Erosional landforms developed as a result of abrasionby wind-carried particles (including, sand, dust, snow/ice crystals) are known from a range of terrestrialenvironments and geomorphic settings including hotand cold deserts, mountains, and coasts (e.g. Greely andIversen, 1985; Pye and Tsoar, 1990; Seppälä, 2004).Erosional landforms can be developed on both soft andhard substrates. Those on such soft, unconsolidatedsubstrates as beach/dune sand and soil can be consideredas transient phenomena whose morphology is capable ofresponding dynamically to small changes in windconditions. Such landforms, therefore, may reflectvariable local winds over relatively short time scales,and have limited preservation potential in the geologicalrecord. Wind-eroded scarps found around the edges ofsand dunes are a typical example (e.g. Jackson andCooper, 1999). By contrast, the wind-eroded formsdeveloped on hard, consolidated land surfaces such as
bedrock or stable surficial clasts are often geomorphi-cally delicate, showing intricate surface details, withvery clear and sharp edges. These landforms, includingyardangs and ventifacts (Grolier et al., 1980; Halimovand Fezer, 1989; Pye and Tsoar, 1990; Laity, 1992,1995), reflect what might be considered mean windconditions over relatively long time periods (oftendecadal or longer scales), and have a much higherpreservation potential in the geological record. Venti-facts and similar ‘hard’ erosional features thereforerepresent only part of a spectrum of forms recording therelationship between wind activity, the presence ofwind-carried abraders, and the land surface.
Ventifacts, developed on these hard land surfaces, areuseful because they can indicate both present and pastwind conditions includingwind direction and, often, windstrength. Their usefulness arises because of their highpreservation potential and because they often remain asrelict (residual) surface forms in hot/cold deserts,mountains and coastal environments, and are often notconcealed by later sediment, soil or vegetation cover. Inaddition, because ventifacts (as geomorphic features)
90 J. Knight / Earth-Science Reviews 86 (2008) 89–105
record the effects of wind-blown abraders in transit, theyare the ‘missing link’ between wind blow activity (asevidenced by wind climate records) and wind-depositedsediment (such as sand dunes, coversand and loess)(Fig. 1).
1.1. Problems with terminology and the identification ofventifacts
Ventifacts are defined on a morphological basis assubaerially-exposed clasts or bedrock surfaces that havebeen abraded by the action of wind-carried particles(Greely and Iversen, 1985). Abrasion of the faces ofthese clasts and rock surfaces gives rise to distinctiveand smooth plano-concave to plano-convex faces(Fig. 2). The number and disposition of these abradedfaces (which are termed facets) was used during theearly Twentieth Century as important criteria to identifyand distinguish different ventifact types. Ventifacts withone significantly abraded facet and one unabraded facewere termed einkanter (forming a Brazil-nut shape inplan view), and those with three opposing edges, whichtogether give rise to a pyramidal ventifact shape in planview, were termed dreikanter (e.g. Wade, 1910; Kuenen,1928). The term ventifact was coined as a replacementfor these morphological terms by Evans (1911, p.335),to mean ‘a general expression… for any wind-shapedstone’. Although genetic, the term ventifact makes noassumptions of the detailed morphology or alignment ofthe ventifact with respect to wind direction, and the termtherefore remains a useful one. This terminology,however, considers only the basal outline of theventifact, and does not consider the different types ofdetailed surface forms present on the wind-abradedfacets themselves (discussed later).
A central problem in the study of ventifacts is that theyare identified on a subjective, visual basis, and that littlework has been done to standardisemethodologies for theiridentification, particularly in field settings. Kuenen (1960,p.447) argued that quartzite ventifacts can be developedand identified visually in hand-specimenwhere there is an
Fig. 1. Illustration of the role of ventifacts as the ‘missing lin
erosive loss of about 6% from wind-abraded surfacescompared to those surfaces that have not been abraded.The calculated time-period over which this ventifactiontakes place depends nonlinearly on wind strength andabrader size (Kuenen, 1960). This figure of 6% (based onweight loss) also assumes that abrasion is consistentacross the surface, which in practice it is not. It is likelythat substantial wind-abrasion effects can be observed onrocks with surface losses below this 6% threshold value,when these surfaces are examined in detail. Equally, evenon rock surfaces with high abrasion loss, ventifacts mightnot be identified if this abrasion is consistent across allsurfaces. Although, therefore, the term ventifact refers tothe external shape of a wind-abraded rock as a whole(when observed in hand specimen), wind-abrasion effectscan be observed at different scales. Therefore the termventifact refers to only one part of a continuumof abrasionscales, and should be considered within a widergeomorphic context. Better resolving these issues ofscale and degree of abrasive loss, through linked field andlaboratory studies, will enable a better understanding ofthe relationships between ventifaction and other processesassociated with rock surface abrasion. In addition,determining threshold values of abrasion loss is likely acritical issue in identifying what is, and what is not, aventifact. It also means that ventifacts have been likelyunder-recorded in the field. A bibliography of aeolianresearch for the period 1647–present, which includesworks on ventifacts, is maintained by Gill et al. (2007).
2. Aims of this paper
This paper is structured into five main parts thatcritically address the past, present and future directionsof ventifact research. In detail the paper aims to (1)review the history of ventifact research from theEnglish-language literature and to identify the keythemes of these studies and their wider geomorpholog-ical significance; (2) outline the geographical andenvironmental controls on the presence and formationof ventifacts; (3) describe the macroscale and detailed
k’ between wind activity and sediment accumulation.
Fig. 2. Photo of a typical ventifact, showing wind-abraded faces andsharp keel (photo by Helene Burningham).
mesoscale surface forms found on ventifacts; (4) explainhow macroscale and mesoscale forms can be used toevaluate present and reconstruct past wind conditions;and (5) outline future research directions, including whyventifacts are important in reconstructing the sedimentdynamics and synoptic-scale climatology of continentalinteriors and coasts.
3. Historical perspectives on ventifact research
The genetic relationship between wind-blown abra-ders (usually sand) and the abrasion of hard rocksurfaces has been known for a long time (e.g. Blake,1855; Travers, 1870; Stowe, 1872; Delo, 1876). In theseearly studies, the geomorphic impacts of wind abrasionon rock surfaces were noted with reference to localpatterns of topography and winds. The key paper byBlake (1855, p.179) evocatively describes the effects ofwind abrasion in the semiarid San Barnardino Moun-tains of southern California, USA: ‘Wherever I turnedmy eyes – on the horizontal tables of rock; or on thevertical faces turned to the wind – the effects of the sandwere visible: there was not a point untouched, the grainshad engraved their track on every stone.’ Here, sand isblown by westerly winds through the San BarnardinoPass, forming asymmetric wind-parallel grooves thatpick out mineralogical variations in the underlyinggranite bedrock. Observations of wind-abraded rocksurfaces, reported from the New Zealand coast aroundthe same time (Travers, 1870; Stowe, 1872), are alsosignificant because they were considered as part of themorphodynamics of sand-dominated coasts as a whole,and were linked to the presence and development ofcoastal sand dunes (cf. Bishop and Mildenhall, 1994).These early observations were made within the context
of routine cataloguing of regional-scale landscapegeology and geomorphology by fledgeling geologicalsurveys, and were therefore sporadic in nature anddistribution.
Such early observations were succeeded by moredetailed field studies that had a clearer morphogenetic,rather than purely descriptive, focus. In these later studies,different types of wind-abraded forms were identified andobservations were made that considered geographicalsetting, rock type and wind pattern as influences overthese abrasive forms (e.g. Woodworth, 1894; Bather,1900; Bosworth, 1910; Wade, 1910; Wentworth andDickey, 1935; King, 1936;Maxson, 1940; Higgins, 1956;Hickox, 1959; Clark and Elson, 1961). For example, thecareful observations of Wade (1910) from the desert ofeastern Egypt showed that ventifacts here do not have aclear alignment with respect to wind direction, and thattheir alignment changes over time as the ventifact'sexternal shape evolves (including wind-reorientation ofclasts; Kuenen, 1928). This concept was expanded on byKing (1936)who argued that abrasion decreases over timeas the ventifact attains a low, flat outline shape that offersleast resistance in the windstream. It is significant that thisview mirrors Davis's ‘cycle of erosion’ for macroscalelandscape development that was in favour at that time.The linked papers by Wentworth and Dickey (1935) andPowers (1936) listed ventifact localities in the USA anddiscussed the overarching conditions required for venti-faction, including climate and geology. A bibliography ofthis early work was brought together by Bryan (1931).
