Paleosols in clastic sedimentary rocks: their geologic applications

Ž .Earth-Science Reviews 47 1999 41–70www.elsevier.comrlocaterearscirevPaleosols in clastic sedimentary rocks: their geologic applicationsMary J. Kraus )Department of Geological Sciences, UniÕersity of Colorado, Boulder, CO, 80309-0399, USAReceived 22 July 1998; accepted 24 March 1999AbstractInterest in pre-Quaternary paleosols has increased over the past decade, in large part, because they have proved to bebeneficial in solving diverse geological problems. The majority of paleosols are described from continental deposits, mostcommonly from alluvial strata. Criteria for recognizing these paleosols have been extensively described; however,classifying them has proved more complicated. Pre-Quaternary paleosols are generally classified according to one or moremodern soil classification systems, although one new classification has been proposed exclusively for paleosols to avoidproblems using the modern soil classifications. In addition to taxonomic classification, paleosols can be categorizedaccording to the interplay among deposition, erosion, and the rate of pedogenesis when they formed. Paleosols can besolitary if they formed during a period of landscape stability following the development of an unconformity. Such paleosolsare commonly thick and extremely well developed. More commonly, paleosols are vertically stacked or multistory becausethey formed in sedimentary systems undergoing net aggradation. If erosion was insignificant and sedimentation was rapidand unsteady, compound paleosols generally formed. If the rate of pedogenesis exceeded the rate of deposition, compositepaleosols developed. Thick, cumulative paleosols indicate that erosion was insignificant and that sedimentation wasrelatively steady. Both autogenic and allogenic processes can influence depositional and erosion patterns and, thus, affect thekinds of soils that form. Consequently, paleosols can help to interpret the history of sediment deposition and the autogenicand allogenic processes that influenced a sedimentary basin. Paleosols are also helpful in stratigraphic studies, includingsequence stratigraphic analyses. They are used for stratigraphic correlations at the local and basinal scale, and some workershave calculated sediment accumulation rates based on the degree of paleosol development. In addition to their stratigraphicapplications, paleosols can be used to interpret landscapes of the past by analyzing paleosol–landscape associations atdifferent spatial scales, ranging from local to basin-wide in scope. At the local scale, lateral changes in paleosol propertiesare largely the result of variations in grain size and topography. At the scale of the sedimentary basin, paleosols in differentlocations differ because of basinal variations in topography, grain size, climate, and subsidence rate. Paleosols are used toreconstruct ancient climates, even to estimate ancient mean annual precipitation Ž MAP.and mean annual temperatureŽ MAT .. Ancient climatic conditions can be interpreted from modern soil analogs or by identifying particular pedogenicproperties that modern studies show to have climatic significance. Stable carbon and oxygen isotopes are also used tointerpret ancient climate, and some effort has been made to estimate MAT from isotopic composition. On the basis of) Tel.: q1-303-492-7251; Fax: q1-303-492-2606; E-mail: mary.kraus@colorado.edu0012-8252r99r$ - see front matter q 1999 Elsevier Science B.V. All rights reserved.Ž .PII: S0012-8252 99 00026-4

42( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70modern soil analogs, paleo-precipitation has been estimated from the depth at which calcic horizons originally formed.Finally, paleosol carbonates have been used to estimate ancient atmospheric CO2values. q 1999 Elsevier Science B.V. Allrights reserved.Keywords: paleosols; clastic rocks; paleoclimatology; lithostratigraphy; terrestrial environment1. IntroductionA paleosol or fossil soil is a soil that formed on alandscape of the past. Soils form because of thephysical, biological, and chemical modification ofsediment or rock exposed at the earth surface. Mostpaleosols are found in sedimentary rocks, and, althoughfirst studied in the Quaternary record, theyare now commonly recognized in strata as old asPrecambrian. Soils and paleosols can form becauseof lengthy episodes of landscape stability, in whichcase they may eventually mark a stratigraphic diastemor unconformity; they can also form in terrestrialdepositional systems that are aggrading as longas the rate of sedimentation does not overwhelm therate of pedogenesis. Soils and paleosols thus reflect acomplex interplay among sedimentation, erosion, andnon-deposition.Paleosols have been described from a variety ofcontinental depositional settings including eolianŽ e.g., Soreghan et al., 1997 ., palustrine Že.g., Tandonet al., 1995; Wright and Platt, 1995; Tandon andGibling, 1997 ., and deltaic Že.g., Fastovsky and Mc-Sweeney, 1987; Arndorff, 1993 .. Paleosols are alsofound in marginal marine strata ŽLander et al., 1991;Wright, 1994.and can appear in marine strata ifsea-level fell to expose marine sediment Že.g., Drieseet al., 1994; Webb, 1994 .. Paleosols are most commonlydescribed from the alluvial record; consequently,they provide many of the examples discussedin this paper. To constrain the scope of thisreview article, discussion centers on, but is not limitedto, paleosols that formed in clastic sedimentarysuccessions that are Phanerozoic but pre-Quaternaryin age. Furthermore, because paleosol studies expandedduring the 1990’s in terms of their number,sophistication, and kinds of geologic applications,most of the examples discussed here are taken fromthese recent investigations.The current interest in pre-Quaternary paleosolshas arisen because paleosols have proven remarkablyuseful for examining an array of geological problems,and the major goal of this paper is to examinesome of the exciting areas of geological research thatbenefit from analysis of paleosols. First, depositionis central to the development of sedimentary paleosols,and processes that control the rate and continuityof sediment accumulation influence the degree ofpedogenic development. Consequently, paleosols canhelp to interpret the history of sediment depositionand the allogenic processes that influenced the sedimentarybasin. Paleosols are also used in stratigraphicanalyses, and one of their more recent applicationsis to sequence stratigraphy. Some workershave also used paleosols to calculate short-term sedimentaccumulation rates, both qualitatively andquantitatively.Other geologic applications of paleosols rangefrom landscape reconstruction to paleoclimatic studies.Paleosols provide detailed insight into ancientlandscapes and landscape evolution because the spatialdistribution of different paleosols reflects theparticular landforms on which they formed and thegeomorphic processes operating in the ancient landscape.The paleoclimatic importance of paleosolsstems from studies of Quaternary soil development,which have shown that some pedogenic features canbe quantitatively related to soil-forming factors suchas climate. Paleosols are being used not only tointerpret ancient climatic regimes but, in some cases,to estimate paleo-precipitation and paleo-temperature.Paleosols are also proving helpful in betterunderstanding the composition of the ancient atmosphereand atmospheric changes over geologic time.Finally, paleosols have importance for studying theadvent and evolution of terrestrial plants and animals.Because criteria for recognizing paleosols and fordistinguishing pedogenesis from diagenesis have beendescribed in detail, the reader is referred to thosesources Že.g., Retallack, 1991, 1997b; Wright, 1992a;Pimentel et al., 1996.rather than providing a synopsishere. One of the noteworthy advances of the lastdecade is that paleosols are increasingly recognized

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 43in cores based on various morphological and geochemicalfeatures Že.g., Leckie et al., 1989; Lander etal., 1991; Platt and Keller, 1992; Caudill et al.,1997 .. Paleosols even have been identified on thebasis of wireline log signatures although this approachis difficult Ž Ye, 1995 .. Classifying paleosolshas received far less attention than recognizing them,and this paper begins by discussing the classificationscommonly used for paleosols and the strengthsand weaknesses of these classifications.2. Classifying paleosolsMost paleosol workers use one or more modernclassification systems including United States SoilTaxonomy Ž Soil Survey Staff, 1975, 1998.and theFAO Ž 1974.classification. The classification ofDuchaufour Ž 1982. has been used by some Že.g.,Besly and Fielding, 1989; Kraus, 1997 .. One classification,that of Mack et al. Ž 1993 ., is specific topaleosols, although it is based on modern soil classifications.The U.S. classification and the FAO classificationare taxonomic systems that use profile characteristicsto classify the soils. The U.S. system relies ondiagnostic horizons that are identified on the basis ofproperties such as texture, color, amount of organicmatter, presence of particular minerals, cation exchangecapacity, and pH. The FAO classificationresembles the U.S. system in using diagnostic horizons;however, as Duchaufour Ž 1982.pointed out,the FAO system is less complicated. Furthermore, itrecognizes hydromorphic soils as a major group, theGleysols, whereas soil saturation is only consideredat the sub-order level in the U.S. system Ž Table 1 ..This difference is important with paleosol classificationbecause floodplain paleosols are common and,like their modern counterparts, many are hydromorphicŽe.g., Fastovsky and McSweeney, 1987; Beslyand Fielding, 1989; Arndorff, 1993; Kraus and Aslan,1993 .. Duchaufour also argued that the U.S. andFAO systems suffer by not considering the soilenvironment more heavily. He proposed an environmentalclassification in which soil properties areconsidered in terms of the particular processes of soilformation that operate under particular environmentalconditions. Although this classification uses thediagnostic horizons found in the U.S. and FAOTable 1Comparison of the Mack et al. Ž 1993.classification of paleosolsand Soil Taxonomy Ž Soil Survey Staff, 1975, 1998 .. Some groupsin the Mack classification have equivalents in Soil Taxonomy;others, such as Calcisols and Gypisols, do not. The DuchaufourŽ 1982.classification is also shown but not compared to the othertwoMack et al. Soil survey DuchaufourstaffProtosol Entisol I. Slightly developed soilsInceptisolVertisol Vertisol II. Desaturated humic soilsHistosol Histosol III. Calcimagnesian soilsGleysol not a great IV. Isohumic Ž steppe.soilsorderExcluded Andisol V. VertisolsOxisol Oxisol VI. Brunified soilsSpodosol Spodosol VII. Podzolised soilsArgillisol Alfisol VIII. Hydromorphic soilsUltisolCalcisol no direct IX. Ferisiallitic soilsequivalentGypisol no direct X. Ferruginous soilsequivalentNo direct Aridisol XI. Ferrallitic soilsequivalentExcluded Mollisol XII. Salsodic soilsNo direct Gelisolsequivalentsystems, it differs from those in that it emphasizesthat classification cannot rely on individual horizons.Rather, all the horizons in a particular paleosol aregenetically related and must be used together inclassification. The Duchaufour classification is interpretativein that the attributes of the soil horizons areused to interpret the processes and environmentalconditions of soil formation. The processes are thenused to classify the soil. For example, his Division IIsoils are characterized by the formation of sesquioxides.Within this division, the kind of weathering andthe degree of weathering serve to distinguish threeclasses: Fersiallitic soils, Ferruginuous soils, andFerrallitic soils Ž Table 1 ..One drawback to using the U.S. or FAO systemsis their dependence on soil properties, such as cationexchange capacity or amount of organic matter, thatare not preserved in paleosols. Modern classificationsalso rely on knowledge of the climatic conditionsunder which a soil formed. Primarily becauseŽ .of these problems, Mack et al. 1993 proposed a

44( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70classification just for paleosols. This system relies onthe presence of stable minerals and morphologicalproperties that tend to be preserved as a soil istransformed to a paleosol. Some of the major soilgroups in the Mack classification are identical tothose in U.S. classification Ž Table 1 .. Some categoriesin the U.S. scheme were excluded because ofdifficulties in recognizing them in the ancient record.For example Aridisols were excluded because thistaxonomic class relies on knowledge of the soilmoisture regime. Several new categories were proposed,some of which have approximate equivalentsin the U.S. scheme Ž e.g., Protosol .. Others, such asCalcisols, which are distinguished on the basis of acalcic horizon, have no direct equivalent. Gelisolswere added to Soil Taxonomy after development ofthe Mack et al. classification ŽSoil Survey Staff,1998 .. These are soils that have permafrost within aparticular depth or have gelic materials, which arethe result of cryopedogenic processes.The Mack et al. Ž 1993.system has been receivedwith some favor because it is designed for fieldidentification of paleosols and because it makes paleosolclassification more objective and simpler and,thus, achievable to a broader group of geologists.Despite these advantages, the system has flaws,which Retallack Ž 1993.outlined. One concern isthat, because of its restriction to paleosols, the use ofthis classification will weaken the communicationbetween soil scientists and paleopedologists. A secondconcern is that, because paleopedologists rely onmodern soil analogs to interpret ancient environmentalconditions, the environmental value of paleosolswill be diminished by a classification that is specificto paleosols.Why classify paleosols? In general, scientificclassifications serve to organize information and tofoster effective communication about a particularsubject. Classifications also provide guidelines forfuture studies in a particular subject by emphasizingwhat factors or properties are important. With paleosols,we want a classification system or systemsthat accomplish these goals. But, because paleosolsare studied in order to help interpret past conditionsand events including climates, depositional conditions,and paleoecological changes, Retallack Ž 1993.is right that the classification must be firmly tied tomodern soil systems. The U.S. and FAO classificationsdo have drawbacks for paleosol classification,primarily the fact that these are taxonomic classificationsthat rely on too many features that are commonlyabsent in the ancient record. In many ways,the Duchaufour Ž 1982.classification is more easilyapplied to paleosols because of its focus on processrather than modern soil properties. Morphologicaland geochemical properties that are preserved in therock record are usually sufficient to determine thesoil forming processes that occurred and then toclassify the paleosol. Another advantage of theDuchaufour classification is that it recognizes soilintergrades, which are soils that occur between twoclasses. For many paleopedologists, the major deterrentto using the Duchaufour classification is probablythat many of the soil categories are so dissimilarfrom the commonly used U.S. categories and thenames themselves are so different Ž Table 1 .. Yet, if aresearcher can get beyond the unfamiliar terminology,the process-oriented approach of this systemcan be effective for classifying paleosols.3. Paleosols and timeThe kind of paleosol that forms in the sedimentaryrecord depends on how rapidly the sedimentaccumulated, whether that accumulation was steadyor discontinuous, and, if pauses occurred, their duration.Sediment accumulation varies through time,producing different kinds of paleosols upwardthrough a vertical succession Ž Fig. 1 .. Sedimentarypaleosols form a continuum, at one end of which aremultiple paleosols, which formed in relatively thickand conformable stratigraphic successions becauseaggradation was relatively continuous Ž Fig. 1B .. Butvarious autogenic processes Žthose that are inherentto the depositional system.and allogenic processesŽ those that are external to the depositional system.produce episodes of landscape stability or erosion.Depending on the particular process and the timescaleover which it operated, stratigraphic gaps ofvarious magnitudes Že.g., diastems and unconformities.will develop in the stratigraphic succession.Many of these surfaces will be marked by thick andvery strongly developed paleosols, which are theother end member in the paleosol continuum ŽFig.1A .. The following sections provide a more detailed

