Soil-Age Relationships on Late Quaternary Moraines, Arrowsmith Range,Southern Alps, New Zealand

D. T. Rodbell

Arctic and Alpine Research, Vol. 22, No. 4. (Nov., 1990), pp. 355-365.

Stable URL:

http://links.jstor.org/sici?sici=0004-0851%28199011%2922%3A4%3C355%3ASROLQM%3E2.0.CO%3B2-%23

Arctic and Alpine Research is currently published by The Regents of the University of Colorado, a body corporate, contractingon behalf of the University of Colorado at Boulder for the benefit of INSTAAR.

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available athttp://www.jstor.org/about/terms.html. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtainedprior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content inthe JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained athttp://www.jstor.org/journals/regentsINSTAAR.html.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printedpage of such transmission.

The JSTOR Archive is a trusted digital repository providing for long-term preservation and access to leading academicjournals and scholarly literature from around the world. The Archive is supported by libraries, scholarly societies, publishers,and foundations. It is an initiative of JSTOR, a not-for-profit organization with a mission to help the scholarly community takeadvantage of advances in technology. For more information regarding JSTOR, please contact support@jstor.org.

http://www.jstor.orgTue Sep 18 22:56:22 2007

Arctic and Alpine Research, Vol. 22, No. 4, 1990, pp. 355-365

SOIL-AGE RELATIONSHIPS ON LATE QUATERNARY MORAINES, ARROWSMITH RANGE, SOUTHERN ALPS, NEW ZEALAND

D. T. RODBELL Department of Geological Sciences, University of Colorado

Boulder, Colorado 80309, U. S.A.

ABSTRACT

Loess on Holocene moraines has accumulated steadily since moraine deposition resulting in soils that generally lack buried horizons and have a degree of development which closely reflects the age of the underlying till. The progressive development of two field indices and five laboratory properties of these soils are best described by logarithmic functions. Indices of pedogenic iron and rubification yield the highest correlation with age of deposit. Rates of soil development are faster than those reported for other parts of the eastern Southern Alps and curves of most soil properties flatten within ca. 10 ka.

In contrast, the use of soils to date pre-Holocene moraines is greatly hindered by loess redistribution which commonly results in eroded and/or buried soils. Radiocarbon dates from charcoal within loess on pre-Holocene moraines suggest that loess reworking has been episodic, resulting in soils that are far younger than underlying tills. Holocene reworking of loess on upland sites in response to fires, frost action, and high winds may serve as an analog for large-scale late Pleistocene loess reworking on the Canterbury Plains.

INTRODUCTION

Although relative age criteria have been used to esti- Mount Cook region and concluded that the data best fit mate ages of surficial deposits in New Zealand, few the power function, Y=AXb, where Y is soil property parameters have been found to be equally useful in dating and X i s time. Knuepfer (1988) also found that power both Holocene and late Pleistocene deposits. The lichen- functions best describe soil development on late Quater- calibration curves of Burrows (1973, 1975) and Gellatly nary stream terraces in the Westland and Marlborough (1982) provide age control for late Holocene deposits, regions. The functions derived by Birkeland (1984a) sug- whereas the weathering-rind calibration curves of Chinn gest that soil development does not attain a steady state (1981) and Whitehouse et al. (1986) enable age estimates in ca. 10 ka, and thus may provide a useful means of dif- of deposits spanning the Holocene. ferentiating older moraines. However, with the exception

There have been numerous studies on soil-time rela- of the work of Webb (1976) and McGregor (1981), little tionships in the South Island, New Zealand. Examples has been done to quantify differences in soil development are the work of Webb (1976), McGregor (1981), Harrison on pre-Holocene moraines, or between Holocene and pre- (1982), Sommerville et al. (1982), Birkeland (1984a) and Holocene moraines. Gellatly (1985) east of the Main Divide; and Mokma et The objectives of this study are (1) to develop addi- al. (1973), Campbell (1975), Ross et al. (1977), Basher tional Holocene soil chronofunctions for a glaciated (1986), and Knuepfer (1988) west of the Main Divide. valley east of the Main Divide, and (2) to determine Birkeland (1984a) calculated chronofunctions for soils whether soil development can be used to discriminate bet- formed on Holocene tills in the Ben Ohau Range and ween early Holocene and pre-Holocene moraines.

@I990 Regents of the University of Colorado D. T. RODBELL/ 355

STUDY AREA AND SITE FACTORS

Soils were studied on moraines in the upper South Ash- burton drainage, Arrowsmith Range, mid-Canterbury, South Island (Figure 1). Bedrock is primarily Permian and Triassic greywacke with a minor component of chert, volcanic rocks, and limestone (Gair, 1967; New Zealand Geological Survey, 1972; Sporli and Lillie, 1974). Mean annual precipitation (MAP) is 3200 mm (Whitehouse et al., 1986), and although Harvey (1974) reports a mean annual temperature (MAT) of 83°C for nearby Paddle Creek drainage, the MAT of the South Ashburton cir- ques is probably lower. Vegetation varies with age of deposit; those less than ca. 3000 yr have little vegetation, whereas tussock grasses (Chionochloa sp.) dominate older surfaces (Burrows, 1975).

Moraines of the South Ashburton Valley are well pre- served in the upper 25 km of the main drainage and in the headwaters of Boundary Creek (Figure 2). Estimated ages of Holocene moraines were determined from weathering-rind measurements and lichenometry, and are constrained by published radiocarbon dates and historical

records from the adjacent Cameron drainage (Table 1). The weathering-rind calibration curve of Chinn (1981) was used for age estimates and has an error of 20% (Chinn, 1981; Whitehouse et al., 1986). The lichen cali- bration curve of Gellatly (1982) was used for age estimates of late Holocene moraines, and although little has been published on the accuracy of this method in New Zealand, a qualitative plus-minus of 20% has been adopted for the purposes of this study, following recommendations of Miller and Andrews (1972) and Calkin and Ellis (1980).

