Geologic mapping of tectonic planets€¦ · ¢ned by a detailed magnetic polarity time scale correlated by radiometric dates. The geologic time scale is continually being re¢ned

Geologic mapping of tectonic planets

Vicki L. Hansen *Department of Geological Sciences, Southern Methodist University, Dallas, TX 75275, USA

Received 30 September 1999; received in revised form 14 January 2000; accepted 14 January 2000

Abstract

Geological analysis of planets typically begins with the construction of a geologic map of the planets' surfaces usingremote data sets. Geologic maps provide the basis for interpretations of geologic histories, which in turn provide criticalrelations for understanding the range of processes that contributed to the evolution. Because geologic mapping shouldultimately lead to the discovery of the types of operative processes that have shaped a planet surface, geologic mappingmust be undertaken in such a way as to allow such discovery. I argue for modifications in current planetary geologicmapping methodology that admit that tectonic processes may have contributed to the formation of a planet surface,and I emphasize a goal of constraining geologic history rather than determining global stratigraphy. To this end, it isimperative that secondary (tectonic) structures be clearly delineated from material units; each record different aspectsof planet surface evolution. Neglecting such delineation can result in geologic maps and interpreted geohistories inwhich both the spatial limits and the relative ages of material units and suites of secondary structures are incorrect.Determination of absolute time is fundamentally difficult to constrain in planetary studies. In planetary geology theonly means to estimate absolute age is the density of impact craters on a surface. Crater density surface ages are moreakin to terrestrial ANd average mantle model ages, which reflect the average time at which all of a rock's componentswere extracted from the Earth's mantle. Rocks, or planetary surfaces, with very different geohistories could yield thesame average model age. Dating tectonic events or determining rates of tectonic events is even more difficult. ß 2000Elsevier Science B.V. All rights reserved.

Keywords: mapping; planets; Venus; tectonics; structural geology

1. Introduction

Advances in space science technology and anincreasing number of missions have revealed un-precedented views of planet surfaces, which formthe fundamental data sets for construction of ex-traterrestrial geologic maps. Geologic maps pro-

vide the basis for interpreting geologic histories,which in turn provide critical relations for under-standing the range of processes that contributedto the evolution of planets and their satellites. Thepurpose of geologic mapping is to determine thegeologic history of a region and to incorporatethat history into an increasingly larger area ofstudy. The larger goal of understanding geologichistories is to understand planetary processes -that is, to understand how a planet works, spa-tially and temporally. The method of investigationmust not sacri¢ce one's ability to constrain geo-

0012-821X / 00 / $ ^ see front matter ß 2000 Elsevier Science B.V. All rights reserved.PII: S 0 0 1 2 - 8 2 1 X ( 0 0 ) 0 0 0 1 7 - 0

* Corresponding author. Tel: 214-768-4179;Fax: 214-768-2701; E-mail: [email protected]

EPSL 5377 2-3-00

Earth and Planetary Science Letters 176 (2000) 527^542

www.elsevier.com/locate/epsl

logic history, and thus sacri¢ce one's ability tounderstand planetary processes - the fundamentalresearch goal. Many planetary geologists cur-rently employ the geologic mapping methodologyoutlined by Wilhelms [1,2]. The method evolvedin the 1960s with the nascent space program anddraws largely from geological analysis of the rel-atively tectonically inactive Moon and Mars. Themethod emerged at a critical time - prior to wide-spread acceptance of plate tectonics, and a widelyaccepted view of a dynamic Earth. As evidence ofthis Wilhelms ([1], p. 219) states: `A major goal ofplanetary geologic mapping, like terrestrial map-ping, is to integrate local stratigraphic sequences(`columns') of geologic units into a stratigraphiccolumn applicable over the whole planet'.Although this method, referred to herein as the`global stratigraphic method', may have proveduseful for tectonically quiescent planets, it re-quires modi¢cation for application to tectonicallyactive planets.

Planetary geologic mapping aimed at determin-ing a global stratigraphy is fundamentally £awedbecause such a method assumes, a priori, thatsuch a stratigraphy exists. However, as stated byGilbert [3], `scienti¢c research consists of the ob-servation of phenomena and the discovery of theirrelations'. In order for a planet to develop a glob-al stratigraphy, it must have evolved by globallysynchronous and spatially continuous processes.If a planet indeed evolved through global pro-cesses, a global stratigraphy could be discoveredthrough study of local geologic histories, and suc-cessively broader correlation of these geologic his-tories. If one assumes a global stratigraphy, butsuch does not exist, any `discovered' global stra-tigraphy must be either grossly incorrect, or sobroadly blurred as to achieve apparent global syn-chroneity. The former derails scienti¢c progress,and the latter greatly inhibits understanding ofplanetary processes because critical details of lo-cal, regional and global planetary processes arelost [3].

Although one might expect that the geologichistory of a nontectonic planet is both less com-plex and simpler to determine than that of a tec-tonic planet, neither is necessarily the case. Infact, because tectonic planets can preserve both

geological material units and secondary (tectonic)structures, the potential exists to determine moredetailed geologic histories, and therefore to under-stand more detailed geological process(es) thanfor nontectonic planets. The surface record of anontectonic planet may lead one to propose anerroneously simple history. Because tectonic plan-ets can record a more detailed history, study ofsuch planets can result in inherently more testablegeologic histories than nontectonic planets.

Thus any planetary geologic mapping method-ology should allow that tectonic processes mayhave played a role in planet evolution, and, assuch, should emphasize geologic history ratherthan global stratigraphy. Section 2 of this contri-bution reviews the historical development ofstratigraphic methodology on Earth and illus-trates how various dating methods worked to in-validate a search for terrestrial global stratigra-phy; £aws in the global stratigraphic methodare illustrated through a geohistory experiment.The body of the paper outlines de¢nitions, map-ping principles, and proposed method modi¢ca-tions aimed at discerning geohistory; this methodis `tested' using the same experiment. A brief dis-cussion of absolute age dating methods with par-ticular application to Venus and implications fortectonic planets concludes the paper. WhereasWilhelms [1] used mostly lunar examples, Venus-ian examples are used herein. The NASA Magel-lan mission (1989^1993) resulted in high-resolu-tion global data including synthetic apertureradar (SAR) imagery, altimetry, emissivity, andgravity sets that can constrain geologic histories.Ford et al. [4] provide an extensive treatment ofMagellan data analysis and cautions. AlthoughEarth and Venus followed di¡erent evolutionarypaths, the methods for understanding local andglobal geological processes share the same philo-sophical underpinnings. For Venus we have onlyremotely collected data sets, which adds to requi-site cautions and resulting challenges.

