and Sun photom-eters during SAFARI 2000. J. Geophys. Res.—Atmos.108, 8471.

Haywood J., Francis P., Osborne S., Glew M., Loeb N.,Highwood E., Tanre D., Myhre G., Formenti P., andHirst E. (2003b) Radiative properties and direct radiativeeffect of Saharan dust measured by the C-130 aircraftduring SHADE. 1: Solar spectrum. J. Geophys. Res.—Atmos. 108, 8577.

Hermann J. R., Bhartia P. K., Torres O., Hsu C., Seftor C.,and Celarier E. (1997) Global distribution of UV-absorbing aerosols from Nimbus 7/TOMS data.J. Geophys. Res.—Atmos. 102, 16911–16922.

Hesse P. P. (1994) The record of continental dust fromAustralia in Tasman Sea sediments. Quat. Sci. Rev. 13,257–272.

Hesse P. P. and McTainsh G. H. (1999) Last Glacial Max-imum to early Holocene wind strength in the mid-latitudes of the Southern Hemisphere from aeolian dustin the Tasman Sea. Quat. Res. 52, 343–349.

Hesse P. P. and McTainsh G. H. (2003) Australian dustdeposits: modern processes and the Quaternary record.Quat. Sci. Rev. 22, 2007–2035.

Holben B., Eck T., Slutsker I., Tanre D., Buis J., Setzer A.,Vermote E., Reagan J., Kaufman Y., Nakajima T.,Lavenu F., Jankowiak I., and Smirnov A. (1998) AERO-NET—a federated instrument network and data archivefor aerosol characterization. Remote Sensing Environ. 66,1–16.

Honjo S., Francois R., Manganini S. J., Dymond J., andCollier R. (2000) Particle fluxes to the interior of theSouthern Ocean in the Western Pacific sector along1701 W. Deep-Sea Res. II 47, 3521–3548.

Houghton J., Ding Y., Griggs D. J., Noguer M., van derLinden P. J., Dai X., Maskell K., and Johnson C. A.(2001) Climate Change 2001: The Scientific Basis. Con-tribution of Working Group I to the Third AssessmentReport of the Intergovernmental Panel on ClimateChange. Cambridge University Press, Cambridge, UK;New York, USA.

House J. I., Brovkin V., Betts R., Constanza R., Silva DiasM. A., Holland E., Que0re0 C. l., Kim Phat N., RiebesellU., and Scholes M. (2006) Air quality andclimate, chapter 13. In Ecosystems & Human Well-Being:Volume 1 Current States and Trends (eds. R. Hassan, R.Scholes, and N. Ash). Island Press, Washington.

Iversen J. D. and White B. R. (1982) Saltation threshold onEarth, Mars, and Venus. Sedimentology 29, 111–119.

Jickells T. D., An Z. S., Andersen K. K., Baker A. R.,Bergametti G., Brooks N., Cao J. J., Boyd P. W., DuceR. A., Hunter K. A., Kawahata H., KubilayN., laRoche J., Liss P. S., Mahowald N., ProsperoJ. M., Ridgwell A. J., Tegen I., and Torres R. (2005)Global iron connections between desert dust, oceanbiogeochemistry, and climate. Science 308, 67–71.

Johnson L. R. (1979) Particle size fractionation of eoliandusts during transport and sampling. Marine Geol. 21,M17–M21.

Kahn R. A., Gaitley B. J., Martonchik J. V., Diner D. J.,Crean K. A., and Holben B. (2005) Multiangle ImagingSpectroradiometer (MISR) global aerosol optical depthvalidation based on 2 years of coincident AerosolRobotic Network (AERONET) observations. J. Geophys.Res.—Atmos. 110, doi:10.1029/2004JD004706.

Kawahata H., Ohkushi K. I., and Hatakeyama Y. (1999)Comparative Late Pleistocene paleoceanographicchanges in the mid latitude Boreal and Austral WesternPacific. J. Oceanogr. 55, 747–761.

Kawahata H., Okamoto T., Matsumoto E., and Ujiie H.(2000) Fluctuations of eolian flux and ocean productivityin the mid-latitude North Pacific during the last 200 kyr.Quat. Sci. Rev. 19, 1279–1292.

Kohfeld K. E. and Harrison S. P. (2001) DIRTMAP: thegeologic record of dust. Earth Sci. Rev. 54, 81–114.

Kohfeld K. E. and Harrison S. P. (2003) Glacial-intergla-cial changes in dust deposition on the Chinese LoessPlateau. Quat. Sci. Rev. 22, 1859–1878.

