Meteoritics amp Planetary Science 41 Nr 2 237ndash246 (2006)Abstract available online at httpmeteoriticsorg

237 copy The Meteoritical Society 2006 Printed in USA

Investigation of Shuttle Radar Topography Mission data of the possible impact structure at Serra da Cangalha Brazil

Wolf U REIMOLD1 Gordon R J COOPER1 Rafael ROMANO2 Duncan R COWAN3 and Christian KOEBERL4

1Humboldt-Universitpermilt zu Berlin Museum fcedilr Naturkunde Institut fuumlr Mineralogie D-10115 Berlin Germany2Department of Geology Federal University of Ouro Preto Ouro Preto Brazil

3Cowan Geodata Services 12 Edna Road Dalkeith Western Australia Australia4Department of Geological Sciences University of Vienna Althanstrasse 14 A-1090 Vienna Austria

Corresponding author E-mail uwereimoldmuseumhu-berlinde

(Received 24 February 2005 revision accepted 18 September 2005)

AbstractndashThe Serra da Cangalha crater structure in northeast Brazil sim13 km in diameter has longbeen widely considered to be a confirmed impact structure based on reports of shatter cone findingsOnly very limited field work has been carried out at this crater structure Landsat Thematic Mapper(TM) and Shuttle Radar Topography Mission (SRTM) data sets for the region around this craterstructure are compared here with regard to their suitability to determine first-order structural detail ofimpact crater structures The SRTM data provide very detailed information regarding drainagepatterns and topography A pronounced central ring of up to 300 m elevation above the surroundingarea two comparatively subdued intermediate rings of 6 and 105 km diameter respectively and thebroad complex crater rim of up to gt100 m elevation can be distinguished in the Serra da Cangalhadata The maximum cratering-related regional deformation (radial and concentric features) seems tobe limited to a radial distance of 16ndash18 km from the center of the structure A first comparison ofmacrostructural information from several impact structures with that from Serra da Cangalha does notyield firm trends but the database is still very small at this stage The varied nature of the targetgeology strongly influences the development of structural features in any impact event

INTRODUCTION AND BACKGROUND

The Serra da Cangalha structure is centered at 8deg05primeS and46deg52primeW in the extreme northeast of Tocantins state in Brazil(Fig 1) Based on findings of shatter cones it is widely heldthat Serra da Cangalha is indeed a confirmed impact structure(Dietz and French 1973 McHone 1979 Santos and McHone1979 CrUcircsta 1987 2004) although bona fide microscopicevidence of shock metamorphism has never been reported(CrUcircsta 2004) Past authors assigned a diameter of 12ndash13 kmto the structure which is comprised of several ring featuresMost prominent of these is a 3 km (5 km according to DeCicco and Zucolotto 2002) inner circular ring of mountains250ndash300 m high that lends this structure an appearancesimilar to that of the Gosses Bluff structure of Australiawhich has a 45 km inner ring and a subdued outer rim featureat 24 km diameter (Milton et al 1996) Similar to GossesBluffrsquos inner ring structure and that of the Libyan impactcrater Oasis (Koeberl et al 2005a) the inner ring at Serra daCangalha is thought to represent the differentially eroded

remnant of a lithologically diverse central uplift featureformed from a layered sequence of target rocks

The Serra da Cangalha structure was formed in theintracratonic ParnaIgraveba basin (formerly known as theMaranhatildeo basin) The basin stratigraphy involves UpperSilurian to Cretaceous sedimentary rocks The geology of theregion and the structure was reviewed by CrUcircsta (1982 1987and references therein) The strata within the structure includeupper Permian sandstones of the Pedra de Fogo FormationPermianCarboniferous sandstones of the 323ndash290 Ma PiauIgraveFormation and the 354ndash323 Ma Poti Formation as well asdark shales of the Upper Devonian Longmiddot Formation Thestructure is surrounded by tabular outliers (mesas) of Triassicsandstones of the SambaIgraveba Formation In the center of thestructure these strata are intensely deformed and displayvertical dips (CrUcircsta 1982) (presumably on bedding surfaces)On the basis of stratigraphic drilling in the region by theGeological Survey of Brazil the original stratigraphic depthof the strata now exposed in the center of the impact structureis estimated between 100 and 1300 m below surface which

238 W U Reimold et al

gives us an estimate on uplift gradient According toAdepelumi et al (2003a GUcirces et al 1993) the sedimentaryrocks in the region showed a preferred NE-SW depositionaldirection (which we interpret to mean that there is a regionalNE-SW-directed fabric)

Serra da Cangalha was first proposed as a possibleimpact structure by Dietz and French (1973) because of thecircular shape recognized in Landsat imagery (Fig 2) theabsence of volcanic rocks in drill core from the central part ofthe structure and because it appeared unlikely that diapirismcould account for the geometry of the structure (no carbonateor salt layers had been recognized in the sedimentary countryrock stratigraphy) Dietz and French (1973) and McHone(1979) referred to shatter cones occurring on quartziteboulders of a conglomerate from the base of the PotiFormation observed in the inner ring structure These authorsalso described intricate fracturing of quartz as well asmicrospherules occurring in microscopic fractures howeverneither of these constitutes proof for the existence of animpact structure CrUcircsta (1987) referred to occurrence ofshock metamorphic features in the form of ldquoshock lamellaerdquoand ldquobrecciardquo but did not provide further detail that could beused to confirm that these features represent bona fide shockdeformation The same author reported in 2004 that to dateno definitive evidence of shock metamorphism has beenreported for this structure (CrUcircsta 2004) Consequently withexception of the early reference to shatter cones no evidencefor impact has been reported from Serra da Cangalha

Consequently we consider Serra da Cangalha as only apossible impact structure still to be confirmed No firmconstraints for the age of the Serra da Cangalha structure havebeen obtained yet either Based only on stratigraphicconsiderations a maximum age of 250 Ma can be estimatedfor the formation of the crater structure (ie the structure wasformed in strata of Triassic or younger age)

Some geophysical analysis of the structure and modelingof the data has been performed in recent years (Adepelumiet al 2003a 2003b 2004 2005a 2005b) Interpretation ofaeromagnetic data (Adepelumi et al 2003b 2004) indicated adiameter of 127 km for the Serra da Cangalha structure andthe magnetotelluric investigation by Adepelumi et al (2004)resulted in the hypothesis that the likely impact-inducedstructural deformation below and in the environs of the craterdid not exceed a depth of 2 km A 2-D resistivity model(Adepelumi et al 2005a 2005b) suggested a 4-layer modelfor the structure a thin resistive layer underlain by aconductive layer weathered basement and ultimatelyresistive crystalline basement The depth to the basement isestimated at 11 km Three-dimensional forward modelingsignificantly reduced the basement resistivity the effect ofwhich was related by the authors to impact-inducedbrecciation fracturing alteration of the shock-deformedzone and the presence of new low-magnetic materials andfluids (see also Abraham et al 2004)

Here we use Shuttle Radar Topographic Mission(SRTM) data for the area of the Serra da Cangalha structure to