2. Background

In the late 1700s to early 1800s an importantgoal of terrestrial geology was the construction of

EPSL 5377 2-3-00

V.L. Hansen / Earth and Planetary Science Letters 176 (2000) 527^542528

a `Standard Stratigraphic Column'. A.G. Werner,following J. G. Lehmann and C. Fu«chsel and thelaw of superposition, believed that strati¢ed rocksoccurred in an invariable order; rock units werecharacterized by their lithology and `correlation'was established by lithologic similarity and se-quential order - that is, the age of rocks every-where was presumed to be directly related to theirlithologic character. Werner divided rocks into¢ve progressively younger groups: primitive, tran-sitional, layered, transported, and volcanic rocks[5]. He believed that all rocks, except volcanicrocks (which were rare and unimportant), formedin water, and that their universal sequence re-corded the Earth's gradual history; each rockgroup re£ected a globally extensive change. Earlyformed crystalline rocks (primitive) precipitatedfrom a deep primeval ocean; as sea level fell,transitional rocks (mostly graywackes) formedby mechanical derivation from earlier formedrocks and continued precipitation. Further subsi-dence produced widespread fossiliferous layeredrocks (and related basalt) ; transported rocks de-veloped as sea level fell yet further; ¢nally vol-canic rocks erupted locally - the result of heatfrom subterranean burning coal.

Although Werner played a major role in elevat-ing geology as a science, general correlationsbased on lithology were shown to be grossly in-correct by W. Smith in 1815 with the recognitionof fossils and faunal assemblages, rather than lith-ology, as tools of correlation [5]. Index fossils,which occur in widely separated rock layers, com-monly with di¡erent lithologies, provided the ¢rstmeans to de¢ne regionally extensive timelines.The utility of index fossils results from their wide-spread geographic distribution and rapid evolu-tionary changes. Faunal assemblages providedclear evidence that similar lithologic units couldbe diachronous, and that lithologically dissimilarunits could be synchronous. By the early 1800'sthe three major stratigraphic principles hademerged: (1) the law of superposition - oldestlayers at the bottom; (2) uniformitarianism - thepresent is the key to the past; and (3) the law offaunal succession - fossils provide temporal con-straints independent of lithology. From theseprinciples, the Standard (stratigraphic) Column

evolved into rock units (lithology) and time units(faunal assemblages).

Williams [6] and the USGS [7,8] objected to therigid parallelism of time and rock units [9], argu-ing that the boundaries of rock units need not be,and typically are not, everywhere the same age,and therefore cannot serve as time units. Theyadvocated a dual and separate classi¢cation sys-tem similar to what is generally accepted today:(1) numerous local rock units based on local sec-tions, independent of time, and (2) relatively fewuniversal time units based on fossils, independentof lithology. The 1933 Stratigraphic Code, how-ever, ignored these arguments, implying exact par-allelism of rock and time units. Through debate,in£uenced in part by the advent of radiometricage dating, the geologic time scale has emergedwith a two-fold classi¢cation based on relativefaunal ages and absolute radiometric ages. Thelast V200 Myr of Earth's history is further de-¢ned by a detailed magnetic polarity time scalecorrelated by radiometric dates. The geologictime scale is continually being re¢ned [10,11] ;time scales di¡er mostly in details speci¢c to re-gion or `time slice'. Therefore, although the orig-inal divisions of type geologic sections were basedon stratigraphic relations of beds (i.e., materialunits), correlations of distinct strata are necessa-rily based on a comparison of fossils (and/or ra-diometric ages), not lithology ([12], p. 165). Thusrock-stratigraphic units should be de¢ned as litho-logic units, without regard to time. These unitsmust be kept distinct from time markers (e.g.,fossils) that are used in correlating local sectionswith each other and with the standard chronology(i.e., the geologic time scale) ([9], p. 293).

In summary, although some unique rock unitsmark speci¢c parts of Earth's history (e.g.,banded iron formations), rock type is generallyuseful only as a local stratigraphic marker, andlithology cannot be uniquely associated withtime. Thus terrestrial geologic units record phys-ical lithospheric processes, not time. Terrestrialtimekeepers are either recorders of broadly unidir-ectional changes within the biosphere or isotopicsystems, or they result from changes within theEarth's internal dynamo (paleomagnetism). Eachof these timekeeping systems is mostly indepen-

EPSL 5377 2-3-00

V.L. Hansen / Earth and Planetary Science Letters 176 (2000) 527^542 529

dent from lithospheric processes recorded by stra-tigraphic sequences, intrusions, and tectonicevents. It is the lack of correlation of these systemswith material units that allows geologists to unrav-el terrestrial lithospheric histories and processes.

The plate tectonics revolution ¢nalized the de-mise of any search for terrestrial global stratigra-phy. Plate tectonics (e.g., [13]) led to a highlymobile view of the Earth, following earlier theo-ries of continental drift [14], and recognition thatgeologic environments vary greatly in time andspace. The plate tectonics revolution resultedfrom technological and scienti¢c advances inbathymetric and magnetic data collection, aidedby faunal and absolute ages of oceanic crust, lead-ing to the realization that the ocean basins aremuch younger than the continents, rather thanolder, as previously assumed. This realizationopened the door for plate tectonics - a mechanismthat successfully explained the hotly debatedtheory of continental drift [15,16]. Conceptually,a global stratigraphic column cannot be recon-ciled with recycling of surface units at convergentmargins or with formation of new crust at diver-gent margins. Studies aimed at discovering a glob-al stratigraphy would contribute little to under-standing Earth processes. Some may argue thatseismic stratigraphy might reveal a global stratig-raphy; however, seismic stratigraphy is generallyrestricted to relatively young marine strata notgreatly disrupted by tectonic processes. Becausemuch of Earth's history involves both continentsand tectonism such a stratigraphy could never betruly global, nor could it record the bulk ofEarth's rich history. Thus, global stratigraphyon any planet is a dangerous assumption.

3. Geologic mapping

One goal of planetary geologists, like terrestrialgeologists, is to unravel geologic histories in orderto understand planet processes. Determination ofthe geologic history of a particular region com-monly begins with construction of a geologicmap, which, for solid planets, begins with identi-¢cation and di¡erentiation of geologic (material)units and geomorphic features [1]. Geomorphic

features can include primary structures (formedduring unit emplacement), secondary (or tectonic)structures, or erosional features (commonly pri-mary structures related to reworking of preexist-ing geologic units by wind, water, or ice (or ana-log equivalents)). Planet surfaces record thedominant processes that operated during theirevolution. Mercury and Moon record widespreadevidence of impact processes, whereas secondarystructures are rare. Mars displays evidence of im-pact and widespread erosional and depositionalprocesses, and localized volcanic processes ; evi-dence for tectonic structures is relatively rare.New evidence of early (s 3.5 Ga) linear magneticanomalies in Mars' southern highland suggeststhat tectonic processes may have been importantin early Mars evolution [17]. Venus shows abun-dant volcanic and tectonic features; sedimentaryand erosional features are rare given the high sur-face temperature (V475³C) and lack of surfacewater. Earth provides examples of all of the aboveprocesses.

Data types and resolution vary from planet toplanet. Earth has the most varied data sets, butnot the most areally extensive. New planetarymissions and advances in technology and scienti¢cknowledge result in new data. The ability to dif-ferentiate material units and tectonic elements re-quires a working knowledge of available datasets; such can be gained from planet- and mis-sion-speci¢c literature.

