Geochemical Climofunctions from North American Soils and ...

692 N. D. SHELDON ET AL.Table 3. Comparison of Mean Annual Precipitation (MAP), Mean Annual Temperature (MAT), Chemical Index ofAlteration without Potassium (CIA-K), and the Molecular Ratio of Bases/Alumina (B/Al) for Alfisols, Mollisols, andUltisols of North AmericaSoil type MAP j MAT j CIA-K j B/Al jAlfisol (n p 27) 991 196 11.1 3.6 69.1 8.1 .61 .19Mollisol (n p 30) 582 181 9.4 3.6 51.4 16.1 1.38 .9Ultisol (n p 51) 1239 138 16.2 2.5 87.4 8.9 .24 .232with low accuracy ( R p 0.27, n p 258). The proportionof weatherable minerals needed for an Alfisolusing this relationship is at least 15% in veryfine sand or coarse silt fractions.Dust and RejuvenationPerhaps the biggest surprise of this study is thatrobust relationships could emerge from such variedparent materials and times for formation. These aresuch strong controls on soil formation that onewould expect uncorrelated scatter plots of geochemicalcompositions and soil forming variables.A variety of processes are rejuvenating and homogenizingNorth American soils. Volcanic ashand eolian dust are the primary agents of renewalin North American soil formation (Brimhall et al.1988; McFadden et al. 1991), especially the voluminousMount St. Helens and Mazama ashes of theAmerican West (Busacca et al. 1992) and the Wisconsinanloess and till of the Midwest and northeasternUnited States (Ruhe 1984; Šibrava et al.1986). Bedrock soils near glacial ice margins arecommonly scraped clean by glacial advances and,even where never glaciated, may have been deeplyfractured and renewed by frost cracking (Ciolkoszet al. 1990). Many paleokarst terranes on Paleozoiclimestones have been renewed by slope wash fromsurrounding sandstone and siltstone as well as fromdustings of postglacial loess (Ruhe and Olson 1980).Even in humid, warm bedrock terraces of the southeasternUnited States remote from glacier terminiand loess, downward weathering into fresh rockprovides an upper limit to soil age of about 3 millionyears (Markewich et al. 1989). Soil formationincludes much alteration of minerals in place, butsoils are homogenized continent-wide by influxesof eolian dust and volcanic ash, by frost crackingand glacial scouring, and by continued downwardweathering.Another surprise is that the best relationshipsemerged from subsurface (Bw or Bt) horizon compositionrather than surface (A) or deep (C) horizonsor profile depth function parameters. This also canbe seen as a validation of eolian models of soilbuilding (Brimhall et al. 1988; McFadden et al.1991), in which B horizon material tends to be of3 5a geological age (some 10 –10 yr), older than materialstill accumulating at the surface and youngerthan materials deeper in the profile that in somecases remains from geologically much earlier periodsof weathering. The B horizon can be consideredto have a mean residence time that averagesout short-term surficial disruptions and additionsas well as deeper residual accumulation.Less surprising is the emergence of taxonomicallyrelevant geochemical criteria from our databecause the Alfisol-Ultisol division was first designedto apply to North American soils (Arnold1994). Nevertheless, temporal and lithological dependenceof United States soil orders is notable inour data (tables 1, 2). In general, Ultisols are foundon old landscapes and felsic, noncalcareous parentmaterials like those of the southeastern UnitedStates (Markewich et al. 1989, 1990), but Alfisolsoccur on young landscapes and mafic, calcareousparent materials like those of the midwesternUnited States (Ruhe 1984; Holliday 1990). Thereare exceptions that reflect individual soil histories.Unlike climofunctions in which competing factorsare minimized, soil taxonomic criteria reflect thecurrent condition of the soil as a compromise ofinteracting factors.Application of Transfer Functions to Eocene andOligocene Paleosols of OregonCentral Oregon has thick sequences of nonmarinevolcaniclastic sedimentary rocks, which includemany hundreds of paleosols ranging in geologicalage back to the Paleocene. These rocks have longbeen famous for their terrestrial fossils, which havebeen important for understanding the evolution ofplants and mammals through the past 45 millionyears, particularly the well-known evolution ofhorses (Retallack et al. 1996). The plant fossils providemuch information on paleoclimate, by virtueof their botanical affinities and foliar physiognomy(Leopold et al. 1992; Manchester 1994; Meyer andManchester 1997). Using such well-known climofunctionsas depth to calcic horizon as a guide topaleoprecipitation (Retallack 1994), we can add in-