Fig 1 Location of the Serra da Cangalha crater structure in the Tocantins state of north-central Brazil

Investigation of Shuttle Radar Topography Mission data 239

investigate the morphology and geometry of the structure aswell as the regional influence exerted by this possible impactevent We also test the recently developed circular sunshadingmethod (Cooper 2003) with this application Finally we setout to investigate whether or not the occurrence of andspacing between specific morphological features in impactstructures follow a definite relationship Several imagesgenerated from the SRTM data over the Serra da Cangalhaarea were recently shown by Almeido-Filho et al (2005)though without any geological interpretation

DATA AND PROCESSING METHODOLOGY

Shuttle Radar Topography Mission (SRTM) data for theregion around the Serra da Cangalha structure were availablefor this study Global SRTM single pass radar interferometrydata (Farr and Kobrick 2000) were obtained by the STS-99space shuttle mission between 11 and 22 February 2000SRTM digital elevation model data have a horizontalresolution of 1 arc second (equivalent to 30 m at the equator)and a vertical resolution of 10 m for the C-band radar The

United States Geological Survey (USGS) is the responsibledata archiving agency data were made available by NASAJet Propulsion Laboratory Global 3 arc second data havebeen released whereas 1 arc second data are only availablefor North America Initial comparison between 3 arc secondSRTM and older GTOPO DEM (Global Topography 30 arcsecond Digital Elevation Model) of the USGS (Cowan andCooper 2003) showed that the resolution of SRTM DEM is asignificant improvement and will be particularly valuable inareas for which limited topographic data are available

The Thematic Mapper (TM) Landsat data available forthe Serra da Cangalha region (Fig 2) provide a means forcomparison with the SRTM data In addition severalenhancement methods were applied in this study Sunshadingtechniques such as fractional order sunshading (Cooper andCowan 2003a 2003b) and a technique that enhances circularanomalies (Cooper 2003) were also employed

A standard filter used to enhance linear features inimages is sunshading It determines the reflectance from thedata of a light source located at infinity Linear features thatlie parallel to the azimuth of the light source (the ldquosunrdquo) are

Fig 2 Landsat image (bands 7-4-2 as RGB) over the Serra da Cangalha impact structure The scene limits are 47deg01prime05primeprimeW to 46deg41prime50primeprimeWand 12prime21primeprimeS to 7deg57prime19primeprime In this and the following images the top of the image is north

240 W U Reimold et al

attenuated while those that lie orthogonal to it are enhanced(Horn 1982) Because sunshading is a form of high-passfiltering it enhances both detail and noise in an image Thesunshading filter uses the first horizontal derivatives of thedata but if noise is a problem then lower-order derivativesmay be used instead These derivatives are of non-integerorder and are best computed in the frequency domain (Cooperand Cowan 2003a) The results from three differentsunshading operations each of which uses derivatives ofdifferent order may be combined to form an RGB imageHigh-frequency features then appear blue while lowerfrequency features appear red (Cooper and Cowan 2003b)Impact structures are mostly near-circular the sunshadingfilter was therefore modified to enhance features that lieeither on or orthogonal to radial vectors that pass through achosen origin position This was achieved by making the sunazimuth a variable over the image rather than a constant(Cooper 2003) As shown below the circularity of this impactstructure is strongly enhanced by this method However inthe application of the circular sunshading technique to thesearch for further impact structures the user should be wellaware that there are a range of other geological features thatmay have similar geometry for example ring structurescaused by differential erosion above an intrusion kimberlitepipes or other volcanic features

RESULTS

Figure 2 shows an RGB color composite image ofLandsat TM bands 7-4-2 Band 7 is in the far infrared and hasa strong response from hydroxyl-bearing minerals and claysspecifically carbonates micas chlorite and amphibolesBand 4 is in the near infrared and has a strong response fromiron absorption and vegetation Band 2 corresponds to greenlight in the visible part of the spectrum It is useful inhighlighting both chlorophyll absorption and iron features Inthe figure the Serra da Cangalha structure is comprised of a3 km wide near-circular central ring around a centraldepression

Adepelumi et al (2004) stated that the structure appearsto be open to the northwest An intermediate ring featurewith a diameter of sim5 km was also indicated Theintermediate ring is surrounded by a broad and apparentlylow topography annulus of sim3ndash4 km width which isterminated by the outer rim of the structure at 6ndash65 km fromthe center The rim itself is well-defined due to the weak colorshading of the interior of the structure but not as apronounced topographic feature

A strong radial (and in some sectors concentric) drainagepattern extends outward from the outer rim Only a few radialdrainage lines originate in the annular trough presumablyexploiting radial faults that breach the outer rim The widerregional drainage pattern is distinctly different from thiscrater-near obviously impact-derived pattern It appears that

cratering-related structure the underlying reason for thecrater-near drainage pattern extends to a maximum distanceof 15ndash2 crater radii from the center of the Serra da Cangalhastructure

A NW-SE directed structural trend is obvious in therather straight geometry of the NE and SW sectors of theinner ring and in fracturing cutting across the SE sector of theouter rim

SRTM Data

The SRTM raw data for the area of the structure areshown in Fig 3 This provides a detailed representation of theintricate drainage pattern providing a much clearer patternthan the Landsat imagery (Fig 2) The regional pattern isobviously different from that at and around the Serra daCangalha structure with strong NW-SE and ENE-WSWcomponents In contrast the crater-related pattern isdominated by radial and annular drainage The SRTMfindings confirm that this trend does not exceed the 15ndash2crater radii limit that can also be estimated from the Landsatimage

Figure 4 gives detailed topographic information obtainedfrom the SRTM raw data along two perpendicular profiles (N-S and E-W) across the structure Despite some variation in thebackground elevation the two intermediate rings can bereadily recognized The crater rim seems to be a broad andstructurally complex annular zone Detailed structuralanalysis of the crater rim zone at the 105 km wide Bosumtwistructure in Ghana (Reimold et al 1998) and the very muchsmaller 2 km diameter BP structure in Libya (Koeberl et al2005a our group unpublished data) also revealed broadstructurally complex crater rims with annular radial andoblique faulting directions The pronounced central ring haselevations of up to 300 m above background The twointermediate rings are within background elevation but areobvious from their sharp local topographic gradients Thecrater rim is broad and complex and seemingly elevated up togt100 m above the crater interior The two elevation profilesdiffer significantly with regard to the definition of the craterrim which is very pronounced on both sides of the N-Sprofile but not so on the eastern side of the crater structureThe reason for this is a distinct breach of the crater rim on thatside as is evident in Figs 2 and 3 Peculiar ENE trendingfeatures extend from the breach in the crater rim outwardsWhile it is not possible to provide an explanation of this fromthe remote sensing data alone it is noteworthy that thisfeature trends parallel to the regional structural grain Inaddition the terrane in the environs of the crater structureshows significant elevation complexity with local occurrenceof mesas

A 3-D image of the SRTM data over the Serra daCangalha region is presented in Fig 5 The SRTMtopographic information is used to represent height and the