Venus' geologic units (or material units; e.g.,volcanic £ows, aeolian deposits, crater deposits)are typically di¡erentiated in Magellan data bypatterns in SAR, emissivity, or root mean square(RMS) slope data that re£ect primary featuressuch as lobate £ows, mottling, or homogeneity.The ¢rst-order task in mapping material units isto determine their spatial distribution and to ex-amine contact relations between adjacent units [1].Several problems must be kept in mind. Availabledata might inhibit unique distinction between dif-ferent material units, or may result in division of asingle geologic unit into two apparently di¡erentunits. For example, spatially separate lava £owsmay show similar radar, emissivity, or RMS slopecharacteristics and hence one might conclude (in-correctly) that these units are time correlative.

EPSL 5377 2-3-00


Alternatively, a single volcanic £ow unit emplacedwithin a single eruptive event may have both pa-hoehoe and aa £ow facies and show di¡erent ra-dar and RMS slope signatures, and therefore theymight be interpreted (incorrectly) as temporallydistinct geologic units. Lumping versus splittingof material units depends in part on the under-standing of process-dependent facies changes(e.g., metamorphic, volcanic, sedimentologic,stratigraphic, magnetic). Fundamentally, thesepotential problems are no di¡erent from thoseencountered in terrestrial mapping, although onEarth one has more tools to address the problem.

Lithologic similarity of units is not required,nor is it su¤cient, for correlation (`correlation'infers temporal equivalence ([5], p. 7;[9], p. 271).Terrestrial correlation requires either faunal orradiometric age data, or both. By contrast, corre-lations of material units in planetary geology are

based, to date, solely on surface crater densities ;this method requires numerous assumptions, re-sulting in gross correlations at best [1,18]. Craterdensity ages are most robust for planets with highcrater densities (very old surfaces). The more lo-calized a planet's surface processes, the less robustare crater density surface ages.

Unidirectional timekeeping systems that aredistinct from geologic units are absolutely neces-sary for correlation studies - without such tools,proposed correlations are simply unsupported as-sumptions. For example, prior to the advent ofgeochronometric techniques all terrestrial gneissicterrains were mapped as `Precambrian' (becausethey were consistently interpreted as the oldestlocal rock unit, and they lacked fossils), leadingsome workers to interpret the Precambrian as atime of global `gneissi¢cation', with further impli-cations for early Earth processes. Geochronomet-

Fig. 1. Geologic history experiment of a simple 10-time-step surface evolution. Sequential £ood-lava £ows (L1^5) darken withtime [19] over three time steps; dotted lines mark £ow boundaries; closely spaced black lines indicate suites of secondary struc-tures (fractures, faults, folds); horizontal lines mark wrinkle ridges. B: Geologic maps and resultant histories determined by:(I) experiment; (II) global stratigraphic method (unit de¢nition includes secondary structures: f, fracture; df, densely fractured;wr, wrinkle ridged), and (III) geohistory method.

EPSL 5377 2-3-00


ric data clearly indicate that gneissic terrains arediachronous, ranging in age from billions to mil-lions of years old; thus concepts of global corre-lation and gneissi¢cation have been discarded.Prudent planetary geologists will not assume tem-poral correlation of spatially separate terrainsthat appear, based on remote data, to comprisesimilar material units.

The construction of a geologic map is partlyfact and partly an interpretation, and interpreta-tions typically include operating assumptions.One must be mindful of assumptions implicit inany method and be careful to ensure that theassumptions do not hamper discovery. A simpleexperiment allows us to test geologic mappingmethods. Fig. 1A illustrates a simpli¢ed ten-step(t1^t10) magmatic and tectonic geohistory of a lo-cal `¢eld area'. In the experiment all material unitsare emplaced as low viscosity lava £ows (L1^5)with straight boundaries. Flow surfaces progres-sively degrade with time as re£ected by a changein radar character and RMS slope data, mimick-ing £ow units on Venus [19]. If £ows do not de-grade with time, some temporal resolution may belost because old and young £ows might be indis-tinguishable. Flows reach a steady-state surfacesignature after three time-steps and are no longerdistinguishable from earlier £ows. Two linearfracture zones form at distinct times (t3 and t6) ;fractures parallel the trend of their respective frac-ture zones. Late wrinkle ridges overprint all pre-existing material units.

The geologic map and `true' geohistory areshown in Fig. 1B. Geologists cannot typically ob-serve the evolution of a planet surface; insteadthey must attempt to reconstruct the surface his-tory by observation of the ¢nal surface (Fig. 1A^t10). Geologists organize observations throughconstruction of geologic maps. Map methodologyin£uences the product and the interpreted geo-logic history. Given that we know the history ofour experimental ¢eld area, we can use it to testmapping methods.

3.1. Global stratigraphic method

The global stratigraphic method has the statedgoal ([1], p. 219) of determining local sequences of

geologic units and integrating these into a coher-ent `global stratigraphy'. It allows secondarystructures (`post-depositional modi¢cations' ([1],p. 214) to be included in the de¢nition of a ma-terial unit (Wilhelms ([1], p. 227) states: `Struc-tures and structural patterns should be relatedto speci¢c rock or time-rock units and thereforeto the evolutionary history of the planet'). Butthis practice assumes a known evolutionary his-tory of a planet independent of, and prior to,mapping secondary structures - reminiscent ofWerner's assumed stratigraphic sequences, geo-logic history, and geologic processes. Determina-tion of global stratigraphy becomes the overridinggoal and its existence cannot be challenged.1

Consider a geologic map of the experimental¢eld area constructed using the global strati-graphic method allowing secondary structures tobe part of a material unit description (Fig. 1B, II)(e.g., [20^23]). The resulting map and history arequite di¡erent from the actual sequence of events(Fig. 1B, I). Wrinkle ridge formation is theyoungest event, and the boundaries of the young-est £ow unit (Lwr2 [L5]) are correctly de¢ned.(Many workers do not actually distinguish theformation of material units and wrinkle ridges[20^23], despite wrinkle ridge spacing of morethan 50 times image resolution). However, neitherthe spatial distribution nor the relative ages ofLwr1, Lf, or Ldf bear any relation to reality.Lwr1 combines all units not a¡ected by a fractureevent and not covered by L5 (parts of L1, L2, L3and L4; t1, t2, t4, and t7). Lf combines parts of

1 Evidence that this goal colors Wilhelms' method is providedin his topical discussion order; he discusses unit correlation(7.2.3) and time-rock units (7.2.4) prior to structures and struc-tural units (7.2.5). Although Wilhelms admits that structurescommonly cut several rock units, and although he argues thatthe Stratigraphic Code does not allow for mapping of unitsbased solely on their structural modi¢cation, in the same para-graph he states (p. 227): `Where the rock units of a highlyfaulted or otherwise deformed terrain are recognizable, theyshould be mapped as separate rock units. Sometimes, however,the deformed rock units are not recognizable. In this case it isbetter to map the structural unit as `fractured plains material'than to ignore the presence of the structures in order to adherestrictly to the code'. This allows secondary structures to char-acterize a material unit.

EPSL 5377 2-3-00


material units L1, L2, L3, and L5 (t1, t10, t4 andt9), and it combines fracture events 1 (t3) and 2(t6). Ldf includes parts of L1 and L2 (t1 and t2)and both fracture events (t3 and t6). Ldf is de-¢ned, therefore, solely based on its deformationfeatures. The youngest interpreted material unit(Lwr2) is di¡erentiated from the next older unit,Lwr1, only based on gray scale (radar character),which would not exist if the experiment were al-lowed to run one more time-step. In that caseLwr1 and Lwr2 would be mapped as a singlematerial unit, combining parts of the oldest andyoungest lava £ow units (L1^5). The proposedgeohistory is grossly incorrect and any processesinterpreted from the purported geohistory mustalso be incorrect.