Kohfeld K. E., Le Quere C., Harrison S. P., and AndersonR. F. (2005) Role of marine biology in glacial-interglacialCO2 cycles. Science 308, 74–78.

Kurtz A. C., Derry L. A., and Chadwick O. A. (2001)Accretion of Asian dust to Hawaiian soils: isotopic,

24 Record of Mineral Aerosols and Their Role in the Earth System

elemental, and mineral mass balances. Geochim. Cosmoc-hem. Acta 65, 1971–1983.

Laurent B., Marticorena B., Bergametti G., Chazette P.,Maignan F., and Schmechtig C. (2005) Simulation ofthe mineral dust emission frequencies from desert areasof China and Mongolia using an aerodynamic roughnesslength map derived from the POLDER/ADEOS 1 sur-face products. J. Geophys. Res.—Atmos. 110, Art. No.D18S04.

Levin Z., Teller A., Ganor E., and Yin Y. (2005) On theinteractions of mineral dust, sea-salt particles, andclouds: a measurement and modeling study from theMediterranean Israeli Dust Experiment campaign.J. Geophys. Res.—Atmos. 110, D20202.

Liao H. and Seinfeld J. H. (1998) Radiative forcing bymineral dust aerosols: sensitivity to key variables.J. Geophys. Res.—Atmos. 103, 31637–31645.

Lu H. and Shao Y. P. (1999) A new model for dust emissionby saltation bombardment. J. Geophys. Res.—Atmos. 104,16827–16841.

Mahowald N., Kohfeld K. E., Hansson M., Balkanski Y.,Harrison S. P., Prentice I. C., Schulz M., and Rodhe H.(1999) Dust sources and deposition during the Last Gla-cial Maximum and current climate: a comparison ofmodel results with palaeodata from ice cores and marinesediments. J. Geophys. Res.—Atmos. 104, 15895–15916.

Mahowald N. M., Baker A. R., Bergametti G., Brooks N.,Duce R. A., Jickells T. D., Kubilay N., Prospero J. M.,and Tegen I. (2005) Atmospheric global dust cycle andiron inputs to the ocean. Global Biogeochem. Cycles 19,GB4025, doi:10.1029/2004GB002402.

Mahowald N. M. and Luo C. (2003) A less dusty future?Geophys. Res. Lett. 30, 1903, doi:1910.1029/2003GL017880.

Mahowald N. M., Muhs D. R., Levis S., Rasch P. J., Yos-hioka M., Zender C. S., and Luo C. (2006) Change inatmospheric mineral aerosols in response to climate: lastglacial period pre-industrial, modern and doubled carbondioxide climates. J. Geophys. Res.—Atmos. 111, D10202,doi:10.1029/2005JD006653.

Mahowald N. M., Rivera G. D. R., and Luo C. (2004)Comment on ‘‘Relative importance of climate and landuse in determining present and future global soil dustemission’’ by I. Tegen et al. Geophys. Res. Lett. 31,doi:10.1029/2004GL021272.

Mahowald N. M., Yoshioka M., Collins W. D., Conley A.J., Filmore D. W., and Coleman D. B. Climate responseand radiative forcing from mineral aerosols during thelast glacial maximum, pre-industrial, current and dou-bled-carbon dioxide climates. Geophys. Res. Lett. (inpress).

Malm W., Schichtel B., Pitchford M., Ashbaugh L., andEldred R. (2004) Spatial and monthly trends in speciatedfine particle concentration in the United States.J. Geophys. Res.—Atmos. 109, Art. No. D03306.

Marticorena B. and Bergametti G. (1995) Modeling theatmospheric dust cycle, 1, Design of a soil-derived dustemission scheme. J. Geophys. Res.—Atmos. 100(D8),16415–16430.

Martin J. (1990) Glacial-interglacial CO2 change: the ironhypothesis. Paleoceanography 5, 1–13.

Martin J. H. and Fitzwater S. E. (1988) Iron deficiencylimits phytoplankton growth in the north-east Pacificsubarctic. Nature 331, 341–343.

Mbourou G. N., Bertrand J. J., and Nicholson S. E. (1997)The diurnal and seasonal cycles of wind-borne dust overAfrica north of the equator. J. Appl. Meteorol. 36, 868–882.

McTainsh G. H., Lynch A. W., and Tews E. K. (1998)Climatic controls upon dust storm occurrence in easternAustralia. J. Arid Environ. 39, 457–466.