Investigation of Shuttle Radar Topography Mission data 241

Fig 3 Shuttle Radar Topography Mission (SRTM) raw data over the Serra da Cangalha structure (90 km E-W 87 km N-S)

Fig 4 N-S and E-W topographic profiles across Serra da Cangalha taken from the SRTM raw data (as shown in Fig 3 90 km E-W 87 kmN-S) Annotations refer to our interpretation of the various images (see detail in text)

242 W U Reimold et al

image intensity is given by a high pass filtered version of thedata set This serves to sharpen the 3-D image without addingany directional bias (as an added light source would havedone) The structure stands out clearly in the DEM imagewhich also shows partially the detail of the regional drainagepattern The crater rim structure resembles that in the terrainblocks (apparently plateaus) surrounding the crater structureThe crater structure is located close to the edge of a stronglydissected and blocky terrain whereas to the east and further tothe southeast of the structure another rather featureless terraintype is noted This latter apparently smoother terrain isidentified as high elevation (compare the elevation profilingin Fig 4) presumably relatively less eroded SambaIgravebaFormation which is comprised of sandstones Thetopography in this area is known to be dominated by plateausand mesas Nothing is known to us about the geology of theterrain to the west and southwest and further detailedgeological analysis is required in this area

Sunshading Results

Sunshading provides a detailed image of the regionalstructure where strong NW-SE and ENE-WSW fault trendsare again prominent In Fig 6 generated after circularsunshading of the SRTM data the crater region is extremelydeformed in contrast to larger areas (plateaus) to the SE NWand NE of the crater area demonstrating the change in terrainnature already commented on in relation to Fig 5 At mostthe cratering-related annular deformation pattern does notexceed 07 crater radius beyond the crater rim Figure 7depicts the result of a calculation in which gradients of 075

100 and 125 were used in the reflectance algorithm (Cooperand Cowan 2003b) The resulting RGB color image was thenoverlain on a 3-D surface whose elevations were based on theoriginal SRTM data This resulted in a detailed DEM imagein which the various structural elements of the crater (centraluplift comprising a prominent inner ring around a smallinferred topographic low surrounded by low-lying annulustwo intermediate ring features of subdued topography outerannulus prominent outer rim) are clearly visible Outside ofthe crater structure several plateaus are prominent and theopening in the eastern crater rim is clearly visible

DISCUSSION

The crater itself is characterized by the followingelements

1 A prominent inner ring with a central low-lying areawhich is considered the collapsed central uplift of thestructure The central topographic low extends to aradius of 11 km and the central ring to a distance of 16km Considering that the structure has a diameter of only134 km it is intriguing to speculate whether this centraluplift feature could represent a peak ring such as onewould otherwise expect to find only at much largerimpact structures

2 At sim3 km radius the first intermediate ring is notedfollowed by a somewhat more prominent thoughincomplete second intermediate ring structure at about55 km radius

3 The outer ring (rim) structure occurs at 67 km from thecenter

Fig 5 SRTM data over the Serra da Cangalha impact structure in 3-D (90 km E-W 87 km N-S) Image intensity is given by a high-pass filteredversion of the same data set

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

Investigation of Shuttle Radar Topography Mission data 239

investigate the morphology and geometry of the structure aswell as the regional influence exerted by this possible impactevent We also test the recently developed circular sunshadingmethod (Cooper 2003) with this application Finally we setout to investigate whether or not the occurrence of andspacing between specific morphological features in impactstructures follow a definite relationship Several imagesgenerated from the SRTM data over the Serra da Cangalhaarea were recently shown by Almeido-Filho et al (2005)though without any geological interpretation

DATA AND PROCESSING METHODOLOGY

Shuttle Radar Topography Mission (SRTM) data for theregion around the Serra da Cangalha structure were availablefor this study Global SRTM single pass radar interferometrydata (Farr and Kobrick 2000) were obtained by the STS-99space shuttle mission between 11 and 22 February 2000SRTM digital elevation model data have a horizontalresolution of 1 arc second (equivalent to 30 m at the equator)and a vertical resolution of 10 m for the C-band radar The

United States Geological Survey (USGS) is the responsibledata archiving agency data were made available by NASAJet Propulsion Laboratory Global 3 arc second data havebeen released whereas 1 arc second data are only availablefor North America Initial comparison between 3 arc secondSRTM and older GTOPO DEM (Global Topography 30 arcsecond Digital Elevation Model) of the USGS (Cowan andCooper 2003) showed that the resolution of SRTM DEM is asignificant improvement and will be particularly valuable inareas for which limited topographic data are available

The Thematic Mapper (TM) Landsat data available forthe Serra da Cangalha region (Fig 2) provide a means forcomparison with the SRTM data In addition severalenhancement methods were applied in this study Sunshadingtechniques such as fractional order sunshading (Cooper andCowan 2003a 2003b) and a technique that enhances circularanomalies (Cooper 2003) were also employed

A standard filter used to enhance linear features inimages is sunshading It determines the reflectance from thedata of a light source located at infinity Linear features thatlie parallel to the azimuth of the light source (the ldquosunrdquo) are

Fig 2 Landsat image (bands 7-4-2 as RGB) over the Serra da Cangalha impact structure The scene limits are 47deg01prime05primeprimeW to 46deg41prime50primeprimeWand 12prime21primeprimeS to 7deg57prime19primeprime In this and the following images the top of the image is north

240 W U Reimold et al

attenuated while those that lie orthogonal to it are enhanced(Horn 1982) Because sunshading is a form of high-passfiltering it enhances both detail and noise in an image Thesunshading filter uses the first horizontal derivatives of thedata but if noise is a problem then lower-order derivativesmay be used instead These derivatives are of non-integerorder and are best computed in the frequency domain (Cooperand Cowan 2003a) The results from three differentsunshading operations each of which uses derivatives ofdifferent order may be combined to form an RGB imageHigh-frequency features then appear blue while lowerfrequency features appear red (Cooper and Cowan 2003b)Impact structures are mostly near-circular the sunshadingfilter was therefore modified to enhance features that lieeither on or orthogonal to radial vectors that pass through achosen origin position This was achieved by making the sunazimuth a variable over the image rather than a constant(Cooper 2003) As shown below the circularity of this impactstructure is strongly enhanced by this method However inthe application of the circular sunshading technique to thesearch for further impact structures the user should be wellaware that there are a range of other geological features thatmay have similar geometry for example ring structurescaused by differential erosion above an intrusion kimberlitepipes or other volcanic features

RESULTS

Figure 2 shows an RGB color composite image ofLandsat TM bands 7-4-2 Band 7 is in the far infrared and hasa strong response from hydroxyl-bearing minerals and claysspecifically carbonates micas chlorite and amphibolesBand 4 is in the near infrared and has a strong response fromiron absorption and vegetation Band 2 corresponds to greenlight in the visible part of the spectrum It is useful inhighlighting both chlorophyll absorption and iron features Inthe figure the Serra da Cangalha structure is comprised of a3 km wide near-circular central ring around a centraldepression