In addition to these studies, a more sophisticatedunderstanding of the temporal variability of windabrasion was gained by identifying abrasion forms insettings no longer affected by strong winds (e.g.Blackwelder, 1929; Powers, 1936; Tremblay, 1961).Two interpretative approaches were considered here.The first focused on the presence of ventifacts inlocations which are not presently affected by winds of astrength sufficient to blow sand in large quantities. Thesecond focused on the absence of a sediment source andpresence of stabilising vegetation across land surfaces inthese areas. Both approaches converge on the view thatventifacts are relict features formed under past environ-mental conditions, and that wind-abrasion was previ-ously more active than at present. Blackwelder (1929),for example, concluded that ventifacts in the SierraNevada, California, have a late Pleistocene origin.
At the same time as these studies, experiments onwind abrasion under laboratory conditions were alsotaking place (e.g. Kuenen, 1928; Schoewe, 1932), andwere aimed at understanding relationships betweenwind-carried particles and different ventifact shapes.
92 J. Knight / Earth-Science Reviews 86 (2008) 89–105
These studies had similar experimental setups, namelydelivery of sand from a hopper into a machine-expelledairstream which blew the sand onto blocks of differentmaterials. The materials used included plaster of Paris,chalk, carbonate putty, and magnesium carbonate.Based on these experimental results, Kuenen (1928)argued that the basal outline of the original clast is astrong determinant of the final shape of the ventifact,and that ventifacts can form by winds from variabledirections (and strengths). By contrast, Schoewe (1932)argued that most ventifacts are formed under a constantwind direction in which the basal outline of the clast isnot important. His experimental work showed that theabrasion rate and angle of the abraded surface bothdecreased over time. Later experimental work usingsimilar apparatus also considered the role of abraders ofdifferent sizes in forming different types of ventifacts(e.g. Kuenen, 1960; Dietrich, 1977).
This initial burst of interest in the morphology andenvironmental significance of ventifacts culminated inthe seminal paper by Sharp (1949) who used detailedfield observations from the semiarid Big Horn Moun-tains, Wyoming (USA), to identify periods of past windabrasion from the morphology of ventifacts at differentlocations. This paper is significant because it integrated,for the first time, many field observational and analyticalmethods that have been used to the present day. Sharpdescribed the spatial distribution of ventifacts on terracesurfaces of different elevation located above mountainriver valleys. Facet facing-direction and mesoscalesurface features were described and some spatialpatterns were identified. Sharp also described likelytrajectories of sediment grains over facets of differentinclinations, and the relationship between wind direc-tion and facet orientation. He interpreted ventifact agewith reference to the age of the river terraces on whichthey are located, but argued that wind-abrasion wasmost active during the late Pleistocene and wasorographically controlled from the northwest (as atpresent). Sharp's (1949) paper is important because itconsidered spatial and temporal variations in the keyenvironmental parameters contributing to ventifaction,and their wider palaeoenvironmental and palaeoclimaticsignificance.
Later studies by Sharp (1964, 1980) in the Coachellavalley of southern California (very close to whereBlake's (1855) observations were made) involved along-term field experiment to identify contemporarywind abrasion rates. In this experiment blocks madefrom different materials, including granite, plaster ofParis, brick and lucite, were placed on an exposedmountain slope. Changes in the size and shape of these
blocks were observed from c. 1948–1965 using fieldmeasurement and repeat photography. A similar meth-odology was used for the period 1951–2003 in acontemporary periglacial setting by Mackay and Burn(2005).
After these studies focusing on ventifact formation inwarm, semiarid environments, attention shifted toAntarctica in the 1970s and 1980s, and the glaciatedmidlatitudes from the 1990s. Work in Antarcticafocused on the high rate of ventifact abrasion associatedwith very strong winds (Selby, 1977; Miotke, 1982;Malin, 1986; Spate et al., 1995), and the possible greatage of some ventifacted surfaces, particularly in extra-glacial east Antarctica (Lindsay, 1973; Selby et al.,1973). This work relates very closely to present ideas oflong-term land surface stability in east Antarctica (e.g.Sugden et al., 1999). Long-term field-based studiesshow wind abrasion rates of 0.05–3.70 mm year− 1,dependent on lithology (Malin, 1986). Short-termlaboratory experiments simulating Antarctic wind con-ditions show episodic abrasion rates of 8.04–16.70 mmyear− 1 (Miotke, 1982). Together, these suggest veryrapid (decadal-scale) ventifact formation.
In the 1990s work on ventifacts was conductedmainly in formerly-glaciated midlatitude Europe andNorth America. Particular attention was paid toventifacts formed in past periglacial environments andextra-glacial forefields, and their implications forpalaeo-wind direction (e.g. Schlyter, 1995; Thorsonand Schile, 1995; Fisher, 1996; Isarin et al., 1997;Christiansen and Svensson, 1998; Hoare et al., 2002;Christiansen, 2004). Here, ventifacts that are geomor-phically associated with sand wedges and lagged glacialoutwash surfaces are common. Fisher (1996) argued,based on such associations, that ventifacts in Saskatch-ewan (Canada) were formed under variable climatic andpermafrost conditions during the late-glacial (c. 13–10 kyr BP). A similar geomorphic association was madebetween ventifacts, dated coversand units, and cryotur-bation structures in Norfolk, eastern England (Hoareet al., 2002). Ventifacts recorded from formerly-glaciated Europe have been used in a spatial sense tobetter understand regional-scale wind patterns duringthe late-glacial to early Holocene, particularly acrosssouthern Sweden and Denmark (Svensson, 1983;Schlyter, 1995; Christiansen and Svensson, 1998;Christiansen, 2004). Ventifacts in these locations werelikely formed under strong winds (which may have beenkatabatically-driven) during the late-glacial when con-tinental ice sheets were still present. Enhanced windactivity also took place during the Younger Dryas undersynoptic conditions similar to present, associated with
the buildup of aeolian sand sheets (Kolstrup, 1997;Hoare et al., 2002; Kolstrup, 2007). These spatial dataon palaeo-wind direction based on the orientation ofventifact facets are also critical as input into synoptic-scale climate models (Isarin et al., 1997).
Work on ventifacts from the 1990s to present hasfocused on modern coastal settings in the glaciatedmidlatitudes, particularly west-facing sites in Scotlandand Ireland (e.g. Wilson, 1991; Braley and Wilson,1997; Knight and Burningham, 2001; Wilson et al.,2002; Knight, 2002, 2003; Knight and Burningham,2003; Knight, 2005). Along these coasts ventifacts arefound mainly in upper intertidal and supratidal posi-tions, and their patterns of wind abrasion are largelyconsistent with present coastline geometry and contem-porary sandy beaches. This infers that these ventifactslikely date from the period of coastal stabilisation andinitial sand dune buildup in western Britain at c. 4 kyrBP (Wilson et al., 2002, 2004). Although most coastalventifacts have clean facets, in some locations theyshow overgrowths of lichen, suggesting they are nolonger being actively abraded (Knight and Burningham,2001). This, amongst other evidence, suggests that mostcoastal ventifacts in western Britain probably corre-spond to enhanced wind action during the Little Ice Age(c. 1550–1850 AD), although there is only indirectdirect dating evidence for this (Braley and Wilson,1997; Wilson et al., 2002). Some rock-built coastaljetties of known age (70–120 year) are also ventifacted,suggesting that ventifaction can take place, even underpresent wind conditions, over very short (decadal) timescales (e.g. Knight, 2003; Knight and Burningham,2003). Finally, these recent coastal studies also showthat wind flow patterns, as reconstructed from venti-facts, are complex and likely reflect sediment supplyand local topography (e.g. Knight, 2005). This may limitthe extent to which ventifacts can be used to reconstructsynoptic-scale climates (discussed later).
3.1. Geographical perspectives on ventifact research
As the summary above indicates, most ventifactshave been reported from the English-speaking world orby English-speaking workers. Earlier perspectives onthe geographical distribution of (mainly midlatitude)ventifacts were discussed by Bryan (1931) and Went-worth and Dickey (1935). The bibliography of Gill et al.(2007) contains 146 references that explicitly discussventifacts, of which some 52% are in English and 26%in German (correct as of June 2007). This lingualdistribution is also a geographical one. Records ofventifact distribution worldwide, compiled from various
sources within the literature in English, are shown inFig. 3. Although this distribution is likely an artefact ofwhere field observations have been undertaken byEnglish-speaking workers, locations where ventifactsappear most common include maritime NW Europe(British Isles, Scandinavia), New Zealand, midwestUSA, and sub-Antarctic islands (such as the Falklands).Common characteristics of these locations include (1)position adjacent to tectonically-active mountains; (2)high-energy coastlines with a long fetch; (3) paraglacialenvironments with readily-available sand-sized sedi-ment; and (4) locations with high rates of physicalweathering in cold/hot deserts. Locations that havesimilar physical characteristics, but where ventifacts areapparently absent or have been less commonly reported,include southern South America, coastal south andcentral Australia, formerly-glaciated eastern Europe,and central Asia. This may hint at locations suitable forfuture research.