3.2. Geohistory method

The geohistory method has the stated goal ofdetermining the geochronology of local regionsand progressively assembling those histories intotestable models of planet evolution. It allows thatgeologic histories record tectonic processes. Themethod begins as Wilhelms [1] outlined, with dif-ferentiation of material units and geomorphic fea-tures, but it also requires di¡erentiation betweenprimary and secondary structures. Material unitsare determined in the manner and with the cau-tions detailed by Wilhelms [1], and outlinedbrie£y above. Geomorphic features, generallythree-dimensional shapes or landforms, can be ei-ther primary or secondary structures. Geomor-phic features can be di¡erentiated on Venus byradar e¡ects resulting from local topography, oraltimetry if su¤ciently large (tens to hundreds ofkilometers). Primary structures commonly pro-vide clues to material unit properties or emplace-ment processes (e.g., £ow morphologies re£ecting£ow viscosity) because they formed during unitemplacement; in addition, they can be unit spe-ci¢c and used as a material unit descriptor. Sec-ondary or tectonic structures formed after materi-al unit deposition or emplacement (e.g., fractures,faults, folds, wrinkle ridges), and thus recordtime(s) and process(es) distinct from the materialunit they deform. Secondary structures absolutelycannot constitute a part of a material unit(s) de-

scriptor or characteristic. Using secondary struc-tures to describe a material unit implies that thematerial unit and the structural element re£ect asingle geologic event; this implication becomesembedded in the data (geologic map) and its via-bility cannot be tested later. In mapping second-ary structures one seeks to understand geometry,kinematics (movement and timing), and ulti-mately dynamic processes (driving forces). Thepitfalls and complexities of such analyses aremany; a few cautions are discussed brie£y.

Secondary structures typically form lineaments

Fig. 2. Idealized models of the surface evolution of a theoret-ical planet. Three arbitrary events (A, B, and C in which ageof AsBsC) and four arbitrary regions (R1^R4) illustratespatial and temporal relations of events: (a) A, B, and C aregenetically linked (must be independently demonstrated), butnot necessarily global synchronous; (b) the broadest of inter-pretations in which A, B, and C are spatially and temporallyunrelated - the most conservative interpretation lacking abso-lute ages; (c) a minimum of four speci¢c absolute ages (boldboxes) could indicate that at R1^R4 A precedes B, whichprecedes C; (d) interpretation following the global strati-graphic method - requires at least 12 speci¢c absolute ages(bold boxes) to document.

EPSL 5377 2-3-00


in two-dimensional data sets such as SAR,although they are three-dimensional planes (e.g.,faults) or curvilinear shapes (e.g., folds). First or-der issues concerning secondary structures in-clude: character (paired, anastomosing, linear),orientation, length and width, density or spacing,wavelength, continuity, spatial distribution andpatterns, and spatial relations with material units.The spatial distribution of secondary structures isnot typically correlative with the spatial limits ofmaterial units. This general lack of spatial associ-ation provides evidence that secondary structuresand material units record fundamentally di¡erentprocesses or parts of processes. Secondary struc-tures result from stresses associated with plane-tary processes, whereas material units record thedistribution (or redistribution) of planet material.Thus, secondary structures and material units re-cord di¡erent events within a geohistory, andthus, di¡erent aspects of geological process(es).A greater ability to delineate material units andsecondary structures in time and space leads to amore detailed geohistory and a more detailedunderstanding of operative processes.

Geohistory analysis requires relative time con-straints. Although faunal succession can help tocorrelate spatially separate terrestrial regions, ingeneral, similar sequences of geologic events re-corded in spatially separate regions do not pro-vide any a priori evidence for correlation. Despitethis limitation, local relative temporal constraints- gleaned from stratigraphic and crosscutting re-lations, and mechanical analysis - are critical tounderstanding fundamental aspects of planetaryprocesses. Stratigraphic analysis deals mainlywith strata in the absence of tectonism; cross-cut-ting relations involve both material units and tec-tonic structures ; mechanical analysis relates totiming among tectonic structures as determinedby material rheology - largely a function of com-position and physical conditions during tecton-ism. Each has its limits and bene¢ts. Because pro-posed geohistories must be able to accommodatetemporal relations determined through each ofthese somewhat independent means, robust geo-history analysis will employ as many methods aspossible because geohistories are fundamentallybuilt on consistency and compatibility arguments.

Stratigraphic relations address local stacking ofgeologic units with older below younger units,and assume original deposition as roughly hori-zontal. Stratigraphic analysis is based the threeprinciples mentioned above: (1) superposition,(2) uniformitarianism, and (3) faunal succession.Stratigraphic analysis assumes that tectonism hasnot omitted or repeated the stratigraphic section.Stratigraphic analysis requires a means to test thisassumption, as well as to correlate strata of sep-arate regions (e.g., 3 above). Means of correlationare rare (presently absent?) for many planets. Inaddition, given two-dimensional remote data sets,it can be di¤cult to robustly determine the stack-ing order, and thus unit superposition [24]. Thepresence of secondary structures, however, mightbe used to constrain the relative temporal rela-tions between two or more stratal units.

Crosscutting relations can provide relative ageconstraints for material units and secondarystructures. The younger unit or secondary struc-ture may overprint the older unit(s) or secondarystructure(s); younger units might contain clasts ofan older unit, or a younger unit may embay topo-graphic features associated with an older structur-al element. If tectonism occurred between the em-placement of material units, tectonic patterndistribution may allow determination of the spa-tial distribution of the younger unit, the area af-fected by the tectonic process, and ¢rm relativetemporal constraints between the two materialunits. Temporal constraints are gained by com-parison of material unit patterns with structuralpatterns. For example, assuming horizontal dep-osition, if a material unit preferentially occupiessynformal fold valleys, and hence the contact be-tween material units correlates with fold crests ortroughs, the material unit that ¢lls the synformalvalleys was emplaced after folding - and foldingpostdates formation of the folded unit(s). If thecontact between material units is not spatially cor-relative with the fold crests or troughs, both ma-terial units likely formed prior to folding. In ad-dition, we must be mindful that a geologicenvironment might not favor formation or pres-ervation of crosscutting relations. For example, ifthe processes or crustal section one is studyingformed below a planet surface, no `classic' cross-

EPSL 5377 2-3-00


cutting relations involving surface material unitswould be possible (e.g., temporal constraints onS^C shear zone fabrics [33]). At the surface `clas-sic' cross-cutting relations will be preserved if andonly if material units deposited on the surfaceformed continuously during the tectonic evolutionand were variably distributed across the maparea. If a material unit has completely covered asurface of interest, the record of earlier geologichistory is commonly lost to the remote investiga-tor.