The radiocarbon dates and historical records are in general agreement with the lichenometric and weathering- rind-derived ages (Table 1). The only exception is that the radiocarbon date of 9520 yr BP, used as a minimum age for the Wildman I1 moraines in the Cameron Valley by Burrows (1975), does not agree with the 8100 yr BP weathering-rind estimate of the Wildman I moraines in the South Ashburton Valley (Rodbell, 1986). In this case the radiocarbon date is used as a minimum age for the Wildman I moraines in the study area.

METHODS

Field descriptions of 22 soils from stable sites on Late Quaternary moraine crests (Rodbell, 1986) were used as a basis for choosing 12 representative soils for laboratory analysis (Table 2). Field descriptions of these soils follow the Soil Survey Staff (1975) format and horizon nomen- clature follows Birkeland (1984b). The soil profile development index (PDI) was used to condense field descriptions into single values that reflect pedological development relative to parent material properties (Harden, 1982; Harden and Taylor, 1983). Values range from 0 (no development) to 1 (maximum development). Similarly, rubification (Harden, 1982) converts Munsell colors into single values that reflect soil redness relative to parent material redness. Because of its value as a soil

development indicator (Birkeland 1984b), soil rubifica- tion was both included in the PDI and used as a separate index. Parent material values for till were estimated from unweathered till adjacent to the modern South Ashbur- ton glacier, whereas those for loess were from the silt frac- tion of modern alluvium.

Standard soil analytical techniques were applied to the <2-mm fraction following the methods of Singer and Janitsky (1986). Organic carbon was determined by the Walkley-Black method, and pH is for both a 1:l ratio of soil to distilled water and a 1 :50 ratio of soil to 1.0 M NaF. Percent sand was determined by sieving, pipette analysis was employed for the silt and clay fractions, and Wentworth size fractions are used. Oxalate extractable

RGURE 1. Location of the study area, South Island, New Zea- land. Stippled pattern refers to braided river channels.

TABLE1 Estimated ages of morainesfrom iichenometry, weathering rrnds, limiting radiocarbon dates and

a historical record in years BPa -- ------- -~~~~~

Radiocarbon dates and Stratigraphic unit -- Lichenometric ageC Weathering-rind ageC

-- historical record

Arrowsmith IIIb 7 5 t 1 5 NAd 100e Arrowsmith IIh 100t 20 NA Arrowsmith Ib 250 t 50 2 0 0 t 40 <537 -t 50 (NZ 687)' Marquee IIb NA 2500+ 500 Marquee Ib NA 2800 t 550 Wildman I11 NA 5750-t 1150 Wildman I1 NA 7600 t 1500 >9520 +95 (NZ 688)' Wildman Ih NA 8100 t 1600 < 11,900t 200 (NZ 1652)f Late Pleistoceneb NA NA > 11,900t 200 (NZ 1652)' Pre-Late Pleistoceneb NA NA >40,900 (NZ 1684)s-- -----

"Lichenometric and weathering-rind data are from Rodbell (1986), and stratigraphic nomenclature is modified from Burrows (1975) and Mabin (1984).

hStratigraphic units represented in this study. CLichenometric ages were derived using curve of Gellatly (1982), and weathering-rind ages were derived

using curve of Chinn (1981). dIndicates that the method is not applicable in that age span. 'Historical record from Burrows (1975). 'Radiocarbon dates from Burrows (1975). gRadiocarbon date from Harvey (1974).

Explanation

Arrowsmith Till

Marquee Till

Wildrnan T~l l

Late Ple~stoceneTill

Pre-Late Pleistocene T~ l l

/L Dra~nageBasin Boundary

Stream

Contact

rg Rock Glacier Deposit

Numbers Refer to Site Localit~es

0 2 4 k m - 91$gJ,Dog's Hill

FIGURE2. General distribution of Late Quaternary tills in the study area (modified from Bur-rows [I9751 and Mabin [1984]).

TABLE2 Selected field and laboratory dataa

Es t~mated S i - Hor i - Depth Color Strut- Con- > Sand S i l t Clay pH NaF Fed Feo Feo/ Organic age o f t e zon (cm) (d ry ) t u r e s i s - 2mn <2mn ~ 6 2 . 5 ~53.9fi pH (%) (%) Fed Carbon

t i L L ( y r B P ) tence (%) 2 6 2 . 5 ~ > 3 . 9 ~ (2 min) ( X ) ( % o f <2mn f r a c t i o n )

2800 17 A 0-10 10YR 6/3 2,m,gr ss,p 0.5 28.1 49.2 22.7 4.1 9.9 1.06 0.10 0.09 4.73 E j Bu 2Coxl 2Cox2

10-26 2.5Y 6.5/4 26-41 lOYR 6/4 41 -56 2.5Y 5/4 56-125+ 5Y 6/4

l,m,sbk 2,c,abk sg sg

S S , ~ 0.4 12.5 s,p 0.8 35.6 ss,ps 65.1 83.5 so,po 75.4 93.7

62.5 44.0

8.7 4.3

25.0 20.4

7.7 2.0

4.5 10.6 4.8 10.9 5.1 10.8 5.3 10.6

1.20 1.28 0.51 0.46

0.94 1.00 0.53 0.54

0.78 0.78 1.04 1.17

3.10 2.32 2.35 0.76

>9520 19 A Buj Bul Bu2 2Coxl 2Cox2 2Cox3

0-22 2.5Y 5/3 22-29 2.5Y 7/6 29-37 lOYR 7/6 37-82 2.5Y 7/6 82-86 2.5Y5/4 86-103 5Y 6/4

103-120+ 5Y 6/4

2,m.abk 2,m,sbk 3,c,abk 3,c,abk l,m,sbk 2,c,abk sg

ss,ps 0.9 76.6 ss,ps 1.1 42.3 s,p 0.3 25.8 s,p 0.1 81.9 s s , p 3 3 . 1 45.2 ss,po 4.4 90.6 so,po 48.3 90.2

