This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the authors institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier’s archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/copyright
Author's personal copy Geomorphology 109 (2009) 236–245
Contents lists available at ScienceDirect
Geomorphology j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / g e o m o r p h
A soil chronosequence in Late Glacial and Neoglacial moraines, Humboldt Glacier, northwestern Venezuelan Andes W.C. Mahaney a,b,⁎, V. Kalm c, B. Kapran a,b, M.W. Milner a,b, R.G.V. Hancock d,e a
Quaternary Surveys, 26 Thornhill Ave., Thornhill, Ontario, Canada L4J 1J4 Geography Department, York University, 4700 Keele St., N. York, Ontario, Canada M3J 1P3 Institute of Ecology and Earth Sciences, Tartu University, Tartu, EE51014, Estonia d Department of Medical Physics and Applied Radiation Sciences, McMaster University Hamilton, Ontario, Canada L8S 4K1 e Department of Anthropology, McMaster University Hamilton, Ontario, Canada L8S 4K1 b c
a r t i c l e
i n f o
Article history: Received 6 December 2008 Received in revised form 10 March 2009 Accepted 11 March 2009 Available online 18 March 2009 Keywords: Little Ice Age Neoglacial Late Glacial Late Pleistocene–Holocene soil chronosequence Humboldt Massif foreland
a b s t r a c t Late Glacial and Neoglacial (Little Ice Age) deposits on the Humboldt Massif were analyzed for relative-age dating parameters, including geomorphic and weathering characteristics, geochemical and soil properties. The soil chronosequence, formed in chemically uniform parent materials, provides an important database to study soil evolution in a tropical alpine environment. Extractable and total Fe and Al concentrations, examined to assess their use in relative-age determination, and as paleoenvironmental indicators, provide an important measure of the accumulation and downward profile movement over time of organically-bound Al, ferrihydrite and other crystalline forms (hematite and goethite) of extractable Fe. Ferrihydrite is particularly useful in determining former perched water levels in soils with relation to paleoclimate. The ratios of most Fe extracts are time dependent. The Fed/Fet ratio, within statistical limits, shows a slow increase from LIA (Little Ice Age) to Late Glacial soils, which closely correlates with other alpine soil studies in the middle latitudes and other tropical alpine locales. Values of Ald (dithionite) and Alo (oxalate extractable) generally do not correlate with time; however, Alp (pyrophosphate extractable) measured against Alt (total) provides insight on the downward translocation over time of organically-bound Al. Low leaching rates in this chronosequence are further supported by clay mineralogy trends and the geochemical data. © 2009 Elsevier B.V. All rights reserved.
1. Introduction Geomorphic characteristics, particularly the terminal positions of Late Glacial and Neoglacial ice and weathering rinds on pebbles, were used to establish relative ages of deposits in the progression on the Humboldt Massif. Soil chronosequences in middle latitude alpine and Arctic areas are rare (Costin, 1955; Mahaney and Fahey, 1988; Karlstrom, 1988; Birkeland et al., 1989; Birkeland, 1999; Mahaney et al., 1999, Evans, 1999; Bockheim et al., 2000; Egli et al., 2001; Tonkin and Basher, 2001; Egli et al., 2003; Kovaleva, 2006), and those in tropical mountain areas (Mahaney, 1990; Mahaney et al., 1994, 2000) are rarer still. In some of these studies, Fe and Al extracts were used, along with the occasional use of geochemical ratios, to elicit information regarding age, paleoclimate, and parent material. While certain problems exist with sodium dithionite and acid ammonium oxalate extracts of Al (Birkeland et al., 1989), the ratio of Alp/Alt (pyrophosphate extractable/total) provides important information on
⁎ Corresponding author. Quaternary Surveys, 26 Thornhill Ave., Thornhill, Ontario, Canada L4J 1J4. E-mail addresses:
[email protected] (W.C. Mahaney),
[email protected] (V. Kalm),
[email protected] (R.G.V. Hancock). 0169-555X/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.geomorph.2009.03.005
the downward movement of organically-complexed Al over time (Parfitt and Childs, 1988). Moreover, geochemical analysis of the parent material ensures a means of testing for uniformity among the sites/profiles examined in the Humboldt foreland and to assess gains and losses related to airfall-influx and leaching. An initial test (Mahaney et al., 2000) of Fe and Al extracts, measured at two sites on the Humboldt Massif ranging from Little Ice Age (LIA) to Late Glacial age, is here expanded to include multiple sites, incorporating the analysis of an additional site dating from the early/middle Neoglacial. Pre-LIA Neoglacial moraines are scarce, as LIA glacial deposits on the Humboldt Massif have overrun earlier Neoglacial deposits. Besides comparison of the Humboldt with other middle latitude and tropical alpine areas, we test the ability of various ratios and arithmetic functions of Fe and Al extracts, and geochemical ratios to subdivide glacial deposits and reconstruct paleoclimate. 2. Regional geology The Humboldt Massif lies along the crest of the eastern cordillera of the northwestern Venezuelan Andes between 8° 30′ and 9° 00′ N and 70° 30′ and 70° 45′ W, at elevations close to 5000 m a.s.l. (Figs. 1 and 2). In more detail, the massif is part of the Sierra de Santo Domingo, the northern extension of the Sierra Nevada de Mérida.
Author's personal copy W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
237
moraine. Cross-valley bedrock bars are spaced every ~ 100 m above the Boconó Fault at ~ 3700 m a.s.l., forming steps covered with either end moraines/outwash bodies or copious amounts of talus (Mahaney et al., 2007). The moraine/outwash complex of Late Glacial/Younger Dryas to LIA age (Mahaney et al., 2008) situated at N4000 m a.s.l. contains the sediments described in this report. The LIA moraines and a reconnaissance study of the Late Glacial deposits is contained in Mahaney et al. (2000). In summary, the Late Glacial is documented as having an age in excess of ~ 12 ka (uncalibrated 14C yr) and the Younger Dryas is dated to between ~ 10 and 11 ka (uncalibrated 14C yr). The Early Neoglacial, a single entity known only in the Humboldt foreland, has not been given formation rank, and is undated by radiocarbon. The Little Ice Age (LIA) is documented as b500 yr by mutually consistent uncalibrated radiocarbon ages (site COR3; Mahaney et al., 2000). 3. Geomorphic-environmental background Fig. 1. Aerial photo imagery of the Humboldt Massif.