Ventifacts have been recorded particularly in north-ern hemisphere midlatitudes where they are associatedwith former (late Pleistocene) glacial and periglacialenvironments. Conditions conducive to ventifaction,and associated with these glacial and periglacialenvironments, include: (1) the erosive effects of glaciersacross granitic and metasedimentary continental cratonsand fold mountains. This helped both generate largeamounts of sand-sized sediments through subglacialabrasion, and transport this sediment to sites ofdeposition on adjacent shelves and coastal lowlandsassociated with ice marginal positions; (2) deglacial andpostglacial changes in relative sea-level (RSL; includingboth glacio-isostatic and eustatic effects). These RSLchanges helped move glacigenic sediments onshore,forming coastal sand dunes, beaches, and estuary-infills;(3) the presence of sediment-rich and vegetation-poorglacial forefields, particularly in areas of low relief; (4)synoptic-scale katabatic circulation associated with thepositioning of high and low pressure cells aroundcontinental ice sheets. Katabatic winds associated withthe late Pleistocene Laurentide and Scandinavian icesheets have been linked genetically to the formation ofventifacts (with a distribution of continental extent)across mid-North America and extra-glacial Europe(e.g. Schlyter, 1995; Thorson and Schile, 1995; Fisher,1996; Isarin et al., 1997). These katabatic winds werelikely also characterised by higher rates of sandtransport, and thus rates of abrasion, because of theincreased air density associated with cooler ambienttemperatures (Seppälä, 2004), and so may thereforeaccount for the co-location of ventifacts with presentand past extra-glacial environments. This factor is also
Fig. 3. Map showing the general global distribution of ventifacts that are reported in the literature, from sources cited in this paper and by interrogationof the Gill et al. (2007) aeolian database.
94 J. Knight / Earth-Science Reviews 86 (2008) 89–105
important in the formation of ventifacts in eastAntarctica and sub-Antarctic islands (e.g. Miotke,1982; Clark and Wilson, 1992; Wilson and Edwards,2004). Although these environmental conditions areimportant in identifying likely settings for the develop-ment of ventifacts (such as in South America), their lackof reportage in the English (and German) literaturesuggests that further work is needed in these areas. Indetail, ventifacts are also distributed unequally acrossthe northern midlatitudes. Ventifacts are most commonon west-facing coasts that are located adjacent to oceanbasins, including the British Isles and western USA.Such coasts are affected by onshore prevailing winds, along fetch, and cyclonic winter storms. Ventifacts arealso found associated with semi-enclosed or shallowepicontinental seas (in coastal areas) and, in inlandlocations, with semi-arid basins and exposed shorelinesof inland seas. These areas are affected by sedimentreworking within (mostly) closed sediment cells(Knight, 2002). This geographical bias is significantbecause it suggests that few ventifact locations havebeen recorded in detail, and that ventifacts represent ageomorphic feature about which little is known.
In addition, ventifacts have been reported from allparts of the geological record. Most (at least 95%)recorded ventifacts are of late Quaternary to earlyHolocene age, but examples are also known from theLittle Ice Age (Knight and Burningham, 2001) andcontemporary coasts (Knight, 2003; Knight and Bur-ningham, 2003). Ventifacts have been reported from
Archaean braided river and high-energy coastal settings(Els, 1998), and from Permo-Triassic sandstone bedswhere bedding planes are exposed (Thompson andWorsley, 1967; Leonard et al., 1982). Similar ventifact-like morphologies, sometimes termed ‘aquafacts’, canalso result from sand abrasion of beachrock by waveprocesses (Shepherd, 2003). In addition, landformsinterpreted as ventifacts (from remotely-operated vehi-cle data) have been identified on Mars (e.g. Bridgeset al., 1999, 2004a, 2005).
3.2. Environmental controls on processes of ventifaction
Interpretation of the geographical spread of venti-facts, discussed above, can be informed by consider-ation of the environmental controls on their formation.Broadly, the environmental conditions required forventifaction (the process by which ventifacts acquiretheir morphology) are: (1) a source of loose sediment ofan appropriate size range; (2) winds of a sufficientstrength and appropriate direction to transport sedimentacross exposed clasts or rock surfaces; and (3) thepresence of clasts or rock surfaces which protrude intothe windstream (Sharp, 1949; Greely and Iversen,1985). These three controls are now discussed in moredetail.
(1) The relationship between past glaciation andventifacts (Seppälä, 2004), noted above, focuseson the generation of sand-sized sediment through
processes of glacial abrasion acting on igneousand metasedimentary bedrocks (characteristic ofcontinental cratons and fold mountains) that aredominated mineralogically by quartz and feldspar.Divergent ice flow vectors, from ice dispersalcentres and towards ice margins located on coastallowlands and continental shelves, helped bothdisperse sediment across glacial forefields, coastallowlands and eustatically-dry continental shelves,and focus this sediment into discrete contempo-raneous (and later) landforms including moraines,outwash spreads, beaches and estuary-fills(Knight, 2002). These concentrations of sand-sized particles acted as source areas for lateruptake by wind (Seppälä, 2004).In addition, sand-sized particles are readilyproduced by mechanical weathering processes inhot/cold deserts (Ritter and Dutcher, 1990). In hotdeserts, diurnal thermal expansion/contraction,frost shattering, salt crystal growth and exfoliationare the main weathering processes (Laity, 1987).In cold, periglacial deserts, frost shattering ofexposed bedrock surfaces and growth of segre-gated ice lenses are the main weathering process-es, and are associated with the formation ofsurficial boulder lags (remanié deposits) (Selby,1977; Miotke, 1982; Malin, 1986). These me-chanical weathering processes in hot/cold desertscontribute to the formation of loose, surficial sandand clasts and generally affect all bedrock types.Sand-sized materials are produced mainly byweathering of igneous and metasedimentary bed-rocks in these environments. Fluvial processesmay be also significant in some locations.
(2) Synoptic- to local-scale wind patterns are con-trolled by hemispheric circulation cells, and thepositioning of high and low pressure centres whichare influenced by land (and sea) surface character-istics. Winds are strong and persistent in areasassociated with the downward-moving limb of theFerrel cell (in hot desert regions), and onshore-moving westerlies (in the midlatitudes). Ventifactsare common in both these regions. Blocking high-pressure cells located over cold continental icesheets (at the present time and in the latePleistocene) are important in the development ofstrong, persistent katabatic winds both outwardsfrom the ice sheet centre, and in circulation aroundice sheet margins (Ashwell, 1966; Bromwich,1989; Parish and Bromwich, 1991). Recently-deglaciated and periglacial environments aretherefore common settings for ventifaction be-
cause of this enhanced synoptic-scale circulation.In addition, ventifacts formed immediately afterthe last deglaciation are often preserved to thepresent day because of the subsequent diminutionof wind strength and/or change in wind directioninto the Holocene (e.g. Schlyter, 1995). Thisargument also holds for ventifacts that were likelyformed during the Little Ice Age, when windpatterns were also enhanced relative to present(Grove, 2004), but which have been preserved tothe present day largely as relict features (Braleyand Wilson, 1997; Knight and Burningham,2001).It is also important to differentiate betweendirection of the prevailing wind and direction ofthe most geomorphically effective wind (Rudberg,1968). The prevailing wind refers to the compassquadrant fromwhich most winds of all magnitudesare directed over a given time period, usually ayear. This can be calculated very easily fromstandard wind data. The most geomorphicallyeffective wind refers to the compass quadrant fromwhich abrading winds are directed. This thereforerequires wind strength to be above the thresholdvelocity for sediment transport (defined broadly bythe Bagnold equation), and for loose sediment tobe available for uptake into the windstream.Direction of the most geomorphically effectivewind can be evaluated semi-quantitatively fromventifact orientation data, and can be matchedmore precisely with standard wind data (split intovelocity categories) where the threshold velocityfor transport is calculated for the loose sedimentsurrounding the ventifact site (e.g. Knight andBurningham, 2001, 2003). In this paper, geomor-phically effective winds are discussed, unlessstated otherwise.
(3) The presence of clasts or rock surfaces of suitablelithologies (discussed above), and which protrudeinto the windstream, is important since these formsubstrates on which ventifacts can potentially bedeveloped. Clasts or rock surfaces are exposedmainly in upland and coastal locations, wheresurface sediment cover is presently thin or absent,and where sediment erosion/transport capacity ishigh. Glaciated environments are in particularcharacterised by exposed clasts or rock surfacesbecause of the capacity of glaciers to (1) strip awayloose surface sediments to rockhead, (2) transporterratic blocks, and (3) increase land surface reliefby differential erosion, particularly in mountain-ous regions (Tomkin and Braun, 2002).