Middleton N. J. (1984) Dust storms in Australia: frequency,distribution and seasonality. Search 15, 46–47.

Middleton N. J. (1986a) Dust storms in the Middle East.J. Arid Environ. 10, 83–96.

Middleton N. J. (1986b) A geography of dust storms insouth-west Asia. J. Clim. 6, 183–196.

Miller R., Perlwitz J., and Tegen I. (2004) Feedback upondust emission by dust radiative forcing through the plan-etary boundary layer. J. Geophys. Res.—Atmos. 109, Art.No. D24209.

Moulin C. and Chiapello I. (2004) Evidence of the control ofsummer atmospheric transport of African dust over theAtlantic by Sahel sources from TOMS satellites (1979–2000). Geophys. Res. Lett. 31, L02107, doi:10.1029/2003GL018931.

Moulin C., Lambert C. E., Dulac F., and Dayan U. (1997)Control of atmospheric export of dust from NorthAfrica by the North Atlantic Oscillation. Nature 387,691–694.

Muhs D. R., Ager T. A., Bettis E. A., III, McGeehin J.,Been J. M., Beget J. E., Pavich M. J., Stafford T. W., Jr.,and Stevens D. A. S. P. (2003) Stratigraphy and palaeo-climatic significance of late quaternary loess–palaeosolsequences of the last interglacial–glacial cycle in centralAlaska. Quat. Sci. Rev. 22, 1947–1986.

Muhs D. R., Aleinikoff J. N., Stafford T. W., Jr., Kihl R.,Been J., and Mahan S. A. (1999) Late Quaternary loessin northeastern Colorado. I: Age and paleoclimatic signi-ficance. Geolog. Soc. Am. Bull. 111, 1861–1875.

Muhs D. R. and Benedict J. B. (2006) Eolian Additions toLate Quaternary Alpine Soils, Indian Peaks WildernessArea, Colorado Front Range. Arctic, Antarct., AlpineRes. 38, 120–130.

Muhs D. R. and Bettis E. A., III (2000) Geochemicalvariations in Peoria Loess of western Iowa indicatepaleowinds of midcontinental North America during lastglaciation. Quat. Res. 53, 49–61.

Muhs D. R. and Bettis E. A., III (2003) Quaternary loess-paleosol sequences as examples of climate-drivensedimentary extremes. In Extreme Depositional Environm-ents: Mega End Members in Geologic Time (eds. M. A.Chan and A. W. Archer), Geological Society of AmericaSpecial Paper 370. Boulder, CO, pp. 53–74.

Muhs D. R., Bush C. A., Stewart K. C., Rowland T. R.,and Crittenden R. C. (1990) Geochemical evidence ofSaharan dust parent material for soils developed onQuaternary limestones of Caribbean and western Atlan-tic Islands. Quat. Res. 33, 157–177.

National Climatic Data Center, http://www4.ncdc.noaa.gov/ol/documentlibrary/datasets.html (accessed June 2,2006).

Nordstrom K. F. and Hotta S. (2004) Wind erosion fromcropland in the USA: a review of problems, solutions andprospects. Geoderma 121, 157–167.

Okin G. S. and Reheis M. C. (2002) An ENSO predictor ofdust emission in the southwestern United States.Geophys. Res. Lett. 29, doi:10.1029/2001GL014494.

Overpeck J., Rind D., Lacis A., and Healy R. (1996) Pos-sible role of dust-induced regional warming in abruptclimate change during the last glacial period. Nature 384,447–449.

Parekh P., Dutkiewicz S., Follows M. J., and Ito T. (2006)Atmospheric carbon dioxide in a less dusty world.Geophys. Res. Lett. 33, L03610.

Peltier W. R. and Marshall S. (1995) Coupled energy-bal-ance/ice-sheet model simulations of the glacial cycle: apossible connection between terminations and terrige-nous dust. J. Geophys. Res. 100, 14269–14289.

Perlwitz J., Tegen I., and Miller R. (2001) Interactive soildust aerosol model in the GISS GCM 1. Sensitivity of thesoil dust cycle to radiative properties of soil dust aero-sols. J. Geophys. Res.—Atmos. 106, 18167–18192.

Petit J. R., Briat M., and Royer A. (1981) Ice age aerosolcontent from East Antarctic ice core samples and pastwind strength. Nature 293, 391–394.