Adepelumi et al (2004) stated that the structure appearsto be open to the northwest An intermediate ring featurewith a diameter of sim5 km was also indicated Theintermediate ring is surrounded by a broad and apparentlylow topography annulus of sim3ndash4 km width which isterminated by the outer rim of the structure at 6ndash65 km fromthe center The rim itself is well-defined due to the weak colorshading of the interior of the structure but not as apronounced topographic feature

A strong radial (and in some sectors concentric) drainagepattern extends outward from the outer rim Only a few radialdrainage lines originate in the annular trough presumablyexploiting radial faults that breach the outer rim The widerregional drainage pattern is distinctly different from thiscrater-near obviously impact-derived pattern It appears that

cratering-related structure the underlying reason for thecrater-near drainage pattern extends to a maximum distanceof 15ndash2 crater radii from the center of the Serra da Cangalhastructure

A NW-SE directed structural trend is obvious in therather straight geometry of the NE and SW sectors of theinner ring and in fracturing cutting across the SE sector of theouter rim

SRTM Data

The SRTM raw data for the area of the structure areshown in Fig 3 This provides a detailed representation of theintricate drainage pattern providing a much clearer patternthan the Landsat imagery (Fig 2) The regional pattern isobviously different from that at and around the Serra daCangalha structure with strong NW-SE and ENE-WSWcomponents In contrast the crater-related pattern isdominated by radial and annular drainage The SRTMfindings confirm that this trend does not exceed the 15ndash2crater radii limit that can also be estimated from the Landsatimage

Figure 4 gives detailed topographic information obtainedfrom the SRTM raw data along two perpendicular profiles (N-S and E-W) across the structure Despite some variation in thebackground elevation the two intermediate rings can bereadily recognized The crater rim seems to be a broad andstructurally complex annular zone Detailed structuralanalysis of the crater rim zone at the 105 km wide Bosumtwistructure in Ghana (Reimold et al 1998) and the very muchsmaller 2 km diameter BP structure in Libya (Koeberl et al2005a our group unpublished data) also revealed broadstructurally complex crater rims with annular radial andoblique faulting directions The pronounced central ring haselevations of up to 300 m above background The twointermediate rings are within background elevation but areobvious from their sharp local topographic gradients Thecrater rim is broad and complex and seemingly elevated up togt100 m above the crater interior The two elevation profilesdiffer significantly with regard to the definition of the craterrim which is very pronounced on both sides of the N-Sprofile but not so on the eastern side of the crater structureThe reason for this is a distinct breach of the crater rim on thatside as is evident in Figs 2 and 3 Peculiar ENE trendingfeatures extend from the breach in the crater rim outwardsWhile it is not possible to provide an explanation of this fromthe remote sensing data alone it is noteworthy that thisfeature trends parallel to the regional structural grain Inaddition the terrane in the environs of the crater structureshows significant elevation complexity with local occurrenceof mesas

A 3-D image of the SRTM data over the Serra daCangalha region is presented in Fig 5 The SRTMtopographic information is used to represent height and the

Investigation of Shuttle Radar Topography Mission data 241

Fig 3 Shuttle Radar Topography Mission (SRTM) raw data over the Serra da Cangalha structure (90 km E-W 87 km N-S)

Fig 4 N-S and E-W topographic profiles across Serra da Cangalha taken from the SRTM raw data (as shown in Fig 3 90 km E-W 87 kmN-S) Annotations refer to our interpretation of the various images (see detail in text)

242 W U Reimold et al

image intensity is given by a high pass filtered version of thedata set This serves to sharpen the 3-D image without addingany directional bias (as an added light source would havedone) The structure stands out clearly in the DEM imagewhich also shows partially the detail of the regional drainagepattern The crater rim structure resembles that in the terrainblocks (apparently plateaus) surrounding the crater structureThe crater structure is located close to the edge of a stronglydissected and blocky terrain whereas to the east and further tothe southeast of the structure another rather featureless terraintype is noted This latter apparently smoother terrain isidentified as high elevation (compare the elevation profilingin Fig 4) presumably relatively less eroded SambaIgravebaFormation which is comprised of sandstones Thetopography in this area is known to be dominated by plateausand mesas Nothing is known to us about the geology of theterrain to the west and southwest and further detailedgeological analysis is required in this area

Sunshading Results

Sunshading provides a detailed image of the regionalstructure where strong NW-SE and ENE-WSW fault trendsare again prominent In Fig 6 generated after circularsunshading of the SRTM data the crater region is extremelydeformed in contrast to larger areas (plateaus) to the SE NWand NE of the crater area demonstrating the change in terrainnature already commented on in relation to Fig 5 At mostthe cratering-related annular deformation pattern does notexceed 07 crater radius beyond the crater rim Figure 7depicts the result of a calculation in which gradients of 075

100 and 125 were used in the reflectance algorithm (Cooperand Cowan 2003b) The resulting RGB color image was thenoverlain on a 3-D surface whose elevations were based on theoriginal SRTM data This resulted in a detailed DEM imagein which the various structural elements of the crater (centraluplift comprising a prominent inner ring around a smallinferred topographic low surrounded by low-lying annulustwo intermediate ring features of subdued topography outerannulus prominent outer rim) are clearly visible Outside ofthe crater structure several plateaus are prominent and theopening in the eastern crater rim is clearly visible

DISCUSSION

The crater itself is characterized by the followingelements

1 A prominent inner ring with a central low-lying areawhich is considered the collapsed central uplift of thestructure The central topographic low extends to aradius of 11 km and the central ring to a distance of 16km Considering that the structure has a diameter of only134 km it is intriguing to speculate whether this centraluplift feature could represent a peak ring such as onewould otherwise expect to find only at much largerimpact structures

2 At sim3 km radius the first intermediate ring is notedfollowed by a somewhat more prominent thoughincomplete second intermediate ring structure at about55 km radius

3 The outer ring (rim) structure occurs at 67 km from thecenter

Fig 5 SRTM data over the Serra da Cangalha impact structure in 3-D (90 km E-W 87 km N-S) Image intensity is given by a high-pass filteredversion of the same data set

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

240 W U Reimold et al

attenuated while those that lie orthogonal to it are enhanced(Horn 1982) Because sunshading is a form of high-passfiltering it enhances both detail and noise in an image Thesunshading filter uses the first horizontal derivatives of thedata but if noise is a problem then lower-order derivativesmay be used instead These derivatives are of non-integerorder and are best computed in the frequency domain (Cooperand Cowan 2003a) The results from three differentsunshading operations each of which uses derivatives ofdifferent order may be combined to form an RGB imageHigh-frequency features then appear blue while lowerfrequency features appear red (Cooper and Cowan 2003b)Impact structures are mostly near-circular the sunshadingfilter was therefore modified to enhance features that lieeither on or orthogonal to radial vectors that pass through achosen origin position This was achieved by making the sunazimuth a variable over the image rather than a constant(Cooper 2003) As shown below the circularity of this impactstructure is strongly enhanced by this method However inthe application of the circular sunshading technique to thesearch for further impact structures the user should be wellaware that there are a range of other geological features thatmay have similar geometry for example ring structurescaused by differential erosion above an intrusion kimberlitepipes or other volcanic features