Underlying bedrock in the area consists of gneiss and granite of the Iglesias Group of Precambrian age (Schubert, 1970). In the Santo Domingo segment, the lithology is divided into two main formations, of which the La Mitisus Banded Gneiss is the oldest unit cropping out in most valleys. The gneiss and minor accompanying schist are often intruded with granitic and quartz dikes. Most cirques, as in the Upper Mucuñuque (for location see Mahaney et al., 2008) and Coromoto valleys (Fig. 2), exhibit lagoons carved in bedrock or dammed by moraines. Pico Humboldt, at the head of the Coromoto catchment, is the top-most summit at 4940 m a. s.l. The high crest of the range stretches along a narrow ridge with arêtes, horns, and steep walls extending downward across talus into bedrock-floored valleys laced with thin, discontinuous ground
The linear fault-controlled Coromoto catchment [all samples abbreviated COR] with headwaters on Pico Humboldt (Fig. 2), trending NW–SE, is similar to other catchments along the west flank of the Range. Bogs are prevalent on the valley floors. Together with alluvial fill, ground and recessional/end moraines, and prominent Nye channels, bogs constitute the only other source materials for soils. The alpine soils, belonging to the Histosol and Entisol orders (NSSC, 1995; Soil Survey Staff, 1999), exhibit udic moisture regimes with no dry period over 90 days, and with lighter than 10YR 5/1 colors, exhibit ochric epipedons. All soils, with the exception of the Histosols, key out in the Orthent and Ochrept sub-orders as Troporthents and Tropochrepts. This is a function of the high altitude sub-tropical locality. Previous research in the area (Mahaney et al., 2000, 2008) with soils focused on deglaciation and LIA advances in the Humboldt Massif area. This later work established tentative chemical differences
Fig. 2. Surficial map of the Humboldt foreland. COR13 is on a recessional LIA moraine; COR3 (Mahaney et al., 2000) is the outer limit of the LIA off the Bonpland Lobe and is considered correlative with COR10 located on a lateral moraine of the northern lobe; COR15 is the Early Neoglacial limit below the Bonpland Lobe, previously unrecognized; COR6 and 7 are the Late Glacial limits in the Humboldt Foreland.
Author's personal copy 238
W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
between soils with similar-appearing profiles, specifically variations in Fe and Al extracts, between Neoglacial and Late Glacial soils. The older Late Glacial soils may in some instances have profile depths and characteristics similar to each other and to older LIA soils, yet their chemical compositions indicate a considerable difference in age. Meteorological data for the ~ 4000-m a.s.l. foreland of the Humboldt Glacier is not known with precision but the mean annual precipitation is estimated at 1 m. Based on extrapolations with height above Mucubaji (3600 m a.s.l.) the mean annual temperature is estimated at 5 °C. There is little seasonal fluctuation in temperature although diurnal fluctuation of 15–20 °C is normal (Mahaney et al., 2007; Servicio Meteorologico de la Fuerza, 1975). 4. Methods Rinds on pebbles, collected from moraine surfaces, were measured to the nearest mm with both maximum and minimum thicknesses recorded (Mahaney, 1990). Rinds indicate the degree to which Fe-bearing minerals oxidize and discolor the outer periphery of clasts, the thickness measured perpendicular to the pebble surface. The thickness of a rind is an estimate of time since deposition (Birkeland, 1973; Kiver, 1974; Mahaney, 1990). Because most workers measure only the maximum thickness of discoloration, neglecting the irregular thickness found on many clasts, we measured both the maximum and minimum rind thicknesses. The maximum rind is defined as the maximum discoloration measured on the outer surface of a population of fifty pebbles of similar lithology, at each site. The minimum rind is the minimum thickness measured to the nearest mm. With gneiss there are varying degrees of internal discoloration of clasts and discolored fracture faces that penetrate the clast but these proved difficult to quantify. The soil descriptions follow guidelines set out by the Soil Survey Staff (1999) and the NSSC (1995). The Cox horizon designation originates from Birkeland (1999), while that of the Cu (unweathered parent material) is from Hodgson (1976). Soil color assessments are based on Oyama and Takehara's (1970) soil chips. Approximately 500 g samples were collected at the sites to allow for particle size, clay mineral, and geochemical analyses. Samples were air dried and treated with H2O2 to decrease organic material content. The materials were then wet sieved, and the b63 µm fraction was subjected to analysis by hydrometer (Day, 1965). Organic carbon was determined by loss on ignition. Samples of very fine to medium grade sizes were selected for routine analysis by Scanning Electron Microscope following procedures outlined by Mahaney (2002). The sands were pretreated for particle size analysis, subjected to light sonication to remove clay and with H2O2 to remove organic contaminants. The Fed and Ald extractions were made from 1 g (b2 mm fraction) sub-samples, with sodium dithionite, releasing crystalline, amorphous, and organically-bound forms of Fe and Al, as well as sodium citrate buffers, following criteria set out by Coffin (1963) and Dormaar and Lutwick (1983). Processed in the dark, acid ammonium oxalate was used to extract Feo (×1.7 = ferrihydrite; Parfitt and Childs, 1988) following McKeague and Day (1966), and Alo. Subsequently, concentrations of Fe and Al were determined using a Perkin-Elmer 373 atomic absorption spectrophotometer (AAS). Instrumental Neutron Activation Analysis (INAA) followed methods identified by Hancock (1978) to extract total Fe and Al in the b2 mm fraction of soil samples. The chemical matrix was determined by INAA using appropriate standards. The XRD analysis is based on procedures outlined by Whittig (1965).
and lateral moraines with respective lobes of ice issuing from the Humboldt Massif (Fig. 2). Some thirty sites were established to test the use of weathering rind and soil profile development to distinguish Neoglacial advances from Late Glacial deposits. Five representative sites of Neoglacial and Late Glacial age were selected for age discrimination from this population of sites. 5.2. Weathering rinds Weathering rinds (Table 1) were measured and studied in an attempt to establish differences in age between the sites (Figs. 1 and 2) in the chronosequence. Samples of felsic gneiss, from source areas in the Bonpland and northern lobes, were collected from moraine surfaces and split on-site with measurements made to the nearest mm. The data for the two LIA sites (COR13, COR10, Table 1) are typical of Late Neoglacial deposits in the tropical mountains, particularly when compared with Mt. Kenya (Mahaney, 1990) where rind counts with similar thicknesses were made on nepheline syenite. Outstanding features in the Humboldt data are the variations in low maximum rinds on the Late Glacial and Early Neoglacial deposits, diminishing on the younger LIA deposits. The nil minimum rind measured on clasts on LIA deposits is similar to what has been reported from other tropical Afroalpine sequences (Mahaney, 1990), and from middle latitude sites in the Rocky Mountains (Mahaney et al., 1999), where similar lithologies have been studied. The Late Glacial maximum rind values of ~2.0 mm correlate closely in thickness with measurements made on 14C-dated deposits in high tropical Afroalpine localities (Mahaney, 1990). The rinds, both maximum and minimum values, at the COR15 site (Fig. 2) lie between the maximum LIA and Late Glacial deposits, the values indicating an age within the Early Neoglacial (estimated at 2–3 ka). The first appearance of a minimum rind thickness from this population of data suggests nearly complete weathering around the clast surface for a minimum number of 2 pebbles (out of 50 measured), though with a variance dependent on the ability of water to cover the entire pebble circumference. The radius curvature of each clast turned downward into the surface soil exhibits the thinnest rind, a positive number lower than the maximum rind found on the upward radius curvature. In some cases, clasts lack any rind at all. The COR15 site, the only deposit related to the Early Neoglacial advance of ice in the Humboldt foreland, is a rarity in the morphosequence of the region in that it survived a succession of LIA advances, the ELA being approximately the same for each advance. Late Glacial sites carry maximum pebble rinds from 1.5 mm to 2.1 mm with minimum rind growth approximately twice as thick as on the Early Neoglacial substrates, making deposit discrimination on a relative-age basis useful in environmental reconstruction. 5.3. Soil profiles Soil profiles in the Humboldt Massif foreland were studied in order to differentiate deposit ages from close-in to the Bonpland and northern lobes of the Humboldt Glacier to Lago Verde (Figs. 1 and 2).