References 25

Petit J. R., Jouzel J., Raynaud D., Barkov N. I., Barnola J.M., Basile I., Bender M., Chappellaz J., Davis M., De-laygue G., Delmotte M., Kotlyakov V. M., Legrand M.,Lipenkov V. Y., Lorius C., Pepin L., Ritz C., Saltzman E.,and Stievenard M. (1999) Climate and atmospheric his-tory of the past 420,000 years from the Vostok ice core,Antarctica. Nature 399, 436–439.

Pettke T., Halliday A. N., Hall C. M., and Rea D. K.(2000) Dust production and deposition in Asia and thenorth Pacific Ocean over the past 12Myr. Earth Planet.Sci. Lett. 178, 397–413.

Prigent C., Tegen I., Aires F., Marticorena B., and Zribi M.(2005) Estimation of the aerodynamic roughness lengthin arid and semi-arid regions over the globe with theERS scatterometer. J. Geophys. Res.—Atmos. 110, Art.No. D09205.

Prospero J. M. (1996) The atmospheric transport of par-ticles to the ocean. In Particle Flux in the Ocean (eds. V.Ittekot, P. Schaefer, S. Honjo, and P. J. Depetris). Wiley,London, pp. 19–56.

Prospero J. M., Ginoux P., Torres O., Nicholson S. E., andGill T. E. (2002) Environmental characterization ofglobal sources of atmospheric soil dust identified with theNimbus 7 Total Ozone Mapping Spectrometer (TOMS)absorbing aerosol product. Rev. Geophys. 40, Art. No.1002.

Prospero J. M. and Lamb P. J. (2003) African droughts anddust transport to the Caribbean: climate change impli-cations. Science 302, 1024–1027.

Pye K. (1987) Aeolian Dust and Dust Deposits. AcademicPress, San Diego.

Pye K. (1995) The nature, orgin and accumulation of loess.Quat. Sci. Rev. 14, 653–667.

Qiang X. K., Li Z. X., Powell C., and Zheng H. (2001)Magnetostratigraphic record of the Late Miocene onsetof the East Asian monsoon, and Pliocene uplift of north-ern Tibet. Earth Planet. Sci. Lett. 187, 83–93.

Rea D. K. (1994) The paleoclimatic record provided byeolian deposition in the deep sea: the geologic history ofwind. Rev. Geophys. 32, 159–195.

Reader M. C., Fung I., and McFarlane N. (2000) Mineralaerosols: a comparison of the last glacial maximum andpreindustrial Holocene. Can. J. Earth Sci. 37, 751–767.

Reheis M. and Kihl R. (1995) Dust deposition in southernNevada and California, 1984–1989: relationship to cli-mate, source area, and source lithology. J. Geophys.Res.—Atmos. 100, 8893–8918.

Reheis M. C. (2003) Dust Deposition in Nevada, California,and Utah, 1984–2002, http://pubs.usgs.gov/of/2003/ofr-03-138/ (accessed 2006). U.S. Geol. Survey Open-FileReport 03-138, Denver, CO.

Reheis M. C., Budahn J. R., and Lamothe P. J. (2002)Geochemical evidence for diversity of dust sources in thesouthwestern United States. Geochim. Cosmochim. Acta66, 1569–1587.

Reynolds R., Neff J., Reheis M., and Lamothe P. (2006)Atmospheric dust in modern soil on aeolian sandstone,Colorado Plateau (USA): variation with landscapeposition and contribution to potential plant nutrients.Geoderma 130, 108–123.

Ridgwell A. and Kohfeld K. E. Dust in the Earth system:the biogeochemical linking of land, air, and sea. InAdvances in Science (ed. P. Sammonds). Imperial CollegePress, London (in press).

Ridgwell A. J., Maslin M. A., and Watson A. J. (2002)Reduced effectiveness of terrestrial carbon sequestrationdue to an antagonistic response of ocean productivity.Geophys. Res. Lett. 29, doi:10.1029/2001 GL014304.

Ridgwell A. J. and Watson A. J. (2002) Feedback betweenaeolian dust, climate, and atmospheric CO2 in glacialtime. Paleoceanography 17, doi:10.1029/2001PA000729.

Roberts H. M., Muhs D. R., Wintle A. G., Duller G. A. T.,and Bettis E. A., III (2003) Unprecedented last-glacial

mass accumulation rates determined by luminescencedating of loess from western Nebraska. Quat. Res. 59,411–419.

Rosenfeld D., Rudich Y., and Lahav R. (2001) Desertdust suppressing precipitation: a possible desertificati-on feedback loop. Proc. Natl. Acad. Sci. USA 98, 5975–5980.