96 J. Knight / Earth-Science Reviews 86 (2008) 89–105
4. Surface morphology of ventifacts
Macro- and mesoscale geomorphic features arepresent across ventifacted surfaces that are observed inthe field and/or in hand specimen. Not only do thesedetailed surface forms aid identification of clasts orbedrock protrusions as ventifacts, they also enable moreaccurate interpretation of wind flow direction, process,timing, sediment characteristics, and geologic/structuralcontrol (Greely and Iversen, 1985). Here, macroscalefeatures (facet, keel) are defined as those that are presentacross the clast or bedrock protrusion as a whole.Developed upon these are mesoscale features (polish,pits, grooves, flutes) that are spatially variable inmorphology, depending on the nature of the macroscalesurface. Finally, microscale features (pits, microfrac-tures) refer to the grain-by-grain variation of the clast orrock surface imparted by the impact of wind-carriedabraders. These microscale features are best observedusing a hand lens or microscope and are not discussed indetail here. The macro- and mesoscale featuresdescribed below are named following Whitney andDietrich (1973). Other wind-abrasion and deflationfeatures found on bedrock surfaces, such as hamadasand yardangs (Laity, 1995), are not discussed here.
4.1. Macroscale features
Macroscale features (b few m in diameter) areassociated with the progressive development of theventifact facet, described above. If winds come fromseveral directions, several opposing facets can developforming, for example, the classic Brazil-nut ventifactshape (Bather, 1900). Facets may be sharply rectilinearto curved, depending on geologic control of the facetsurface, and the angle of wind attack. Winds thatapproach the facet strike at an oblique angle areassociated with the development of curved facets(Wade, 1910) on which the abrasion rate is variableacross the facet. In this case, facet margins becomecurved in the direction of wind flow (Anderson, 1986;Mackay and Burn, 2005). Many laboratory studies haveshown that the abrasion rate is highest on steep facets,but decreases as the facet is abraded down (e.g. Schoewe,1932; King, 1936; Bridges et al., 2005), a classicexample of negative feedback. The smaller-scale formsdeveloped on facets (discussed below) also varyaccording to facet steepness, with steep facets char-acterised by polish or pits, and shallow facets by grooves(Maxson, 1940; Sharp, 1949; Knight and Burningham,2001). In addition, the number of facets present on aventifact can vary depending on the size/shape of the
original clast or bedrock protrusion, and variability ofwind direction (Kuenen, 1928; Suzuki and Takahashi,1981). Clark and Elson (1961) found that geomorphi-cally effective winds that were present through 26% ofthe year were able to shape ventifacts that show one(34% of all recorded), two (65%) or three (1%) facets(n=250). A very similar result was obtained by Knightand Burningham (2001) where geomorphically effectivewinds (during 25–28% of the year) formed ventifactswith one (24%), two (36%), three (36%) or four (4%)facets (n=50). On small ventifacts, the number of facetspresent can decrease over time as abrasion continues(Lindsay, 1973), particularly where ventifacts areundercut and reoriented by the wind (Kuenen, 1928;Sharp, 1949; Mackay and Burn, 2005).
As the facet develops through continued abrasion,the uppermost edge of the facet often becomes sharperwith respect to the (unabraded) leeward facet. Thisangular break of slope between adjacent faces or facets,resembling an upturned boat, is termed a keel. In manycases, the ventifact keel is straight, can be knife-sharpparticularly when developed between opposing facets,and generally extends across the length of the ventifactat its uppermost point (Fig. 4). Some keels may also besinuous in plan view depending on the original shape ofthe ventifact. Most ventifact studies measure the strikeof the keel, which may be difficult to do in practicewhere the keel is curved in plan view. Fewer studiesmeasure the facing direction of the facet(s) on which thekeel is developed, but this is important since it oftenreflects the direction of the most geomorphicallyeffective wind.
Interpretation of these field measurements is prob-lematic for several reasons. A ventifact keel can have analignment that may be parallel, or perpendicular, to winddirection. Wind-parallel keels are usually associatedwith low, elongate ventifacts with shallow, opposingfacets that are convex in section (e.g. Fig. 5). Wind-perpendicular keels are usually associated with upstand-ing ventifacts that only have one steep, concave orconvex facet (Fig. 6). In these cases, the keel maybecome oversteepened or even overhanging due towindflow separation over the crest of the ventifact, andreattachment on the ventifact leeside (Fisher, 1996;Knight and Burningham, 2003). The size and structuralalignment of the original clast or bedrock protrusion isalso a strong determinant of keel orientation (Schoewe,1932). This means that keel alignment cannot be useduncritically as a palaeo-wind indicator (Knight, 2005).The facing direction of a ventifact facet can moreusefully estimate palaeo-wind direction, but this isdifficult where facets are convex or rounded.
Fig. 4. Sketch showing the main morphological characteristics of aventifacted boulder, and its relationship to wind direction (adaptedfrom Knight, 2005).
Fig. 5. ‘Brazil-nut’ shaped ventifact showing a sharp, wind-parallelkeel separating two convex facets. Wind direction is bottom right totop left (photo by Helene Burningham).
Fig. 6. Ventifacted bedrock outcrop showing a wind-perpendicularkeel separating the wind-abraded and non-abraded sides of theoutcrop. Wind direction is from bottom to top across the strike of theventifact, which therefore shows no geological control. Trowel forscale is 28 cm long.
Most observations of ventifact keels recorded in theliterature are on ventifacts developed on detached clastsrather than on bedrock protrusions, and which are oftenless than about 1 m in diameter. Studies on erosive lossfrom facet surfaces suggest that ventifacts are morelikely to develop on smaller clasts than larger ones, andtherefore that larger clasts or bedrock protrusions aremore likely to preserve any structural control over theirshape and are less likely to show large-scale ventifaction(Kuenen, 1928; Schoewe, 1932; Kuenen, 1960). This isborne out by the relative rarity with which largeventifacts are reported in the literature (e.g. Mackayand Burn, 2005).
Such large ventifacts, however, are present in theestuary of Loughros More, County Donegal, NWIreland (54°46′N, 8°28′W), where joint-controlledgranite surfaces crop out in the supratidal zone of theinner estuary and are partially overlain by sand dunes.Here, rock surfaces have a relief of 0.2–1.5 m, and formasymmetric ridges that have an approximate NE–SWorientation. West-facing ridge sides show evidence forlarge-scale wind abrasion (Fig. 7). This evidenceconsists of (1) a smoothed, planar west-facing ridgeside which has a low relief, contrasting with the east-facing ridge side which is convex and highly weathered;and (2) a notable absence of lichen overgrowths on theabraded, west-facing surface compared to east-facingones. At this location also, some large, equant graniteboulders (1.6–2.0 m dimensions) are present, inter-preted as glacial erratics. Some of the boulders arelocated on the granite surface, some on the beach. West-facing boulder faces are prominently abraded to the fullheight of the boulder (Fig. 8).
These large ventifacts are unusual for severalreasons. Ventifacts are infrequently reported on granitebedrock because the coarse rock texture is notconducive to forming, and thus identifying, smallerscale abrasion features. The ventifact facets reportedhere, therefore, show the significance of wind abrasion
on coarse bedrock and to an extent previouslyunrecorded. The vertical height to which wind abrasionis recorded on granite boulders may suggest eitherexhumation of boulders following beach burial, orblown sand activity to a much greater height in thewindstream than is commonly recorded (e.g. Wilshireet al., 1981). In terms of their scale, the large, ventifactedsurfaces developed on bedrock may be consideredtransitional to yardangs (Halimov and Fezer, 1989).
Another very unusual macroscale feature, notrecorded in the literature, is observed on the south sideof the Loughros More estuary. Here, rocks outcroppingin the supratidal zone and at the cliff foot comprisediabase dikes that cut through Dalradian-age schist.Remarkable undercut wind abrasion forms are devel-oped in the diabase dikes (Figs. 9, 10). These forms,
Fig. 7. Mesoscale ventifact developed on granite in the estuary ofLoughros More, County Donegal, NW Ireland. The ventifactedsurface (left side) is smooth, planar and free of lichen. This contrastswith the non-abraded surface (right side) which is convex and lichen-covered. Trowel for scale is 28 cm long.
Fig. 8. Steep, ventifact facet developed on a large granite boulder(1.5 m high) within the intertidal zone of the Loughros More estuary,County Donegal, NW Ireland. The ventifacted surface is pitted,convex, and has a sharp upper keel (near the trowel). Trowel for scaleis 28 cm long.
Fig. 9. Unusual hammerhead ventifacts (upstanding features) in theLoughros More estuary, County Donegal, NW Ireland, developed indiabase at the boundary with subjacent schist. Wind direction is bottomto top in the photo. Trowel for scale is 28 cm long.