16.5 45.1 57.4 17.9 13.4 8.3 7.5

6.8 12.6 16.8 0.2

41.4 1.1 2.8

4.8 10.2 0.50 0.36 4.9 10.1 0.92 0.81 4.5 9.7 1.90 1.68 5.1 11.0 1.71 1.69 4 . 9 - - - - 0 . 5 5 0 . 2 7 5.0 10.5 0.21 0.12 5.2 10.2 0.20 0.25

0.72 0.88 0.88 0.99 0.49 0.57 1.25

2.88 2.71 2.20 1.81 1.19 0.58 0.55

211,900 65 A 0-17 2.5Y5/3 Bu 17-26 10YR6/3 Coxl 26-52 2.5Y 6/4 Cox2 52-83 2.5Y 7/4 Bub 83-103 10YR 6/4 Coxlb 103-150 2.5Y 7/4 Cox2b 150-175 2.5Y 6/5 2Cox3b 175-200+ 5Y 7/4

>11,900 72 A Buj Bu 2Coxl 2Cox2

0-20 2.5Y 5/2.5 20-44 2.5Y7/5 44-78 2.5Y 7/7 78-95 2.5Y 7/4 95-120+ 5Y 8/3.5

2.5,c.gr 2,c.abk 2.5,c,abk 2,f ,sbk l , f , sbk

ss,ps s,ps s,p ss,po so,po

3.0 2.5 1.6

70.0 62.8

33.9 25.8 25.5 62.3 74.1

36.2 37.3 43.5 23.0 18.2

29.9 36.9 31.0 14.7 7.7

4.8 9.0 5.1 10.3 5.0 10.6 5.1 10.6 5.1 10.1

1.08 0.72 1.380.91 1.54 1.01 0.51 0.42 0.26 0.15

0.67 0.66 0.66 0.82 0.58

3.67 2.12 1.36 0.69 0.37

240,900 77 A Coxl Cox2 2Cox3

0 -9 2.5Y 9-44 2.5Y

44-70 2.5Y 70-120+ 2.5Y

6/3 7/4 4/5 6/4

2, f ,gr l , f , sbk 2,c,abk sg

s s , p ss,ps s,ps ss,po

1.4 72.3 12.7 66.8 4.0 46.2

34.2 74.1

22.4 20.9 35.9 18.2

5.3 12.3 17.9 7.7

4.6 9.0 4.9 9.1 5.0 10.2 5.3 10.1

0.84 0.37 0.990.49 0.76 0.22 0.91 0.50

0.44 0.49 0.29 0.55

2.02 1.92 0.48 0.68

>40,900 91 A 0-15 2.5Y 5/2 E 15-62 5Y 6/2 Cox 62-73 2.5Y 7/3 Ab 73-82 2.5Y 7/2 E j b 82-88 5Y 7/3 Coxlb 88-104 2.5Y 7.5/4 2Cox2b 104-115+ 5Y 8/3.5

2,m,gr 2,c.abk 3,vc.abk 3,vc.abk 3,vc,abk 3,vc,abk 3,c,sbk

ss,ps s,ps s,p vs,vp vs,p s,p s,p

0.7 19.0 2.9 15.5 4.5 19.0 5.3 19.6 7.7 32.4 1.1 40.8

11.4 61.5

48.0 49.6 60.1 26.9 47.1 48.8 30.3

17.3 35.0 20.9 53.5 20.5 10.4 8.2

5.0 9.6 4.8 10.3 5.1 10.3 5.0 10.2 5.2 10.0 5.1 10.1 5.3 10.1

1.01 0.35 0.87 0.50 0.59 0.95 0.55

0.72 0.23 0.71 0.37 0.39 0.72 0.36

0.71 0.66 0.82 0.74 0.66 0.76 0.65

4.54 3.12 1.38 1.44 0.74 0.84 0.37

V a r e n t ma te r ia t s are Loess, o r mixed loess and t i l t , and t i l l (2)

iron (Fe,) was measured on samples free of magnetic minerals (Walker, 1983), and amounts of Fe, and dithionite-citrate-bicarbonate extractable iron (Fed)were determined using a spectrophotometer.

NaF pH is a relative measure of the reactive aluminum components in a soil (Fieldes and Perrott, 1966; Neall and Paintin, 1986). ECDAM (exchange complex domi-nated by amorphous material) is present if NaF pH ex-ceeds 9.4 in 2 min (Neall and Paintin, 1986). Fedis con-sidered to be the total pedogenic Fe, whereas Fe, is amorphous Fe oxide (mostly ferrihydrite) and organically complexed Fe (Parfitt and Childs, 1988). The ratio of Fe, to Fed, termed the Fe activity ratio (Blume and Schwertmann, 1969), indicates the degree of Fe oxide crystallinity with lower values reflecting a more crystalline Fe oxide composition (McFadden and Hendricks, 1985).

Several methods have been used to illustrate the varia-tion in chemical properties with time. The index of pro-file anisotropy (IPA) quantifies the anisotropic vertical distribution of soil chemical properties (Walker and Green, 1976). It assumes that at the time of deposition there was no chemical variability with depth in the parent material, and that with time the soil becomes pedogeni-cally layered and chemical properties become anisotropic

with depth. The IPA as modified by Birkeland (1984b) was used in this study:

where D is the difference between a horizon property and that of the parent material (PM). For profile values, indi-vidual values are multiplied by horizon thickness, summed, and the latter divided by the thickness of the profile. Another index used in this study is the weighted mean percentage, in which horizon values are multiplied by horizon thickness, summed, and the latter divided by the thickness of the profile.

Linear and nonlinear regression analyses were applied to PDI, rubification, mIPA, and weighted mean data from the eight soils on Holocene moraines in Table 2 in order to determine whether rates of property development are most closely modeled by linear, exponential, power, or logarithmic functions. Models yielding the highest cor-relation coefficients were chosen as the best chronofunc-tion following Bockheim (1980), Levine and Ciolkosz (1983), Birkeland (1984a), and McFadden and Hendricks (1985).