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 45Fig. 1. Schematic diagram showing the range of paleosols that can form in a thick vertical section depending on whether sedimentaccumulation was steady or discontinuous, and, if pauses occurred, their duration. Ž A.A thick and strongly weathered paleosol formed on anunconformable surface because of a lengthy period of landscape stability and soil development. Ž B.A thick sequence of multiple paleosolsformed on floodplain deposits because erosion was insignificant and sedimentation was steady. Ž C.A moderately long pause insedimentation related to valley incision produced a paleosol that is more strongly developed than the multiple paleosols but not as intenselyweathered as the paleosol at the unconformity. This paleosol has partly overlapped one of the underlying multiple paleosols.description of paleosols of different temporal magnitudesand the processes by which they form.3.1. Paleosols in aggradational systemsPaleosols can be classified according to the balancebetween sediment accumulation and the rate ofpedogenesis Že.g., Morrison, 1978; Marriott andWright, 1993; Wright and Marriott, 1996.Ž Fig. 2 .. Iferosion is insignificant and sedimentation is rapidand unsteady, compound paleosols usually form ŽFig.2A .. These are weakly developed, vertically stackedprofiles that are separated by minimally weatheredsediment. If the rate of pedogenesis exceeds the rateof deposition, vertically successive profiles maypartly overlap, giving rise to composite paleosols. Incontrast, if erosion is insignificant and sedimentationis steady, thick cumulative soils can form Ž Fig. 2B ..

46( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Fig. 2. Vertical profiles of sediments and soils Ž paleosols. reflecting varying rates of pedogenesis and sedimentation for Ž A.non-steady andŽ B. steady depositional conditions. In Ž C ., sedimentation was interrupted by a period of erosion. Compound paleosols are likely whensedimentation is non-steady. Weakly developed cumulative profiles form when sedimentation is steady but rapid; better developedcumulative profiles form when sedimentation rates are slow relative to rates of pedogenesis. With erosion, a scour surface either separatestwo distinct paleosols or is incorporated into the paleosol. See text for more details. Agsgleyed A horizon; Bgsgleyed B horizon;BwsB horizon showing color or structure development but little if any illuvial accumulation; BtsB horizon showing accumulation ofclays; Btjsincipient development of a Bt horizon; Cgsgleyed C horizon Žafter Morrison, 1978; Bown and Kraus, 1981; Marriott andWright, 1993 ..

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 47These profiles reflect the deposition of successive,thin increments of sediment accompanied by pedogenesis.Using fluvial systems as an example, channeldeposits may show little evidence of pedogenesis ormay contain compound soils because sedimentationis so rapid Že.g., Marriott and Wright, 1993; Krausand Aslan, 1999 .. As the channel migrates laterallyover time, composite or well-expressed soils with Bthorizons may form in the upper deposits. Overbankdeposition is usually steady but slow, commonly onthe order of 1–10 mmryear Že.g., Walling et al.,1992 ., and overbank deposits tend to thin and decreasein grain size away from the active channelŽ Guccione, 1993; Marriott, 1996 .. As a result, compoundsoils with weakly expressed profiles tend toform on natural levees Ž Fig. 3A ., whereas cumulativeprofiles develop in floodbasin areas distal to theactive channel Ž Fig. 3B .. Compound soils can alsoform in crevasse deposits because of rapid and unsteadydeposition Ž Fig. 3C ..Autogenic and allogenic processes that operateŽ3 4over intermediate time-scales 10 to 10 years.canproduce relatively lengthy pauses in deposition aswell as local erosion, both of which can influencesoil development. An example of the effects of anautogenic mechanism is provided by avulsion in afluvial system. By causing channels to shift theirpositions on the floodplain, avulsion can terminatesedimentation on a particular part of the floodplainfor periods of time that are probably on the order of310 years, which is the periodicity of avulsion Že.g.,Bridge and Leeder, 1979 .. A well-developed paleosolcan form in floodplain areas cut off from sedimentationby this autogenic process. In the PaleogeneWillwood Formation, episodic avulsions haveproduced stratigraphic intervals that are meters thickand that have two kinds of paleosols: Ž. 1 relativelywell developed cumulative paleosols, which formedon floodbasin deposits and Ž. 2 very weakly developedcompound or cumulative paleosols, whichformed on the sediment deposited by the avulsion ofthe trunk channel ŽKraus and Aslan, 1993; Kraus,1996; Kraus and Gwinn, 1997.Ž Fig. 3B .. Verticalpaleosol sequences show alternations of the weaklydeveloped and more strongly developed paleosols,and no allogenic controls need be invoked to generatesuch an alternation.Turning to extrinsic mechanisms, climatic changescan also influence soil development at an intermediatetime scale by initiating and terminating sedimentation.Many Quaternary loess successions show avertical alternation between pedogenically-unmodifiedsediment and relatively well developed paleosolsŽe.g., Kukla and An, 1989; Frakes and Sun,1994; Pecsi, 1995 .. In fact, these alternations providea loess-paleosol stratigraphy Ž e.g., Pecsi, 1995 .. Thepaleosols are believed to have formed during periodsof reduced sediment input associated with more humidclimates Ž Fig. 4 .. During drier times, sedimentinput overwhelmed pedogenesis, and no soil formed.Although few pre-Quaternary examples of loessiteand loess-paleosols have been described, an exceptionis provided by Soreghan et al. Ž 1997.whoobserved alternations between paleosols and unmodifiedloessite horizons in the Late Paleozoic MaroonFormation. In this example, the paleosols show significantlyhigher magnetic susceptibilities than dothe loessite horizons, which aids in their recognition.As with the Quaternary cases, the paleosols werelinked to more humid climatic conditions and theunaltered loessite layers to drier climates Ž Fig. 4 ..Although some processes merely halt sedimentation,other intermediate-scale mechanisms cause incision.Mechanisms like climate change can producetruncated soil profiles due to erosion of the upperpart of a developing soil Že.g., Marriott and Wright,1993.Ž Fig. 2C .. In the sedimentary record, episodesof truncation are eventually followed by renewedsedimentation. If that sedimentation is slow andsteady, a cumulative-truncated profile can result;however, if sedimentation is so rapid that the truncatedsoil becomes buried, a compound-truncated orcomposite-truncated set can develop ŽMarriott andWright, 1993 .. In the former, the truncated paleosolis separated by unweathered sediment from an overlyingpaleosol; in the latter, the older, truncatedpaleosol and the younger paleosol partially overlap.In a study of upper Silurian and lower Devonianpaleosols, Marriott and Wright showed how thesedifferent kinds of paleosols can be used to understandthe complex processes that operate in ancientfluvial systems. They compared and contrasted twostratigraphic sections, one dominated by cumulativepaleosols and one with numerous truncated paleosols.They inferred that the first section had a stable

48( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 49floodplain that underwent slow but relatively continuousdeposition. The other section had an unstablefloodplain subject to periods of erosion, which mighthave been triggered by changes in climatic conditionsor vegetation cover.In fluvial systems, climatic change can also causefloodplain incision and terrace development. Althoughcommon in Quaternary fluvial systems, pre-Quaternary incised floodplains and terraces are ofteninferred, rather than directly observed Že.g., Wright,1992b .. This is presumably a problem of scale.Many outcrops are probably too laterally restrictedto allow large cut-and-fill features to be recognized.Paleosols may be especially useful for recognizingancient episodes of incision and terracingŽ Figs. 1C and 3D .. For example, in a study ofEocene–Oligocene alluvial paleosols, BestlandŽ 1997.recognized several episodes of terrace formation,because each is marked by a strongly developedpaleosol indicating lengthy pauses in sedimentation.As discussed in a later section, paleosols are also ofvalue for recognizing incised valleys and terracesthat resulted from eustatic changes.Finally, Retallack Ž 1998.has recognized sawtoothpatterns of development in sequences of paleosols.The paleosols in the sequence show a gradual upwardincrease in their degree of development until amajor decline in development occurs. He provides anexcellent discussion of the mechanisms that couldproduce this kind of pattern including: tectonic activity,sea-level fluctuations, and climatic changes.3.2. Paleosols associated with unconformitiesUnconformities usually represent significant hiatusesthat may last millions or even tens of millionsof years. They form when a period of landscapedegradation andror an episode of landscape stabilityis followed by sediment deposition. At least someperiod of landscape stability is needed for a soil todevelop on the unconformable surface. In some examples,paleosols marking regional unconformitiesare exceptionally thick and well developed, indicatinglengthy periods of soil development and landscapestability Ž Fig. 1A .. For example, a PaleogeneOxisol described by Abbott et al. Ž 1976.is 30 mthick and this profile thickness was regarded as aminimum because the upper contact is an erosionalsurface.The formation of unconformities is controlled byallogenic factors such as sea level fluctuations, globalor regional climate change, and regional tectonics,processes that influence geomorphic systems over5 7time intervals of 10 –10 years Že.g., Summerfield,1991 .. For example, the thick Paleogene paleosoldescribed above formed on a tectonically-generatedunconformable surface Ž Abbott et al., 1976 ., andDriese et al. Ž 1994. and Webb Ž 1994.documentedpaleosols that mark unconformities developed onlimestones exposed during regressions in the LatePaleozoic. Although Driese et al. found that thesubsequent transgression partially eroded the soilthat developed, enough was preserved to distinguishthe unconformity. As described in a later section,paleosols are being increasingly used to identifysequence boundaries and to subdivide a particularstratigraphic succession into sequences.Unconformities are commonly regional in scale,and they can be highly irregular surfaces along whichthe amount of missing time varies considerably Že.g.,Wheeler, 1958 .. Consequently, the paleosol associatedwith an unconformity can show lateral changeson a regional scale, and those changes can be used tointerpret lateral variations in topography and missingFig. 3. Ž A.Levee deposits with weakly developed paleosols. Although the rock shows reddening and root traces, relict bedding is apparentand indicates that sedimentation was so rapid that the parent material did not undergo mixing or homogenization. Lens cap at top of photo is6 cm in diameter. Ž B. Alternations of strongly developed cumulative paleosols Ž S. and weakly developed compound paleosols Ž W .. Thecumulative paleosols are distinguished by gray A horizons overlying thick red B horizons; they formed on floodbasin deposits. The weakpaleosols have pale colors and formed on what Kraus Ž 1996.interpreted to be avulsion deposits. Ridge containing the paleosols is 18 mhigh. Ž C.Two vertically stacked compound paleosols formed on what are interpreted as distal crevasse splay deposits. The paleosols showreddening but are separated by rock that shows little pedogenic modification and is represented by white bands. The upper compoundpaleosol Ž orange and underlying white band. is 47 cm thick and the lower is 50 cm thick. Ž D.Floodplain incision. Well-developed redpaleosols have been scoured and the scour filled with more weakly developed grayish paleosols. The scour cuts approximately 10 m downinto the red paleosols. All examples from the lower Eocene Willwood Formation, Bighorn Basin, Wyoming.

50( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Fig. 4. Schematic diagram showing succession of steps that produce alternations of unmodified loess and paleosols in thick sequences ofloessite. Formation of the paleosols has been attributed to reduced sediment input related to more humid climates. During drier periods,sediment input is sufficient to inhibit pedogenic development.time along the ancient landscape. A good example,described by Joeckel Ž 1995 ., is a Pennsylvanianpaleosol that changes laterally from a thick andwell-developed profile to a thinner profile withweaker pedogenic development. Joeckel used thepaleosol changes to understand topographic variationsand differences in exposure time across thePennsylvanian landscape.3.3. Stratigraphic analysisRecognizing that many individual paleosols aremorphologically distinctive and are really extensive,Quaternary geologists have long used them as stratigraphicmarkers for subdividing Quaternary depositsand for local and regional correlations Že.g., Rich-mond, 1962; Morrison, 1967; Pecsi, 1995 .. Althoughgeologists working in Tertiary and older rocks alsohave a relatively long record of using paleosols tosolve stratigraphic relationships, prior to about 1980,the focus was on the stratigraphic significance ofstrongly developed paleosols that marked unconformitiesŽe.g., Ritzma, 1955; Schultz et al., 1955;Abbott et al., 1976 .. More recent studies demonstratethe utility of paleosols for subdividing thick continentalsuccessions into genetic sequences and solvinglocal and even global correlation problems. Theyalso show that pedogenic carbonates are a potentiallyimportant new means of absolute dating continentalrocks and that floodplain paleosols can be a usefulcomponent of sequence stratigraphic models and fieldstudies.

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 51The Bozeman Group is a thick continental successionin MT, USA, which spans the Cenozoic. Hannemanand Wideman Ž 1991.and Hanneman et al.Ž 1994.showed that problems with the stratigraphicrelations of this group could be clarified using calcretesto delimit basin-wide unconformities and thusto subdivide the succession into five sequences. Ofparticular interest is that Hanneman et al. Ž 1994.integrated surface geologic, seismic, and well logdata and showed that the calcretes produced brightseismic reflections that could be traced in the subsurface.This study not only highlights the potential ofpaleosols for solving stratigraphic problems, but alsoshows that, in some cases, paleosols can be recognizedin the subsurface using data other than coredata and the paleosols can then be used for regionalcorrelations. Similarly, Ye Ž 1995.used paleosols tocorrelate Miocene strata in the subsurface and toconstrain the position of a petroleum reservoir. Thepaleosols were identified in cores primarily on thebasis of color, the presence of roots and calcretenodules, and mineralogy. Although Ye was also ableto use log characteristics to distinguish paleosolsfrom rock that was not pedogenically modified, incontrast to Hanneman et al., he found it very difficultto identify, and thus correlate, individual paleosolsfrom log characteristics alone.Going beyond local or regional correlation usingpaleosols, a study by Koch et al. Ž 1992.showed thatsome paleosols have the potential to make globalcorrelations. A sharp decrease in the 13 Cr 12 C ratiosof biogenic marine carbonates has been well documentedat approximately the PaleocenerEoceneboundary and has been attributed to global marinewarming Že.g., Shackleton, 1986; Zachos et al.,1993 .. Koch et al. were able to use pedogenic carbonatesto identify this sharp isotopic change inPaleogene alluvial rocks in the Bighorn Basin ofWyoming and, thus, to correlate the continentalrecord to the marine isotopic record. Through thiscorrelation, they showed that the PrE boundary inthe marine record corresponds to a particular biostratigraphicboundary in the Bighorn Basin and werethen able to relate changes in mammalian faunas atthis boundary to global climate change.As anyone working in the stratigraphic record isaware, direct absolute dating of the sedimentaryrecord is difficult. Yet, recent work indicates thatUrPb dating of pedogenic carbonates offers a newapproach to obtaining absolute dates for continentalrocks Ž Rasbury et al., 1998 .. Those authors estimatedthe U–Pb ages of pedogenic carbonates from Pennsylvanian–Permianstrata. Because they argued thatthose strata were deposited rapidly and show noevidence of later diagenetic modification, they concludedthat the time of formation of the calcretes is aproxy for the time of deposition of the strata thatenclose them. This technique is potentially powerfulbecause pedogenic carbonates are found in numerousstratigraphic successions throughout the Phanerozoic.Finally, various authors have emphasized the sequencestratigraphic importance of paleosols, bothfrom a modeling approach Že.g., Wright and Marriott,1993.and, more commonly, through field studiesŽe.g., Aitken and Flint, 1996; McCarthy andPlint, 1998 .. Wright and Marriott Ž 1993.developed amodel that predicts how the degree of pedogenicmaturity and soil drainage of coastal plain paleosolsvaries at different times in the sea level cycle. Theirmodel suggests that, during sea-level lowstands, matureand well-drained soils should form on terracesthat are produced by channel incision because thewater table drops and the floodplains receive nosediment Ž Fig. 5 .. When sea level first begins to rise,hydromorphic soils develop because baselevel is rising.As sea level continues to rise, accommodationspace is created and floodplain sedimentation is rapid,leading to weakly developed paleosols. Later, duringearly highstand, the rate of aggradation begins todecrease and more strongly developed soils shouldform. Finally, when highstand is fully achieved,accommodation space is low. Mature soils form;however, their preservation potential is probably lowbecause aggradation rates are so low that sedimentreworking by channel combing is intense.A key feature of the Wright and Marriott model isthat, contrary to other models of sequence stratigraphy,it predicts that the highest rates of aggradationare characteristic of the transgressive systems tractŽ TST ., and that, because of low accommodationspace, the highstand systems tract Ž HST.had lowrates of aggradation. Consequently, abundant overbankdeposits with weakly developed paleosolsshould characterize the TST, whereas lower amountsof overbank deposits with well-developed paleosolsare typical of the HST. In contrast some field studies

52( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Fig. 5. Schematic diagram showing Wright and Marriott Ž 1993. model of pedogenic development related to the sea-level cycle. Ž A.Thelowstand systems tract Ž LST.is characterized by channel incision and strongly developed and well-drained paleosols that form on terraces.Ž B.Early in the TST, hydromorphic paleosols may form because base level is rising; channel sandstones may overlap becauseaccommodation space is still low. Ž C.Increased accommodation space in the TST produces rapid sediment accumulation and weaklydeveloped soils. Ž D.The HST is characterized by low accommodation space, thus, well-developed paleosols Žmodified from Wright andMarriott, 1993 ..Že.g., figure supplied by M. Uliana and L. Legarreta,cited in Shanley and McCabe, 1994.have shown thatthe HST, rather than the TST, is dominated byoverbank deposits and that paleosols are most commonin the HST. Shanley and McCabe Ž1994, p.562 ., suggested that this discrepancy may have arisenbecause the Wright and Marriott model ‘‘takes noaccount of the basinal position of systems tracts northe importance of recognition of the landward equivalentof marine condensed intervals’’. Nonetheless,the model shows that paleosols are an important areaof future research in sequence stratigraphy.Sequence stratigraphic field studies generallystress the importance of paleosols for identifyinginterfluvial sequence boundaries and for distinguishingincised valley fills from major channel sandstones.For example, Aitken and Flint Ž 1996.foundpaleosols to be critical for identifying interfluvialsequence boundaries in Pennsylvanian rocks, because,beyond the incised valley fills, the boundariesare subtle features. The paleosols, which developedon delta plain deposits, are typically hydromorphicor gleyed paleosols; however, the presence of compositepaleosols indicates that many of the soils wereinitially well drained when water tables were low,and then they underwent a secondary gleyed stagewhen the water table rose in response to transgression.Despite their potential value in identifying sequenceboundaries, paleosols like these probably havelow preservation potential because the ensuing transgressioncan obliterate them Ž Aitken and Flint, 1996 ..Interfluvial sequence boundaries with associated paleosolshave also been described from entirely fluvialsuccessions Ž e.g., Leckie et al., 1989 .. The presenceof numerous, closely spaced paleosols in the Cretaceousstrata suggested to Leckie et al. that sediment

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 53supply to the floodplain was cut off and they attributedthis to channel incision caused by loweredbase level.Despite the progress using paleosols to recognizeinterfluvial sequence boundaries, McCarthy and PlintŽ 1998.noted that more attention should be paid todetails of the paleosols. In particular, those authorsshowed how examining the micromorphologic attributesof the paleosols in conjunction with fieldmorphology and stratigraphic relationships leads to amore complete understanding of the sequence ofevents that produced paleosols at a sequence boundary.In a study of Cretaceous paleosols that formedon coastal plain deposits, they found that grey soilcolors and the presence of siderite and preservedorganic matter indicated that the soils first formedunder poorly drained conditions Ž Fig. 6A .. The presenceof clay coatings and various ferruginous featuressuggest that drainage then improved, andpapules Ž fragmented clay coatings.and pedorelictsindicate that the interfluves underwent an episode oferosion Ž Fig. 6B .. They attributed these features tolowered base level, which caused the channels toincise and water tables to drop. Silty layers directlyabove the sequence boundary within the paleosolsreflect renewed sedimentation on the interfluves becausebase level rose again. Finally, wet soil indicators,present above the sequence boundary, show thatwater tables also rose, although multiple organic-richbeds indicate that water tables fluctuated somewhatŽ Fig. 6C ..Although sea level fluctuations are an importantcontrol on coastal plain rivers and the developmentand morphology of paleosols associated with thoserivers, the current opinion is that sea-level effectsprobably do not extend inland more than about 100–150 km from the shoreline Že.g., Shanley and Mc-Cabe, 1994; North, 1996 .. In alluvial systems moredistant from the sea or in closed basins, climate andregional tectonic activity are the major controls onFig. 6. Schematic diagram showing sequential development of an interfluve paleosol that marks a sequence boundary. Ž A.Siderite and greysoil colors show that initial soil drainage was poor. Ž B.Base level dropped, the floodplain was incised, and the water table dropped. Claycoatings and ferruginous nodules formed in the better-drained soil. Ž C.Base level rose and new sediment was deposited. The water tablealso rose, producing hydromorphic features including multiple organic horizons. See text for more details Žmodified from McCarthy andPlint, 1998 ..

54( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Ž .rivers e.g., Blum, 1994; Blum and Valastro, 1994 ,and paleosol development reflects those effects ratherthan sealevel fluctuations.3.4. Sediment accumulation ratesSediment accumulation rates are generally calculatedby dividing the thickness of a particular stratigraphicsection by its known or estimated time span.Sediment accumulation rates estimated for continentalstratigraphic successions are usually severelytime-averaged because radiometric dates are notavailable or widely spaced in time and because thetemporal resolution of paleomagnetic dating or continentalbiostratigraphy is generally relatively coarseŽ e.g., Kraus and Bown, 1993 .. Because the developmentalhistory of a sedimentary paleosol reflects therate of pedogenesis relative to the total time of soildevelopment and how steady or unsteady that depositionwas Ž Fig. 2 ., the kind of paleosol that developsis a good indicator of sediment accumulation ratesfor thin stratigraphic intervals. The relative degree ofpedogenic development is used as a proxy for therelative rate of sediment accumulation, and, as discussedin a later section, paleosols have been used tocompare and contrast sediment accumulation rates,in a qualitative sense, in different parts of a depositionalbasin Že.g., Atkinson, 1986; Platt and Keller,1992. and through time Že.g., Kraus and Aslan,1993; Kraus, 1997; Soreghan et al., 1997; examplesin Kraus and Aslan, 1999 ..In addition to this qualitative approach, severalworkers have attempted to use paleosols to quantitativelyestimate short-term sediment accumulationrates, using somewhat different approaches. In aseminal paper on this topic, Leeder Ž 1975.compiledages of Quaternary calcretes to establish ages offormation for particular stages of calcrete development.The ages were then incorporated into a modelfor estimating floodplain accretion rates. RetallackŽ 1983, 1984.also took a quantitative approach toaccumulation rates by estimating development timesfor different kinds of paleosols through analogy tomodern soils.Bown and Kraus Ž 1993.and Kraus and BownŽ 1993.took a different approach. On the basis ofmorphologic criteria including the combined thicknessof the A and B horizons, geochemical trendswithin a paleosol profile, and degree of clay translocation,those workers assigned floodplain paleosolsin the Paleogene Willwood Formation to seven stagesof pedogenic development. Bown and Kraus Ž 1993.constructed a composite stratigraphic section extendingfrom the base to the 600-m level of the formationfrom numerous measured sections and then subdividedthe composite section into 25-m-thick intervals.The paleosols found within a particular 25-minterval were assigned to a stage of development andthen averaged to yield a relative degree of pedogenicmaturity for each interval Ž Fig. 7A .. The relativeamount of time represented by each interval wasthen estimated from the relative degree of develop-Fig. 7. Ž A.Plot of relative degree of pedogenesis vs. meter levelin composite section of the lower Tertiary Willwood Formation.Relative degree of pedogenic development was calculated for25-m-thick intervals by assigning paleosols in each interval to astage of pedogenic development outlined by Bown and KrausŽ 1993 .. Ž B.That information was combined with the absolute ageof the 600-m-thick section to estimate sediment accumulationrates for each 25-m-thick interval. See text for more detailsŽ modified from Bown and Kraus, 1993; Kraus and Bown, 1993 ..