Table 1 Weathering rind dataa from a morphosequence of Late Glacial/Neoglacial deposits in the Humboldt foreland, northwestern Venezuelan Andes. Site
Maximum rind
Minimum rind
Projected age
5.1. Surficial geology
COR13 COR10 COR15 COR6 COR7
0.56 0.68 0.96 1.54 2.14
Nil Nil 0.08 0.16 0.14
LIA LIA Early Neoglacial Late Glacial Late Glacial
Deposits in the glacial sequence were mapped from aerial photography (Fig. 1) to show relationships between end, recessional
a Sample populations are n = 50 for each site. Measurements are in mm. For additional information regarding rind counts on Neoglacial deposits in the tropical mountains see Mahaney, 1990, chapter 8.
5. Results
Author's personal copy W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
Soil properties of use in reconstructing the soil chronosequence are: depth of weathering, horizon thickness, color, texture, structure, plasticity, stickiness, and organic matter (roots) in the surface and subsurface horizons. Another application of this analysis is the reconstruction of paleoclimate in the alpine region. The profiles showing development properties from youngest to oldest are depicted in Fig. 3. The COR13 profile, nested in the inner Neoglacial moraine ridge, is the least developed, with only a thin C/Cu horizon distribution. The C horizon, from 0–5 cm depth, is pebbly loamy sand in texture that registers 10YR 6/2 in color. It is massive, loose, nonplastic, and nonsticky. The Cu begins at 5 cm depth, and is similar in texture, though with a change in color to 2.5Y 6/2. The penetration of roots at this site is limited to the C horizon, and these are up to 1 mm in diameter. The COR10 profile, somewhat older than COR13, is Late Neoglacial in age, and bears a close similarity to site COR3 reported upon by Mahaney et al. (2000). The COR10 soil is slightly more developed, with a thin Ah horizon from 0–3 cm depth. Its weak granular structure is friable, nonsticky and nonplastic, with a color of 10YR 3/2. The Cox (3–10 cm) and Cu (10+ cm) horizons, on the other hand, are massive, loose, as well as nonplastic and nonsticky. The colors change downward through the horizons, transitioning from
239
10YR 5/3 in the Cox to 2.5Y 6/2 in the Cu. The colors of the till in both Neoglacial profiles are of the same strength. The COR15 profile, considered to be of Early Neoglacial age, is located on a segmented moraine ridge extending across the river valley (Fig. 2), formed in a deposit with a high boulder frequency. The soil is a developing Inceptisol with vivid colors and coarse textured horizons throughout. The Ah horizon, notably thicker than in the younger profiles, extends from 0–15 cm depth, and is pebbly sandy loam throughout, with a stronger granular structure than the Ah of COR10; it is also friable, nonplastic and nonsticky, with a color of 7.5YR 2/3. The Bw horizon (15–32 cm) is also pebbly, with friable consistence, nonplastic and nonsticky characteristics. The blocky structure is probably the result of slightly greater clay content and higher organic carbon. The horizon has an overall color of 7.5YR 4/4, with 7/5YR 3/4 in the darker pockets. The Cox (32–54 cm) and the Cu horizons (54+ cm) have similar properties; both are pebbly, massive, loose, nonplastic, and nonsticky. The Cox exhibits a color of 10YR 5/6, but is mottled, with patches of 2.5Y 5/3, the color of the Cu. The mottling is an expected product of shifts in water saturation over time. The two older profiles, COR6 and COR7, are both strong Inceptisols, and are considered to be Late Glacial in age, formed in recessional
Fig. 3. Soil profiles (a, b, c, d, e) of the Humboldt foreland showing variations in pedon development from the LIA (COR13, 10) through the Early Neoglacial (COR15) to Late Glacial/ Younger Dryas time (COR6, 7).