RESULTS

Figure 2 shows an RGB color composite image ofLandsat TM bands 7-4-2 Band 7 is in the far infrared and hasa strong response from hydroxyl-bearing minerals and claysspecifically carbonates micas chlorite and amphibolesBand 4 is in the near infrared and has a strong response fromiron absorption and vegetation Band 2 corresponds to greenlight in the visible part of the spectrum It is useful inhighlighting both chlorophyll absorption and iron features Inthe figure the Serra da Cangalha structure is comprised of a3 km wide near-circular central ring around a centraldepression

Adepelumi et al (2004) stated that the structure appearsto be open to the northwest An intermediate ring featurewith a diameter of sim5 km was also indicated Theintermediate ring is surrounded by a broad and apparentlylow topography annulus of sim3ndash4 km width which isterminated by the outer rim of the structure at 6ndash65 km fromthe center The rim itself is well-defined due to the weak colorshading of the interior of the structure but not as apronounced topographic feature

A strong radial (and in some sectors concentric) drainagepattern extends outward from the outer rim Only a few radialdrainage lines originate in the annular trough presumablyexploiting radial faults that breach the outer rim The widerregional drainage pattern is distinctly different from thiscrater-near obviously impact-derived pattern It appears that

cratering-related structure the underlying reason for thecrater-near drainage pattern extends to a maximum distanceof 15ndash2 crater radii from the center of the Serra da Cangalhastructure

A NW-SE directed structural trend is obvious in therather straight geometry of the NE and SW sectors of theinner ring and in fracturing cutting across the SE sector of theouter rim

SRTM Data

The SRTM raw data for the area of the structure areshown in Fig 3 This provides a detailed representation of theintricate drainage pattern providing a much clearer patternthan the Landsat imagery (Fig 2) The regional pattern isobviously different from that at and around the Serra daCangalha structure with strong NW-SE and ENE-WSWcomponents In contrast the crater-related pattern isdominated by radial and annular drainage The SRTMfindings confirm that this trend does not exceed the 15ndash2crater radii limit that can also be estimated from the Landsatimage

Figure 4 gives detailed topographic information obtainedfrom the SRTM raw data along two perpendicular profiles (N-S and E-W) across the structure Despite some variation in thebackground elevation the two intermediate rings can bereadily recognized The crater rim seems to be a broad andstructurally complex annular zone Detailed structuralanalysis of the crater rim zone at the 105 km wide Bosumtwistructure in Ghana (Reimold et al 1998) and the very muchsmaller 2 km diameter BP structure in Libya (Koeberl et al2005a our group unpublished data) also revealed broadstructurally complex crater rims with annular radial andoblique faulting directions The pronounced central ring haselevations of up to 300 m above background The twointermediate rings are within background elevation but areobvious from their sharp local topographic gradients Thecrater rim is broad and complex and seemingly elevated up togt100 m above the crater interior The two elevation profilesdiffer significantly with regard to the definition of the craterrim which is very pronounced on both sides of the N-Sprofile but not so on the eastern side of the crater structureThe reason for this is a distinct breach of the crater rim on thatside as is evident in Figs 2 and 3 Peculiar ENE trendingfeatures extend from the breach in the crater rim outwardsWhile it is not possible to provide an explanation of this fromthe remote sensing data alone it is noteworthy that thisfeature trends parallel to the regional structural grain Inaddition the terrane in the environs of the crater structureshows significant elevation complexity with local occurrenceof mesas

A 3-D image of the SRTM data over the Serra daCangalha region is presented in Fig 5 The SRTMtopographic information is used to represent height and the

Investigation of Shuttle Radar Topography Mission data 241

Fig 3 Shuttle Radar Topography Mission (SRTM) raw data over the Serra da Cangalha structure (90 km E-W 87 km N-S)

Fig 4 N-S and E-W topographic profiles across Serra da Cangalha taken from the SRTM raw data (as shown in Fig 3 90 km E-W 87 kmN-S) Annotations refer to our interpretation of the various images (see detail in text)

242 W U Reimold et al

image intensity is given by a high pass filtered version of thedata set This serves to sharpen the 3-D image without addingany directional bias (as an added light source would havedone) The structure stands out clearly in the DEM imagewhich also shows partially the detail of the regional drainagepattern The crater rim structure resembles that in the terrainblocks (apparently plateaus) surrounding the crater structureThe crater structure is located close to the edge of a stronglydissected and blocky terrain whereas to the east and further tothe southeast of the structure another rather featureless terraintype is noted This latter apparently smoother terrain isidentified as high elevation (compare the elevation profilingin Fig 4) presumably relatively less eroded SambaIgravebaFormation which is comprised of sandstones Thetopography in this area is known to be dominated by plateausand mesas Nothing is known to us about the geology of theterrain to the west and southwest and further detailedgeological analysis is required in this area

Sunshading Results

Sunshading provides a detailed image of the regionalstructure where strong NW-SE and ENE-WSW fault trendsare again prominent In Fig 6 generated after circularsunshading of the SRTM data the crater region is extremelydeformed in contrast to larger areas (plateaus) to the SE NWand NE of the crater area demonstrating the change in terrainnature already commented on in relation to Fig 5 At mostthe cratering-related annular deformation pattern does notexceed 07 crater radius beyond the crater rim Figure 7depicts the result of a calculation in which gradients of 075

100 and 125 were used in the reflectance algorithm (Cooperand Cowan 2003b) The resulting RGB color image was thenoverlain on a 3-D surface whose elevations were based on theoriginal SRTM data This resulted in a detailed DEM imagein which the various structural elements of the crater (centraluplift comprising a prominent inner ring around a smallinferred topographic low surrounded by low-lying annulustwo intermediate ring features of subdued topography outerannulus prominent outer rim) are clearly visible Outside ofthe crater structure several plateaus are prominent and theopening in the eastern crater rim is clearly visible

DISCUSSION

The crater itself is characterized by the followingelements

1 A prominent inner ring with a central low-lying areawhich is considered the collapsed central uplift of thestructure The central topographic low extends to aradius of 11 km and the central ring to a distance of 16km Considering that the structure has a diameter of only134 km it is intriguing to speculate whether this centraluplift feature could represent a peak ring such as onewould otherwise expect to find only at much largerimpact structures

2 At sim3 km radius the first intermediate ring is notedfollowed by a somewhat more prominent thoughincomplete second intermediate ring structure at about55 km radius

3 The outer ring (rim) structure occurs at 67 km from thecenter

Fig 5 SRTM data over the Serra da Cangalha impact structure in 3-D (90 km E-W 87 km N-S) Image intensity is given by a high-pass filteredversion of the same data set

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

Investigation of Shuttle Radar Topography Mission data 241

Fig 3 Shuttle Radar Topography Mission (SRTM) raw data over the Serra da Cangalha structure (90 km E-W 87 km N-S)