Some understanding of structural topography(three-dimensional character) can also be gainedby detailed mapping of a `lava-line', or the em-bayment patterns marked by the boundary be-tween deformed crust and £ood lava £ows. The`shoreline' de¢ned by £ows can provide a detailedpicture of topographic highs and lows. Compar-ison of the lava line and structural orientationsnear the `shoreline' can provide strong constraintson the detailed topography de¢ned by secondarystructures that predated the lava £ow.

Mechanical analysis of relative age addressesthe interaction and development of tectonic struc-tures, and depends, in part, on kinematic analysis.Kinematic analysis endeavors to identify distinctstructural elements, and to address geometrical,rheological and temporal relations recorded byeach suite of elements with the goal of under-standing the deformation history. Kinematic anal-ysis involves several, typically iterative, tasks: (1)identify and di¡erentiate structural suites ; (2) de-termine geometry, orientation, and spatial distri-bution of each suite; (3) determine temporal rela-tions between and among structural suites,recognizing that they could have formed synchro-nously, that they could be partially or wholly re-activated, and that determination of temporal re-lations requires knowledge of material propertiesand deformation mechanisms, and (4) develop akinematic model that accommodates documentedconstraints. Structural principles and their rangeof applications, the topic of many textbooks andnumerous journal articles, are discussed here onlyin the broadest terms.

Identi¢cation and di¡erentiation of structuralelements is a ¢rst, and critical, step. It is imper-ative that secondary structures be delineated from

primary structures and mapped independent ofmaterial units. Further, if distinct structural ele-ments are not di¡erentiated, subsequent geometricand temporal analysis may be £awed. Primarystructures, genetically (and thus temporally) re-lated to unit emplacement, may prove useful asmapping criteria, and provide clues to unit em-placement processes. In contrast, secondary struc-tures record tectonic processes; they postdate em-placement of units they deform. In somesituations the character, orientation, and distribu-tion relative to an associated material unit willindicate whether structural elements are primaryor secondary. If a suite of structures is di¤cult toidentify as primary or secondary, it is generallybest to map the suite as secondary because thisinterpretation can be tested with continuedmapping. If the structures are mapped as pri-mary, this implies temporal equivalence with the`host' material unit, an interpretation that be-comes embedded in the material unit descriptionand thus impossible to test with continued map-ping.

Geometric analysis requires three-dimensionalexposure, or, in the case of remotely sensed two-dimensional data, a means to constrain three-di-mensional geometry. For two-dimensional datageometric models are commonly non-unique,and in many cases unique geometric determina-tion is a nontrivial exercise. Furthermore, geome-try, no matter how carefully de¢ned, cannotuniquely constrain relative timing. Geometricconstraints are necessary but not su¤cient forkinematic analysis ; relative temporal constraintsare also required.

Relative temporal clues can be gleaned frominteractions between material units and structuralsuites, or from rheological considerations. In eachcase, temporal constraints depend on deformationor emplacement mechanisms. For example, deter-mining the relative age of a layer-parallel bodyrequires knowledge of whether it was emplacedas a volcanic £ow unit or an intrusive sill ; inde-pendently determined temporal data could favorone mechanism over the other, or detailed obser-vation might indicate the mechanism and henceconstrain the relative timing. Unfortunately, com-monly the deformation mechanism is precisely

EPSL 5377 2-3-00


what one is attempting to understand; hence theiterative nature of kinematic analysis.

Knowledge of rheology and rheologic structurecan provide clues about deformation mechanisms.Clues to rheology can be gained from the wave-length of regularly spaced structural elementssuch as folds, fractures, joints, or boudins - rela-tions that can commonly be extracted from two-dimensional data sets. Empirical, analytical, andanalog modeling studies consistently illustratethat a dominant wavelength:thickness ratio (l :d)exists for both contractional and extensionalstructures, with average l :d V5.5 for folds,V2^6 for brittle extensional features, and up to10^20 for ductile extensional features, respectively(e.g., [25,26]). Layer instability analysis can con-strain planet layer thickness during deformation[27,28] that can, in turn, provide clues for defor-mation mechanisms and timing [29].

Clues to the dip, relative timing, and depth ofstructures might be inferred from their spacing,their interactions with topography, and/or theirinteraction with other suites of structures. Steeplydipping structures cut across topography; shallowdipping structures follow topography. Terminatedfractures are typically younger than through-going fractures, because a through-going fractureacts as a free surface that blocks propagation ofyounger fractures [30,31]. However, a youngerfault could truncate an earlier formed fracture(or fault), resulting in a similar `T' geometry butwith opposite temporal relations. Interpretingtemporal relations from crossing fractures canbe more di¤cult: if an older through-going frac-ture is ¢lled, it might no longer act as a free sur-face, allowing younger fractures to propagateacross it. In addition, if a young fracture initiatesat a greater depth than an old fracture, it maypropagate beneath and around the older fracture.Fracture spacing can provide an independent con-straint on timing because fracture spacing re£ects,in part, the depth of the fracture set [29^31]. Twointersecting fractures could form synchronouslyunder the right conditions [32].

Temporal constraints derived from crosscuttingrelations or mechanical analysis must also consid-er possible structural reactivation-renewed slipalong previously formed mechanical discontinu-

ities. Early formed secondary structures can bereactivated any time after their formation givenfavorable conditions. Reactivation can occur afterdeposition or emplacement of a younger materialunit that covers the previously formed tectonicstructure; thus preliminary temporal analysismay be misleading. Once formed, a secondarystructure has the potential to be a material weak-ness. Whether or not a pre-existing structureforms a weakness that could be later reactivateddepends on the character of the structure and onits orientation with respect to future (younger)principal stresses [34]. Careful mapping with at-tention to delineation of material units and sec-ondary structures can provide evidence of reacti-vation.

Because all geologic features are three-dimen-sional, geologic mapping should include the con-struction of cross-sections and block diagrams inorder to explore possible three-dimensional rela-tions [1]. Cross-sections and block diagrams pro-vide a critical reminder that geologic elements arespatially three-dimensional ; they also help toidentify unstated assumptions, previously unrec-ognized problems or solutions, and they encour-age workers to ask additional questions of thedata.

Using these principles we can construct ageologic map of the experimental ¢eld area(III of Fig. 1B). The resulting map correctly de-lineates each lava £ow unit (LF4 [L5], LF3 [L4],LF2 [L3]), except LFb, which combines L1 andL2 as an undi¡erentiated basal unit. The geohis-tory also matches `reality' except for the lumpingof L1 and L2 into a basal lava £ow unit (LFb).This method results in robust determination ofboth the spatial and relative temporal relationsof all post-L2 events (t2^t10), from which it ispossible to hypothesize processes responsible forsurface formation. Further degradation of £owsurfaces (t11T) would result in the combining ofLF3 and LF4 into a single young £ow unit.This simpli¢ed experiment does not considerthree-dimensional relations that could provideadditional clues to geohistory and hence, opera-tive processes. Additional clues might also begleaned from mechanical analysis of tectonicstructures.