Author's personal copy 240
W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
moraines; correlative deposits nearby were overrun by Younger Dryas ice (Mahaney et al., 2008). On the basis of this cross-correlation, sites COR6 and COR7 have probable 14C ages of ~ 12 ka. The COR6 profile, located near Lago Verde, is an iron-rich soil with a thick Ah horizon 0–20 cm thick. Much like the other sites, this horizon has a 10YR 3/4 color, a pebbly sandy loam texture, with a granular structure, friable consistence, and nonplastic and nonsticky characteristics. The Bw horizon (20–29 cm), Cox (29–52 cm), and Cu (52+ cm) horizons are very similar, though massive rather than granular in the lower horizons. The Bw has 7.5YR 3/4 and 4/4 colors, while the Cox is at 10YR 4/4 to 4/6 and the Cu is 2.5Y 5/3. Roots at this site penetrate to a greater depth, with lower root density below 30 cm. The high iron content and coloration indicate water penetration into the Bw from higher precipitation in the area at some time in the past, perhaps in the Altithermal (5–8 ka). The COR7 profile is a clast-supported breccia, with pebbles and cobbles found throughout. The Ah is only slightly thicker here compared with the younger profiles, extending to 21 cm, with a granular structure, friable, and nonplastic and nonsticky properties. The Bw horizon (21–29 cm) is weak blocky in structure, friable, nonplastic and nonsticky, while the Cox (20–62 cm) is massive, friable, and nonplastic and nonsticky. The soil colors are similar to those of COR6, probably as a function of similar age, with the Ah at 10YR 2/3, the Bw at 7.5YR 3/4, and the Cox at 10YR 4/4. Root penetration is also deep, with the frequency and diameter dropping off only into the Cox horizon. 5.4. Particle size Particle size analysis (Table 2) contributed to the study of weathering and relative age in the chronosequence. The clay content in the younger profiles increases slightly in COR6 and 7, a function of age in the older pedons and related weathering processes. The increase is most evident in the Ah horizons, and in COR7 there is a slight downward movement of clay over time. Silt distributions, although quite variable across the sequence, increase downward in some profiles. The relatively consistent sand content throughout all profiles is highest in the C horizons, probably as a result of surface weathering processes. 5.5. Scanning Electron Microscopy Selected grains were sub-sampled from surface horizons of soil profiles in the chronosequence and imaged using the SEM. Samples of Little Ice Age soils were free of secondary coatings of any kind and many carried a number of percussion cracks (Fig. 4a) indicative of
Table 2 Particle size distributions and organic carbon determinations for soils in the soil chronosequence, Humboldt foreland, northwestern Venezuelan Andes. Site
Horizon
Depth (cm)
Sand
Silt
Clay
Organic carbon (%)
COR13
C Cu Ah C Cu Ah Bw Cox Cu Ah Bw Cox Cu Ah Bw Cox
0–5 5+ 0–3 3–10 10+ 0–15 15–32 32–54 54+ 1–20 20–29 29–52 52+ 0–21 21–29 29–62
61.8 61.9 79.3 73.8 78.6 77.8 77.9 79.8 77.4 69.7 40.7 40.7 82.3 69.2 63.9 86.4
33.0 32.9 16.1 21.3 16.2 15.5 17.5 17.3 19.5 22.6 56.1 56.1 15.4 21.2 28.0 10.2
5.2 5.2 4.6 4.9 4.8 6.7 4.6 2.9 3.1 7.7 3.2 3.2 2.3 9.6 8.1 3.4
0.96 0.33 1.40 1.27 0.43 7.76 3.64 0.33 0.41 8.80 2.06 1.46 0.31 9.80 3.31 1.85
COR10
COR15
COR6
COR7
Organic carbon was determined by loss on ignition.
transport in water (Mahaney, 2002), which is not surprising given the meltwater content in tropical glaciers. Grains of Early Neoglacial age (COR15) carry partial coatings of Fe and very partial coatings of Si, along with incipient etching zones which cover at most 10–15% of grain surfaces (Fig. 4b and c). Lastly, grains in the Ah/Bw horizons of sites COR6 and 7 carry irregular thin coatings of Fe as indicated in Fig. 4d (light tonal contrast indicates Fe coating). 5.6. Clay minerals The clay mineralogy (Table 3) of the samples was analyzed to determine the influence of weathering and lithology. It is likely that some primary minerals, such as quartz, plagioclase, and muscovite, originated in the parent material. The data show a substantial content of 1:1 (Si:Al = 1:1) clay minerals, such as kaolinite, with a slight decrease in the older profiles suggesting a detrital rather than a pedogenic effect. Chlorite, a 2:1:1 clay mineral, exemplifies a similar trend. Interestingly, the amounts of 2:1 clay minerals (Si:Al = 2:1), particularly illite and vermiculite, are uncharacteristically low, with little evidence of conversion of illites into other clays, such as chlorite and vermiculite. This indicates low weathering and leaching across sites in the chronosequence. There is a greater presence of randomly-interstratified illite–smectite in the upper horizons, but with little difference compared with the parent materials. The presence of sepiolite in the parent material of COR6 and phlogopite mica in two other profiles is unexplained, but presumably results from airfall-influx or reworking of materials from a previous weathering episode. 5.7. Fe/Al extractions Extractable Fe and Al (Fig. 5) contents increase with time and in some profiles the highest values are at depth. The Fed content increases over time, with the greatest amounts seen in the upper horizons as a function of wetting depth. The concentrations of Feo also increase, though to a smaller extent. A similar trend can be seen in the Ald and Alo. Alp, organically-bound Al, also gradually increases with age, as predicted by Parfitt and Childs (1988). A comparison of Alo and Ald shows a consistently higher concentration of Alo with regard to Ald, a progression observed in other alpine areas. As discussed in Mahaney (1990) and Birkeland et al. (1989), this is probably a result of the inefficiency of sodium dithionite to extract crystalline Al. The change in downward movement of ferrihydrite (Feo/Fed) is only slight between the profiles, with minute increases in the Cu horizons. Other ratios, set out by Parfitt and Childs (1988) and Mahaney et al. (1999), were calculated to determine: (a) the prevalence of crystalline Fe oxihydrites (Fed/Fet), (b) the changes in ferrihydrite over time (Feo × 1.7), (c) the hematite and goethite content (Fed–Feo), and (d) the reduction of lattice Fe with weathering over time (Fet–Fed). The growth of ‘free’ crystalline Fe (Fed/Fet) is a slow but steady process, with the lowest ratios in the younger soils, and the higher in the older COR6 and COR7 profiles. This indicates slight gradual oxidation, consistent with the low water penetration in the sites and Fe-skins determined by SEM investigations of sands in various horizons across the chronosequence. The ferrihydrite trend (Feo × 1.7) is one of slow increases with time, with the content doubling or more in the older soils. The high concentrations of Feo and Fet in the COR7-Bw horizon are due to a concentration of magnetite, the rest of the samples have nil content. Since acid ammonium oxalate reacts strongly with magnetite, it is estimated that approximately 50% of the increase shown in Fig. 4 is due to the magnetite content; the other 50% is due to perched ground water as indicated by the ferrihydrite concentration. The hematite and goethite concentrations were calculated (Fed–Feo), and display an increase over time that is particularly noticeable in the A horizons. The
Author's personal copy W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
241
Fig. 4. a, Microcline grain free of etched microfeatures and secondary coatings in the COR13 profile. Light area on top is charging slightly; b, Early Neoglacial angular quartz from the Ah horizon in COR15 with conchoidal fractures, sharp edges and partial etching on the upper surface coated with very minor secondary Fe (light areas); c, Enlargement of b with minor Fe coatings (light areas) similar to b; d, a weakly cemented assemblage of orthoclase and quartz grains (left) with an intact orthoclase grain (right) in COR6-Bw. The grain to the right is variably coated (~ 50–100 nm) in places with Fe and Mn judging by tonal contrast.