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 55ment for each interval. For example, the lowest 25-minterval has the highest relative degree of maturity at12.0 and occupied 9% of the total time representedby the total section Ž Fig. 7A .. Because absolute ageshad been established for the lower and upper boundariesof the 600-m-thick section, the time span foreach 25-m interval could be calculated. From thatcalculation, the sediment accumulation rate for each25-m-thick interval was estimated Ž Fig. 7B .. Krausand Bown Ž 1993.then used these short-term accumulationrates to examine up-section changes insedimentology and paleontology.Using paleosols to estimate sediment accumulationrates in a quantitative fashion suffers from severallimitations. The Leeder Ž 1975.and RetallackŽ 1983, 1984.approach assumes that times of formationof modern soil analogs can be reliably determined.The pitfalls of this assumption for calcreteshave been thoroughly discussed by Wright Ž 1990 ..He concluded that this approach is not justifiedbecause Quaternary calcretes actually show considerablevariability in their rates of formation dependingon local rainfall and the availability of airbornecarbonate dust.A second potential problem is erosion or slowsedimentation rates, which can generate compositepaleosols. With erosion, accumulation rates can beunderestimated. If the paleosols are composite, asthey are in numerous stratigraphic successions, accumulationrates can be overestimated. Finally, RetallackŽ 1998.has pointed out that studies like those ofBown and Kraus Ž 1993. and Kraus and Bown Ž 1993.assume that there are no major discontinuities in thepaleosol section and that all of the time contained ina vertical section is represented by paleosols. Henoted that this is probably an unrealistic assumptionfor stratigraphic sections that span several millionsof years.4. Paleolandscape reconstructionIn the Quaternary record, soil–landscape studiesare important for interpreting landscape evolutionŽ e.g., McFadden and Knuepfer, 1990 .. Similarly, inpre-Quaternary strata, paleosol–landscape studies canhelp in reconstructing ancient landscapes Že.g., Bownand Kraus, 1987; Besly and Fielding, 1989; Platt andKeller, 1992 .. Paleosolrlandscape studies focus onthe spatial distribution of different kinds of paleosolsand the different landscape elements and processesthat produced those paleosols. The study of JoeckelŽ 1995 ., described above, is a good example of usinglateral changes in a paleosol associated with anunconformity to analyze regional topographic variationsalong a specific degradational landscape. Furthermore,thick sedimentary sequences with multiplepaleosols provide a record of land surfaces over timeso that landscape changes and evolution can beexamined.Paleosol–landscape associations can be studied atdifferent spatial scales that range from local Že.g., achannel and associated floodplain, an eolian duneand associated interdune.to basin-wide in scopeŽ e.g., a drainage system, a sand sea .. At the localscale, lateral changes in paleosol properties arelargely the result of local variations in grain size andtopography. At the scale of the sedimentary basin,paleosols in different locations differ because ofbasinal variations in topography, grain size, climate,and subsidence rate. The following discussion providesexamples of how paleosols have been used tointerpret landscapes of the past, starting at the localscale and building to the basin-wide scale. For amore thorough discussion of paleosol–landscape associationsat different spatial and temporal scales,Ž .see Kraus and Aslan 1999 .4.1. Local landscape reconstructionPaleosol–landscape associations at this scale havebeen studied primarily in alluvial rocks Že.g., Bownand Kraus, 1987; Kraus, 1987, 1997; Besly andFielding, 1989; Platt and Keller, 1992 .. Catenas andpedofacies are common paleosol–landscape associationsthat develop at the scale of the channel andassociated floodplain. The two are not mutually exclusive,and floodplain paleosols can show a combinationof the two.In a catena, better-drained soils form on channelmarginaldeposits Ž levees, crevasse splays.becausethe deposits are elevated relative to the surroundingfloodbasin and usually consist of relatively permeablefine sands and silts Ž Fig. 8 .. The soils may showevidence of oxidized conditions, including brown Aand Bw Ž weathered or structural B.horizons and

56( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Fig. 8. Schematic diagram showing a paleocatena, which developed because of changes in grain size and topography away from the ancientchannel. A moderately or well-drained soil generally forms on the levee, although subsurface horizons can be gleyed Ž Bg horizon.becauseof proximity to the ground water table. Poorly to very poorly drained soils are more typical of the floodbasin Žmodified from Kraus andAslan, 1999 ..sparse organic matter. Grey soil colors are moreabundant lower in the profile ŽBg and Cg horizons,which are gleyed B and C horizons.because it iscloser to the ground water table and more prone toreduced conditions Ž e.g., Duchaufour, 1982 .. Awayfrom the channel, the soils become progressivelymore poorly drained because the topographic positionis lower and the sediment is finer-grained andless permeable. Waterlogged, reduced conditions favorthe accumulation and preservation of organicmatter in an Ag Ž gleyed A.horizon and the developmentof a thick, grey Bg horizon with many redoximorphicfeatures. Fluctuations of the water table alsoproduce intersecting slickensides in the clayey sediments.Ancient catenas have been described from Paleogenefluvial rocks Že.g., Fastovsky and McSweeney,1987. and from Jurassic deltaic plain deposits Že.g.,Arndorff, 1993 .. Those catenary relationships havebeen used to reconstruct the ancient floodplain landscape.Arndorff, for example, found that the distributionof different paleosols reflected their position onthe deltaic plain. Relatively well drained, light brownpaleosols with a yellowish to orange subsurface horizonformed on sandy natural levees and crevassesplays. In contrast, dark-colored silty claystones interpretedas gleyed alluvial soils formed in the backswamps.Pedofacies focus on lateral changes in the degreeof pedogenic development with increased distance

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 57from the ancient channel Ž Bown and Kraus, 1987 ..Bown and Kraus observed weakly developed paleosolson channel-marginal deposits and increasinglywell developed paleosols away from the channelsandstone Ž Fig. 9 .. These changes were primarilyattributed to sediment accumulation rates, which tendto decrease away from an active channel Že.g., Guccione,1993 .. Pedofacies have also been inferredfrom cores on the basis of the degree of paleosoldevelopment and the proportion of sandstone presentŽ Platt and Keller, 1992. Ž Fig. 10 .. Weakly developedpaleosols Ž stage 1 of Platt and Keller.were found incores dominated by meander belt sandstones Ž25% oftotal core thickness.and crevasse-splay sandstonesŽ 30% of core thickness .. Cores where sandstones ofany kind were less common were interpreted to bedistal to major channels. These were dominated bymore strongly developed paleosols Ž stages 2–3 ., aspredicted by the pedofacies model.Although pedofacies relationships have been recognizedin these and other ancient alluvial sequencesŽ e.g., Wright and Robinson, 1988; Smith, 1990 ., thepedofacies model, as it is currently understood, doesnot satisfactorily explain lateral relationships in allfloodplain paleosol successions Ž e.g., Wright, 1992b ..Kraus Ž 1997. and Kraus and Aslan Ž 1999.havediscussed limitations of the pedofacies model. Inaddition, North Ž 1996.pointed out that the pedofaciesmodel may be appropriate only for thick, aggradationalsuccessions that are dominated by compositeand cumulative paleosol profiles.Paleosols in other sedimentary environments alsoshow lateral changes that reflect different locationsin the depositional landscape. Although studies ofpre-Quaternary paleosols associated with eolian environmentsare sparse, Quaternary deposits providenumerous examples. Loess deposits become thinnerand finer-grained away from source areas, and theseFig. 9. Schematic diagram showing pedofacies relationships. Compound paleosols tend to form adjacent to the channel becausesedimentation was rapid and episodic. Farther from the active channel, more strongly developed, cumulative paleosols formed becausesedimentation was slow and steady Ž modified from Bown and Kraus, 1987 ..

58( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Fig. 10. Paleosol–landscape associations in alluvial rocks at both the local and basinal scales. Pedofacies developed over a distance of -2km, with weakly developed paleosols Ž Entisols or Inceptisols. grading into better-developed paleosols Ž Alfisols and Vertisols .. Over adistance of ;75 km, paleosols became even more strongly developed Ž Stage 4.in response to basinal changes in the rate of aggradation.See text for more details Ž modified from Platt and Keller, 1992 ..lateral changes in parent material produce lateralchanges in the cumulative soils that form on theloess Ž e.g., Smith, 1942; Ruhe, 1983 .. For exampleMcDonald and Busacca Ž 1990.found that, with increasingproximity to the sediment source, a singlesoil bifurcated into two well-developed soils. HollidayŽ 1990.also documented lateral variations insoils developed on eolian deposits and attributedthese primarily to variations in the grain size of thesediment, although local thickness variations alsoplayed a role.Catenas and pedofacies are produced by variationsin topography and grain size related to landscapeposition. Modern soil studies show thatcompositional differences are linked to grain sizedifferences and suggest that those compositional differencesalso influence soil development. In fluvialsystems, for example, sands and coarse silts thataccumulate in channel-marginal environmentsŽ levees, splays.tend to be dominated by quartz,feldspar, and lithic fragments Že.g., Schumacher etal., 1988 .. In contrast, clay and fine silt, whichtypically accumulate in distal floodbasins, consistprimarily of clay minerals such as smectite, illite,kaolinite, and chlorite. Aslan and Autin Ž 1998.concludedthat, in Mississippi River floodplains, thesedepositionally controlled compositional differenceshad a greater impact on the chemistry of alluvialsoils than did weathering processes. Similarly, withQuaternary loess deposits, compositional changes,including an increase in clay minerals, are foundwith increasing distance from source areas, Že.g.,

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 59.Smith, 1942; Kleiss and Fehrenbacher, 1973 . Thesestudies indicate that compositional variations associatedwith different parts of the depositional landscapeshould also be considered when analyzing andinterpreting paleosols.4.2. Basinal landscape reconstructionAt the basin scale, analyzing paleosolrlandscapeanalysis can provide information about global orregional climate change, sea level fluctuations, andregional tectonics. At a truly basinal scale, AlonsoZarza et al. Ž 1992.described Miocene paleosols thatdeveloped in a spectrum of depositional settings,including alluvial fans, lake margins, and river floodplains.Stratigraphic intervals approximately 100 mthick were examined in two contrasting areas ofthe Madrid Basin. Paleosols in the two areas differ,and those differences were related to local climaticconditions and sediment accumulation rates. Consequently,the authors could use the paleosols tointerpret basinal variations in climate and basin subsidence,which controlled sediment accumulationrates.In fluvial basins, changes in the kinds of paleosolshave been documented in the direction of regionalpaleoslope. For example, in the Eocene Capella Formation,Atkinson Ž 1986.observed more mature andbetter-drained paleosols proximal to the source area.Over a distance of nearly 30 km down paleoslope,progressively less mature and more poorly drainedpaleosols were found. Atkinson attributed thesechanges to a decline in topographic relief and increasedaccumulation rates away from the source.Over an even greater downslope distance Ž ;75 km .,Platt and Keller Ž 1992.also documented an increasein the maturity and hydromorphy of floodplain paleosolsŽ Fig. 10 .. Consistent with the increased maturity,the stratigraphic interval thins downslope, indicatingslower sediment accumulation in more distalparts of the alluvial basin.McCarthy et al. Ž 1997.recognized floodplain paleosolsof varying degrees of pedogenic developmentand related the paleosols to landscape position on abasinal scale. In this Lower Cretaceous example, thepaleosols formed in a foreland basin adjacent to theepicontinental seaway that developed in North Americaduring that time. Not surprisingly, the authorsasserted that pedogenic development should be directlydependent on floodplain stability and that thetectonic hinge zone is the most stable area of thelandscape. Consequently, they argued that ancienthinge zones can be located by examining regionalpatterns of paleosol development.Finally, a study of early Eocene fluvial rocksshows that basinal variations in grain size, related tothe basin geography, can also produce different kindsof paleosols Ž Kraus and Gwinn, 1997 .. Paleosols inthe northern part of the Bighorn Basin are betterdrained than paleosols 75 km to the south, in thecentral part of the basin. The differences were attributedto grain size differences, and those, in turn,were linked to basin position, in particular, distancefrom a local sediment source. Because the northernarea is directly adjacent to a local sediment source,the sediment is coarser and more permeable, producingbetter-drained paleosols. The southern area ismore distal to any of the sediment sources aroundthe basin, resulting in clay-rich sediment that producedwaterlogged soils.Despite their potential for helping interpret thehistory of alluvial basins, studies of paleosol variabilityat the basin scale are few, in part, because theyare time and labor intensive. Kraus Ž 1992.demonstratedthat these difficulties can be overcome byincluding remote sensing analysis of satellite or airbornespectral data in a field and laboratory study ofpaleosols. Remote sensing is potentially valuablebecause it can provide data not easily obtained onthe ground and gathers some data more efficientlythan field study. Yet, remote sensing is rarely used tosolve basic stratigraphic or sedimentologic problems.Because remote sensing imagery is becoming morereadily available and because data analysis now canbe done on personal computers, this tool should bemore widely used in future paleosol studies, whereexposures are good.5. Interpreting paleoclimatesPaleosols are used both to interpret the paleoclimaticregime and to quantitatively estimate ancientmean annual precipitation Ž MAP.and mean annual

60( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70temperature Ž MAT .. Ancient climatic conditions canbe interpreted by classifying the paleosols and usingmodern analogs to infer the paleoclimatic regime orby identifying particular pedogenic properties thatmodern studies show to have climatic significance.Stable carbon and oxygen isotopes are also used tointerpret ancient climate and some effort has beenmade to use isotopic composition to estimate MATŽ e.g., Mack et al., 1991; Koch et al., 1995 .. Finally,the depth at which calcic horizons originally formedhas been used to estimate paleo-precipitation. Thefollowing sections provide overviews of these approachesas well as their limitations.5.1. Modern soil analogsThe paleosol literature provides numerous examplesof the classification approach to paleoclimaticinterpretation. For example, Mack Ž 1992.used alluvialpaleosols to interpret a climate change across theLower Cretaceous–Upper Cretaceous boundary inNew Mexico. Older red, calcic paleosols, interpretedas Aridisols, were attributed to semi-arid or aridconditions. These are overlain by Inceptisols andAlfisols that have somber colors and lack carbonate,suggesting a subhumid or humid climate. Similarly,Bestland Ž 1997.used paleosols to interpret a climatechange across the Eocene–Oligocene boundary. Theolder paleosols were interpreted as ‘Ultisol-like’ paleosolsthat formed under humid, subtropical conditions,whereas the younger smectitic paleosols werejudged to be ‘Alfisol-like’ and indicative of humid,temperate conditions.In a more unifying approach to paleoclimaticinterpretation, Mack and James Ž 1994.generated apaleosol analog to the Soil Map of the World Že.g.,FAO, 1974 ., which links modern soils to particularclimatic zones. Using the Mack et al. Ž 1993.paleosolclassification Ž Table 1 ., they suggested that highlyweathered paleosols, such as Oxisols and Argillisols,should be characteristic of wet equatorial paleoclimatesin which MAP and MAT are high and havelittle seasonal variation. Argillisols, Spodosols, andGleysols are more apt to form in humid ŽMAP)1000mm.midlatitude climates, whereas Calcisols are indicativeof a dry Ž MAP-1000 mm.subtropicalpaleoclimatic zone. One caveat to this global approachis that it is not appropriate for paleosols thatformed prior to the advent of vascular plants.The second approach to climate reconstruction—focusing on a particular soil property or group ofproperties—is based on Quaternary soil studies,which have shown that particular features, includingclay and carbonate accumulations and the depth ofsoil oxidation can be quantitatively related to soilformingfactors, including climate Že.g., Bockheim,1980; Harden and Taylor, 1983; Birkeland, 1984 ..Molecular weathering ratios Že.g., silicarsesquioxidesor basesralumina.provide a good illustration.In some cases, the ratios indicate intense leaching ofbase cations and the loss of silica, which are characteristicsof modern soils that formed in humid, tropicalclimates Že.g., Arndorff, 1993; Retallack andGerman-Heins, 1994; Gill and Yemane, 1996 .. Claymineralogy has also been used to detect and interpreta climatic change. Robert and Kennett Ž 1994 ., forexample, found that clay minerals on the Antarcticcontinent showed a dramatic increase in smectiteŽ and corresponding decrease in illite.during thelatest Paleocene. The clay change corresponds to awell-documented isotopic shift that marks the latestPaleocene thermal maximum Že.g., Zachos et al.,1993 .. Robert and Kennett concluded that the increasein smectite was the result of increased chemicalweathering as a result of warmer temperaturesand greater rainfall at that time.In some cases, the paleosols are parts of cyclothemsin which alternations of particular kinds ofpaleosols are linked to climatic cycles. Good examplesare provided by the wetrdry cycles linked topedogenically-unmodified sediment and paleosols inloess and loessite. In some, the cyclic climaticchanges are linked to either sea-level changes ortectonic activity. For example, working in LowerPermian cyclothems of the U.S. mid-continent, Milleret al. Ž 1996.described cyclic changes from verticpaleosols to Alfisols with calcic horizons and attributedthese to fluctuations between semi-aridrsubhumidconditions and humid monsoonal conditions.In a subsequent study Ž McCahon and Miller, 1997 .,these paleosolrclimate cycles were linked to glaciallycontrolled sea-level changes. The change to increasedprecipitation followed by a rapid changeback to arid conditions was attributed to gradualregression caused by glacial activity followed by a

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 61Fig. 11. Schematic diagram showing development of paleosol cycles described by Tandon and Gibling Ž 1994 .. Ž A.Hydromorphic paleosolsformed during marine highstand when climates were wet. Ž B. Climates remained humid as bayfill deposits accumulated. Ž C.Calcareouspaleosols formed because of change to dry conditions during lowstand and early transgression.rapid transgression. Tandon and Gibling Ž 1994.describedCarboniferous cyclothems that are 20 to 30m thick and that show an alternation between hydromorphicpaleosols, including Histosols, and calcareouspaleosols. They invoked a climatic change fromhumid to strongly seasonal conditions to explain thepaleosol cycles, and they too concluded that theclimatic cycles were linked to sea-level fluctuationsŽ Fig. 11 .. Relatively humid climates accompanied amarine highstand, and conditions became drier duringthe lowstand and early transgression.5.2. Depth to calcic horizonA particular soil property that has been used toestimate MAP is depth to a calcic horizon in soilscontaining a zone of calcareous nodules. This tech-nique has been developed by Retallack Ž 1994.whoexpanded on earlier efforts by Jenny Ž 1941.to linkMAP Ž mean annual temperature.to depth to thecalcic horizon. Jenny produced a scatter plot of dataobtained from soils on the Great Plains of the US, towhich Retallack added data from calcareous soilsfound in different soil-forming environments all overthe world. The empirical relationship he determinedfrom the modern soils was then applied to calcareouspaleosols of Eocene and Oligocene age to estimatepaleo-precipitation. The advantages to this approachare obvious. Many paleosols contain calcareous nodules,and this approach provides an expedient meansof estimating paleo-precipitation.Nonetheless, as Retallack Ž 1994.indicated, thisapproach has several potential problems. First, depthto the calcic horizon varies with atmospheric CO 2

62( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Ž Cerling, 1991, 1992 .. Cerling Ž 1991.and othersŽ e.g., Driese and Mora, 1993.have shown that levelsof CO2in the atmosphere have varied over time andhave differed significantly from the modern levelsunder which the empirical curve was constructedŽ see Section 6 .. A second potential problem is thatMAP can be underestimated if the upper part of apaleosol has been truncated. Retallack Ž 1994.hassuggested ways to avoid these two problems, and thereader is referred to his discussion.A more contentious problem is the effects ofsediment compaction, especially if the rocks haveundergone significant burial. Retallack Ž 1994.hastended to calculate depth to the carbonate horizonafter decompacting the strata using the compactioncurves of Baldwin and Butler Ž 1985 .. Yet, severalstudies suggest that these curves are not appropriatefor paleosols and, in fact, that many paleosols arelittle affected by compaction. In an innovative study,Caudill et al. Ž 1996.used vertical micro-cracks inPaleozoic Vertisols to show that the paleosols underwentno more than 10% compaction despite thedepth of burial. Caudill et al. concluded that, becauseVertisols have initial bulk densities that are quitehigh, they undergo little compaction with burial. Intheir case, they calculated 650 mm of precipitationon the basis of depth to the carbonate horizon in theVertisols, which is comparable to the precipitationunder which modern Vertisols form. Evidence fromboth modern and ancient floodplain deposits alsoindicates that, because paleosols require surface exposure,they underwent significant dewatering andcompaction prior to burial Ž Nadon and Issler, 1997 ..These studies suggest that depth to the calcic horizoncan be used directly to estimate MAP as long asproblems with truncation and paleoatmospheric CO 2can be reconciled.5.3. Isotopic analysisFor a thorough discussion of the factors controllingthe isotopic composition of soil minerals and theproblems encountered in interpreting isotopic compositions,the reader is referred to Cerling and Quade18Ž 1993 .. In brief, the d O value of soil carbonatedepends on the isotopic composition of the meteoricwaters from which it precipitated. The isotopic compositionof the meteoric waters is, in turn, influencedby the mean annual temperature Ž MAT. ŽCerling,1984; Cerling and Quade, 1993 .. Furthermore, thecarbon isotopic composition of modern soil carbonatestends to be higher in warmer areas because thekind of vegetation Ž C vs. C plants.3 4 is temperaturedependent. C 3 plants, which include most trees,shrubs, and cool-season grasses, produce soil carbonated 13 C values around y27% whereas C 4 plants,which include most of the summer grasses andsedges, produce a value around y12%. On the basisof studies that show a major change from C 3 to C 4vegetation between 8 and 7 Ma ŽCerling, 1992;Cerling et al., 1993; Quade et al., 1994, 1995; Latorreet al., 1997 ., most paleoclimatic interpretationsof pre-Miocene paleosols assume that the vegetationconsisted entirely of C 3 plants. Because of theserelationships between isotopic composition and temperature,paleosol carbonates have been analyzed toreconstruct paleoclimates Že.g., Mack et al., 1991;Koch et al., 1995 .. Other soil minerals, particularlyclay minerals, are also potentially useful Že.g., Sternet al., 1997 ..Reconstructing paleoclimates using the isotopiccomposition of paleosol minerals suffers from severalproblems. First is climatic overprinting, whichcan occur if sediment accumulates very slowly andthe pedogenic carbonate precipitates under more thanone climatic regime Ž Cerling, 1984 .. In those circumstances,the carbonate will contain an isotopicmix. Second, the oxygen isotopic composition of soilminerals is more prone to diagenetic modificationthan is the carbon isotope composition. Studies ofLate Paleozoic pedogenic carbonates by Driese andMora Ž 1993.and Eocene carbonates by CerlingŽ 1991.showed little if any diagenetic alteration ofthe carbon isotopes; however, the d 18 O in both exampleswas depleted as a result of recrystallization.To understand more thoroughly the effects of diagenesison oxygen isotopes, Mora et al. Ž 1998.analyzedthe isotopic compositions of pedogenic calciteand illite in three vertic paleosols that underwentdifferent burial depths and burial temperatures. Theirresults clearly demonstrate that the oxygen isotopesin this example were affected by burial diagenesisand are not appropriate for paleoenvironmental andpaleoclimatic interpretations.Cognizant of the limitations to oxygen isotopes,Mack et al. Ž 1991.used the isotopic composition of

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 63Permian pedogenic carbonates to interpret a climaticchange over the 18 m.y. that the stratigraphic intervalspanned. In particular, Mack et al. modified theapproach of Cerling Ž 1984 ., which was devised forenvironments dominated by C 4 plants, to examinePermian carbonates, which formed in a C 3 plantworld. They found that d 18 O values increase up-section,from which they concluded that MAT increasedfrom 158C to308C. They also concluded that precipitationdecreased as the temperature rose. d 13 C alsoincreased up-section, an isotopic change that theauthors attributed to a decrease in plant productivity,and a change that they suggested is consistent withincreased temperatures and decreased precipitation.An important aspect of this study is that Mack et al.did not analyze the isotopes in isolation; rather theyevaluated the reliability of their isotopic conclusionsin light of paleomagnetic data, which provided thepaleolatitudinal position of the study area, and macroandmicromorphological properties of the paleosols.Koch et al. Ž 1995.developed a different approachthat is based on empirical equations, developed byFriedman and O’Neil Ž 1977 ., that relate soil temperatureto the d 18 O value of pedogenic carbonate andthe d 18 O value of the soil water from which thecarbonate precipitated. The Paleogene WillwoodFormation was used to demonstrate this method. Toestimate the d 18 O values of the Paleogene meteoricwaters, Koch et al. measured the d 18 O values ofapatite from fossil teeth Žwhich are abundant in theWillwood paleosols.and aragonite in fossil freshwaterbivalves, because the oxygen in biogenic mineralscomes primarily from meteoric waters. Then, theycombined the estimated d 18 O values of meteoricwaters with measured d 18 O values of pedogeniccarbonates from the same stratigraphic levels to determineancient soil temperatures. Because BradyŽ 1990.showed that, at depths more than about 30 cmin the soil profile, air temperatures are similar to soiltemperatures, the calculated soil temperatures wereused to estimate MAT. The MAT values rangedbetween about 08C and 358C, temperatures that arereasonable when compared to MAT values estimatedfor the Willwood Formation on the basis of plantfossils. The authors, however, acknowledged that theresults were mixed and that this technique suffersfrom significant problems, which they discussed indetail. Because the Willwood Formation accumu-lated rapidly Ž e.g., Kraus and Bown, 1993 ., climaticoverprinting should not have been a problem. Moreproblematic with the Willwood paleosols is diageneticoverprinting because, as mentioned above, CerlingŽ 1991.found the Willwood carbonates to bedepleted in d 18 O due to recrystallization.Because pedogenic carbonate are not found in allsoils, Stern et al. Ž 1997.explored the possibility ofreconstructing paleoclimates using d 18 O values fromclay minerals, which, like carbonates, have beenshown to reflect the isotopic composition of meteoricwaters Ž see Refs. in Stern et al., 1997 .. They measuredthe isotopic compositions of smectite andkaolinite from strata deposited between ;12 Maand 2 Ma as part of the Siwalik succession inPakistan. When plotted against time, d 18 O values forsmectite showed similar trends to those obtainedfrom pedogenic carbonates in the same strata byQuade et al. Ž 1989 .. Both data sets show increases ind 18 Oat;7.5 Ma, and Stern et al. attributed theisotopic shift to a climatic change, probably eitherthe result of increased evaporation caused by drierclimates or the development of a rainshadow causedby uplift of the Tibetan plateau. In contrast, thekaolinite isotopes showed no significant trend upwardthrough the section, and the authors suggestedthis may reflect contamination with detrital kaolinite.Nonetheless, their results indicate that the isotopiccomposition of pedogenic clay minerals shows considerablepromise for paleoclimatic reconstructionwith the same cautions that apply to pedogenic carbonate.5.4. LimitationsSome workers have put forth caveats to the interpretationof paleoclimates from paleosols. The majorcaution is the impact of hydrologic conditions in thedepositional area. In a study of Paleogene alluvialrocks in Portugal, Pimentel et al. Ž 1996.found earlydiagenetic groundwater alteration features that mimichydromorphic soil features such as pseudogley mottlingand calcic horizons. Pimentel et al. providedcriteria for distinguishing between the diagenetic andpedogenic features; however, they cautioned that thediagenetic features could be mistaken for pedogenicfeatures and lead to incorrect paleoclimatic interpretations.The features do not have climatic significancein this case.

64( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Although Caudill et al. Ž 1996.concluded thatcalcareous Vertisols might provide an excellent toolfor estimating ancient MAP, Aslan and Autin Ž 1998.have suggested that caution should be used whenestimating paleo-precipitation from the depth of calcichorizons, even in Vertisols. They observed calciczones Ž Bk and Ck horizons.in Mississippi Riverfloodplain soils, many of which are Vertisols. Applyingthe Retallack equation to both natural levee soilsand backswamp soils, they calculated MAP between545 and 908 mm, figures that seriously underestimatethe actual MAP of 1500 mm. Aslan and Autinconcluded that the position of the calcite was notcontrolled by the climate. Rather the depth to calcitein these soils is controlled by water-table levels andthat calcite precipitates from ascending, not descending,groundwaters during the wet season. Consequently,they suggested that paleosols that formed inaggradational floodplain settings are more likely topreserve information on floodplain sedimentation andpaleohydrology than on paleoclimates.A similar concern about carbonates formed fromascending vs. descending waters was sounded bySlate et al. Ž 1996 .. Those authors compared andcontrasted the morphology and isotopic compositionsof carbonates that formed in well-drained paleosolsand hydromorphic paleosols. They emphasized thatthe Cerling model should be restricted to soil carbonatethat formed in the vadose zone because only hereare all oxidized carbon species in isotopic equilibrium.Carbonates from hydromorphic paleosols areunsuitable for isotopic analysis because ascendingground waters probably contaminated their isotopiccomposition. This paper is also important because itprovides criteria for distinguishing between soil carbonatesthat formed in well-drained paleosols fromthose that formed in hydromorphic paleosols. Similarly,Pimentel et al. Ž 1996.concluded that carbonatesin Paleogene alluvial rocks were the product ofgroundwater diagenesis, although the carbonates resembledthose resulting from pedogenesis. The authorswarned that such carbonates have no particularclimatic interpretation, nor should they be used toestimate sediment accumulation rates on the basis oftheir degree of development.Finally, depending on the nature of the exposureand amount of data available, it may be difficult todistinguish climatic controls from other mechanismsthat can produce vertical changes in paleosol properties.The uncertainty that may exist is demonstratedby a study of Lower Carboniferous paleosols byWright et al. Ž 1991 .. An upward change from calcrete-bearingVertisols to paleosols interpreted asferrolytic Vertisols indicates a change in soil moisturefrom better drained to more poorly drainedconditions. The ferrolytic paleosols also contain anunusual clay assemblage that was attributed to intenseleaching. The authors suggested that either achange in local drainage conditions Žan intrinsicmechanism.or a climate change to more prolongedprecipitation could have been responsible. Althoughthey favored the climatic interpretation, they admittedthat this interpretation was tenuous because itwas based on only two paleosol profiles from onlytwo localities.6. Interpreting ancient atmospheres6.1. Isotopic analysisCerling Ž 1991, 1992.provided a model for estimatingancient atmospheric rŽ CO .2 by showing thatthe isotopic composition of soil carbonate is directlyrelated to soil CO 2, which depends on the concentrationof CO2in the atmosphere. As noted above, thecarbon isotopic composition of soil carbonates isalso sensitive to the kind of vegetation ŽC 3 vs. C 4plants .. The model assumes that the pre-Miocenevegetation consisted entirely of C 3 plants, becausevarious isotopic studies ŽCerling, 1992; Cerling etal., 1993; Quade et al., 1994, 1995; Latorre et al.,1997.have shown a major change from C 3 to C 4vegetation between 8 and 7 Ma. Also of importancefor paleosols studies is the depth of the pedogeniccarbonate because d 13 CO2values become more negativedownward in a soil profile to a depth of ;20cm.Using paleosol carbonates and the Cerling model,atmospheric CO2values have been estimated forvarious times in the Phanerozoic Ž Fig. 12 .. In a studyof pedogenic carbonates of Late Silurian throughPennsylvanian age, Mora et al. Ž 1996.concluded thatatmospheric CO2levels were high for the Late SilurianŽ 3200–5200 ppm V ., then steadily declined

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 65Fig. 12. Changes in atmospheric CO2based on isotopic analyses of paleosol carbonates. Paleozoic values are ranges from Mora et al.Ž 1996 .; Late Cretaceous value from Ghosh et al. Ž 1995 .; other values from Cerling et al. Ž 1992 ..through the Paleozoic Ž700–2050 ppm V for theDevonian; 450–1000 ppm V for the Mississippian;450–800 ppm V for the Pennsylvanian .. By Permiantime, atmospheric CO2levels dropped to 150–200ppm V Ž Mora et al., 1996 .. Retallack Ž 1997a.interpreteda similar drop in atmospheric CO2frompaleosols. This significant decline was linked to theexpansion of land plants and to the global climatechange that produced extensive late Paleozoic glaciationŽ Mora et al., 1996 ..Cerling Ž 1991.determined that atmosphericrŽ CO .2 then rose during the Late Triassic to EarlyJurassic Ž 2000–3000 ppm V .. Following high EarlyCretaceous atmospheric CO Ž 2500–3300 ppm V .2,values then fell through the Cenozoic. In a study ofpedogenic carbonates from central India, Ghosh etal. Ž 1995. concluded that rŽ CO .2 in the Late Cretaceousatmosphere was 800–1200 ppm V. Levels fellfurther in the Cenozoic with 600 ppm V estimatedfrom Eocene paleosols, and 400–700 ppm V determinedfrom Miocene paleosols Ž Cerling, 1991 .. Thecurrent value is approximately 300 ppm V.6.2. LimitationsTo yield reliable pŽ CO .2 estimates, the carbonatecollected from paleosols must be of indisputablepedogenic origin and cannot have undergone postpedogenicmodification. Wright and Vanstone Ž 1991.emphasized that groundwater carbonates pose a potentialproblem to this method. Groundwater carbonateshave different isotopic compositions than overlyingpedogenic carbonates but they can be difficultto distinguish from true pedogenic carbonates. Anotherproblem associated with groundwaters is that,with continued sedimentation, a soil becomes buriedand moved below the water table. Consequently,carbonate that first formed in the rooted zone can beoverprinted by groundwater precipitation. Burial diagenesisis also a potential problem, although studiesof lower Eocene paleosols Ž Cerling, 1991.and Devonianpedogenic carbonates Ž Driese and Mora, 1993.both show that the d 13 C of paleosol carbonates wasonly minimally affected by diagenesis.The particular soil environment in which the carbonateprecipitated can also affect its isotopic compositionand, thus, the pŽ CO .2 value it yields. In astudy of Devonian Vertisols, Driese and Mora Ž 1993.examined carbonate from two different sources inthe paleosols: rhizoliths and pedogenic nodules. Theyfound that the nodule carbonates were isotopicallyheavier than the rhizolith carbonates. The authorsattributed this to the depth at which the carbonatesprecipitated. The nodules precipitated in the zone ofsoil cracking, and, because Vertisols commonly developcracks of 1 m in depth, atmospheric CO2mayhave penetrated deep into the developing soil, resultingin nodules with isotopic values that overestimateatmospheric CO 2. The rhizoliths formed deeper inthe soil profile and apparently below the zone ofcracking. Consequently, the authors concluded thatthey provided a more accurate estimate of paleoatmosphericCO2than the nodules.Finally, as noted above, the proportion of C 3 orC vegetation influences the d 13 4C values of paleosolcarbonates. The Cerling model Ž 1991, 1992.assumesthat C 4 vegetation did not appear until late Miocenetime. This assumption has been questioned by Wright

66( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70and Vanstone Ž 1991.who argued that both paleobotanicalstudies Ž e.g., Spicer, 1989.and heavy isotopicvalues obtained from pre-Miocene pedogeniccarbonates Ž e.g., those of Mora et al., 1991.indicatedthat the early vegetation contained C 4 as wellas C 3 plants. The reader is referred to discussion andreply on this topic Ž Cerling et al., 1992 ..And what of the change to vascular plants in theearly Paleozoic? As mentioned earlier, Mora et al.Ž 1996.suggested that atmospheric CO2levelsdropped markedly during mid to late Paleozoic timein response to the expansion of terrestrial vegetation.On the basis of an isotopic study of goethite takenfrom Ordovician ironstones, Yapp and Poths Ž 1994.suggested that the productivity of pre-vascular plantswas similar to that of modern plants. Thus, theybelieve that, when using the isotopic analysis ofpaleosol carbonates to estimate paleoatmosphericlevels of CO 2, no assumptions or corrections need tobe made.7. SummaryIt now appears that soil development, due to theepisodic nature of sediment accumulation, is a normalpart of the continental sedimentary regime andthat many ancient continental deposits contain verticallystacked or multistory paleosols. Because sedimentaryrocks comprise 75% of the rocks exposed atthe Earth surface Ž Pettijohn, 1975 ., and becausemany of those rocks formed in terrestrial environmentssubject to pedogenic modification, paleosolsare abundant in the geologic record. Their geologichistory is also lengthy, extending well back into thePrecambrian Ž e.g., Driese et al., 1995 ..The study of pre-Quaternary paleosols and theways in which the paleosols are being used to helpsolve various geologic problems have grown in thepast decade. Because many ancient soil profiles areeasily recognized, exposed over broad areas, andformed almost instantaneously in terms of geologictime Ž in the range of 2000–30,000 years ., they offera nearly ideal method of correlating deposits in thecontinental realm, both at a local and at a regional orbasinal scale. Paleosols are still an under-utilizedaspect of sequence stratigraphy, and this is an area inwhich future research should focus. Continental successionswith multistory paleosols can also provide acontinuous record of ancient climatic conditions andclimatic changes through time. Additionally, paleosol–landscapeanalysis can produce a clearer, morecomplete picture of the environmental conditions andprocesses operating in ancient continental basins. Inparticular, recognizing and analyzing paleosol variabilityat different spatial and temporal scales isimportant for evaluating how landscapes evolvedover time and for assessing the relative significanceof autogenic and allogenic controls on landscapeevolution.AcknowledgementsResearch contributing to this paper was supportedby National Science Foundation Grants EAR-9303959 and EAR-9706115 to MJK. P.D. Gingerichprovided invaluable logistical support for fieldworkin the Bighorn Basin. Field assistance during studiesof paleosols in the Bighorn Basin was provided byMaureen McHugh, Brian Gwinn, Roberta Yuhas.Constructive reviews for the journal were providedby Greg Retallack and Paul Wright.ReferencesAbbott, P.L., Minch, J.A., Peterson, G.L., 1976. Pre-Eocene paleosolsouth of Tijuana, Baja California, Mexico. J. Sediment.Petrol. 46, 355–361.Aitken, J.F., Flint, S.S., 1996. Variable expressions of interfluvialsequence boundaries in the Brathitt Group Ž Pennsylvanian .,eastern Kentucky, USA. In: Howell, J.A., Aitken, J.F. Ž Eds. .,High Resolution Sequence Stratigraphy: Innovations and Applications.Geol. Soc. London, Spec. Publ. 104, 193–206.Alonso Zarza, A.M., Wright, V.P., Calvo, J.P., Garcia del Cura,M.A., 1992. Soil–landscape and climatic relationships in themiddle Miocene of the Madrid Basin. Sedimentology 39,17–35.Arndorff, L., 1993. Lateral relations of deltaic palaeosols from theLower Jurassic Ronne Formation on the Island of Bornholm,Denmark. Palaeogeogr., Palaeoclimatol., Palaeoecol. 100,235–250.Aslan, A., Autin, W.J., 1998. Holocene flood–plain soil formationin the southern lower Mississippi Valley: implications forinterpreting alluvial paleosols. Geol. Soc. Am. Bull. 110,433–449.