EPSL 5377 2-3-00


4. Absolute time

Even if one is able to determine a unique se-quence of geologic events within a speci¢c ¢eldarea, several critical questions with regard toplanetary processes remain unconstrained, per-haps the most pressing of which deals with abso-lute time. The absolute time required for terres-trial processes is commonly debated for twofundamental reasons: (1) its importance, and (2)its di¤culty to constrain. Methods for measuringabsolute time are continually improving. In thelate 1800s the Earth was considered to be V100Myr old based on stratigraphic arguments, salin-ity calculations, and cooling models of the Earthand Sun. The discovery of radioactive decay anddevelopment of radiometric techniques tentativelyestablished Earth's age as 2 byr old by the 1930s.Once the age of the Earth was relatively well es-tablished, e¡orts to determine absolute time weredirected toward the geologic time scale and deter-mining rates of tectonic processes [35]. Rate deter-mination requires the ability to date both the on-set and cessation of an event. Currently availabledating techniques can date speci¢c minerals, or, inthe case of planetary studies, surfaces comprisedof material units. Tectonic processes typically re-sult in secondary structures, the dating of whichrequires bracketting by two datable materialunits. The accuracy relates, in part, to how closelythe material units temporally bracket the tectonicevent in question. This process is nontrivial evenon Earth.

Workers must be ever mindful of the potentialproblems with absolute age determinations. Themost robust of geohistories will be built upon arelative temporal framework punctuated with ro-bustly determined absolute age constraints. Cor-relation of di¡erent regions with similar relativegeologic histories or a sequence of geologic eventsis not valid in the absence of an independent timemarker (Fig. 2). `Correlations' based only on sim-ilar relative geohistories assume - but do notdocument - global synchroneity, following Wern-er. Fig. 2 illustrates interpretations of three theo-retical geologic events (A, B and C) identi¢ed infour spatially separate regions (R1^R4). At eachlocation, A predates B, which predates C. The

three events could be completely unrelated, orthey could record a common geologic process. Ifthe events are unrelated, documentation of globalsynchroneity of all three events would requiredocumentation of the absolute age of each eventin each region (a minimum of 12 absolute ages). Ifone could establish that events A^C record thesame type of sequential geologic process at eachlocation, then synchroneity might be establishedwith only four absolute age determinations of thesame phase of the A^C sequence at each of thefour locations. (Of course this does not establishthat the A^C sequence is globally synchronous,only that the sequence occurred at the sametime at these four locations). Lacking absoluteage data, the similar sequence of events at eachlocation in no way requires, or even suggests, syn-chroneity. Although an interpretation of globalsynchroneity of each of the three events mightbe the simplest model to envision (e.g., Werner),such a result actually requires the most highlyconstrained model of planet evolution and themost detailed supporting data. For example, con-tinental rifting results in a common succession ofrock types and deformation, the sequence ofwhich re£ects the fundamental process of rifting;but the stages of continental rifting need not oc-cur synchronously in di¡erent locations, and themore general interpretation is that they do not.To establish that continental rifting in separatelocations was indeed globally synchronous wouldrequire individual absolute age constraints foreach location. Establishing global synchroneity re-quires detailed data sets including numerous inde-pendently determined absolute ages.

Determination of absolute time is fundamen-tally di¤cult to constrain in planetary studies.Because absolute time and rates are so criticalto understanding the process one must be partic-ularly vigilant about the robustness of all absoluteages used to constrain models. Signi¢cantly incor-rect temporal data are much worse than no tem-poral data at all. In planetary geology the onlymeans to estimate absolute age is the density ofimpact craters on a surface [18]. In the best ofcircumstances this might allow one to constrainthe absolute age of a particular surface. This tech-nique requires several assumptions (in addition to

EPSL 5377 2-3-00


an assumed impactor £ux rate), and ages are onlyas valid as the assumptions.

Although crater density might place limited ab-solute age constraints on Lunar and Martian sur-faces, the assumptions inherent in this datingtechnique must be critically evaluated for Venusand other tectonic planets. Venus hosts V950 rel-atively evenly globally distributed impact craters[36,37]. The statistics of small numbers requirethat any datable surface on Venus be greaterthan 20 000 000 km2 [38] - a region larger thanNorth America. Additionally one must assumethat the entire surface developed quickly (that is,that it formed synchronously). These assumptionsare circular. Furthermore, surface ages estimatedby this method cannot be extrapolated in thesame way as radiometric dates, which typicallyrecord the cooling of speci¢c minerals throughknown closure temperatures. Crater density sur-face ages are more akin to terrestrial ANd averagemantle model ages, which re£ect the average timeat which all of a rock's components were ex-tracted from the Earth's mantle. Rocks withvery di¡erent geohistories could yield the sameaverage mantle model age [39]. Similarly, craterdensity determinations yield average surface mod-el ages, and therefore surfaces with extremely dif-ferent geohistories could yield the same craterdensity; a surface could be comprised of manyunits with a wide range in ages, or a single unitwith the mean crater density age, or an in¢nitenumber of other possibilities [40,41].

Several workers have used crater density in at-tempts to discern relative or absolute ages of geo-morphic provinces across Venus [42^44]. Thesestudies should be reevaluated with regard to issuesof crater ages [40,41], and the implications ofthese studies should not be incorporated intomodels without careful consideration of theunderlying assumptions. These studies lack ro-bustness because the true mean crater density ispoorly constrained, and any crater density can beaccommodated by a host of rate functions rang-ing from simple to complex; thus, currently, cra-ter counts o¡er little constraint on the relativetime of duration of geologic units on Venus [41].

Dating tectonic events or determining rates ofevents are even more perilous. In addition to the

signi¢cant problems outlined above, one must beable to robustly establish whether each individualcrater is younger or older than the relevant sec-ondary structures. This is di¤cult because bothsecondary structures and impact craters arespaced elements - that is, they do not occur every-where, even within a surface that predated theirformation.

In the case of dating tectonic elements on Ve-nus the best candidate might be wrinkle ridges,because their distribution broadly correlates withglobal gravity^topography relations, indicatingthat wrinkle ridge formation might have been re-stricted to a somewhat narrow (recent) timeframe[45]. Any attempt to test this hypothesis requiresunique determination of the relative ages of indi-vidual wrinkle ridges and impact craters.Although most Venusian impact craters are wellpreserved, various features distinguish relativelyyoung (fresh) from relatively old (degraded) im-pact craters (Fig. 3A). Radar-bright (rough)£oored craters with halo deposits are young,whereas radar-dark (£ood-lava-¢lled) £oored cra-ters with little or no halo deposits are old; inaddition, impact crater formation and interior£ooding represent two distinct events [46]. Mostcraters with interior ¢ll lack evidence of £oodingfrom outside the crater [36,47], and thus at leastone possible model for crater ¢lling is £oodingfrom below (Fig. 3B (see also [48])). Impact cra-ters and wrinkle ridges must each postdate thesurfaces they deform. Therefore, if a surface hostscraters and wrinkle ridges, both craters and wrin-kle ridges are younger than the surface materialunit(s). How much younger is unconstrained. Fur-thermore, given that interior £ooding postdatescrater formation (impact), and that craters andwrinkle ridges are spaced, unique determinationof crater^wrinkle ridge temporal relations is rarelypossible. In SAR data, material units or tectonicelements are delimited by brightness, a function ofroughness and orientation relative to incident ra-dar. Because it is impossible to determine uniquerelative temporal relations between spatially over-lapping radar bright crater ejecta or radar brightcrater interiors with radar-bright wrinkle ridges(Fig. 3C, k,l), the relative ages of wrinkle ridgesand young craters (lacking interior radar-dark ¢ll)

EPSL 5377 2-3-00


is indeterminate. For craters with £ooded interi-ors (Fig. 3C, m^p), one must consider three dis-tinct geologic events following formation of thebase surface. The relative age of wrinkle ridge^crater formation can only be determined in the

infrequent case in which a wrinkle ridge cuts,and therefore postdates, the radar-dark crater¢ll (Fig. 3C, p). In that case, the wrinkle ridgepostdates the crater, but it also postdates crater£ooding - therefore an indeterminate period of

Fig. 3. (A) Crater degradation model illustrating sequential stages of crater morphology [46]; radar-dark crater ¢ll (presumablylava £ows) are temporally distinct from impact event. (B) Block diagram of subsurface crater £ooding to some pressure surfacelevel. (C) Geologic maps (k^p) and resultant geologic histories illustrating possible crater^wrinkle ridge temporal relations (oldestbelow youngest).