lattice Fe (Fet–Fed) is consistently high, with only slight variations in the upper soil horizons. 5.8. Geochemistry Before assessing the variability of the elemental concentration data in the sample suite, data for each profile were normalized for C content, using data in Table 2, since this was an effect that ranged from 0.3% to 10%. The sixteen COR samples, from five locations, are relatively homogeneous in their elemental compositions and are equivalent to the dispersion of elemental concentrations in ancient, coarse-ware ceramics (e.g. Harrison and Hancock, 2005), the latter considered as an index of element uniformity. A check of the REE (Rare Earth Elements) concentrations indicated that the sample from the Ah horizon at COR15 was highly enriched (2–3 times) in light REES relative to the rest of the sample suite. This specific surface horizon enrichment probably arises from the inclusion of small particles of monazite, a mineral rich in light REEs and Th. Because of this REE concentration anomaly, the means of the ele-
mental concentrations for the Ah horizons were calculated both with and without the COR15 sample data. Hence, mean C-corrected, elemental concentrations, together with standard deviations for each group of horizons, are presented in Table 4, with 9 C horizon samples, 3 Bw horizon samples, and 4 and 3 (without COR15) Ah horizon samples. In Table 4, the major elements are presented first in alphabetical order, followed by an alphabetical listing of minor and trace elements, and ending with the Rare Earth Elements (REEs) in their atomic number order. Elements that were predominantly below detection limits, such as As, Cl, Ga, I, Ni, and Sb, were not included in this table. Where some detection-limit data were included in the calculation of the means, the values are preceded by “b”. The anomalously high Ah elemental means (Th and light REEs) are given in bold. In an attempt to analyze the mean data of Table 4, inter-horizon ratios of the means were calculated, with data presented in Table 5. For a homogenous material, the data would scatter closely around 1.0. But, akin to the coarse-ware pottery analogy (above), the ratios were
Author's personal copy 242
W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
Table 3 Primary and secondary mineralogya of the b2 µm fraction of soils in the chronosequence, Humboldt foreland, northwestern Venezuelan Andes. Site
Horizon K
COR13 C Cu COR10 Ah C COR15 Ah Bw Cox Cu COR6 Ah Bw Cox Cu COR7 Ah Bw Cox
xx x – x tr. tr. x x – x tr. x tr. – x
H Dickite I
I–S V
– x – – – – – – – – – – – – –
x tr. – tr. – – x x – tr. tr. x – – x
– tr. – – – – x x – – – – – – –
x xx – x – – x x – – tr. x x – x
– – – tr. – – x – tr. – – – – – tr.
Chl Sepiolite Q x x tr. x x x x x tr. – tr. tr. x tr. –
– – – – – – – – – – – x –
xx xx x x x x xx x x x x x x x x
P
M O Phl
x x x x x – x x x x x tr. x x –
– – x – – x x – x x x x x x x
– – – x – – – – – – – – – x –
– – – – – – – x – – – x – – –
a Clay minerals include kaolinite (K), meta-halloysite (H), dickite (D), illite (I), illite– smectite (I–S), vermiculite (V), chlorite (Chl), sepiolite (S). Primary minerals include quartz (Q), plagioclase (P), muscovite (M), orthoclase (O), phlogopite (Phl). Semiquantitative estimate of amounts of individual mineral species is based on relative peak height: –nil, tr.—trace, x = small amount, xx = moderate amount, and xxx = abundant (N 50%).
expected to be within the ± 10–20% relative zone. Potentially anomalous data are highlighted in bold for ratios far greater than 1.00 and in italics for ratios far less than 1.00. 6. Discussion Weathering rind development shown in Fig. 6 positively correlates with the growth of secondary Fe [∑ = hematite + goethite + ferrihydrite]. Thus, the liberation of oxides to form weathering encrustations on pebbles occurs with a similar rate to that of the genesis of secondary Fe-oxides in soils. This means that the oxidizing potential on pebble surfaces is about equal to that within the wetting depth of soils in the chronosequence. There is no correlative example like this in the literature but other researchers might consider testing rinds against secondary Fe in soil genesis studies carried out in other areas.
Despite differences between the Humboldt Massif (felsic gneiss) and Mount Kenya (syenite), rind growth follows a similar trend in both tropical alpine areas, suggesting that the bracketing figures for rind growth from LIA → Early Neoglacial → Late Glacial deposits reported here might be used elsewhere or at least tested in areas with comparable lithologies. Inter-valley contrasts of soil profile expression between the Humboldt Massif (COR6 and COR7) and Mucuñuque Canyon (MUM7 and MUM7B; see Mahaney et al., 2008 for location of the MUM7 complex of deposits), where deposits of similar Late Glacial age exhibit profoundly different soil profiles, highlight the importance of determining absolute ages or at least very stringently determined relative ages by independent weathering criteria (weathering pits, rinds etc.). In these two cases, deposits are either dated directly by 14C (MUM7B) or by cross-valley correlation (COR6 and 7), with dated moraines (COR21 in Mahaney et al., 2008). Differences in the relative development of the two soils is attributed to relative stability in the case of COR6 and COR7 and deflation processes occurring at MUM7 and MUM7B, possibly caused by a drier microclimate, although direct confirmation with instrumental meteorological data is not available. Considering that both suites of sites have similar biota, topography, lithology and time, only climate, and probably wind-generated processes, serve to explain the weak soil expression at MUM7 and MUM7B in the upper Mucuñuque Basin. The A/Cu and A/C/Cu profiles at MUM7 and MUM7B, respectively, may well reflect deflation during the Neoglacial, when anabatic and katabatic winds during colder intervals may have deflated surface soils providing an unweathered surface in which subsequent pedogenesis occurred over the last several centuries. The development and movement of extractable Fe and Al are of great importance in determining age relationships and reconstructing the paleoclimate of these alpine soil systems. While concerns persist with the method, mostly regarding the Ald and Alo data, the increase in Alp clearly coincides with an increase in age. However, the alpine environment here notably permits little downward movement of organically-bound Al within the profiles, just as in other areas such as the Rocky Mountains (Mahaney et al., 1999). Stemming from the influences of vegetation and low leaching in the profiles, as demonstrated by the clay mineralogy, the Fe extracts follow a similar trend.
Fig. 5. Extractable Fe and Al data, including ratios and arithmetic functions, for soils in the Humboldt chronosequence.
Author's personal copy W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
243
Table 4 Carbon-corrected, summary analytical data for the C (both Cu and Cox), Bw, and Ah horizons.