SOILS ON HOLOCENE MORAINES

For Holocene soils, there is a progressive change in pro-file development with age of moraine (Table 2). Soils on Arrowsmith I11 and I1 moraines (ca. 75 to 100 yr BP) have little pedogenic development except for thin (<5 cm) A horizons, thin (<10 cm) incipient Cox horizons with a maximum color development of 7.5Y 8/2. Arrowsmith I moraines (ca. 250 yr BP) have thicker (ca. 10 cm) and darker A horizons, and slightly redder Cox horizons (5Y 7/3). Soils on all Arrowsmith moraines are classified as Cryorthents (Soil Survey Staff, 1975). Marquee moraines (ca. 2500-3000 yr BP) are the youngest moraines with subsurface horizon hues of 2.5Y and redder; these are considered Bw horizons, and the soils are classified as Typic Cryochrepts. Soils on U'ildman moraines (ca. 9500 yr BP) have a distinct Bw horizon with a lOYR hue and high chroma, and are classified as either Typic Cryum-brepts or Typic Cryochrepts, depending on the thickness of the epipedon.

Loess thickness typically increases with age of deposit and is the parent material for both A and B horizons of mid- and early-Holocene deposits (Figure 3; Table 2). Loess is recognized by material composed of 60 to 80% silt plus clay, and virtually no material greater than 2 mm (Table 2). The lack of buried soils within loess on Holocene soils suggests that loess deposition has been more or less continuous. Furthermore, the absence of buried soils at loess-till boundaries suggests either that loess deposition closely followed moraine deposition or that buried soils, if they existed, were masked by later pedogenesis.

PDI and rubification curves show a clear relationship with age of moraine, and both curves are quite similar (Table 3; Figure 4). Moreover, the curves indicate that field properties in this area develop faster than those in either the Mount Cook area or the Ben Ohau Range.

Pleistocene late- pre-lale-

200 1 - 1 I Moraine n=33 I I

o Terrace n=32 1 . 1 I I

150 I I

Estimated Age (yr BP)

FIGURE3. Loess thickness on Holocene moraines and fluvial terraces (data from Table 2 and Rodbell [1986]).

TABLE3 PDI, rubification, mIPA (I), and weighted mean percent data (IiJ

- ---- --

Organic Fed Fe, NaF pH carbonEstimated age PH -

(year BP) Site PDI Rubification I I1 I I1 I II I I I1 --.- ---

i+ :hts study

hlt Cook area

lZ5 0 BenOhauRange

0.4 -

I 0 I

A I I I I I

t thls study I r ~t cook area I

3 0 BenOhauRange I 0.0,

Estimated Age (yr BP) Estimated Age (yr BP)

FIGURE 4. Plots of rubification and PDI with estimated moraine FIGURE5 . Plots of mIPA pH, NaF pH and weighted mean NaF age. Data plotted to the right of the dashed line are for soils pH with estimated moraine age. Data plotted to the right of on pre-Holocene moraines, and those from the Mount Cook the dashed line are from soils on pre-Holocene moraines and area and Ben Ohau Range are from Birkeland (1984a). The loga- those from the Mount Cook area and Ben Ohau Range are from rithmic curves are based on data from the eight Holocene soils Birkeland (1984a). The logarithmic curves are based on data in this study area. from the eight Holocene soils in this study area.

Laboratory properties also show the predicted change with time. Moraines estimated at 250 yr BP and older meet the criteria for ECDAM (Table 2), and data for both mIPA and weighted mean result in very similar curves (Table 3; Figure 5). Data for distilled-water pH also result in a similar curve and, moreover, further suggests that pedogenic development in this area is faster than at Mount Cook and the Ben Ohau Range (Figure 5).

Data for the two iron extracts suggest both a close rela- tionship with age of deposit and relatively rapid rates of pedogenesis in this area (Table 3; Figure 6). Fe, and Fed data yield nearly identical curves for both mIPA and weighted mean percent plots, and virtually all mIPA data from this study lie above that for the Mount Cook area and the Ben Ohau Range. The weighted mean percent plots indicate that by ca. 2500-3000 yr BP, Fe concen- trations have reached their Holocene maxima, and older soils do not show appreciably greater Fe concentrations. Of particular interest is the Fe data for site 10 (ca. 2500 yr BP) which has the highest percent Fe of any soil in the Holocene chronosequence (Tables 2, 3). The presence of an Ej horizon and of high organic carbon and NaF pH at depth suggests, unlike other sites, podzolization may account for the high Fe values.

Fe activity ratios generally fall between 0.75 and 1.0 and, although there is a general trend for higher values with depth in individual profiles, there is no apparent trend with age of Holocene moraine (Table 2). Seven horizons, with low percent Fed and Fe,, yield Fe activity ratios greater than 1 .O, and these probably represent ana- lytical errors in Fe detection (Parfitt and Childs, 1988). Inasmuch as Fe activity ratios are similar for all Holocene soils (Table 2), it is apparent that the gradual transforma- tion of ferrihydrite to more crystalline forms of Fe ox- ide, such as goethite, is slow (>10 ka), and/or ferri- hydrite is being produced at a rate equal to the rate that it is being transformed to goethite (McFadden, 1988). Alternatively, the relatively rapid release of Fe (Figure 6), and high organic carbon content with depth on mid- Holocene and older soils (Table 2) may cause ferrihydrite to be the stable Fe species for Holocene soils in this area (Birkeland 1984b; McFadden and Hendricks, 1985).

With the exception of weighted mean percent organic carbon, the progressive development of all the field and laboratory properties considered are best described by logarithmic functions (Table 4). Correlation coefficients for weighted mean percent organic carbon indicate that a power function (r2 =0.73) is an equally appropriate model as a logarithmic function (r2 =0.72). Rubification, PDI, Fe, and Fed yield r2 values in excess of 0.9. Weighted mean percent data yield lower r Zvalues than do mIPA data, except Fed (Table 4). Thus, this data support the conclusions of Bockheim (1980), Levine and Ciolkosz (1983), and McFadden and Hendricks (1985) that logarithmic functions most appropriately model the development of many soil properties with time. More- over, Reheis et al. (1989) conclude that although pedo- genic properties influenced by CaC03- and salt-rich dust may proceed at a linear rate, those indicative of chemical

weathering are best portrayed by logarthmic models. Birkeland et al. (1989) have shown that soils in the

Southern Alps develop faster than in other arctic and alpine areas, and suggested that pedogenic gradients follow regional climate with fastest pedogenesis occur- ring in the warmest and wettest areas. This study sup- ports the above conclusions, and indicates that, with the exception of organic carbon, soil development in this area is faster than heretofore recognized in the eastern South- ern Alps (Birkeland, 1984a).