EPSL 5377 2-3-00


time elapsed between crater and wrinkle ridge for-mation. In another rare case one might argue thatan individual wrinkle ridge spans the width of a¢lled crater (Fig. 3C, o), and that it is likely (butnot required) that the wrinkle ridge predated cra-ter £ooding; however, the wrinkle ridge couldhave formed either before or after the impactevent. In each other case the order of crater for-mation, crater £ooding, and wrinkle ridge forma-tion are indeterminate (Fig. 3C, k^n). Thus, inonly one unique case out of six can the relativeage between individual crater and wrinkle ridgeformation be determined (Fig. 3C, p). This uniquecase allows only determination of wrinkle ridgeformation after crater formation and £ooding.Therefore, absolute age constraints (much lessrates) of tectonic processes on Venus are not cur-rently possible.

Some prominent current tectonic models forVenus fall victim to these uncertainties. For exam-ple, models that propose early globally extensivehigh strain deformation resulting in formation ofglobal tessera represent a non-unique interpreta-tion of limited data sets lacking temporal con-straints [20^23,49,50]. These models, conceptuallysimilar to abandoned terrestrial Precambriangneissi¢cation models, assume global distributionand synchroneity of tessera, as well as high strainduring tessera formation; yet, none of these as-sumptions are documented, and are, in fact, in-consistent with detailed structural analysis[29,51,52].

5. Summary

Construction of a geologic map is a critical ¢rststep in unraveling geologic history for any solidplanet. Because geologic mapping is itself an in-terpretation that results in a database to be usedfor interpretation of geologic processes, mappingmust be conducted in such a fashion as to ensurethat any operative process can be discovered -that is, the mapping method must not predeter-mine the geologic map. Any method should allowthat tectonic processes could be important in theevolution of a planet surface. It is imperative thatsecondary structures be distinguished from mate-

rial units because material units and secondarystructures record di¡erent `time slices' in the evo-lution of a planet surface and re£ect di¡erent as-pects of planetary processes. Clear di¡erentiationof primary and secondary structures might not becritical for tectonically inactive planets; however,one should not assume tectonic inactivity on un-investigated planets. For Venus and other tectonicplanets, failure to clearly di¡erentiate secondarystructures from material units leads to erroneousinterpreted geologic histories, and can inhibit rec-ognition of critical aspects of the planet's historyand therefore critical aspects of planetary evolu-tion processes. In addition, time is a fundamentalfactor in understanding processes of planet evolu-tion in general, and tectonic processes in particu-lar. Time must be robustly constrained, and can-not be assumed. To assume time is to assumefundamental aspects of the processes one is at-tempting to understand.

Acknowledgements

This work was supported by NASA GrantsNAGW-4562 and NAGW-4609. Discussionswith B. Banks, L. Bleamaster, H. DeShon, R.Ghent, J. Goodge, L. McAlester, D. Oliver, andD. Young helped clarify the ideas and presenta-tion. C. Teyssier and J. Zimbelman provided help-ful formal reviews.[RV]

References

[1] D.E. Wilhelms, Geologic mapping, in: R. Greeley, R.M.Batson (Eds.), Planetary Mapping, Cambridge Univ.Press, NY (1990) 208^260.

[2] K.T. Tanaka, H.J. Moore, G.G. Schaber, M.G. Chap-man, E.R. Stofan, D.B. Campbell, P.A. Davis, J.E. Guest,G.E. McGill, P.G. Rogers, R.S. Saunders, J.R. Zimbel-man, The Venus geologic mappers' handbook, USGSOpen-File Report 94^438, 1994, 50 pp.

[3] G.K. Gilbert, Inculcation of the scienti¢c method, Am. J.Sci. 31 (1886) 284^299.

[4] J.P. Ford, J.J. Plaut, C.M. Weitz, T.G. Farr, D.A. Senske,E.R. Stofan, G. Michaels, T.J. Parker, Guide to MagellanImage Interpretation, JPL Publ. 93^24, 1993, 148 pp.

[5] J.M. Weller, Stratigraphic Principles and Practice, Harperand Brothers, NY, 1960, 725 pp.

EPSL 5377 2-3-00


[6] H.S. Williams, Dual nomenclature in geological classi¢ca-tion, J. Geol. 2 (1894) 145^160.

[7] C.D. Walcott, Nomenclature and classi¢cation for thegeological atlas of the United States, US Geol. Surv.Annu. Rep. 24 (1903) 21^27.

[8] J.W. Powell, Conference on map publication, US Geol.Surv. Annu. Rep. 10 (1890) 56^79.

[9] C.O. Dunbar, J. Rogers, Principles of Stratigraphy, JohnWiley and Sons, NY, 1957, 356 pp.

[10] F.M. Gradstein, F.P. Agterberg, J.G. Ogg, J. Hardenbol,P. van Veen, J. Thierry, Z. Huang, A Triassic, Jurassicand Cretaceous time-scale, in: W.A. Berggren, D.V. Kent,M.-P. Aubry, J. Hardenbol (Eds.), Geochronology, timescales and global stratigraphic correlation, SEPM (1995)95^129.

[11] C.E. Isachsen, S.A. Bowring, E. Landing, S.D. Samson,New constraint on the division of Cambrian time, Geol-ogy 22 (1994) 496^498.

[12] J. Gilluly, A.C. Waters, A.O. Woodford, Principles ofGeology, W.H. Freeman, San Francisco, CA, 1951, 631pp.

[13] J.T. Wilson, Static or mobile earth: the current scienti¢crevolution, Am. Phil. Soc. Proc. 112 (1968) 309^320.

[14] A. Wegner, The Origin of Continents and Oceans, Me-thuen and Co. LTD, London, 1924, 212 pp.

[15] A.M.C. Sengo«r, Plate tectonics and orogenic researchafter 25 years: A Tethyan perspective, Earth-Sci. Rev.27 (1990) 1^201.

[16] A.M.C. Sengo«r, J.F. Dewey, Terranology: vice or virtue?,Phil. Trans. R. Soc. London 331 (1990) 457^477.