Fig 4 N-S and E-W topographic profiles across Serra da Cangalha taken from the SRTM raw data (as shown in Fig 3 90 km E-W 87 kmN-S) Annotations refer to our interpretation of the various images (see detail in text)

242 W U Reimold et al

image intensity is given by a high pass filtered version of thedata set This serves to sharpen the 3-D image without addingany directional bias (as an added light source would havedone) The structure stands out clearly in the DEM imagewhich also shows partially the detail of the regional drainagepattern The crater rim structure resembles that in the terrainblocks (apparently plateaus) surrounding the crater structureThe crater structure is located close to the edge of a stronglydissected and blocky terrain whereas to the east and further tothe southeast of the structure another rather featureless terraintype is noted This latter apparently smoother terrain isidentified as high elevation (compare the elevation profilingin Fig 4) presumably relatively less eroded SambaIgravebaFormation which is comprised of sandstones Thetopography in this area is known to be dominated by plateausand mesas Nothing is known to us about the geology of theterrain to the west and southwest and further detailedgeological analysis is required in this area

Sunshading Results

Sunshading provides a detailed image of the regionalstructure where strong NW-SE and ENE-WSW fault trendsare again prominent In Fig 6 generated after circularsunshading of the SRTM data the crater region is extremelydeformed in contrast to larger areas (plateaus) to the SE NWand NE of the crater area demonstrating the change in terrainnature already commented on in relation to Fig 5 At mostthe cratering-related annular deformation pattern does notexceed 07 crater radius beyond the crater rim Figure 7depicts the result of a calculation in which gradients of 075

100 and 125 were used in the reflectance algorithm (Cooperand Cowan 2003b) The resulting RGB color image was thenoverlain on a 3-D surface whose elevations were based on theoriginal SRTM data This resulted in a detailed DEM imagein which the various structural elements of the crater (centraluplift comprising a prominent inner ring around a smallinferred topographic low surrounded by low-lying annulustwo intermediate ring features of subdued topography outerannulus prominent outer rim) are clearly visible Outside ofthe crater structure several plateaus are prominent and theopening in the eastern crater rim is clearly visible

DISCUSSION

The crater itself is characterized by the followingelements

1 A prominent inner ring with a central low-lying areawhich is considered the collapsed central uplift of thestructure The central topographic low extends to aradius of 11 km and the central ring to a distance of 16km Considering that the structure has a diameter of only134 km it is intriguing to speculate whether this centraluplift feature could represent a peak ring such as onewould otherwise expect to find only at much largerimpact structures

2 At sim3 km radius the first intermediate ring is notedfollowed by a somewhat more prominent thoughincomplete second intermediate ring structure at about55 km radius

3 The outer ring (rim) structure occurs at 67 km from thecenter

Fig 5 SRTM data over the Serra da Cangalha impact structure in 3-D (90 km E-W 87 km N-S) Image intensity is given by a high-pass filteredversion of the same data set

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

242 W U Reimold et al

image intensity is given by a high pass filtered version of thedata set This serves to sharpen the 3-D image without addingany directional bias (as an added light source would havedone) The structure stands out clearly in the DEM imagewhich also shows partially the detail of the regional drainagepattern The crater rim structure resembles that in the terrainblocks (apparently plateaus) surrounding the crater structureThe crater structure is located close to the edge of a stronglydissected and blocky terrain whereas to the east and further tothe southeast of the structure another rather featureless terraintype is noted This latter apparently smoother terrain isidentified as high elevation (compare the elevation profilingin Fig 4) presumably relatively less eroded SambaIgravebaFormation which is comprised of sandstones Thetopography in this area is known to be dominated by plateausand mesas Nothing is known to us about the geology of theterrain to the west and southwest and further detailedgeological analysis is required in this area

Sunshading Results

Sunshading provides a detailed image of the regionalstructure where strong NW-SE and ENE-WSW fault trendsare again prominent In Fig 6 generated after circularsunshading of the SRTM data the crater region is extremelydeformed in contrast to larger areas (plateaus) to the SE NWand NE of the crater area demonstrating the change in terrainnature already commented on in relation to Fig 5 At mostthe cratering-related annular deformation pattern does notexceed 07 crater radius beyond the crater rim Figure 7depicts the result of a calculation in which gradients of 075

100 and 125 were used in the reflectance algorithm (Cooperand Cowan 2003b) The resulting RGB color image was thenoverlain on a 3-D surface whose elevations were based on theoriginal SRTM data This resulted in a detailed DEM imagein which the various structural elements of the crater (centraluplift comprising a prominent inner ring around a smallinferred topographic low surrounded by low-lying annulustwo intermediate ring features of subdued topography outerannulus prominent outer rim) are clearly visible Outside ofthe crater structure several plateaus are prominent and theopening in the eastern crater rim is clearly visible

DISCUSSION

The crater itself is characterized by the followingelements

1 A prominent inner ring with a central low-lying areawhich is considered the collapsed central uplift of thestructure The central topographic low extends to aradius of 11 km and the central ring to a distance of 16km Considering that the structure has a diameter of only134 km it is intriguing to speculate whether this centraluplift feature could represent a peak ring such as onewould otherwise expect to find only at much largerimpact structures

2 At sim3 km radius the first intermediate ring is notedfollowed by a somewhat more prominent thoughincomplete second intermediate ring structure at about55 km radius

3 The outer ring (rim) structure occurs at 67 km from thecenter

Fig 5 SRTM data over the Serra da Cangalha impact structure in 3-D (90 km E-W 87 km N-S) Image intensity is given by a high-pass filteredversion of the same data set

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

Investigation of Shuttle Radar Topography Mission data 243

Fig 6 Circular sunshaded SRTM data over the Serra da Cangalha region (90 km E-W 87 km N-S)

Fig 7 Fractional order horizontal derivative sunshaded SRTM data over the Serra da Cangalha impact structure overlain on the SRTM dataitself (in 3-D) The sun inclination used was 30deg from the horizontal and the sun azimuth was northwest Image measures 317 km E-W and346 km N-S Exact vertical exaggeration unknown as determined by software

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

244 W U Reimold et al

Very intricate structural detail in the form of inferredfaults of radial to oblique (with regard to the center)orientation is imaged outside of the outer rim but thesestructures do not extend into the plateau regions further outThus the impact-macrodeformed zone is limited to amaximum radial distance of 10ndash11 km from the centerSeveral radial faults do extend further to a maximum distanceof 16ndash18 km from the center about 15 times as far as theextent of the crater-related drainage pattern Thisinterpretation is further supported by the two elevationprofiles shown in Fig 4 They also demonstrate that the craterhas a high degree of symmetry The complex structuralterrane indicated in these data strongly suggests Serra daCangalha as a prominent target for detailed ground-basedstructural geological analysis to supplement the limitedstructural data base for impact craters of such moderate (10ndash20 km) diameter