98 J. Knight / Earth-Science Reviews 86 (2008) 89–105
which are consistently west-facing and which resemblea hammerhead in profile, are developed throughabrasive undercutting of the diabase near its contactwith the subjacent softer schist (Fig. 9). The hammer-head forms (30 cm long, 20–25 cm high) are undercutfrom their base to about two thirds of the way up theirtotal height, and show vague wind-parallel groovesextending around their lateral margins (Fig. 10).Adjacent schist surfaces are consistently polished butdo not show any other ventifact forms.
The macroscale features commonly described onventifacts (facet, keel) are important and quantifiablemorphological properties that can, with care, revealmacroscale data on wind direction and abrasion process.More revealing, however, are mesoscale forms that aredeveloped across facet surfaces, because they areclosely related to both detailed wind direction andparticle transport and abrasion processes.
4.2. Mesoscale features
In terms of its spatial extent across facet surfaces, themost common mesoscale (b few tens of cm diameter)feature observed is polish. This refers to the develop-ment of a light-reflecting sheen across the facets of, inparticular, fine-grained rocks that are polished throughthe abrasive action of quartz sand grains (and also ice/snow crystals or dust particles) blown across theirsurfaces (Whitney and Dietrich, 1973; Dietrich, 1977;Greely and Iversen, 1985; Schlyter, 1994; Mackay andBurn, 2005). Polish is developed in particular on finegrained sedimentary and metasedimentary bedrocks that
are rich in quartz and feldspars (Christiansen andSvensson, 1998), and the polish is independent of theorientation of grain/crystal faces within the rock mass.Polish is less well developed on coarser rocks such asgranite because of its granular surface texture, and itsmechanism of surface abrasion (by crystal-by-crystaldetachment, leading to pits (discussed below), ratherthan across-crystal smoothing, leading to polish). Polishis also developed progressively as ventifact facets areshaped, and is often found consistently across all facetsirrespective of their facing direction (Christiansen,2004). This may suggest the role of meso-/microscale
Fig. 10. Close-up of a hammerhead ventifact form showing pits andvague grooves on the undercut surface. Wind direction is left to right inthe photo. Trowel for scale is 28 cm long.
Fig. 11. Small pits developed on quarried diabase boulders in coastalOregon, western USA (from Knight and Burningham, 2003). Trowelfor scale is 28 cm long.
boundary layer turbulence in recirculating fine particlesaround the margins and leesides of obstacles.
Pits are also common features on ventifact facets andare defined as mesoscale (and sometimes microscale)enclosed depressions on the facet surface formed bygrain impaction (Fig. 11). This impact process acts todislodge, or lever out, individual crystals or grains fromthe host rock, and therefore exploits pre-existingweaknesses or crystalline/mineralogical variabilitywithin the rock. Pits may have a variable width (lessthan 1 mm to 4 mm) and depth (less than 1 mm to 6 mm)but are generally circular in plan (Whitney and Dietrich,1973). Pitting of ventifact surfaces is caused bysaltation. Experimental studies show that sedimenttransport by saltation is concentrated in the lowermost0.5 m or so of the windstream (e.g. Dong et al., 2002;Bridges et al., 2004a, 2005), thus that saltation impactshould be focused at this level. Bridges et al. (2005) alsodescribe results from wind tunnel experiments whereparticles are impacted on targets with different steep toshallow facets. Particle trajectories across these facetsvary considerably, with steep facets (90°) resulting insaltating particles rebounding directly away from thefacet; intermediate facets (30–45°) where particlesrebound at an angle away from the facet (in bothforward and backward directions); and shallow facets(15°) where particles rebound up the length of the facet.These results confirm field observations in which pitsare best developed on steep facets that are alignedperpendicular to wind direction (Knight and Burning-ham, 2001). Pits on these steep facets are generallysmall, deep, and extend into the rock facet at right angles(Fig. 11). Conversely, pits are least well developed onshallowly-dipping facets that are aligned oblique to
windflow. Here, pits are elongate, shallow, asymmetric,and reflect blown-sand transport across the ventifactfacet. Such elongate pits are transitional to grooved andfluted abrasion forms (discussed below).
Wind-parallel grooves that are symmetric in longprofile are often developed on ventifact facets. Groovescan form alternating ridge and furrow-like features orcorrugations that extend up the facet, often changing inspacing upwards and giving a divergent, fan-likeappearance in plan view (Sharp, 1949; Bishop andMildenhall, 1994; Knight and Burningham, 2003).Grooves can also form discontinuous and symmetric‘scoops’ up the facet, giving a net-like appearance inplan view (Knight and Burningham, 2001). The spacingbetween grooves is consistent across the facet. Groovesare generally 2–6 cm long, up to 4 mm deep, and have aspacing of 4–8 mm (Braley and Wilson, 1997; Wilsonand Edwards, 2004; Knight, 2005). Similar wind-parallel forms termed flutes are also sometimesobserved. Flutes can be distinguished from groovesbecause they are asymmetric in long profile, are moremorphologically diverse (Maxson, 1940), and tend tobroaden and flatten in a downwind direction (Sharp,1949; Svensson, 1983; Greely and Iversen, 1985; Hoareet al., 2002). Fisher (1996) described asymmetric forms(that he termed rat-tails), morphologically similar toflutes, on both windward and leeside ventifact faces.The leeside forms are interpreted as due to reattachmenteddies formed and moving up the leeward face. Inpractice, grooves and flutes are difficult to distinguishfrom one another (Whitney and Dietrich, 1973), and theformer term is preferred in this paper to describe allwind-parallel concavities developed on ventifact facets.
Grooves also develop by the exploitation ofweak spotson a rock surface, including relatively softer feldspars, anddifferential erosion can pick out irregularities such as
Fig. 12. Wind-parallel grooves (developed from bottom to top in thephoto) on the upper surface of a ventifact. Trowel for scale is 28 cmlong.
100 J. Knight / Earth-Science Reviews 86 (2008) 89–105
quartz veins (Fig. 12) and harder, upstanding mineralssuch as garnet. In addition, asymmetric grooved formsthat resemble fluvial scallops can also be developed whenweathering pits widen (Maxson, 1940; Selby, 1977;Miotke, 1982). Wind processes associated with theformation of grooves have not been studied in detail,but helicoidal vortices that develop up medium-slopingfacets are likely important (Whitney, 1979; Fisher, 1996;Knight and Burningham, 2003). These are closelyanalogous to bedrock forms and flow processes knownfrom fluvial and subglacial settings (e.g. Gray, 1981;Allen, 1982 ch7; Dreimanis, 1993).
A further surface mesoform, similar to grooves butuncommonly reported, is that of asymmetric teardrop-shaped forms which are developed on steep windwardfacets. This form has been reported from two contem-porary coastal locations (in Northern Ireland andOregon, USA) where quarried boulders have anelevated position on a rock-built jetty adjacent to abroad, sandy beach. The teardrop-shaped forms com-prise downward-facing and asymmetric protrusions,which have a positive relief, and which are up to 5 mmdeep and 5–8 mm across (Knight, 2003; Knight andBurningham, 2003). These forms contrast with grooves,which have a negative relief. The teardrop-shaped formsalso have a very variable spacing across the facet,leading to a ‘knobbly’ surface texture. The downward-facing nature of these forms strongly suggests that sandis blown upwards from the adjacent beach and so is ableto impact the boulder face from below.
Macro- and mesoscale features together form a widerange of ventifact morphologies. They can occurtogether or singly, be superimposed, or succeed eachother over time (Sharp, 1949; Whitney and Dietrich,1973). In addition, these features are key to identifyingventifact-shaping processes, and therefore have impor-tant palaeoenvironmental implications in the interpreta-tion of ventifacts in the geological record.
4.3. Microscale features
Microscale features are those that are clearly visibleonly under hand lens, microscope or scanning electronmicroscope (SEM), including sub-millimetre-scale pitsand microfractures. These features reflect the action ofwind-carried abraders across a ventifact facet. Pits,which have a morphological range identical to thosedeveloped on the mesoscale, can be formed by twoprocesses: impaction and cavitation. Impaction of coarseabraders against a rock surface can help dislodgeindividual crystals or grains from the surface, leavinga ‘hole’ or pit. Impaction can also result in the chipping
away of the rock surface, leading to the formation ofpercussion fractures and crescentic gouges that reflectthe direction of abrader transport across the surface(Sellier, 2006). Such impaction, by both dislodgementand chipping, helps increase surface micro-roughness.Increased roughness and the formation of pits can leadto cavitation, by a process of positive feedback, due todifferential air pressure between the top and bottom ofthe pit (Whitney, 1979). These microscale processes arenot considered in detail here because they do not clearlyreflect characteristics of wind direction and speed, andare not easily observed in the field or hand specimen.