[17] J.E.P. Connerney, M.H. Acuna, P.J. Wasilewski, N.F.Ness, H. Reme, C. Mazelle, D. Vignes, R.P. Lin, D.L.Mitchell, P.A. Cloutier, Magnetic lineations in the ancientcrust of Mars, Science 284 (1999) 794^798.

[18] W.B. McKinnon, K.J. Zahnle, B.A. Ivanov, H.J. Melosh,Cratering on Venus: Models and observations, in: S.W.Bouger, D.M. Hunten, R.J. Phillips (Eds.), Venus II,Univ. AZ Press (1997) pp. 969^1014.

[19] R.E. Arvidson, R. Greeley, M.C. Malin, R.S. Saunders,N. Izenberg, J.J. Plaut, E.R. Stofan, M.K. Shepard, Sur-face modi¢cation of Venus as inferred from Magellanobservation of plains, J. Geophys. Res. 97 (1992)13303^13318.

[20] M.A. Ivanov, J.W. Head, Tessera terrain on Venus: Asurvey of the global distribution, characteristics, and rela-tion to surrounding units from Magellan data, J. Geo-phys. Res. 101 (1996) 14861^14908.

[21] A.T. Basilevsky, J.W. Head, Evidence for rapid and wide-spread emplacement of volcanic plains on Venus: Strati-graphic studies in the Baltis Vallis region, Geophys. Res.Lett. 23 (1996) 1497^1499.

[22] A.T. Basilevsky, J.W. Head, The geologic history of Ve-nus: A stratigraphic view, J. Geophys. Res. 103 (1996)8531^8544.

[23] J.W. Head, A.T. Basilevsky, Sequence of tectonic defor-mation in the history of Venus: Evidence from globalstratigraphic relations, Geology 26 (1998) 35^38.

[24] J. Zimbelman, Planetary geologic mapping exercise, in:Geol. Soc. Am. workshop, Toronto, 1998.

[25] N. Price, J. Cosgrove, Analysis of Geological Structures,Cambridge Univ. Press, Cambridge, 1990, 502 pp.

[26] P. Hudleston, L. Lan, Rheological information from geo-logical structures, PAGEOPH 145 (1995) 607^620.

[27] M.T. Zuber, Constraints on the lithospheric structure ofVenus from mechanical models and tectonic surface fea-tures, J. Geophys. Res. 92 (1987) E541^E551.

[28] W.B. Banerdt, M.P. Golombek, Deformation models ofrifting and folding on Venus, J. Geophys. Res. 93 (1988)4759^4772.

[29] V.L. Hansen, J.J. Willis, Ribbon terrain formation, south-western Fortuna Tessera, Venus Implications for litho-sphere evolution, Icarus 132 (1998) 321^343.

[30] T. Engelder, P. Geiser, On the use of regional joint sets astrajectories of paleostress ¢elds during the development ofthe Appalachian Plateau, New York, J. Geophys. Res. 85(1980) 6319^6341.

[31] D.D. Pollard, A. Aydin, Progress in understanding joint-ing over the past century, Geol. Soc. Am. Bull. 100 (1988)1181^1204.

[32] J.J. Willis, V.L. Hansen, Conjugate shear fractures at `KiCorona', southeast Parga Chasma, Venus, Lunar Planet.Sci. 25 (1996) 1443^1444.

[33] D. Berthe, P. Choukroune, P. Jegouzo, Orthogneiss, my-lonite and noncoaxial deformation of granite: The exam-ple of the South Armoricain shear zone, J. Struct. Geol. 1(1979) 31^42.

[34] R.H. Sibson, A note on fault reactivation, J. Struct. Geol.7 (1985) 751^754.

[35] R.L. Armstrong, Sevier orogenic belt in Nevada andUtah, 79 (1968) 429^458.

[36] R.J. Phillips, R.F. Raubertas, R.E. Arvidson, I.C. Sarkar,R.R. Herrick, N. Izenberg, R.E. Grimm, Impact cratersand Venus resurfacing history, J. Geophys. Res. 97 (1992)15923^15948.

[37] G.G. Schaber, R.G. Strom, H.J. Moore, L.A. Soderblom,R.L. Kirk, D.J. Chadwick, D.D. Dawson, L.R. Gaddis,J.M. Boyce, J. Russell, Geology and distribution of im-pact craters on Venus: What are they telling us?, J. Geo-phys. Res. 97 (1992) 13257^13302.

[38] R.J. Phillips, The age spectrum of the Venusian surface,Eos 74 (16) (1993) 18.

[39] G.L. Farmer, D.J. DePaolo, Origin of Mesozoic and Ter-tiary granite in the Western United States and implica-tions for the pre-Mesozoic crustal structure 1. Nd and Srisotopic studies in the geocline of the northern Great Ba-sin, J. Geophys. Res. 88 (1983) 3379^3401.

[40] S.A. Hauck, R.J. Phillips, M.H. Price, Venus: Crater dis-tribution and plains resurfacing models, J. Geophys. Res.103 (1998) 13635.

[41] B.A. Campbell, Surface formation rates and impact craterdensities on Venus, J. Geophys. Res. 104 (1999) 21951^21955.

[42] N. Namiki, S.C. Solomon, Impact crater densities on vol-

EPSL 5377 2-3-00


canoes and coronae on Venus: implications for volcanicresurfacing, Science 265 (1994) 929^933.

[43] M. Price, J. Suppe, Mean age of rifting and volcanism onVenus deduced from impact crater densities, Nature 372(1994) 756^759.

[44] M.H. Price, G. Watson, C. Brankman, Dating volcanismand rifting on Venus using impact crater densities, J. Geo-phys. Res. 101 (1996) 4637^4671.

[45] D.T. Sandwell, C.L. Johnson, J. Suppe, Driving forces forlimited tectonics on Venus, Icarus 129 (1997) 232^244.

[46] N.R. Izenberg, R.E. Arvidson, R.J. Phillips, Impact craterdegradation on Venusian plains, Geophys. Res. Lett. 21(1994) 289^292.

[47] R.R. Herrick, V.L. Sharpton, M.C. Malin, S.N. Lyons,K. Feely, Morphology and morphometry of impact cra-ters, in: S.W. Bouger, D.M. Hunten, R.J. Phillips (Eds.),Venus II, Univ. Ariz. Press (1997) 1015^1046.

[48] R.W. Wichman, Internal crater modi¢cation on Venus:Recognizing crater-centered volcanism by changes in £oormorphology and £oor brightness, J. Geophys. Res. 104(1999) 21957^21977.

[49] S.C. Solomon, The geophysics of Venus, Phys. Today 46(1993) 48^55.

[50] R.E. Grimm, Recent deformation rates on Venus, J. Geo-phys. Res. 99 (1994) 23163^23171.

[51] R.R. Ghent, V.L. Hansen, Structural and kinematic anal-ysis of eastern Ovda regio. Venus implications for crustalplateau formation, Icarus 139 (1999) 116^136.

[52] V.L. Hansen, B.K. Banks, R.R. Ghent, Tessera terrainand crustal plateaus, Venus, Geology 27 (1999) 1071^1074.

EPSL 5377 2-3-00


Documents

Geologic mapping of tectonic planets€¦ · ¢ned by a detailed magnetic polarity time scale correlated by radiometric dates. The geologic time scale is continually being re¢ned