The SRTM data provide a powerful tool for the detailedstructural investigation of geological features includinglikely impact structures In comparison to the Landsat TMdata impact-related deformation can be imaged andinterpreted in greater detail from the SRTM data Thisallowed us to better define the various structural elements ofthe crater and their respective diameters While it can beargued that some of the information had been obtained fromLandsat imagery already the SRTM data provide a newmeans to investigate crater morphology at a very goodresolution

In Table 1 the first-order structural features of Serra daCangalha are compared against those for Gosses Bluff(Australia) after Milton et al (1996) the BP and Oasisstructures in Libya after Koeberl et al (2005a) SteinheimBasin (Ivanov and Stˆffler 2005) Aorounga and Gweni-Fadain Chad (Koeberl et al 2005b) Ries (Wcedilnnemann et al

2005) and Chicxulub (Morgan et al 1997 2002) Note theuncertainties on many of these data as shown in the table anddiscussed in associated footnotes In several cases hardly anyground-based geological information is available to confirmthese estimates based on interpretation of remote sensinginformation It is obvious that this first compilation of suchdata does not yield any obvious trends the existing values forapparent crater rim diameters and the ratios of crater rimdiametercentral uplift diameter give a weak indication ofpositive correlation as does plotting of crater rim diameterversus ratio of crater rim diameterinner intermediate ringdiameter for the craters up to 24 km apparent diameter butwith much scatter Clearly more data of this nature arerequired to test the validity of such spurious correlations

It must also be cautioned that any such comparison ofdata for terrestrial impact structures must be subject to carefulconsideration of a number of factors such as targetstratigraphy and particularly variation in relative levels oferosion The values compiled are generally apparentdiameters (cf Turtle et al [2005] for a detailed discussion ofthis problem) In addition the nature of outer ring features isstill a subject of debate as for example discussed by Wagneret al (2002)

Comparison of the crater structures of very similar smallsize (BP and Steinheim Basin) suggests similarity of cratergeometry within the limitations of topographic analysisBosumtwi Serra da Cangalha Oasis Gweni-Fada andAorounga are all traditionally listed with similar size (105 to17 km diameters) but have different diameters of centraluplifts and first intermediate rings These differences could becaused by uncertainty in where to take the actual diameter ofa collapsed uplift Using the recently derived (Koeberl et al2005b) larger crater diameters of 22 and 17 km for Gweni-Fada and Aorounga respectively yields better agreement in a

Table 1 Apparent diameters of structural features of several eroded impact structures

Crater BP Steinheima SdCa Oasisb Bosumtwi Gweni-Fada AoroungaGoss Bluff

Ries crater

Chicxulubcrater

Crater diameterc 2d 34 134 115 [18] 105 13 [18e] 13 [24]e 24 24 sim180Central ringuplift 05 1 32 5ndash6 18 5 4 [6] 45 4f 80First intermediate ring

ndash ndash 6 ndash 8 6 7g 10 ndash

Second intermediate ring

ndash ndash 105 ndash ndash ndash 11ndash16 ndash ndash ndash

Crater rim 2 34 134 115 [18] 105 13 [18] 13 [24] 24 24 180Outer ring feature(detachment fault)h

28 18 18ndash20 240

aSdC = Serra da Cangalha Steinheim = Steinheim Basin (data from Ivanov and Stˆffler [2005])bThe remote sensing observations are not conclusive and ground truth is scarce there are two alternatives regarding the actual crater diameter 115 km as

favored by Koeberl et al (2005a) and 18 km as favored by McHone et al (2002)cEstimate of apparent (eroded) crater diameterdA further low-topography feature outside of this rim was interpreted by Koeberl et al (2005a) as a detachment fault dipping at low angle toward the crater

interioreLarger estimate based on SRTM topographic data (Koeberl et al 2005b)fRies data from Wcedilnnemann et al (2005 and references therein) central uplift is considered ldquoincipientrdquo Chicxulub data after Morgan et al (1997 2002)gOur interpretation of Fig 2 of Milton et al (1996) hSee Wagner et al (2002) and Koeberl et al (2005a)

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

Investigation of Shuttle Radar Topography Mission data 245

comparison of these craters with the Ries and Gosses Bluff interms of crater diameter central upliftring and intermediatering spacing Whether or not this is a possible indication thatthe diameters of both Gweni-Fada and Aorounga haveoriginally been underestimated (as suggested by Koeberl et al2005b) remains to be tested

Naturally studies like the present one are important withregard to further identification of impact structures on Earththrough remote sensing and regarding their differentiationfrom volcanic crater structures As described in some detailby for example Greeley (1994) volcanic structures can be ofa wide range of morphologies from simple bowl-shapedstructures to large complex caldera collapse structures Inparticular collapsed crater walls may be extensively faultedresembling the complex deformation of impact crater rimsSmall maar-type volcanic craters are very similar inappearance to simple bowl-shaped impact craters Similarlydrainage and fault patterns around volcanic structures mayresemble the annular and radial patterns formed aroundimpact structures To date however we have not found anyevidence of multiple ring structures or distinct central upliftfeatures in volcanic crater structures with the exception ofvolcanic craters characterized by repeated eruptions resultingin a further cone or crater form in the central part Wherevolcanic craters display raised rims this would be the result ofaccumulation of ejecta and in a fresh case such a rim appearsrather continuous and comparatively smooth as comparedwith the heavily faulted and folded upturned rims of impactstructures Erosion of such volcanic landforms may ofcourse result in more complex structural morphologiesCrater collapse structures (calderas and caldera structures)are for example reviewed by Cole et al (2005)

Distinguishing volcanic features from impact structuresis not easy as both types have circular outlines and a rim andcentral structure Digital elevation models possibly generatedfrom SRTM or other SAR data can help to discriminatebetween the two types of features as impact structuresgenerally have steeper slopes and larger rim to surroundingsheight as was observed in tests of automatic craterrecognition programs (eg Earl et al 2005 J Earl andC Koeberl unpublished observations) Thus high-resolutionremote sensing data that allow to recognize central peakforms and multiple ring structures and their morphologicalnature will be extremely useful for further recognition ofimpact structures

CONCLUSIONS

The SRTM data have been shown to be superior to theolder Landsat TM data with regard to detail that can bediscerned in drainage patterns and structural analysis of craterstructures The following macro-structural division of theSerra da Cangalha impact structure in NE Brazil wascharacterized a central ring (diameter = 32 km) two

intermediate rings of limited elevation (diameters = 6 and105 km respectively) and a structurally complex crater rim(diameter = sim134 km) It appears that impact-producedmacrostructural deformation in the environs of the crater islimited to a maximum of about 95 km (sim08 crater radii) fromthe crater rim Serra da Cangalha would be a rewarding targetfor detailed structural geological investigation A firstcomparison of structural data for impact structures of a widerange of apparent crater diameters suggests that some of thesedata may have been estimated wrongly in the past Furtherdetailed remote-sensing studies in particular based on thenow accessible global SRTM data together with comparativefield-based analysis as well as numerical modeling arerequired to provide a more extensive data set on cratermorphological features for other complex terrestrial impactstructures