5. The use of ventifacts to evaluate present andreconstruct past wind climates
5.1. Ventifacts in modern environmental settings
The relationship between the macro- and mesoscalefeatures found on ventifacts, and wind climate, can beevaluated with most confidence from modern-daysettings where observations on ventifacts in the fieldcan be matched directly with numerical wind data. Onlya handful of studies have taken this approach (e.g. ClarkandWilson, 1992; Knight and Burningham, 2001, 2003;Knight, 2003, 2005), largely because ventifacts oftenappear to be relict and not related to present-day windclimate.Whilst this is an interpretive andmethodologicalproblem, it is also true to say that the likely time periodrequired for ventifact development in modern fieldsettings is very much longer than that of available windrecords. In a sense, therefore, our interpretive power isweakened by this mismatch in scale of observation.
This problem of linking ventifacts to winds can bepartly solved, however, where the age of ventifacts in
Table 1Table of calculated ventifact abrasion rates described from field studies
Location Material on which ventifacts are developed Abrasion rate(mm year− 1)
Abrasion period Reference
Oregon, USA Sandstone and diabase boulders 0.24–1.63 70 years Knight andBurningham (2003)
Northern Ireland Basalt boulders 0.01–0.05 b 100 years Knight (2003)Victoria Land,
AntarcticaVarious lithologies 0.05–3.70 407 days Malin (1986)
Princess ElizabethLand, Antarctica
Various lithologies 0.015–0.022 4 years Spate et al. (1995)
Victoria Land,Antarctica
Various lithologies 8.04–23.40 ‘a few decades or withina few centuries at most’
Miotke (1982)
California, USA Various lithologies 0.09–6.77 6–10 years Sharp (1964)California, USA Various lithologies 1.86–36.0 15 years Sharp (1980)Victoria Land,
AntarcticaVarious lithologies 0.1–0.3 c. 1000 years French and Guglielmin (1999)
the field can be temporally constrained. Currently, thereare two approaches that are able do this: (1) Ventifactsare found on boulder-built coastal jetties of known age,located at the mouths of rivers in Northern Ireland (UK)and Oregon (USA) (Knight, 2003; Knight and Burning-ham, 2003). Estimates of rock loss from bouldersurfaces can help evaluate abrasion rates over thesewell-defined time scales (Table 1). (2) Repeated fieldobservations of ventifact development over decadal timescales have been made at locations in California (USA)and Northwest Territories (Canada) (Sharp, 1964, 1980;Mackay and Burn, 2005), and on shorter time scales inAntarctica (Miotke, 1982; Malin, 1986; Spate et al.,1995). Together, these observations also enable somelong-term abrasion rates to be calculated (Table 1). Onlyone of the above studies, however (Knight andBurningham, 2003), discusses wind climate in anydetail. In addition, one further field study has shownthat, even if contemporary wind patterns are known,there is a wide variation in reconstructed wind flowdirections (by up to 90° from the prevailing wind)caused mainly by local topographic response ofboundary-layer winds (Knight, 2005). This uncertaintysuggests that wind reconstruction from ventifacts maybe scale-dependent, and may be most accurate at themesoscale where local topographic effects are minimal.
Laboratory-based wind tunnel studies have also beenused to quantify abrasion rates over very short (sub-hourly) time scales (e.g. Kuenen, 1928; Schoewe, 1932;Kuenen, 1960; Suzuki and Takahashi, 1981; Bridgeset al., 2004a, 2005). These studies are useful becausethey show abrasion under controlled environmentalconditions in which sediment supply and wind speed areknown, but their results should be interpreted asmaximum (and transient) abrasion values since they
do not take into account turbulent effects, sedimentpulses, natural grain size spectra, or variable winds.These studies are most useful in their comparison withAntarctic ventifacts, which are associated with verystrong (N 100 mph), sustained (over several hours) andunidirectional winds (e.g. Miotke, 1982; Malin, 1986).What these studies do not provide information on,however, is long-term (centennial to millennial-scale)and episodic abrasion, which is likely to characterisemany ventifacts in past (late Pleistocene and Holocene)environmental settings (e.g. Clark and Wilson, 1992).
5.2. Ventifacts in past environmental settings
Spatial data from relict ventifacts distributed acrosslarge areas have been used to reconstruct past regional-scale wind patterns (e.g. Schlyter, 1995; Christiansen,2004). The likely accuracy of these wind maps dependson the size of the area examined and its physicalgeography, the density of data points, and the accuracywith which former wind direction can be reconstructed.Wind flows reconstructed by this process are often justshown as straight lines on a map with little sense ofdynamic behaviour, the positioning of high and lowpressure cells, interaction with the nature of the landsurface (including topography), or relationship to otherwind-formed features such as loess, coversand, dunesetc. A further interpretive challenge specific to perigla-cial environments is that it is often difficult todisentangle synoptic-scale wind patterns that are relatedto the positioning of high and low pressure cells, jetstream and storm tracks (Rudberg, 1968); and katabaticwinds that are controlled mainly by proximity tocontinental ice sheets (Bromwich, 1989; Thorson andSchile, 1995) and which tend to overwhelm the synoptic
102 J. Knight / Earth-Science Reviews 86 (2008) 89–105
signature. Since it is the ‘background’ synoptic-scalewinds (developed over oceans) that are most importantin terms of global palaeoclimate reconstruction, this is asignificant problem. The role of topography, on differentscales, in determining wind direction also means thatreconstructed winds from mountain settings show a highdegree of variability imparted by high land relief (e.g.Christiansen, 2004) and thus may reveal little aboutsynoptic wind patterns. This contrasts with winds in flat,proglacial settings which might be most reliable in termsof regional directional consistency but which are likelydominated by katabatic processes.
6. Discussion
Ventifacts are a somewhat obscure and neglectedlandform but one that has a long history of investigation(since the 1850s). Whilst our basic understanding ofwind abrasion processes over this investigative periodhas not progressed markedly, the environmental signif-icance of ventifacts (in both palaeo and contemporarysettings), and their power as interpretive tools, is nowbeing realised. One key outcome from ventifact studies(in combination with unrelated advances in dating) isthe identification of relatively stable and relict landsurfaces that are of regional extent, such as in Antarctica(e.g. Sugden et al., 1999). Conversely, studies also nowreveal that wind abrasion is an active contemporaryprocess in the midlatitudes (e.g. Knight and Burning-ham, 2003).
These studies also highlight the global-scale mis-match between the locations of (1) presently-activewind-deposited features (including active desert dunes);(2) relict surface features indicating past wind deposi-tion (including loess and coversand); and (3) active andrelict ventifacts (cf. Fig. 3). This mismatch is evidencedby the fact that ventifacts have only infrequently beenfound in association with wind-deposited sediments(e.g. Bishop and Mildenhall, 1994; Wilson, 1991; Hoareet al., 2002). It is still not clear why this mismatch exists,but it is a problem that may be best addressed using asediment systems approach that focuses on identifyingand quantifying sediment transport vectors, fluxes andbudgets within closed sediment cells such as beach-dunesystems (Knight, 2002). An advantage of this approachis that ventifacts can be considered as part of acontinuum of evidence for wind activity that includesboth erosional and depositional landforms and sedi-ments. It also means that more complex interpretationsof past wind direction can be made, such as links withloess (sedimentary structures, directional indicators,mineralogy) and desert dune dynamics (Laity, 1992).
In addition, it suggests that the ventifact record alone isinsufficient to draw conclusions on wind strength anddirection, and should be informed by evidence fromwind-depositional landforms and sediments, and GCMs(Isarin et al., 1997).
More widely, understanding of past wind direction,particularly during the last glacial–interglacial transi-tion, has implications for the understanding of globalclimate reorganisation, and links between atmosphericcirculation forcing and land/biosphere response. Includ-ed in this are topics as diverse as patterns of postglacialhuman migration; seed dispersal and colonisation; andavian fauna migration. This understanding also impactson present environmental issues such as patterns ofpollution and contaminant transport and dispersal,monitoring and modelling. Ventifacts therefore providecritical data on past and contemporary boundary layerclimates that can input into these wider issues, and yielda better understanding of the dynamics of climatically-sensitive, sediment-rich environments.
6.1. Future directions in ventifact research
The wider significance of ventifact studies, outlinedabove, and technological advances, point to excitingfuture directions in ventifact research. A clear researchimperative is the need for basic field observations inlocations that are likely susceptible to wind abrasion,including glaciated and high-energy coasts and on bothpassive and constructive sand-dominated continentalmargins and interiors. This includes South America,central Asia, and Australasia. There is also a need tointerrogate the non-English literature on ventifacts andother associated phenomena, particularly in French andGerman.
Further understanding of ventifact development canalso be achieved using techniques such as laser scanning(e.g. Bridges et al., 2004b) and repeat SEM analysis inorder to quantify microscale processes and abrasionrates (Whitney, 1979; Sellier, 2006). Dating (particular-ly K/Ar and cosmogenic techniques) should also beapplied to ventifacted surfaces. This can be done inconcert with studies in related areas of desert geomor-phology such as development of weathering rinds andvarnish (Laity, 1992, 1995). The role of erosion bybiological processes is also a developing area ofresearch that impacts on issues of rock surface stability(e.g. Naylor et al., 2002).