Ruddiman W. F. (1997) Tropical Atlantic terrigenous flux-es since 25,000 yrs B.P. Marine Geol. 136, 189–207.

Schneider S. (2002) Berichte-Fachbereich GeowissenschaftenUniversitat Bremen. Universitat Bremen, Bremen,pp. 134.

Seinfeld J., Carmichael G., Arimoto R., Conant W., BrechtelF., Bates T., Cahill T., Clarke A., Doherty S., Flatau P.,Huebert B., Kim J., Markowicz K., Quinn P., Russell L.,Russell P., Shimizu A., Shinozuka Y., Song C., TangY., Uno I., Vogelmann A., Weber R., Woo J., and ZhangX. (2004) ACE-ASIA—regional climatic and atmosphericchemical effects of Asian dust and pollution. Bull. Am.Meteorol. Soc. 85, 367.

Seinfeld J. A. and Pandis S. N. (1997) Atmospheric Chem-istry and Physics: From Air Pollution to Climate Change.Wiley, New York.

Siegel D. A. and Armstrong R. A. (2002) Corrigendum to‘‘Trajectories of sinking particles in the Sargasso Sea:modeling of statistical funnels above deep-ocean sedi-ment traps’’ [Deep-Sea Res. I 44, 1519–1541]. Deep-SeaRes. I 49, 1115–1116.

Siegel D. A. and Deuser W. G. (1997) Trajectories ofsinking particles in the Sargasso Sea: modeling of statis-tical funnels above deep-ocean sediment traps. Deep-SeaRes. I 44, 1519–1541.

Smalley I. J. and Krinsley D. H. (1978) Loess deposits as-sociated with deserts. Catena 5, 53–66.

Smalley I. J. and Smalley V. (1983) Loess material and loessdeposits: formation distribution and consequences. InEolian Sediments and Processes (eds. M. E. Brookfieldand T. S. Ahlbrandt), Developments in Sedimentology.Elsevier, Amsterdam, pp. 51–68.

Sokolik I., Winker D., Bergametti G., Gillette D., CarmichaelG., Kaufman Y., Gomes L., Schuetz L., and Penner J.(2001) Introduction to special section: outstanding prob-lems in quantifying the radiative impacts of mineral dust.J. Geophys. Res.—Atmos. 106, 18015–18027.

Steffensen J. P. (1997) The size distribution of micropar-ticles from selected segments of the Greenland Ice CoreProject ice core representing different climatic periods.J. Geophys. Res.—Atmos. 102, 26755–26763.

Sun J., Kohfeld K. E., and Harrison S. P. (2000) Records ofaeolian dust deposition on the Chinese Loess Plateauduring the Late Quaternary. Max Planck Institute forBiogeochemistry, Jena, Germany, pp. 318.

Sun Y. B. and An Z. S. (2005) Late Pliocene-Pleistocenechanges in mass accumulation rates of eolian depositson the central Chinese Loess Plateau. J. Geophys. Res.—Atmos. 110, doi:10.1029/2005JD006064.

Svensson A., Biscaye P. E., and Grousset F. E. (2000)Characterization of late glacial continental dust in theGreenland Ice Core Project ice core. J. Geophys. Res.—Atmos. 105, 4637–4656.

Swap R., Garstang M., Greco S., Talbot R., and KallbergP. (1992) Saharan dust in the Amazon Basin. Tellus,Series-B l44B, 133–149.

Tegen I. (2003) Modeling the mineral dust aerosolcycle in the climate system. Quat. Sci. Rev. 22, 1821–1834.

Tegen I., Harrison S., Kohfeld K., Prentice I., Coe M., andHeimann M. (2002) Impact of vegetation and preferen-tial source areas on the dust aerosol cycle. J. Geophys.Res.—Atmos. 107, 4576, doi:10.1029/2001JD000963.

Tegen I., Lacis A., and Fung I. (1996) The influence onclimate forcing of mineral aerosols from disturbed soils.Nature 380, 419–422.

26 Record of Mineral Aerosols and Their Role in the Earth System

Tegen I., Werner M., Harrison S. P., and Kohfeld K. E.(2004a) Relative importance of climate and land usein determining present and future global soil dust emis-sion. Geophys. Res. Lett. 31, doi:10.1029/2003GL019216.

Tegen I., Werner M., Harrison S. P., and Kohfeld K. E.(2004b) Reply to comment by N. M. Mahowald et al. on‘‘Relative importance of climate and land use in determi-ning present and future global soil dust emission’’. Geophys.Res. Lett. 31, doi:10.1029/2004GL021560.