Inasmuch as both the mIPA and weighted mean per- cent curves for most properties have little slope at ca. 10 ka, it is apparent that soils here approach a steady state in as little as 10 ka. In contrast, Birkeland (1984a) found that most curves of soil property development in the Ben

8 I

+ Fe, this study I

X Fed thlS Study 0 I

A Fe, Mt Cook area I

0 2000 4000 6000 8000 10000

Estimated Age (yr BP)

FIGURE6. Plots of mIPA Fe, and Fe,, and weighted mean per-cent Fe,, Fe, and organic carbon with estimated moraine age. Data plotted to the right of the dashed line are from soils on pre-Holocene moraines and those from the Mount Cook area and Ben Ohau Range are from Birkeland (1984a). The logarith- mic curves are based on data from the eight Holocene soils in this study area.

- - -- - - - - --

- -- - - --- -- - -

- - - -- - - -- - -

- - - ---

- - - - -- - -

TABLE4 Correlation coefficients of exponential (I), h e a r (11), power (111), and logarithmic (IV) equations

for various soil property versus time relationships

Property

Field Properties Rubification PDI

Lab Properties mIPA

FeO Fed PH NaF pH Organic carbon

Weighted Mean Percent Fed Fed NaF pH Organic carbon

-

Correlation Coefficients ( r 2 ) d

I I I 111 IV - -

0.50 0.70 0.87 0.98 0.45 0.57 0.80 0.91

0.51 0.66 0.86 0.97 0.14 0.61 0.24 0.87 0.31 0.38 0.63 0.78 0.28 0.46 0.60 0.87 0.30 0.30 0.67 0.75

0.53 0.64 0.87 0.96 0.45 0.56 0.77 0.90 0.48 0.50 0.78 0.80 0.34 0.30 0.73 0.72 -

Solution to logarithmic equation"

a b

-77.89 40.08 -0.18 0.15

-3.35 2.30 -2.47 1.72 -0.19 0.12 -0.22 0.12 -9.80 5.89

-0.26 0.24 -0.23 0.24 6.21 1.16

-0.99 0.67

"r2 values that exceed 0.76 are significant at the 0.001 level. k and b are constants in the logarithmic function Y = a + blogX, where X is time and Y is soil property.

Ohau Range and Mount Cook area still have measurable rates of change at ca. 10 ka and are best modeled by power functions.

The relatively rapid rate at which soils in this study area approach a steady state is not easily explained. One ex- planation is that it may be due to the combination of a relatively high MAP (3200 mm) and a relatively high MAT (8.j°C). The Mount Cook area is wetter (MAP = 4000 mm) but colder (MAT = 7S°C), and the Ben Ohau Range is drier (MAP = 1000 mm) and colder (MAT = 4"C), and data points for both these areas gen- erally fall well below those for this study area (Figures 4-7). However, it seems unlikely that the small difference in MAT between the Mount Cook area and this study area can explain the large difference in soil development. A second explanation for the relatively rapid rate of soil

development in this study area is that here loess comprises a greater proportion of the soil parent material than in either the Ben Ohau Range or the Mount Cook area. Loess influences soil development in that because it has a high surface area per unit volume it weathers rapidly and, where it is sufficiently thick, most pedogenesis is restricted to the loess rather than extending into the underlying till. Furthermore, although it is possible that the age control used in these studies is insufficient to discriminate between the true rates of pedogenesis for the three areas, it is unlikely that the ages used here are in error by an amount large enough to change the shape of the curves. Finally, variations in bedrock composition between the three study areas, and microclimatological differences between individual sites may have influenced the shape of the curves and more data points are needed.

SOILS ON PRE-HOLOCENE MORAINES

Whereas it is usually possible to find soils on stable parts of all Holocene moraines that seem to reflect the total time that pedogenesis has taken place, soils on pre- Holocene moraines do not show a close developmental relationship with age. PDI and rubification plots for pre- Holocene soils indicate that only data from site 65 plot above that for the early Holocene (Table 3; Figure 4). In-dices of pH and organic carbon for pre-Holocene mo- raines all plot below those for the early Holocene, and only the indices of Fe from site 65 plot appreciably above those of the early Holocene (Figures 5-7). Weathering- rind data also does not reflect pre-Holocene ages for these moraines, as the rinds are not thicker than those on early Holocene moraines (Rodbell, 1986).

The lack of soil and weathering rind trends continu-

ing beyond the Holocene are likely the result of episodic erosion and deposition of loess on pre-Holocene sites. Evidence from the present study indicates that loess deposition has been episodic; there may have been times when it was stripped from parts of the landscape. Buried soils within loess on pre-Holocene deposits record the episodic nature of loess deposition. In this area, buried soils are relatively common where loess is thick (> 100 cm) and uncommon where loess is thin (<50 cm) (Rodbell, 1986). In the former case, buried soils become isolated from later pedogenesis, whereas in the latter, buried soils, if initially present, may have been masked by subsequent pedogenesis to form a composite soil. Buried soil recognition is best where buried A horizons are present (i.e., site 91, Table 2), however in many places

such A horizons are not present and buried soil recogni- tion has been based on reversals in down-profile rubifica- tion and/or organic carbon, pH, and NaF pH trends, such as at site 65 (Table 2).

Soil development on pre-Holocene moraines, therefore, does not reflect the total amount of pedogenesis since moraine deposition. When loess is present, pedogenesis takes place within it, but when loess is eroded, the evidence for pedogenesis is partly lost, and subsequent erosion of the underlying till can then erase still more evidence of pedogenesis. Pre-Holocene moraines appear to have been subject to more frequent wind erosion dur- ing the Holocene than their Holocene counterparts. This is perhaps due to their more wind-exposed locations, and results in soil-property development that is generally less than that on Holocene moraines.