Processes of ventifact abrasion can also be examined inmore detail. Although laboratory models of aeoliansediment fluxes are highly developed (e.g. Dong et al.,2002), these have not commonly been linked to aspects of
either wind abrasion or boundary-layer issues of micro-topography, turbulence, and grain–grain interaction.These issues can be explored in more detail using particleimaging velocimetry (PIV) in which the movement ofparticles is mapped over very short time intervals usingpulsed laser or photography/video techniques (e.g. Will-ert, 1997; Adrian, 2005; Zhang et al., 2007) and, to somedegree, field observations of wind-blown sedimentdynamics (e.g. Anderson, 1986; Arens, 1996; Namikas,2003). These laboratory and field studies also need to belinked more closely with numerical models of particle/aerosol dispersal, turbulent diffusion, and fluid interactionwith boundary layers of different macro-/mesotopogra-phy; and be informed by studies involving fractal (scaleinvariant) processes of particle/boundary layer interac-tion, and artificial neural networks. Future integration ofventifact studies with these research strands in differentfields demonstrates the potential of ventifacts in aidingunderstanding of sediment-wind interactions and dynam-ics over different scales.
7. Conclusions
Ventifacts are important because they provideevidence for sand and other abraders (such as dust andsnow/ice crystals) that are in transit between source andsink. As such, they represent the ‘missing link’ betweenthe process of wind-blow, and depositional evidence forthis process in the form of sand dunes, loess etc (Fig. 1).Ventifacts therefore offer an unprecedented and uniqueinsight into past wind processes, and are useful in thereconstruction of former wind flow conditions (includ-ing wind strength, direction and permanence) for whichthere may be otherwise little or no evidence.
Investigations of ventifact characteristics over thepast 150 years have evolved from purely descriptive tointerpretive and related to wider issues of sedimentbudgets and dynamics, particularly in periglacial andparaglacial environments. The macro- and mesoscaleforms developed on ventifacts can be used to identifythe direction of geomorphically effective winds and thedetailed processes of wind abrasion. This informationcan also help identify those locations that are activelyabrading at the present time, and can provide vital inputinto studies of sediment dynamics in present-day hot/cold deserts, coastal and periglacial settings.
Acknowledgements
I thank the editor and reviewers Nathan Bridges(JPL) and an anonymous reviewer for their constructivecomments on an earlier version of this paper.
References
Adrian, R.J., 2005. Twenty years of particle image velocimetry.Experiments in Fluids 39, 159–169.
Allen, J.R.L., 1982. Sedimentary Structures — their character andphysical basis. Developments in Sedimentology 30B, vol II.Elsevier, Amsterdam.
Anderson, R.S., 1986. Erosion profiles due to particles entrained bywind: application of an eolian sediment-transport model. Geolog-ical Society of America Bulletin 97, 1270–1278.
Arens, S.M., 1996. Rates of aeolian transport on a beach in a temperatehumid climate. Geomorphology 17, 3–18.
Ashwell, A.Y., 1966. Glacial control on wind and of soil erosion inIceland. Annals of the Association of American Geographers 56,529–540.
Bather, F.A., 1900. Wind-worn pebbles in the British Isles. Proceed-ings of the Geologists' Association 16, 396–420.
Bishop, D.G., Mildenhall, D.C., 1994. The geological setting ofventifacts and wind-sculpted rocks at Mason Bay, Stewart Island,and their implications for late Quaternary paleoclimates. NewZealand Journal of Geology and Geophysics 37, 169–180.
Blackwelder, E., 1929. Sandblast action in relation to the glaciers ofthe Sierra Nevada. Journal of Geology 37, 256–260.
Blake, W.P., 1855. On the grooving and polishing of hard rocks andminerals by dry sand. American Journal of Science 20, 178–181.
Bosworth, T.O., 1910. Wind erosion on the coast of Mull. GeologicalMagazine 7, 353–355.
Braley, S.M., Wilson, P., 1997. Ventifacts from the coast ofNorthumberland. Proceedings of the Geologists' Association108, 141–147.
Bridges, N.T., Greeley, R., Haldemann, A.F.C., Herkenhoff, K.E.,Kraft, M., Parker, T.J., Ward, A.W., 1999. Ventifacts at thePathfinder landing site. Journal of Geophysical Research 104,8595–8615.
Bridges, N.T., Laity, J.E., Greeley, R., Phoreman, J., Eddlemon, E.E.,2004a. Insights on rock abrasion and ventifact formation fromlaboratory and field analog studies with applications to Mars.Planetary and Space Science 52, 199–213.
Bridges, N.T., Razdan, A., Greeley, R., Laity, J.E., 2004b. Highresolution laser scanning techniques for rock abrasion and textureanalyses onMars and Earth. Lunar and Planetary Science 35, 1897.
Bridges, N.T., Phoreman, J., White, B.R., Greeley, R., Eddlemon, E.E.,Wilson, G.R., Meyer, C.J., 2005. Trajectories and energy transferof saltating particles onto rock surfaces: application to abrasion andventifact formation on Earth and Mars. Journal of GeophysicalResearch 110 (E12004). doi:10.1029/2004JE002388.
Bromwich, D.H., 1989. An extraordinary katabatic wind regime atTerra Nova Bay, Antarctica. Monthly Weather Review 117,688–695.
Bryan, K., 1931. Wind-worn stones or ventifacts — a discussion andbibliography. Report of the Committee on Sedimentation, 1929–1930. Reprint and Circular Series of the National Research Council98, 29–50.
Christiansen, H.H., 2004. Windpolished boulders and bedrock in theScottish Highlands: evidence and implications of Late Devensianwind activity. Boreas 33, 82–94.
Christiansen, H.H., Svensson, H., 1998. Windpolished boulders asindicators of a Late Weichselian wind regime in Denmark inrelation to neighbouring areas. Permafrost and PeriglacialProcesses 9, 1–21.
Clark, T.H., Elson, J.A., 1961. Ventifacts and eolian sand at Charette,P.Q. Transactions of the Royal Society of Canada 55, 1–11.
indicators in north-west Saskatchewan, Canada, 11 ka to 9.9 ka BP.Permafrost and Periglacial Processes 7, 391–408.
French, H.M., Guglielmin, M., 1999. Observations on the ice-marginal, periglacial geomorphology of Terra Nova Bay, northernVictoria Land, Antarctica. Permafrost and Periglacial Processes 10,331–347.
Gill, T.E., Warren, A., Stout, J.E., 2007. Bibliography of AeolianResearch. Available from http://www.lbk.ars.usda.gov/wewc/bib-lio/bar.htm (accessed 11 June 2007).
Gray, J.M., 1981. p-forms from the Isle of Mull. Scottish Journal ofGeology 17, 39–47.
Greely, R., Iversen, J.D., 1985. Wind as a geological process on Earth,Mars, Venus and Titan. Cambridge University Press, Cambridge.
Grolier, M.J., McCauley, J.F., Breed, C.S., Embabi, N.S., 1980.Yardangs of the Western Desert. Geographical Journal 146, 86–87.
Grove, J.M., 2004. Little Ice Ages: ancient and modern. Routledge,London.
Halimov, M., Fezer, F., 1989. Eight yardang types in Central Asia.Zeitschrift für Geomorphologie NF 33, 205–217.
Hickox Jr., C.F., 1959. Formation of ventifacts in a moist, temperateclimate. Bulletin of theGeological Society ofAmerica 70, 1489–1490.
Higgins, C.G., 1956. Formation of small ventifacts. Journal ofGeology 64, 506–516.
Hoare, P.G., Stevenson, C.R., Godby, S.P., 2002. Sand sheets andventifacts: the legacy of aeolian action in west Norfolk, UK.Proceedings of the Geologists' Association 113, 301–317.
Isarin, R.F.B., Renssen, H., Koster, E.A., 1997. Surface wind climateduring the Younger Dryas in Europe as inferred from aeolianrecords and model simulations. Palaeogeography, Palaeoclimatol-ogy, Palaeoecology 134, 127–148.
Jackson,D.W.T.,Cooper, J.A.G., 1999. Formation of ephemeral bedformturrets in coastal foredunes. Journal of Geology 107, 633–639.
King, L.C., 1936. Wind-faceted stones from Marlborough, NewZealand. Journal of Geology 44, 201–213.
Knight, J., 2002. Wind abrasion (ventifaction) on Donegal and Oregoncoasts and implications for the sediment dynamics of coastalsystems. Journal of Coastal Research Special Issue 36, 441–449.
Knight, J., 2003. A note on the formation of ventifacts at Castlerock,Northern Ireland coast. Irish Journal of Earth Sciences 21, 39–45.