AcknowledgmentsndashWe thank NASAJPL for the SRTM andthe Landsat TM data This research is supported by a grantfrom the National Research Foundation of South Africa toW U R C K is supported by the Austrian ScienceFoundation R R is grateful to the Barringer Crater Companyfor financial support that facilitated his attendance of the 35thLunar and Planetary Science Conference Critical andconstructive reviews by Steve Prevec and an anonymousreviewer are appreciated This is University of theWitwatersrand Impact Cratering Research GroupContribution No 92

Editorial HandlingmdashCarlEgrave Pieters

REFERENCES

Abraham A A Flexor J M and Fontes S L 2004 Geophysicalsignature of Serra da Cangalha impact crater Brazil (abstract)Meteoritics amp Planetary Science 39A11

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003a Imaging Serra da Cangalha impact crater using Eulerdeconvolution method (abstract) Society of ExplorationGeophysicists Technical Program Expanded Abstracts 22584ndash587

Adepelumi A A Flexor J M Fontes S L and Schnegg P A2003b Interpretation of the aeromagnetic signature of the Serrada Cangalha impact crater Brazil (abstract) Expanded abstractsof the 8th International Congress of the Brazilian GeophysicalSociety CD-ROM

Adepelumi A A Fontes S L and Flexor J M 2004 Environmentalperturbations caused by the Serra da Cangalha impact craterstructure northeastern Brazil Environtropica 158ndash71

Adepelumi A A Fontes S L Schnegg P A and Flexor J M2005a An integrated magnetotelluric and aeromagneticinvestigation of the Serra da Cangalha impact crater BrazilPhysics of the Earth and Planetary Interiors 150159ndash181

Adepelumi A A Flexor J M and Fontes S L 2005b An appraisalof the Serra da Cangalha impact structure using the Eulerdeconvolution method Meteoritics amp Planetary Science 401149ndash1157

Almeida-Filho R Moreira F R S and Beisl C H 2005 The Serra

246 W U Reimold et al

da Cangalha astrobleme as revealed by ASTER and SRTMorbital data International Journal of Remote Sensing 26833ndash838

Cole J W Milner D M and Spinks K D 2005 Calderas andcaldera structures A review Earth-Science Reviews 691ndash26

Cooper G R J 2003 Feature detection using sunshading Computersamp Geosciences 29941ndash948

Cooper G R J and Cowan D R 2003a The application of fractionalcalculus to potential field data Exploration Geophysics 3451ndash56

Cooper G R J and Cowan D R 2003b Sunshading geophysicaldata using fractional order horizontal gradients The LeadingEdge 22204ndash205

Cowan D R and Cooper G R J 2003 The Shuttle RadarTopography Mission A new source of near-global digitalelevation data Australian Society of Exploration GeophysicistsPreview 10727ndash29

CrUcircsta A P 1982 Estruturas de impacto no Brasil Uma sIgraventese doconhecimento atual Anais do XXXII Congresso Brasileiro deGeologia Salvador Bahia 41372ndash1377

CrUcircsta A P 1987 Impact structures in Brazil In Research interrestrial impact structures edited by Pohl J WiesbadenGermany Vieweg and Sons pp 30ndash38

CrUcircsta A P 2004 Impact craters in Brazil How far wersquove gotten(abstract) Meteoritics amp Planetary Science 39A27

De Cicco M and Zucolotto M E 2002 Appraisal of Brazilianastroblemes and similar structures (abstract) Meteoritics ampPlanetary Science 37A40

Dietz R S and French B M 1973 Two possible astroblemes inBrazil Nature 244561ndash562

Farr T G and Kobrick M 2000 Shuttle Radar Topography Missionproduces a wealth of data EOS Transactions of the AmericanGeophysical Union 81583ndash585

GUcirces A M O Travassos W A S and Nunes K C 1993 ProjetpParnaIgraveba ReavaliaAacuteatildeo da bacia e perspectivas exploratUcircriasPetrobrmiddots Internal Report 97 p

Greeley R 1994 Planetary landscapes 2nd ed New York Chapmanand Hall 286 p

Horn B K P 1982 Hill shading and the reflectance map Geo-Processing 265ndash146

Ivanov B A and Stˆffler D 2005 The Steinheim impact craterGermany Modeling of a complex crater with central uplift(abstract 1443) 36th Lunar and Planetary Science ConferenceCD-ROM

Koeberl C Reimold W U and Plescia J 2005a BP and Oasisimpact structures Libya Remote sensing and field studies In

Impact tectonics edited by Koeberl C and Henkel H Berlin-Heidelberg Springer-Verlag pp 161ndash190

Koeberl C Reimold W U Cooper G R J Cowan D R andVincent P M 2005b Aorounga and Gweni-Fada impactstructures Chad Remote sensing petrography and geochemistryof target rocks Meteoritics amp Planetary Science 401455ndash1471

McHone J F 1979 Riachao Ring Brazil A possible meteoritecrater discovered by the Apollo astronauts In Apollo-SoyuzTest Project summary science report vol II Earthobservations and photography NASA Special Publication SP-412 pp 193ndash202

McHone J F Greeley R Williams K K Blumberg D G andKuzmin R O 2002 Space shuttle observations of terrestrialimpact structures using SIR-C and X-SAR radars Meteoritics ampPlanetary Science 37407ndash420

Milton D J Glikson A Y and Brett R 1996 Gosses BluffmdashAlatest Jurassic impact structure central Australia Part 1Geological structure stratigraphy and origin AGSO Journal ofAustralian Geology and Geophysics 16453ndash486

Morgan J Warner M and the Chicxulub Working Group 1997 Sizeand morphology of the Chicxulub impact crater Nature 390472ndash476

Morgan J V Christeson G L and Zelt C A 1997 Testing theresolution of a 3D velocityn tomogram across the Chixculubcrater Tectonophysics 355215ndash226

Santos U P and McHone J F 1979 Field report on Serra daCangalha and Riachatildeo circular features Instituto Nacional dePesquisas Espacias (INPE) Relatorio Interno 1458-NTE15313 p

Turtle E P Pierazzo E Collins G S Osinski G R Melosh H JMorgan J V and Reimold W U 2005 Impact structures Whatdoes crater diameter mean In Meteorite impacts and planetaryevolution III edited by Kenkmann T Hˆrz F and Deutsch AGSA Special Paper 384 Boulder Colorado Geological Societyof America pp 1ndash24

Wagner R Reimold W U and Brandt D 2002 Bosumtwi impactcrater Ghana A remote sensing investigation In Impacts inPrecambrian shields edited by Plado J and Pesonen L Berlin-Heidelberg Springer-Verlag pp 189ndash210

Wuumlnnemann K Morgan J V and Jˆdicke H 2005 Is Ries cratertypical for its size An analysis based upon old and newgeophysical data and numerical modeling In Large meteoriteimpacts III edited by Kenkmann T Hˆrz F and Deutsch AGSA Special Paper 384 Boulder Colorado Geological Societyof America pp 67ndash83