Radiocarbon dates from charcoal within loess from soil profiles 65 and 91 (Figure 8; Table 2) provide age con- trol for some of the loess erosion/deposition events on pre-Holocene moraines. Charcoal in a buried A horizon underlying approximately 82 cm of loess on a pre-late Pleistocene moraine (site 91) yielded a radiocarbon date of 6940 + 150 yr BP (NZ 6810). The buried soil is formed in approximately 30 cm of loess overlying at least 15 cm of slightly oxidized till. Field and laboratory properties (Table 2) suggest that the buried soil could have formed in ca. 2500 yr. Thus, at this site, two Holocene soils have formed in two loess sheets which in turn overlie slightly weathered (5Y hue) till; the till, on the basis of a radiocar- bon date from charcoal overlying it, is dated older than 40,900 yr BP (NZ 1684; Harvey, 1974). At site 65, ap- proximately 175 cm of loess overlie slightly oxidized (5Y

I 0 I

+ this study I

A Mt Cook area

0 Ben Ohau Range I I I I I I I

0 I 0 I

C--H

O 0 h 0 1 + A

m l * A I +

I I

Estimated Age (yr BP)

FIGURE7. Plot of mIPA organic carbon with estimated moraine age. Data plotted to the right of the dashed line are from soils on pre-Holocene moraines and those from the Mount Cook area and Ben Ohau Range are from Birkeland (1984a). The logarith- mic curve is based on data from the eight Holocene soils in this study area.

hue) till. Field and laboratory data of the surface soil sug- gest a mid-Holocene age. Charcoal from 26-52 cm depth yielded a radiocarbon date of 2180+ 100 yr BP (NZ 6808), and that from 83-103 cm depth yielded a date of 5240+ 110 (NZ 6803) (Figure 8). If one assumes a linear rate of loess deposition, an average rate of slightly less than 18 cm/1000 yr for the last ca. 5200 yr is obtained from site 65, and one slightly less than 12 cm/1000 yr for the last ca. 7000 yr is obtained from site 91 (Figure 8). Although field and laboratory trends show reversals indicative of the presence of a buried soil below 83 cm (Table 2), the assumption of a linear rate of loess deposi- tion may still be valid. The deposition of weathered loess, reworked by wind from other soils in the area may result in the appearance of a buried soil in an otherwise cumulic profile. The numerous nearby areas that contain little or no loess may have been sources for the reworked loess at this site. Rubification and Fe data for this site plot sub- stantially above those for the early Holocene (Figures 4, 6), thus supporting the hypothesis that the loess was weathered prior to deposition at this site. Such a phe- nomenon has been recognized in other New Zealand alpine areas (Harrison, 1982) and in the southern High Plains, U.S.A. (Holliday, 1989).

Most pre-Holocene moraines studied have undergone a similar process of episodic loess erosion, deposition, and soil formation. The depth of loess on many pre- Holocene moraines is variable (Figure 3). In places where nearly fresh till is exposed at the surface, erosion has removed most of the soil. Areas with thick loess cover, with or without buried soils, probably have been loess traps throughout much of the Holocene, and contain

Radiocarbon Age (yr BP)

120

FIGURE8. Radiocarbon age and depth of buried charcoal at sites 65 and 91. Age error bars represent a plus-minus of two sigma, and the dashed lines through the dates represent a linear rate of loess deposition of ca. 18 cm/1000 yr.

complex cumulic soils, similar to the polygenetic soils of Harrison (1982). In nearby Paddle Creek drainage, Harvey (1974) concluded that where vegetation is lack- ing on upland sites, soils erosion is enhanced by two pro- cesses. One is raindrop impact followed by transport, and the other is needle ice which, on melting and drying, leaves a low bulk density surface very susceptible to subse- quent wind erosion. This is especially true for wind- exposed upland sites which include the crest sites of most pre-Holocene moraines. This complex history of loess erosion and deposition greatly hinders the use of soils to estimate ages of pre-Holocene moraines in this region of the eastern Southern Alps. In contrast, soil development has been used as a criteria for discriminating between pre- Holocene tills near Lake Tekapo, New Zealand (Webb, 1976; McGregor, 1981), and the success of these studies

is not easily explained. Large-scale "pedospheric stripping" of loess during the

Pleistocene has been recognized on the Canterbury Plains and southern South Island, and has been attributed to widespread natural fires, needle-ice formation and per- sistent high winds (Bruce, 1973; Ives, 1973; Tonkin et al., 1974). Whereas large-scale loess erosion and redeposition are no longer widespread on the Canterbury Plains (Ives, 1973), results from this study indicate that it has been prevalent throughout the Holocene in this region of the eastern Southern Alps. The abundance of fossil charcoal in other areas (Molloy et al., 1963) suggests that in the alpine zone of Canterbury, similar to the Canterbury Plains during the Pleistocene, periodic fires, persistent winds, and needle-ice formation have caused widespread loess reworking over at least the last 10 ka.

CONCLUSIONS

Holocene soil chronofunctions for the Arrowsmith Range are best modeled by logarithmic functions and in- dicate that soil development in this area is faster than that in either the Mount Cook area or the Ben Ohau Range. This is due in part to a thicker loess cover in this study area and to climatic differences between the three areas, and it complicates the use of soils as a means of cor- relating between Holocene moraines in the eastern Southern Alps.

The use of soils as a means of estimating ages of pre- Holocene moraines is hindered by episodic loess erosion

and redeposition, and the result is that soils may be as much as an order of magnitude younger than underlying tills. The presence of Holocene charcoal in loess on pre- Holocene moraines suggests that fires, as well as high wind speeds and frost action, have been responsible for widespread loess reworking throughout the Holocene on these sites. Present environmental conditions on upland pre-Holocene sites in this area may be similar to those that existed on the Canterbury Plains during the late Pleistocene.