Knight, J., 2005. Controls on the formation of coastal ventifacts.Geomorphology 64, 243–253.
Knight, J., Burningham, H., 2001. Formation of bedrock-cut ventifactsand late Holocene coastal zone evolution, County Donegal,Ireland. Journal of Geology 109, 647–660.
Knight, J., Burningham, H., 2003. Recent ventifact development onthe Oregon coast, western USA. Earth Surface Processes andLandforms 28, 87–98.
Kolstrup, E., 1997. Wind blown sand and palynological records in pastenvironments. Aarhus Geoscience 7, 91–100.
Kolstrup, E., 2007. Lateglacial older and younger coversand innorthwest Europe: chronology and relation to climate andvegetation. Boreas 36, 65–75.
Kuenen, P.H., 1928. Experiments on the formation of wind-wornpebbles. Leidsche Geologische Medellinger 3, 17–28.
Laity, J.E., 1992. Ventifact evidence for Holocene wind patterns in theeast-central Mojave Desert. Zeitschrift für Geomorphologie NFSupplementbund 84, 73–88.
Laity, J.E., 1995. Wind abrasion and ventifact formation in California.In: Tchakerian, V.P. (Ed.), Desert Aeolian Processes. Chapman andHall, London, pp. 295–321.
Leonard, A.J., Moore, A.G., Selwood, E.B., 1982. Ventifacts from adeflation surface marking the top of the Budleigh Salterton PebbleBeds, east Devon. Proceedings of the Ussher Society 5, 333–339.
Lindsay, J.F., 1973. Ventifact evolution in Wright Valley, Antarctica.Geological Society of America Bulletin 84, 1791–1798.
Mackay, J.R., Burn, C.R., 2005. A long-term field study (1951–2003)of ventifacts formed by katabatic winds at Paulatuk, westernArctic coast, Canada. Canadian Journal of Earth Sciences 42,1615–1635.
Malin, M.C., 1986. Rates of geomorphic modification in ice-free areassouthern Victoria Land, Antarctica. Antarctic Journal of the UnitedStates 20, 18–21.
Maxson, J.H., 1940. Fluting and faceting of rock fragments. Journal ofGeology 48, 717–751.
Miotke, F.-D., 1982. Formation and rate of formation of ventifacts inVictoria Land, Antarctica. Polar Geomorphology 6, 98–113.
Namikas, S.L., 2003. Field measurement and numerical modelling ofaeolian mass flux distributions on a sandy beach. Sedimentology50, 303–326.
Naylor, L.A., Viles, H.A., Carter, N.E.A., 2002. Biogeomorphologyrevisited: looking towards the future. Geomorphology 47, 3–14.
Parish, T.R., Bromwich, D.H., 1991. Continental-scale simulationof the Antarctic katabatic wind regime. Journal of Climate 4,135–146.
Powers, W.E., 1936. The evidences of wind abrasion. Journal ofGeology 44, 214–219.
Pye, K., Tsoar, H., 1990. Aeolian Sand and Sand Dunes. UnwinHyman, London.
Ritter, D.F., Dutcher, R.R., 1990. Geomorphic controls on the originand location of the Tolman Ranch ventifact site, Park County,Wyoming, USA. Journal of Geology 98, 943–954.
Rudberg, S., 1968. Wind erosion — preparation of maps showing thedirection of eroding winds. Biuletyn Peryglacjalny 17, 181–193.
Schlyter, P., 1994. Paleo-ventifact formation by suspended silt or snow—site studies in south Sweden. Geografiska Annaler 76A, 187–201.
Schlyter, P., 1995. Ventifacts as palaeo-wind indicators in southernScandinavia. Permafrost and Periglacial Processes 6, 207–219.
Schoewe, W.H., 1932. Experiments on the formation of wind-facetedpebbles. American Journal of Science 140, 111–134.
Selby, M.J., 1977. Transverse erosional marks on ventifacts fromAntarctica. New Zealand Journal of Geology and Geophysics 20,949–969.
Selby, M.J., Palmer, R.W.P., Smith, C.J.R., Rains, R.B., 1973.Ventifact distribution and wind directions in the Victoria Valley,Antarctica. New Zealand Journal of Geology and Geophysics 16,303–306.
Sellier, D., 2006. Contribution à la caractérisation des marques de lacorrosion éolienne par l'analyse exoscopique de ventifacts duMassifarmoricain. Zeitschrift für Geomorphologie NF 50, 393–420.
Seppälä, M., 2004. Wind as a Geomorphic Agent in Cold Climates.Cambridge University Press, Cambridge.
Sharp, R.P., 1949. Pleistocene ventifacts east of the Big HornMountains, Wyoming. Journal of Geology 57, 175–195.
Sharp, R.P., 1964. Wind-driven sand in Coachella Valley, California.Geological Society of America Bulletin 75, 785–804.
Sharp, R.P., 1980. Wind-driven sand in Coachella Valley, California:further data. Geological Society of America Bulletin 91,724–730.
Shepherd, M., 2003. Wave-faceted pebbles, Sovi Bay Beach, VitiLevu, Fiji. Zeitschrift für Geomorphologie, N.F. 47, 439–448.
Spate, A.P., Burgess, J.S., Shevlin, J., 1995. Rates of rock surfacelowering, Princess-Elizabeth-Land, eastern Antarctica. EarthSurface Processes and Landforms 20, 567–573.
Stowe, E., 1872. On the effects of wind-blown sand as a cutting agent.Transactions and Proceedings of the Royal Society of NewZealand 5, 105–106.
Sugden, D.E., Summerfield, M.A., Denton, G.H., Wilch, T.I.,McIntosh, W.C., Marchant, D.R., Rutford, R.H., 1999. Landscapedevelopment in the Royal Society Range, southern Victoria Land,Antarctica: stability since the mid-Miocene. Geomorphology 28,181–200.
Suzuki, T., Takahashi, K., 1981. An experimental study of windabrasion. Journal of Geology 89, 23–36.
Svensson, H., 1983. Ventifacts as paleowind indicators in a formerperiglacial area of southern Scandinavia. Proceedings of the 4thInternational Conference on Permafrost. National Academy Press,Washington DC, pp. 1217–1220.
Thompson, D.B., Worsley, P., 1967. Periods of ventifact formation inthe Permo-Triassic and Quaternary of the north east CheshireBasin. Mercian Geologist 2, 279–298.
Tomkin, J.H., Braun, J., 2002. The influence of alpine glaciation on therelief of tectonically active mountain belts. American Journal ofScience 302, 169–190.
Travers, W.T.L., 1870. On the sandworn stones of Evans' Bay.Transactions of the New Zealand Institute 2, 247–248.
Tremblay, L.P., 1961. Wind striations in northern Alberta andSaskatchewan, Canada. Geological Society of America Bulletin72, 1561–1564.
Wade, A., 1910. On the formation of dreikante in desert regions.Geological Magazine 7, 394–398.
Wentworth, C.K., Dickey, R.I., 1935. Ventifact locality in the UnitedStates. Journal of Geology 43, 97–105.
Whitney, M.I., 1979. Electron micrography of mineral surfaces subjectto wind-blast erosion. Geological Society of America Bulletin 90,917–934.
Whitney, M.I., Dietrich, R.V., 1973. Ventifact sculpture by windblowndust. Geological Society of America Bulletin 84, 2561–2582.
Willert, C., 1997. Stereoscopic digital particle image velocimetry forapplication in wind tunnel flows. Measurement Science andTechnology 8, 1465–1479.
Wilshire, H.G., Nakata, J.K., Hallet, B., 1981. Field observationsof the December 1997 wind storm, San Joaquim Valley, California.In: Péwé, T.L. (Ed.), Desert Dust: Origin, Characteristics, andEffect on Man. Geological Society of America Special Paper,vol. 186, pp. 233–251.
Wilson, P., 1991. Sediment clasts and ventifacts from the north coast ofNorthern Ireland. Irish Naturalists' Journal 23, 446–450.
Wilson, P., Christiansen, H.H., Ross, S.M., 2002. The geomorpho-logical context and significance of wind-abraded gravels, bouldersand outcrops from the coast of Scotland. Scottish GeographicalJournal 118, 41–57.
Wilson, P., Edwards, E.J., 2004. Further examples of ventifacts andunusual patterned ground from the Falkland Islands, SouthAtlantic. Geografiska Annaler 86A, 107–115.
Wilson, P., McGourty, J., Bateman, M.D., 2004. Mid- to late-Holocenecoastal dune event stratigraphy for the north coast of NorthernIreland. Holocene 14, 406–416.
Woodworth, J.B., 1894. Post-glacial eolian action in southern NewEngland. American Journal of Science 47, 63–71.
Zhang, W., Wang, Y., Lee, S.-J., 2007. Two-phase measurements ofwind and saltating sand in an atmospheric boundary layer.Geomorphology 88, 109–119.