Al Ca Fe K Na Ba Br Co Cr Cs Hf Mn Rb Sc Ta Th Ti U V La Ce Nd Sm Eu Tb Dy Yb Lu
% % % % % ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm ppm
C
Bw
Ah
Ah
9 samples
3 samples
4 samples
3 samples (COR15 sample removed)
7.3 ± 0.5 1.6 ± 0.3 3.2 ± 0.3 1.7 ± 0.4 2.3 ± 0.3 610 ± 90 b 7 ± 8.0 11.3 ± 1.1 52 ± 10 4.0 ± 1.5 5.4 ± 0.7 970 ± 120 77 ± 18 14.8 ± 1.2 0.9 ± 0.3 10.2 ± 1.3 2200 ± 900 2.5 ± 1.0 80 ± 13 34.9 ± 3.6 65 ± 5 64 ± 19 5.6 ± 0.4 0.76 ± 0.07 0.91 ± 0.10 b 3.2 ± 1.4 2.7 ± 0.4 0.41 ± 0.04
7.3 ± 0.8 1.2 ± 0.4 4.3 ± 1.6 1.4 ± 0.2 1.8 ± 0.8 460 ± 30 37 ± 21 11.2 ± 1.2 68 ± 18 4.1 ± 1.4 5.1 ± 0.9 830 ± 70 61 ± 6 15.7 ± 1.3 1.0 ± 0.4 9.1 ± 1.8 2300 ± 1200 2.5 ± 1.5 96 ± 19 30.6 ± 4.0 60 ± 9 48 ± 13 5.4 ± 0.8 0.83 ± 0.07 0.89 ± 0.11 3.9 ± 0.5 2.8 ± 0.1 0.41 ± 0.02
6.7 ± 0.7 1.5 ± 0.2 3.5 ± 1.0 1.2 ± 0.3 1.8 ± 0.3 530 ± 90 39 ± 24 9.1 ± 0.8 51 ± 12 4.1 ± 2.0 6.4 ± 2.4 820 ± 80 57 ± 8 14.4 ± 2.6 1.0 ± 0.5 15.9 ± 9.9 3100 ± 1300 3.0 ± 0.9 94 ± 30 52.1 ± 29.2 97 ± 53 59 ± 33 7.6 ± 3.3 0.88 ± 0.21 1.11 ± 0.33 4.6 ± 1.6 2.9 ± 0.8 0.43 ± 0.08
6.4 ± 0.4 1.6 ± 0.3 3.5 ± 1.2 1.3 ± 0.2 1.8 ± 0.4 520 ± 110 32 ± 24 9.0 ± 0.9 52 ± 15 4.5 ± 2.2 6.7 ± 2.8 840 ± 80 59 ± 8 14.1 ± 3.1 1.0 ± 0.6 11.0 ± 1.8 3300 ± 1600 2.9 ± 1.0 100 ± 34 37.6 ± 4.1 71 ± 12 43 ± 8 6.0 ± 1.0 0.80 ± 0.17 0.98 ± 0.21 3.9 ± 0.5 2.6 ± 0.7 0.41±0.08
Major element data are followed by minor and trace elements (all alphabetical), and then by REEs in order of atomic number. “b“signs are in front of means that included detection-limit data. The highest means are in bold.
Table 5 Carbon-corrected, mean elemental concentration ratios between horizons Ah, Bw and C.
Al Ca Fe K Na Ba Br Co Cr Cs Hf Mn Rb Sc Ta Th Ti U V La Ce Nd Sm Eu Tb Dy Yb Lu
Ah4/Bw3
Ah4/C9
Bw3/C9
Ah3/B3 (COR19 Ah data removed)
Ah3/C9 (COR19 Ah data removed)
0.92 1.27 0.81 0.83 1.03 1.15 1.05 0.81 0.76 1.02 1.25 0.99 0.94 0.92 1.01 1.75 1.30 1.20 0.98 1.71 1.61 1.23 1.41 1.06 1.26 1.19 1.03 1.06
0.92 0.95 1.08 0.72 0.80 0.87 N 5.9 0.80 0.99 1.03 1.18 0.85 0.74 0.97 1.10 1.56 1.37 1.18 1.17 1.49 1.49 0.93 1.36 1.15 1.23 N 1.43 1.05 1.05
1.00 0.75 1.33 0.87 0.77 0.76 N 5.7 0.99 1.31 1.01 0.95 0.86 0.78 1.06 1.09 0.90 1.05 0.99 1.20 0.88 0.92 0.76 0.96 1.09 0.98 N 1.20 1.02 0.99
0.88 1.29 0.82 0.89 1.00 1.14 0.88 0.80 0.77 1.12 1.31 1.01 0.98 0.90 1.08 1.21 1.40 1.14 1.04 1.23 1.18 0.90 1.12 0.97 1.10 1.00 0.93 1.01
0.88 0.97 1.09 0.77 0.78 0.86 N5.0 0.80 1.01 1.12 1.24 0.87 0.76 0.95 1.17 1.08 1.47 1.13 1.25 1.08 1.08 0.68 1.07 1.05 1.07 N1.20 0.96 1.00
The numbers after the horizon designations are the numbers of samples included in each calculation. Values above 1.0 may indicate additions (highest ratios in bold). Values below 1 may indicate losses (lowest quotients in italics). The numbers after each horizon indicate the population (n=) for each calculation. The N sign results from the denominator containing a b sign.
Fig. 6. Weathering rind—Fed regression showing a strong positive correlation between rind growth and secondary Fe accumulation in the Late Glacial/Holocene deposit/soil sequence.