ACKNOWLEDGMENTS

This research was conducted as part of my MS thesis under the supervision of Dr. P. W. Birkeland, University of Colorado. I am grateful to Dr. Birkeland for his invaluable guidance throughout all stages of this research. I am greatly appreciative of Dr. P . J. Tonkin for providing the radiocarbon dates, P . J. Hall for assistance in the field, J. B. J. Harrison and M. Flanagan for kindly providing a place for me to live while in Christchurch, the Thorndycrofts of Mount Somers village for

great food and accommodations, and J. White of Hakatere sta- tion for allowing me to work on his land and use his huts. I wish to thank Drs. P . W. Birkeland, A. Werner, M. Litaor, J. B. Benedict, and J . T. Andrews for critically reviewing an earlier draft of this manuscript. This manuscript greatly benefitted from critical reviews by Drs. L. R. Basher and E. T. Karlstrom. Funding for this research was provided by Sigma Xi, the Atlantic Richfield Foundation, and the University of Colorado.

REFERENCES CITED

Basher, L. R., 1986: Pedogenesis and erosion history in a high rainfall, mountainous drainage basin -Cropp River, New Zealand. Ph.D. dissertation, University of Canterbury, New Zealand. 494 pp.

Birkeland, P . W., 1984a: Holocene soil chronofunctions, Southern Alps, New Zealand. Geoderma, 34: 115-134.

, 1984b: Soils and Geomorphology. New York: Oxford University Press. 372 pp.

Birkeland, P . W., Burke, R. M., and Benedict, J. B., 1989: Pedogenic gradients for iron and aluminum accumulation and phosphorus depletion in arctic and alpine soils as a function of time and climate. Quaternary Research, 32; 193-204.

Blume, H. P. and Schwertmann, U., 1969: Genetic evaluation of profile distribution of aluminum, iron, and manganese ox- ides. Soil Science Society of America Proceedings, 33: 438-444.

Bockheim, J . G., 1980: Solution and use of chronofunctions in studying soil development. Geoderma, 24: 7 1-85.

Bruce, J. G. , 1973: Loessial deposits in southern South Island, with a definition of Stewarts Claim Formation. New Zealand Journal of Geology and Geophysics, 16: 533-548.

Burrows, C. J., 1973: Studies on some glacial moraines in New Zealand-2. New Zealand Journal of Geology and Geophysics, 16: 831-835.

, 1975: Late Pleistocene and Holocene moraines of the Cameron Valley, Arrowsmith Range, Canterbury, New Zealand. Arctic and Alpine Research, 7: 125-140.

Calkin, P . E. and Ellis, J. M., 1980: A lichenometric dating curve and its application to Holocene glacier studies in the central Brooks Range, Alaska. Arctic and Alpine Research, 12: 245-264.

Campbell, A. S., 1975: Chemical and mineralogical properties

of a sequence of terrace soils near Reefton, New Zealand. Ph.D. dissertation, Lincoln College, New Zealand. 447 pp.

Chinn, T. J. H., 1981: Use of rock weathering rind thickness for Holocene absolute age-dating in New Zealand. Arctic and Alpine Research, 13: 33-45.

Fieldes, M. and Perrott, K. W., 1966: The nature of allophane in soils. Part 3-Rapid field and laboratory test for allophane. New Zealand Journal of Science, 9: 623-629.

Gair, H. S., 1967: Sheet 20Mt. Cook, GeologicalMap of New Zealand (1:250,000). Wellington, Department of Scientific and Industrial Research.

Gellatly, A. F., 1982: Lichenometry as a relative-age dating method in Mount Cook National Park, New Zealand. New Zealand Journal of Botany, 20: 343-353.

,1985: Phosphate retention: relative dating of Holocene soil development. Catena, 12: 227-240.

Harden, J . W., 1982: A quantitative index of soil development from field descriptions: examples from a chronosequence in central California. Geoderma, 28: 1-28.

Harden, J. W. and Taylor, E. M., 1983: A quantitative com- parison of soil development in four climatic regimes. Quater-nary Research, 20: 342-359.

Harrison, J. B. J., 1982: Soil periodicity in a formerly glaciated drainage basin, Ryton Valley, Craigieburn Range, Canter- bury, New Zealand. MS thesis, Lincoln College, New Zealand. 189 pp.

Harvey, M. D., 1974: Soil studies in a high country catchment -Paddle Creek, South Canterbury. MS thesis, Lin- coln College, New Zealand. 193 pp.

Holliday, V. T., 1989: The Blackwater Draw Formation (Quater- nary): A 1.4-plus-m.y. record of eolian sedimentation and soil formation on the Southern High Plains. Geological So- ciety of America Bulletin, 101: 1598-1607.

Ives, D., 1973: Nature and distribution of loess in Canterbury, New Zealand. New Zealand Journal of Geology and Geophysics, 16: 587-610.

Knuepfer, P. L. K., 1988: Estimating ages of late Quaternary stream terraces from analysis of weathering rinds and soils. Geological Society of America Bulletin, 100: 1224-1236.

Levine, E. R. and Ciolkosz, E. J., 1983: Soil development in till of various ages in northeastern Pennsylvania. Quaternary Research, 19: 85-99.

Mabin, M. C. G., 1984: Late Pleistocene glacial sequence in the Lake Heron Basin, mid Canterbury. New Zealand Jour- nal of Geology and Geophysics, 27: 191-202.

McFadden, L. D., 1988: Climatic influences on rates of soil development in Quaternary deposits of southern California. In Reinhardt, J , and Sigleo, W. J. (eds.), Paleosols and Weathering Through Geologic Time: Principles and Applica- tions. Geological Society of America Special Paper, 216: 153-177.

McFadden, L. D. and Hendricks, D. M., 1985: Changes in the content and composition of pedogenic iron oxyhydroxides in a chronosequence of soils in southern California. Quater-nary Research, 23: 189-204.

McGregor, G. R., 1981: Weathering characteristics of late Pleistocene tills. New Zealand Journal of Geology and Geophysics, 24: 107-113.