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 67Atkinson, C.D., 1986. Tectonic control on alluvial sedimentationas revealed by an ancient catena in the Capella FormationŽ Eocene. of northern Spain. In: Wright, V.P. Ž Ed. ., Paleosols:Their Recognition and Interpretation. Blackwell ScientificPublications, Oxford, pp. 139–179.Baldwin, B., Butler, C.O., 1985. Compaction curves. Am. Assoc.Pet. Geol. Bull. 69, 622–626.Besly, B.M., Fielding, C.R., 1989. Palaeosols in Westphaliancoal-bearing and red-bed sequences, central and northern England.Palaeogeogr., Palaeoclimatol., Palaeoecol. 70, 303–330.Bestland, E.A., 1997. Alluvial terraces and paleosols as indicatorsof early Oligocene climate change ŽJohn Day Formation,Oregon .. J. Sediment. Res. 67A, 840–855.Birkeland, P.W., 1984. Soils and Geomorphology, Oxford UniversityPress, New York, 372 pp.Blum, M.D., 1994. Genesis and architecture of incised valley fillsequences: a Late Quaternary example from the ColoradoRiver, Gulf Coastal Plain of Texas. In: Weimer, P. Posamentier,H. Ž Eds. ., Siliciclastic Sequence Stratigraphy: RecentDevelopments and Applications, Am. Ass. Petrol. Geol. Mem.58, 259–283.Blum, M.D., Valastro, S. Jr., 1994. Late Quaternary sedimentation,lower Colorado River, Gulf Coastal Plain of Texas. Geol.Soc. Am. Bull. 106, 1002–1016.Bockheim, J.G., 1980. Solution and use of chronofunctions instudying soil development. Geoderma 24, 71–85.Bown, T.M., Kraus, M.J., 1981. Lower Eocene alluvial paleosolsŽ Willwood Formation, northwest Wyoming, USA.and theirsignificance for paleoecology paleoclimatology, and basinanalysis. Palaeogeogr., Palaeoclimatol., Palaeoecol. 34, 1–30.Bown, T.M., Kraus, M.J., 1987. Integration of channel and floodplainsuites: I. Developmental sequence and lateral relations ofalluvial paleosols. J. Sediment. Petrol. 57, 587–601.Bown, T.M., Kraus, M.J., 1993. Time-stratigraphic reconstructionand integration of paleopedologic, sedimentologic, and bioticevents ŽWillwood Formation, Lower Eocene, northwestWyoming, USA .. Palaios 8, 68–80.Brady, N.C., 1990. The Nature and Properties of Soils. Macmillan,New York, 621 pp.Bridge, J.S., Leeder, M.R., 1979. A simulation model of alluvialstratigraphy. Sedimentology 26, 599–623.Caudill, M.R., Driese, S.G., Mora, C.I., 1996. Preservation of apaleo-Vertisol and an estimate of late Mississippian paleoprecipitation.J. Sediment. Res. 66A, 58–70.Caudill, M.R., Driese, S.G., Mora, C.I., 1997. Physical compactionof vertic palaeosols: implications for burial diagenesisand palaeo-precipitation estimates. Sedimentology 44, 673–685.Cerling, T.E., 1984. The stable isotopic composition of modernsoil carbonate and its relationship to climate. Earth Planet. Sci.Lett. 71, 229–240.Cerling, T.E., 1991. Carbon dioxide in the atmosphere: evidencefrom Cenozoic and Mesozoic paleosols. Am. J. Sci. 291,377–400.Cerling, T.E., 1992. Use of carbon isotopes in paleosols as anindicator of the rŽ CO .2 of the paleoatmosphere. GlobalBiochem. Cycles 6, 307–314.Cerling, T.E., Quade, J., 1993. Stable carbon and oxygen isotopesin soil carbonates. Climate Change in Continental IsotopicRecords. American Geophysical Union, pp. 217–231.Cerling, T.E., Wright, V.P., Vanstone, S.D., 1992. Further commentson using carbon isotopes in palaeosols to estimate theCO2content of the palaeo-atmosphere. J. Geol. Soc. London149, 673–676.Cerling, T.E., Wang, Y., Quade, J., 1993. Expansion of C4ecosystems as an indicator of global ecological change in thelate Miocene. Nature 361, 344–345.Driese, S.G., Mora, C.I., 1993. Physico-chemical environment ofpedogenic carbonate formation in Devonian vertic palaeosols,central Appalachians, USA. Sedimentology 40, 199–216.Driese, S.G., Srinivasan, K., Mora, C.I., Stapor, F.W., 1994.Paleoweathering of Mississippian Monteagle Limestone precedingdevelopment of a lower Chesterian transgressive systemstract and sequence boundary, middle Tennessee andnorthern Alabama. Geol. Soc. Am. Bull. 106, 866–878.Driese, S.G., Simpson, E.L., Eriksson, K.A., 1995. Redoximorphicpaleosols in alluvial and lacustrine deposits, 1.8 GaLochness Formation, Mount Isa, Australia; pedogenic processesand implications for paleoclimate. J. Sediment. Res.65A, 675–689.Duchaufour, P., 1982. Pedology. Allen and Unwin, London, 448pp.Fastovsky, D.E., McSweeney, K., 1987. Paleosols spanning theCretaceous–Paleogene transition, eastern Montana and westernNorth Dakota. Geol. Soc. Am. Bull. 99, 66–77.Food and Agriculture OrganisationrUnited Nations Scientific andCultural Organization, 1974. Soil Map of the World. Vol. I,Legend. Unesco, Paris.Frakes, L.A., Sun, J., 1994. A carbon isotope record of the upperChinese loess sequence; estimates of plant types during stadialsand interstadials. Palaeogeogr., Palaeoclimatol., Palaeoecol.108, 183–189.Friedman, I., O’Neil, J.R., 1977. Compilation of stable isotopefractional factors of geochemical interest. In: Fleischer, M.Ž Ed. ., Data of Geochemistry. U.S. Geol. Surv. Prof. Pap.440-KK, 12 pp.Ghosh, P., Bhattacharya, S.K., Jani, R.A., 1995. Palaeoclimateand palaeovegetation in Central India during the Upper Cretaceousbased on stable isotope composition of the palaeosolcarbonates. Palaeogeogr., Palaeoclimatol., Palaeoecol. 114,285–296.Gill, S., Yemane, K., 1996. Implications of a Lower PennsylvanianUltisol for equatorial Pangean climates and early, oligotrophic,forest ecosystems. Geology 24, 905–908.Guccione, M.J., 1993. Grain-size distribution of overbank sedimentand its use to locate channel positions. In: Marzo, M.,Puigdefabregas, C. Ž Eds. ., Alluvial Sedimentation. Int. Ass.Sediment., Spec. Publ. 17, 185–194.Hanneman, D.L., Wideman, C.J., 1991. Sequence stratigraphy ofCenozoic continental rocks, southwestern Montana. Geol. Soc.Am. Bull. 103, 1335–1345.Hanneman, D.L., Wideman, C.J., Halvorson, J.W., 1994. Calcicpaleosols: their use in subsurface stratigraphy. Am. Assoc. Pet.Geol. Bull. 78, 1360–1371.

68( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70Harden, J.W., Taylor, E.M., 1983. A quantitative comparison ofsoil development in four climatic regimes. Quat. Res. 20,342–359.Holliday, V.T., 1990. Soils and landscape evolution of eolianplains: the Southern High Plains of Texas and New Mexico.In: Knuepfer, P.L.K., McFadden, L.D. Ž Eds ., Soils and LandscapeEvolution. Geomorphology 3, 489–515.Jenny, H., 1941. Factors of Soil Formation. McGraw-Hill, NewYork, 281 pp.Joeckel, R.M., 1995. Paleosols below the Ames Marine UnitŽ Upper Pennsylvanian, Conemaugh Group.in the AppalachianBasin, U.S.A.; variability on an ancient depositional landscape.J. Sediment. Res. 65A, 393–407.Kleiss, H.J., Fehrenbacher, J.B., 1973. Loess distribution as revealedby mineral variation. Soil Sci. Soc. Am. Proc. 37,291–295.Koch, P.L., Zachos, J.C., Gingerich, P.D., 1992. Correlationbetween isotope records in marine and continental carbonreservoirs near the PaleocenerEocene boundary. Nature 358,319–322.Koch, P.L., Zachos, J.C., Dettman, D.L., 1995. Stable isotopestratigraphy and paleoclimatology of the Paleogene BighornBasin Ž Wyoming, USA .. Palaeogeogr., Palaeoclimatol.,Palaeoecol. 115, 61–89.Kraus, M.J., 1987. Integration of channel and floodplain suites: II.Lateral relations of alluvial paleosols. J. Sediment. Petrol. 57,602–612.Kraus, M.J., 1992. Alluvial response to differential subsidence:sedimentological analysis aided by remote sensing, WillwoodFormation Ž Eocene ., Bighorn Basin, Wyoming, USA. Sedimentology39, 455–470.Kraus, M.J., 1996. Avulsion deposits in lower Eocene alluvialrocks, Bighorn Basin, Wyoming. J. Sediment. Res. 66B, 354–363.Kraus, M.J., 1997. Lower Eocene alluvial Paleosols; pedogenicdevelopment, stratigraphic relationships, and Paleosolrlandscapeassociations. Palaeogeogr., Palaeoclimatol., Palaeoecol.129, 387–406.Kraus, M.J., Aslan, A., 1993. Eocene hydromorphic paleosols:significance for interpreting ancient floodplain processes. J.Sediment. Petrol. 63, 453–463.Kraus, M.J., Aslan, A., 1999. Paleosol sequences in floodplainenvironments: a hierarchical approach. In: Thiry, M. Ž Ed. .,Palaeoweathering, Palaeosurfaces and Related Continental Deposits.Int. Assoc. Sedimentol., Spec. Publ. 27, 303–321.Kraus, M.J., Bown, T.M., 1993. Short-term sediment accumulationrates determined from Eocene alluvial paleosols. Geology21, 743–746.Kraus, M.J., Gwinn, B., 1997. Facies and facies architecture ofPaleogene floodplain deposits, Willwood Formation, BighornBasin, Wyoming, USA. Sediment. Geol. 114, 33–54.Kukla, G.J., An, Z., 1989. Loess stratigraphy in Central China.Palaeogeogr., Palaeoclimatol., Palaeoecol. 72, 203–225.Lander, R.H., Bloch, S., Mehta, S., Atkinson, C.D., 1991. Burialdiagenesis of paleosols in the giant Yacheng gas field, People’sRepublic of China: bearing on illite reaction pathways. J.Sediment. Petrol. 61, 256–268.Latorre, C., Quade, J., McIntosh, W.C., 1997. The expansion ofC 4 grasses and global change in the late Miocene; stableisotope evidence from the Americas. Earth Planet. Sci. Lett.146, 83–96.Leckie, D., Fox, C., Tarnocai, C., 1989. Multiple paleosols of thelate Albian Boulder Creek formation, British Columbia,Canada. Sedimentology 36, 307–323.Leeder, M.R., 1975. Pedogenic carbonates and flood sedimentaccretion rates: a quantitative model for alluvial arid-zonelithofacies. Geol. Mag. 112, 257–270.Mack, G.H., 1992. Paleosols as an indicator of climatic change atthe Early–Late Cretaceous boundary, southwestern New Mexico.J. Sediment. Petrol. 62, 483–494.Mack, G.H., James, W.C., 1994. Paleoclimate and the globaldistribution of paleosols. J. Geol. 102, 360–366.Mack, G.H., Cole, D.R., Giordano, T.H., Schaal, W.C., Barcelos,J.H., 1991. Paleoclimatic controls on stable oxygen and carbonisotopes in caliche of the Abo Formation Ž Permian ., southcentralNew Mexico, U.S.A. J. Sediment. Petrol. 61, 458–472.Mack, G.H., James, W.C., Monger, H.C., 1993. Classification ofpaleosols. Geol. Soc. Am. Bull. 105, 129–136.Marriott, S.B., 1996. Analysis and modelling of overbank deposits.In: Anderson, M.G., Walling, D.E., Bates, P.D. Ž Eds. .,Floodplain Processes. Wiley, Chichester, pp. 63–93.Marriott, S.B., Wright, V.P., 1993. Paleosols as indicators ofgeomorphic stability in two Old Red Sandstone alluvial suites,South Wales. J. Geol. Soc. London 150, 1109–1120.McCahon, T.J., Miller, K.B., 1997. Climatic significance of natrichorizons in Permian Ž Asselian.palaeosols of north-centralKansas, USA. Sedimentology 44, 113–125.McCarthy, P.J., Plint, A.G., 1998. Recognition of interfluve sequenceboundaries: integrating paleopedology and sequencestratigraphy. Geology 26, 387–390.McCarthy, P.J., Martini, I.P., Leckie, D.A., 1997. Anatomy andevolution of a Lower Cretaceous alluvial plain; sedimentologyand palaeosols in the upper Blairmore Group, south-westernAlberta, Canada. Sedimentology 44, 197–220.McDonald, E.V., Busacca, A.J., 1990. Interaction between aggradinggeomorphic surfaces and the formation of a Late Pleistocenepaleosol in the Palouse loess of eastern Washingtonstate. In: Knuepfer, P.L.K., McFadden, L.D. Ž Eds ., Soils andLandscape Evolution. Geomorphology 3, 449–470.McFadden, L.D., Knuepfer, P.L.K., 1990, Soil geomorphology;the linkage of pedology and surficial processes. In: Knuepfer,P.L.K., McFadden, L.D. Ž Eds ., Soils and Landscape Evolution.Geomorphology 3, 197–205.Miller, K.B., McCahon, T.J., West, R.R., 1996. Lower PermianŽ Wolfcampian.Paleosol-bearing cycles of the U.S. Midcontinent;evidence of climatic cyclicity. J. Sediment. Res. 66A,71–84.Mora, C.I., Driese, S.G., Seager, P.G., 1991. Carbon dioxide inthe Paleozoic atmosphere: evidence from carbon-isotope compositionsof pedogenic carbonate. Geology 19, 1017–1020.Mora, C.I., Driese, S.G., Colarusso, L.A., 1996. Middle to latePaleozoic atmospheric CO2levels from soil carbonate andorganic matter. Science 271, 1105–1107.Mora, C.I., Sheldon, B.T., Elliott, W.C., Driese, S.G., 1998. An

( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–70 69oxygen isotope study of illite and calcite in three AppalachianPaleozoic vertic paleosols. J. Sediment. Res. 68, 456–464.Morrison, R.B., 1967. Principles of Quaternary soil stratigraphy.In: Morrison, R.B., Wright, H.E. Ž Eds. ., Means of Correlationof Quaternary Successions, Int. Assoc. Quat. Res. Ž INQUA .,VII Congr., Proc. 9, 1–69.Morrison, R.B., 1978. Quaternary soil stratigraphy—concepts,methods, and problems. In: Mahaney, W.C. Ž Ed. ., QuaternarySoils, Geo Abstracts, Norwich. pp. 77–108.Nadon, G.C., Issler, D.R., 1997. The compaction of floodplainsediments: timing, magnitude and implications. Geosci. Can.24, 37–43.North, C.P., 1996. The prediction and modelling of subsurfacefluvial stratigraphy. In: Carling, P.A., Dawson, M.R. Ž Ed. .,Advances in Fluvial Dynamics and Stratigraphy. Wiley,Chichester, pp. 395–508.Pecsi, M., 1995. The role of principles and methods in loesspaleosolinvestigations. GeoJournal 36, 117–131.Pettijohn, F.J., 1975. Sedimentary Rocks, Third edn. New York,Harper and Row, 628 pp.Pimentel, N.L., Wright, V.P., Azevedo, T.M., 1996. Distinguishingearly groundwater alteration effects from pedogenesis inancient alluvial basins: examples from the Palaeogene ofsouthern Portugal. Sediment. Geol. 105, 1–10.Platt, N.H., Keller, B., 1992. Distal alluvial deposits in a forelandbasin setting—the lower freshwater Molasse Ž lower Miocene .,Switzerland: sedimentology, architecture and palaeosols. Sedimentology39, 545–565.Quade, J., Cerling, T.E., Bowman, J.R., 1989. Development ofAsian monsoon revealed by marked ecological shift during thelatest Miocene in northern Pakistan. Nature 342, 163–165.Quade, J., Solounias, N., Cerling, T.E., 1994. Stable isotopicevidence from Paleosol carbonates and fossil teeth in Greecefor forest or woodlands over the past 11 Ma. Palaeogeogr.,Palaeoclimatol., Palaeoecol. 108, 41–53.Quade, J., Cater, J.M.L., Ojha, T.P., Adam, J., Harrison, T.M.,1995. Late Miocene environmental change in Nepal and thenorthern Indian subcontinent; stable isotopic evidence frompaleosols. Geol. Soc. Am. Bull. 107, 1381–1397.Rasbury, E.T., Hanson, G.N., Meyers, W.J., Holt, W.E., Goldstein,R.H., Saller, A.H., 1998. U–Pb dates of paleosols:constraints on Late Paleozoic cycle durations and boundaryages. Geology 26, 403–406.Retallack, G.J., 1983. A paleopedological approach to the interpretationof terrestrial sedimentary rocks: the mid Tertiaryfossil soils of Badlands National Park, South Dakota. Geol.Soc. Am. Bull. 94, 823–840.Retallack, G.J., 1984. Completeness of the rock and fossil record:some estimates using fossil soils. Paleobiology 10, 59–78.Retallack, G.J., 1991. Untangling the effects of burial alterationand ancient soil formation. Ann. Rev. Earth Planet. Sci. 19,183–206.Retallack, G.J., 1993. Classification of paleosols: discussion andreply. Discussion Geol. Soc. Am. Bull. 105, 1635–1636.Retallack, G.J., 1994. The environmental factor approach to theinterpretation of paleosols. In: Amundson, R., Harden, J.,Singer, M. Ž Eds. ., Factors of Soil Formation: a Fiftieth AnniversaryRetrospective. Soil Sci. Soc. Am., Madison, WI, pp.31–64.Retallack, G.J., 1997a. Early forest soils and their role in Devonianglobal change. Science 276, 583–585.Retallack, G.J., 1997b. A Colour Guide to Paleosols. Wiley,Chichester, 175 pp.Retallack, G.J., 1998. Fossil soils and completeness of the rockand fossil records. In: Donovan, S.K., Paul, C.R.C. Ž Eds. ., TheAdequacy of the Fossil Record. Wiley, Chichester, pp. 133–163.Retallack, G.J., German-Heins, J., 1994. Evidence from paleosolsfor the geological antiquity of rain-forest. Science 265, 499–502.Richmond, G.M., 1962. Quaternary stratigraphy of the La SalMountains, Utah. U.S. Geol. Surv., Prof. Pap. 324, 135 pp.Ritzma, H.R., 1955. Late Cretaceous and Early Cenozoic structuralpattern, southern Rock Springs Uplift Wyoming. Wyo.Geol. Assoc., 10th Ann. Field Conf. Guidebook, 135–137.Robert, C., Kennett, J.P., 1994. Antarctic subtropical humidepisode at the Paleocene–Eocene boundary: clay–mineral evidence.Geology 22, 211–214.Ruhe, R.V., 1983. Aspects of Holocene pedology in the UnitedStates. In: Wright, H.E., Jr. Ž Ed. ., Late-Quaternary Environmentsof the United States, Vol. 2. The Holocene. Univ. ofMinnesota Press, Minneapolis,Schultz, C.B., Tanner, L.G., Harvey, C., 1955. Paleosols of theOligocene of Nebraska. Bull. Univ. Neb. State Mus. 4, 1–16.Schumacher, B.A., Day, W.J., Amacher, M.C., Miller, B.J., 1988.Soils of the Mississippi River alluvial plain in Louisiana.Louisiana Agri. Exp. Sta. Bull. 796, 275 pp.Shackleton, N.J., 1986. Paleogene stable isotope events. Palaeogeogr.,Palaeoclimatol., Palaeoecol. 57, 91–102.Shanley, K.W., McCabe, P.J., 1994. Perspectives on the sequencestratigraphy of continental strata. Am. Assoc. Pet. Geol. Bull.78, 544–568.Slate, J.L., Smith, G.A., Wang, Y., Cerling, T.E., 1996. Carbonate-paleosolgenesis in the Plio-Pleistocene St. David Formation,southeastern Arizona. J. Sediment. Res. 66A, 85–94.Smith, G.D., 1942. Illinois loess—variations in its properties anddistribution: a pedologic interpretation. Univ. Illinois Agri.Exp. Sta. Bull. 490, 137–184.Smith, R.M.H., 1990. Alluvial paleosols and pedofacies sequencesin the Permian Lower Beaufort of the southwestern KarooBasin, South Africa. J. Sediment. Petrol. 60, 258–276.Soil Survey Staff, 1975. Soil taxonomy. U.S. Dept. Agri. Handbook436, 754 pp.Soil Survey Staff, 1998. Keys to Soil Taxonomy, 8th edn. U.S.Dept. Agri. Nat. Res. Conserv. Serv., 327 p.Soreghan, G.S., Elmore, R.D., Katz, B., Cogoini, M., Banerjee,S., 1997. Pedogenically enhanced magnetic susceptibility variationspreserved in Paleozoic loessite. Geology 25, 1003–1006.Spicer, R.A., 1989. Physiological characteristics of land plants inrelation environment through time. Trans. R. Soc. Edinburgh:Earth Sci. 44, 321–329.Stern, L.A., Chamberlain, C.P., Reynolds, R.C., Johnson, G.D.,

70( )M.J. KrausrEarth-Science ReÕiews 47 1999 41–701997. Oxygen isotope evidence of climate change from pedogenicclay minerals in the Himalayan molasse. Geochim.Cosmochim. Acta. 61, 731–744.Summerfield, M.A., 1991. Global Geomorphology: an Introductionto the Study of Landforms. Wiley, New York, 537 pp.Tandon, S.K., Gibling, M.R., 1994. Calcrete and coal in lateCarboniferous cyclothems of Nova Scotia, Canada: climateand sea-level changes linked. Geology 22, 755–758.Tandon, S.K., Gibling, M.R., 1997. Calcretes at sequence boundariesin upper Carboniferous cyclothems of the Sydney Basin,Atlantic Canada. Sediment. Geol. 112, 43–67.Tandon, S.K., Good, A., Andrews, J.E., Dennis, P.F., 1995.Palaeoenvironments of the dinosaur-bearing Lameta BedsŽ Maastrichtian ., Narmada Valley, Central India. Palaeogeogr.,Palaeoclimatol., Palaeoecol. 117, 153–184.Walling, D.E., Quine, T.A., He, O., 1992. Investigating contemporaryrates of floodplain sedimentation. In: Carling, P.A.,Petts, G.E. Ž Eds. ., Lowland Floodplain Rivers: GeomorphologicalPerspectives. Wiley, Chichester, pp. 165–184.Webb, G.E., 1994. Paleokarst, paleosol, and rocky-shore depositsat the Mississipian–Pennsylvanian unconformity, northwesternArkansas. Geol. Soc. Am. Bull. 106, 634–648.Wheeler, H.E., 1958. Time-stratigraphy. Am. Assoc. Pet. Geol.Bull. 42, 1047–1063.Wright, V.P., 1990. Estimating rates of calcrete formation andsediment accretion in ancient alluvial deposits. Geol. Mag.127, 273–276.Wright, V.P., 1992a. Paleosol recognition; a guide to early diagenesisin terrestrial settings. In: Wolf, K.H., Chilingarian, G.V.Ž Eds. ., Diagenesis, III. Elsevier., Amsterdam, pp. 591–619.Wright, V.P., 1992b. Paleopedology: stratigraphic relationshipsand empirical models. In: Martini, I.P., Chesworth, W. Ž Eds. .,Weathering, Soils and Paleosols. Elsevier, Amsterdam, pp.475–499.Wright, V.P., 1994. Paleosols in shallow marine carbonate sequences.Earth-Sci. Rev. 35, 367–395.Wright, V.P., Marriott, S.B., 1993. The sequence stratigraphy offluvial depositional systems: the role of floodplain storage.Sediment. Geol. 86, 203–210.Wright, V.P., Marriott, S.B., 1996. A quantitative approach to soiloccurrence in alluvial deposits and its application to the OldRed Sandstone of Britain. J. Geol. Soc. London 153, 907–913.Wright, V.P., Platt, N.H., 1995. Seasonal wetland carbonate sequencesand dynamic catenas; a re-appraisal of palustrinelimestones. Sediment. Geol. 99, 65–71.Wright, V.P., Robinson, D., 1988. Early Carboniferous floodplaindeposits from South Wales: a case study of the controls onpalaeosol development. J. Geol. Soc. London 145, 847–857.Wright, V.P., Vanstone, S.D., 1991. Assessing the carbon dioxidecontent of ancient atmospheres using palaeo-calcretes: theoreticaland empirical constraints. J. Geol. Soc. London 148,945–947.Wright, V.P., Vanstone, S.D., Robinson, D., 1991. Ferrolysis inArundian alluvial palaeosols, evidence of a shift in the EarlyCarboniferous monsoonal system. J. Geol. Soc. London 148,9–12.Yapp, C.J., Poths, H., 1994. Productivity of pre-vascular continentalbiota inferred from the FeŽ CO .3 OH content of goethite.Nature 368, 49–51.Ye, L., 1995. Paleosols in the Upper Guantae Formation Ž Miocene.of the Gudong oil field and their application to the correlationof fluvial deposits. Am. Assoc. Pet. Geol. Bull. 79, 981–988.Zachos, J.C., Lohmann, K.C., Walker, J.C.G., Wise, S.W., 1993.Abrupt climate change and transient climates during the Paleogene:a marine perspective. J. Geol. 101, 191–213.Mary Kraus has been a professor at theUniversity of Colorado since 1983. Sheholds degree from Yale University ŽB.S.,Biology ., the University of WyomingŽ M.S., Geology.and the University ofColorado Ž Ph.D., Geology .. Her researchand publications focus on Ž. 1 understandingthe physical processes bywhich ancient river floodplains wereconstructed; Ž. 2 using floodplain paleosolsto reconstruct ancient landscapesat different spatial and temporal scales;and Ž. 3 integrating the study of paleosols with the more traditionalstudy of channel sandstone bodies to understand the factors thatinfluence alluvial architecture.