There is a steady increase in Fed from the Early Neoglacial to the Late Glacial profiles, while the Fed/Fet ratio allows for an even more accurate analysis of age differences, similar to what was used in the Teton Range (Mahaney et al., 1999). In fact, the Fe concentrations are not very different between the Teton and Humboldt chronosequences, which is surprising given climatic differences between the two areas. Additionally, the relative stability of the Feo/Fed ratio indicates little groundwater fluctuation over time, with the exception of COR7. The presence of perched groundwater, as measured in COR7, is comparable to results published elsewhere (Parfitt and Childs, 1988; Mahaney and Fahey, 1988; Mahaney et al., 2007) and bears certain similarities to deposition of oxihydroxides in stream beds (Childs et al., 1982). The discovery of an Early Neoglacial moraine with advanced soil/ profile properties compared with soils in LIA substrates is a real improvement over previous work in tropical alpine localities (see Mahaney, 1990; Mahaney et al., 1994, 2000). This, coupled with rock rind measurements that show accelerated weathering in this Early Neoglacial deposit, is a major addition to our previous knowledge of the tropical alpine glacial record. Evidence of the existence of a pre-LIA Neoglacial record begs the question as to its precise age and relation to middle latitude records (Mahaney et al., 1999; Mahaney et al., 2003/ 2004). The Ah horizon in profile COR10 is similar to the Ah horizon previously described (Mahaney et al., 2000) at Site COR3, the oldest of the LIA soils dated by 14C to ~450 yr BP or 1500 AD (Mahaney, unpublished data). The oldest LIA soil (COR3, Mahaney et al., 2000) in Bonpland/Northern Lobe deposits contains thin but recognizable Ah horizons, whereas the younger LIA deposits, e.g. COR13, display only C/Cu profiles. This is a departure from the middle latitude alpine, in particular the Rocky Mountains (Mahaney et al., 1999), and the tropical Afroalpine of Mt. Kenya, where LIA deposits are devoid of any appreciable humus accumulation in surface horizons. No doubt the absence of winter and the somewhat lower elevation of sites in the Humboldt foreland favor the development of humic constituents and the resulting Ah horizon. Highlighting the value of processes of soil evolution and soil stratigraphy in relative dating, these data are a major addition to the study of alpine chronosequences in tropical alpine regions. The relatively low degree of weathering seen in the profiles is typical of alpine regions in general, but slightly unexpected here, given the absence of winter and the +1 m of precipitation. Nonetheless, aside from equivocal clay mineral trends, the main weathering development entails Fe and Al extracts and geochemical gains and losses. Though there is overall low water penetration and leaching, stronger coloration in profiles COR6 and COR7 may indicate higher levels of precipitation in the past, a hypothesis supported by higher concentrations of
Author's personal copy 244
W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245
ferrihydrite. In part, the stronger coloration in COR7 may result from higher magnetite concentrations in the Bw horizon. Age-related Inceptisol characteristics are evident, namely, an increase in horizon depth, development of a Bw horizon in the Early Neoglacial soil and increases in organic carbon in the older pedons. While silt increases across the chronosequence may be age related as in other alpine areas (Dahms, 1992, 1993), the silt bulge in the lower horizons of COR7 may be related to glacial deposition during moraine construction. There is little change in sand content due to the similar lithology of the profiles. Indeed, the progression from the youngest Neoglacial soil in COR13 to the oldest Late Glacial profiles (COR6 and COR7) is clearly documented in the soil dynamics, of which profile characteristics and extractable Fe and Al movements are an important part. Elemental increases in the surface horizons shown in Tables 4 and 5 probably result from a number of processes including: recycling by plants (e.g. Br and Ca), affinity of radioactive elements U and Th with organic constituents, aeolian influx of material rich in light REEs, Th (as monazite?), and Ti, and Hf, the latter possibly associated with zircon. Elemental quotients for the Ah/Cu (Table 5) depict a trend close to values calculated for the surface horizon group which is consistent with the relatively homogeneous composition of all 16 samples. Losses from the Ah are limited to K, probably leached from orthoclase during hydrolysis, and small amounts of Fe, possibly organically-complexed. Analysis of Bw/Cu ratios shows a near-constant trend in the concentration of Al across all horizons including the parent material, and an increase of Fe in Bw horizons weighted somewhat by higher concentrations in the older pedons, followed by losses of alkaline and alkali metals downward in the profiles. The geochemical data highlight the homogeneity of the material across the chronosequence, aeolian influx of foreign material, and translocation to the B group of horizons. In particular, the build-up of Fe, certain trace elements and individual heavy and light REEs relative to the parent materials underscores the cumulicative nature of alpine soils which receive varying amounts of airfall-influxed material. 7. Conclusions Deposits of early to mid-Neoglacial age in the tropical mountains are not commonly preserved, as they are often destroyed by later and more expansive ice advances during the LIA. Here, evidence is presented for an Early Neoglacial advance occurring sometime prior to 1.0 ka. Previous section descriptions in the Humboldt area show finely disseminated peat in alluvial sands that may be related to limited periods of soil development that were terminated by deposition of coarse gravels. These weathering periods, dated at 0.97 ka and 1.5 ka (Mahaney et al., 2000) are probably too young to correlate with the COR15 site because the soil profile characteristics argue for a more advanced age of ~ 3.0 ka. Particle size calculations show slight increases of clay in Ah and Bw horizons in the older middle Holocene and Late Glacial profiles, sometimes with downward movement as evidenced by the clay/ organic films across Ah/Bw horizon contacts, and at other times with clay confined to the surface horizon. Clay mineralogy across the chronosequence depicts a detrital pattern, which outlines the effect of withdrawals and surges of ice over time, demonstrated in part by the highest kaolinite concentration in the youngest soil. The presence of minor amounts of dickite, a super-ordered form of kaolinite, only in the Neoglacial profiles is a further substantiation of this inference as neither of these two minerals could form in a short time in an alpine environment. The occurrence of illite in all the parent materials, or deepest soil mineralic horizons, with slight depletions higher in the pedon, suggests conversion to other secondary clays, perhaps vermiculite in rare cases and chlorite in others. However, the latter is confounded by the presence of hydrothermal minerals in the gneiss, including Fe-chlorites.
Geochemistry shows a relatively uniform material with minor anomalies that may be explained by assorted airfall-influx and leaching processes. However, the data reported here in an inter-valley comparison underscores the problem of using soils as relative-age indicators to date deposits in a sequence of deposits. Iron and Al extracts show variable trends with time, which parallel weathering rind development. Among these, the most important include Fed/Fet, which depicts the slow accumulation of goethite/ hematite/ferrihydrite with time; a trend supported by the slow development of Fe coatings on sands in the Ah/B horizons. In addition, the gradual decrease in lattice Fe upward in the older profiles provides a measure of Fe+2 conversion to Fe+3 over time. Also noteworthy is the accumulation of Feo in the Bw horizon of COR7, which indicates perched water presumably for long periods in the past, perhaps during various stages of Neoglaciation. Iron coatings overprint grains of different lithologies which remove parent material differences that tend to confound rind age assessments. Thus, Fe coatings, especially Fed/Fet, provide a close assessment of relative age as depicted in Fig. 6. Acknowledgements This research was funded by Quaternary Surveys, Toronto. Financial support to VK by Estonian State Target Foundation Project No. 0180048s08 is acknowledged. We thank M. Bezada (UPAL, Caracas) for assistance in the field, and two anonymous reviewers for critical assessments of a draft manuscript.