Thompson L. G. (1977) Variations in microparticle con-centration, size distribution and elemental compositionfround in Camp Century, Greenland and Byrd station,Antarctica, deep ice cores. In Proceedings of the GrenobleSymposium on Isotopes and Impurities in Snow and Ice.IAHS-AISH, pp. 351–364.

Tiedemann R., Sarnthein M., and Shackleton N. J. (1994)Astronomic timescale for the Pliocene Atlantic d18O anddust flux records of Ocean Drilling Program site 659.Paleoceanography 9, 619–638.

Tiessen H., Hauffe H. K., and Mermut A. R. (1991) Dep-osition of Harmattan dust and its influence on basesaturation of soils in northern Ghana. Geoderma 49,285–299.

Visible Earth, http://visibleearth.nasa.gov (accessed 2006).Vitousek P. M., Kennedy M. J., Derry L. A., and Chad-

wick O. A. (1999) Weathering versus atmosphericsources of strontium in ecosystems on young volcanicsoils. Oecologia 121, 255–259.

Washington R., Todd M., Lizcano G., Tegen I., FlamantC., Koren I., Ginoux P., Engelstaedter S., Bristow C.,Zender C., Goudie A., Warren A., and Prospero J. (2006)Links between topography, wind, deflation, lakes anddust: the case of the Bodele Depression, Chad. Geophys.Res. Lett. 33, L09401.

Watson A. J., Bakker D. C. E., Ridgwell A. J., Boyd P. W.,and Law C. S. (2000) Effect of iron supply on SoutherOcean CO2 uptake and implications for glacial atmos-pheric CO2. Nature 407, 730–734.

Werner M., Tegen I., Harrison S. P., Kohfeld K. E., Pren-tice I. C., Balkanski Y., Rodhe H., and Roelandt C.

(2003) Seasonal and interannual variability of themineral dust cycle under present and glacial climateconditions. J. Geophys. Res.—Atmos. 108, doi:10.1029/2002JD002365.

Wheaton E. E. and Chakravarti A. K. (1990) Dust stormsin the Canadian prairies. Int. J. Climatol. 10, 829–837.

Woodward S., Roberts D. L., and Betts R. A. (2005) Asimulation of the effect of climate change–induceddesertification on mineral dust aerosol. Geophys. Res.Lett. 32, doi:10.1029/2005GL023482.

Yu H., Kaufman Y., Chin M., Feingold G., Remer L.,Anderson T., Balkanski Y., Bellouin N., Boucher O.,Christopher S., DeCola P., Kahn R., Koch D., Loeb N.,Reddy M., Schulz M., Takemura T., and Zhou M. (2006)A review of measurement-based assessments of the aer-osol direct radiative effect and forcing. Atmos. Chem.Phys. 6, 613–666.

Zarate M. A. (2003) Loess of southern South America.Quat. Sci. Rev. 22, 1987–2006.

Zender C. S., Miller R. L., and Tegen I. (2004) Quantifyingmineral dust mass budgets: terminology, constraints, andcurrent estimates. Eos 85, 509–512.

Zhao C., Dabu X., and Li Y. (2004) Relationship betweenclimatic factors and dust storm frequency in InnerMongolia of China. Geophys. Res. Lett. 31, doi:10.1029/2003GL018351.

Zheng H. B., An Z. S., and Shaw J. (1992) New contribu-tions to Chinese Plio-Pleistocene magnetostratigraphy.Phys. Earth Planet. Interiors 70, 146–153.

Zheng H. B., Powell C. M., Rea D. K., Wang J., and WangP. X. (2004) Late Miocene and mid-Pliocene enhance-ment of the East Asian monsoon as viewed from the landand sea. Global Planet. Change 41, 147–155.

Zielinski G. A. and Mershon G. R. (1997) Paleoenviron-mental implications of the insoluble microparticle recordin the GISP2 (Greenland ice core during the rapidlychanging climate of the Pleistocene–Holocene transition.Geol. Soc. Am. Bull. 109, 547–559.

r 2007, Elsevier Ltd. All rights reserved. Treatise On Geochemistry

No part of this publication may be reproduced, stored in a retrieval system or

transmitted in any form or by any means, electronic, mechanical, photocopying,

recording or otherwise, without prior written permission of the Publisher.

ISBN (set): 0-08-043751-6

pp. 1–26