Miller, G. H. and Andrews, J. T., 1972: Quaternary history of northern Cumberland Peninsula, east Baffin Island, NWT, Canada Part VI: Preliminary lichen growth curve for Rhizocarpon geographicum. Geological Society of America Bulletin, 83: 1133-1138.

Mokma, D. L., Jackson, M. L., and Syers, J . K., 1973: Mineralogy of a chronosequence of soils from greywacke and

micaschist alluvium, Westland, New Zealand. New Zealand Journal of Science, 16: 769-797.

Molloy, B. P. J., Burrows, C. J . , Cox, J. E., Johnston, J. A, , and Wardle, P., 1963: Distribution of subfossil forest remains, eastern South Island, New Zealand. New Zealand Journal of Botany, 1: 68-77.

Neall, V. E, and Paintin, I. K., 1986: Rates of weathering of I4C-dated late Quaternary volcaniclastic deposits in the western United States. In Colman, S. M. and Dethier, D. P. , (eds.), Rates of Chemical Weathering of Rocks and Minerals. Orlando: Academic Press, 331-350.

New Zealand Geological Survey, 1972: Geological Map of New Zealand (1:1,000,000). Wellington, Department of Scientific and Industrial Research.

Parfitt, R. L. and Childs, C. W., 1988: Estimation of forms of Fe and Al: A review, and analysis of contrasting soils by dissolution and Mossbauer Methods. Australian Journal of Soil Research, 26: 121-144.

Reheis, M. C., Harden, J. W., McFadden, L. D., and Shroba, R. R., 1989: Development rates of late Quaternary soils, Silver Lake Playa, California. Soil Science Society of America Jour- nal, 53: 1127-1140.

Rodbell, D. T., 1986: The use of lichenometry, rock weather- ing and soil development to estimate ages of moraines and fluvial terraces in the upper South Branch, Ashburton Valley, South Island, New Zealand. MS thesis, University of Colo- rado. 147 pp.

Ross, C. W., Mew, G., and Searle, P. L., 1977: Soil sequences on two terrace systems in the North Westland area, New Zealand. New Zealand Journal of Geology and Geophysics, 20: 23 1-244.

Singer, M. J. and Janitsky, P.(eds.), 1986: Field and laboratory procedures used in a soil chronosequence study. U.S. Geological Survey Bulletin, 1648. 49 pp.

Soil Survey Staff, 1975: Soil Taxonomy. A Basic System of Soil Classification for Making and Interpreting Soil Surveys. Soil Conservation Service, Agricultural Handbook No. 436. Wash- ington, U.S. Department of Agriculture. 754 pp.

Sommerville, P., Mark, A. F., and Wilson, J. B., 1982: Plant succession on moraines of the upper Dart Valley, southern South Island, New Zealand. New Zealand Journal of Botany, 20: 227-244.

Sporli, K. B. and Lillie, A. R., 1974: Geology of the Torlesse supergroup in the northern Ben Ohau Range, Canterbury. New Zealand Journal of Geology and Geophysics, 17: 115-141.

Tonkin, P. J , Runge, E. C. A., and Ives, D. W., 1974: A study of late Pleistocene loess deposits, south Canterbury, New Zealand. Quaternary Research, 4: 21 7-23 1.

Walker, A. L., 1983: The effects of magnetite on oxalate and dithionite-extractable iron. Soil Science Society of America, 47: 1022- 1026.

Walker, P. H. and Green, P. , 1976: Soil trends in two valley fill sequences. Australian Journal of Soil Research, 14: 291-303.

Webb, T. H., 1976, Pedological studies of soils of the Tekapo Set in east Lake Pukaki region, South Canterbury, New Zealand. MS thesis, Lincoln College, New Zealand. 183 pp.

Whitehouse, I. E., McSaveney, M. J . , Kneupfer, P. L. K., and Chinn, T. J. H., 1986: Growth of weathering rinds on Torlesse sandstone, Southern Alps, New Zealand. In Colman, S. M. and Dethier, D. P. (eds.), Rates of Chemical Weathering of Rocks and Minerals. Orlando: Academic Press, 419-438.

MS submitted February 1990

You have printed the following article:

Soil-Age Relationships on Late Quaternary Moraines, Arrowsmith Range, Southern Alps,New ZealandD. T. RodbellArctic and Alpine Research, Vol. 22, No. 4. (Nov., 1990), pp. 355-365.Stable URL:

http://links.jstor.org/sici?sici=0004-0851%28199011%2922%3A4%3C355%3ASROLQM%3E2.0.CO%3B2-%23

This article references the following linked citations. If you are trying to access articles from anoff-campus location, you may be required to first logon via your library web site to access JSTOR. Pleasevisit your library's website or contact a librarian to learn about options for remote access to JSTOR.

References Cited

Late Pleistocene and Holocene Moraines of the Cameron Valley, Arrowsmith Range,Canterbury, New ZealandC. J. BurrowsArctic and Alpine Research, Vol. 7, No. 2. (Spring, 1975), pp. 125-140.Stable URL:

http://links.jstor.org/sici?sici=0004-0851%28197521%297%3A2%3C125%3ALPAHMO%3E2.0.CO%3B2-B

A Lichenometric Dating Curve and Its Application to Holocene Glacier Studies in the CentralBrooks Range, AlaskaParker E. Calkin; James M. EllisArctic and Alpine Research, Vol. 12, No. 3. (Aug., 1980), pp. 245-264.Stable URL:

http://links.jstor.org/sici?sici=0004-0851%28198008%2912%3A3%3C245%3AALDCAI%3E2.0.CO%3B2-G

Use of Rock Weathering-Rind Thickness for Holocene Absolute Age-Dating in New ZealandT. J. H. ChinnArctic and Alpine Research, Vol. 13, No. 1. (Feb., 1981), pp. 33-45.Stable URL:

http://links.jstor.org/sici?sici=0004-0851%28198102%2913%3A1%3C33%3AUORWTF%3E2.0.CO%3B2-Z

http://www.jstor.org

LINKED CITATIONS- Page 1 of 1 -