References Birkeland, P.W., 1973. Use of relative age dating methods in a stratigraphic study of rock glacier deposits, Mt. Sopris, Colorado. Arctic and Alpine Research 5, 401–416. Birkeland, P.W., 1999. Soils and Geomorphology. Oxford University Press, Oxford, UK. 430 pp. Birkeland, P.W., Burke, R.M., 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. Bockheim, J.G., Birkeland, P.W., Bland, W.L., 2000. Carbon storage and accumulation rates in alpine soils: evidence from the Holocene chronosequences. In: Lal, R., Kimble, J.M., Stewart, B.A. (Eds.), Global Climate Change and Cold Regions Ecosystems. CRC Press, pp. 185–195. Childs, C.W., Downes, C.J., Wells, N., 1982. Hydrous iron oxide minerals with short range order deposited in a spring stream system, Tongariro National Park, N.Z. Australian Journal Soil Research 20, 119–129. Coffin, D.E., 1963. A method for the determination of free iron in soils and clays. Canadian Journal of Soil Science 43, 7–17. Costin, A.B., 1955. Alpine soils in Australia with reference of conditions in Europe and N.Z. Journal of Soil Science 6, 35–50. Dahms, D.E., 1992. Comment on ‘Origin of Silt-Enriched Surface Mantles, Indian Basin, Wyoming. Soil Science Society of America Journal 56, 991–992. Dahms, D.E., 1993. Mineralogical evidence for Eolian sediments in soils on Late Quaternary Moraines, Wind River Mountains, Wyoming. Geoderma 59, 175–196. Day, P., 1965. Particle fractionation and particle size analysis. In: Black, C.A. (Ed.), Methods of Soil Analysis, pp. 545–567. Madison Wi., Amer. Soc. Agron. Dormaar, J.F., Lutwick, L.E., 1983. Extractable Fe and Al as an indicator for buried soil horizons. Catena 10, 167–173. Egli, M., Fitze, P., Mirabella, A., 2001. Weathering and evolution of soil formed on granitic glacial deposits: results from chronosequences of Swiss alpine environments. Catena 45, 19–47. Egli, M., Mirabella, A., Fitze, P., 2003. Formation rates of smectites derived from two Holocene chronosequences in the Swiss Alps. Geoderma 117, 81–98. Evans, D.J.A., 1999. A soil chronosequence from Neoglacial moraines in western Norway. Geografisker Annaler, Series A 81, 47–62. Hancock, R.G.V., 1978. Some aspects of the analysis of ancient artefacts by neutron activation. Journal of International Institute of Conservation-Canadian Group 8, 21–27. Harrison, T.P., Hancock, R.G.V., 2005. Geochemical analysis and sociocultural complexity: a case study from Early Iron Age Megiddo, (Israel). Archaeometry 47, 705–722. Hodgson, J.M., 1976. Soil Survey Field Handbook—Soil Survey Tech. Monograph, no. 5, Rothamsted Experimental Stn., Harpenden, Herts, 99p. Karlstrom, E.T., 1988. Soil development in Quaternary glacial deposits, Waterton Park, southwestern Alberta. Geographie Physique et Quaternaire 42, 147–162. Kiver, E.P.,1974. Holocene glaciation in the Wallowa Mountains, Oregon. In: Mahaney, W.C. (Ed.), Quaternary Environments: Proceedings of a Symposium. Geographical Monographs No. 5. York University, Toronto, pp. 69–95. Kovaleva, N., 2006. Northern Tian-Shan paleosol sedimentary sequences as a record of major climatic events in the last 30,000 years. Geophysical Research Abstracts 8, 4885–4886.
Author's personal copy W.C. Mahaney et al. / Geomorphology 109 (2009) 236–245 Mahaney, W.C., 1990. Ice on the Equator. Wm Caxton, Ellison Bay, Wi., 387 pp. Mahaney, W.C., 2002. Atlas of Sand Grain Surface Textures and Applications. Oxford University Press, Oxford, UK. 237 pp. Mahaney, W.C., Fahey, B.D., 1988. Extractable Fe and Al in Late Pleistocene and Holocene paleosols on Niwot Ridge, Colorado Front Range. Catena 15, 17–26. Mahaney, W.C., Hancock, R.G.V., Sanmugadas, K., 1994. Extractable Fe, Al and Mn in paleosols of Late Quaternary age in the Virunga Mountains, northwestern Rwanda. Catena 21, 27–36. Mahaney, W.C., Sanmugadas, K., Hancock, R.G.V., 1999. Extractable Fe and Al of soils in the Middle Teton chronosequence, western Wyoming, USA. Zeitschrift für Geomorphologie 43, 393–407. Mahaney, W.C., Milner, M.W., Sanmugadas, K., Kalm, V., Bezada, M., Hancock, R.G.V., 2000. Late Quaternary deglaciation and Neoglaciation of the Humboldt Massif, northern Venezuelan Andes. Zeitschrift für Geomorphologie 122, 209–226. Mahaney, W.C., Falen, A.L., Fosberg, M.A., Hancock, R.G.V., 2003/2004. Later Pleistocene and Holocene soils in Mammoth Basin and Green River areas, Wind River Mountains, Wyoming. Zeitschrift für Gletscherkunde und Glazialgeologie 39, 161–181. Mahaney, W.C., Dirszowsky, R., Milner, M.W., Harmsen, R., Finkelstein, S., Kalm, V., Bezada, M., Hancock, R.G.V., 2007. Soil stratigraphy and plant–soil interactions on a Late Glacial–Holocene terrace sequence, Sierra Nevada National Park, northern Venezuelan Andes. Journal of South American Earth Sciences 23, 46–60. Mahaney, W.C., Milner, M.W., Kalm, M., Dirszowsky, R.W., Hancock, R.G.V., Beukens, R.P., 2008. Evidence for a Younger Dryas glacial advance in the Andes of northwestern Venezuela. Geomorphology 96, 199–211.
245
McKeague, A., Day, J., 1966. Dithionite and oxalate extractable Fe and Al as aids in differentiating various classes of soils. Canadian Journal of Soil Science 46, 13–22. National Soil Survey Center (NSSC), 1995. Soil Survey Laboratory Information Manual. Soil Survey Investigations Report no. 45, version 1.00, U.S.D.A., 305 pp. Oyama, M., Takehara, H., 1970. Standard Soil Color Charts. Japan Research Council for Agriculture, Forestry and Fisheries. Parfitt, R.L., Childs, C.W., 1988. Estimation of forms of Fe and Al: a review and analysis of contrasting soils by dissolution and Moessbauer methods. Australian Journal Soil Research 26, 121–144. Schubert, C., 1970. Glaciation of the Sierra de Santo Domingo, Venezuelan Andes. Quaternaria 13, 225–246. Servicio Meteorologico de la Fuerza, 1975. Meteorologica data area Venezolana. Boletin Meterorologico del MANAR, Venezuela. Soil Survey Staff, 1999. Soil Taxonomy, Agricultural Handbook, No. 436, Second edition, Washington, USDA, US GPO. Tonkin, P.J., Basher, L.R., 2001. Soil chronosequences in subalpine superhumid Cropp Basin, western southern Alps, N.Z. N.Z. Journal of Geology and Geophysics 44, 37–45. Whittig, L.D., 1965. X-ray diffraction techniques for numerical identification and mineralogical composition. In: Black, C.A. (Ed.), Methods of Soil Analysis. American Society of Agronomy, Madison, Wisc., pp. 671–696.
