Mineralogy and Magnetic Properties of Basaltic
Substrate Soils: Kaho’olawe and Big Island, Hawaii
Remke L. Van Dam*
Dep. of Geological Sciences
Michigan State Univ.
206 Natural Science
East Lansing, MI 48824
J. Bruce J. Harrison
Dep. of Earth and Environmental Science
New Mexico Tech
801 Leroy Place
Socorro, NM 87801
Deidre A. Hirschfeld
Dep. of Materials Engineering
New Mexico Tech
801 Leroy Place
Socorro, NM 87801
Todd M. Meglich
Yaoguo Li
SOIL MINERALOGY
Dep. of Geophysics
Colorado School of Mines
1500 Illinois St.
Golden, CO 80401
Magnetic behavior of soils can seriously hamper the performance of geophysical sensors.
Currently, we have little understanding of the types of minerals responsible for the magnetic
behavior, as well as their distribution in space and evolution through time. This study investigated the magnetic characteristics and mineralogy of Fe-rich soils developed on basaltic
substrate in Hawaii. We measured the spatial distribution of magnetic susceptibility (Flf )
and frequency dependence (Ffd%) across three test areas in a well-developed eroded soil on
Kaho’olawe and in two young soils on the Big Island of Hawaii. X-ray diffraction spectroscopy, x-ray fluorescence spectroscopy (XRF), chemical dissolution, thermal analysis, and
temperature-dependent magnetic studies were used to characterize soil development and
mineralogy for samples from soil pits on Kaho’olawe, surface samples from all three test areas,
and unweathered basalt from the Big Island of Hawaii. The measurements show a general
increase in magnetic properties with increasing soil development. The XRF Fe data ranged
from 13% for fresh basalt and young soils on the Big Island to 58% for material from the
B horizon of Kaho’olawe soils. Dithionite-extractable and oxalate-extractable Fe percentages
increase with soil development and correlate with Flf and Ffd%, respectively. Results from the
temperature-dependent susceptibility measurements show that the high soil magnetic properties observed in geophysical surveys in Kaho’olawe are entirely due to neoformed minerals.
The results of our studies have implications for the existing soil survey of Kaho’olawe and
help identify methods to characterize magnetic minerals in tropical soils.
Abbreviations: DTA, differential thermal analysis; Fed, dithionite-extractable Fe; Feo, oxalate-extractable
Fe; TEM, time-domain electromagnetic; TGA, thermogravimetric analysis; UXO, unexploded ordnance;
VRM, viscous remanent magnetization; XRD, x-ray diffraction spectroscopy; XRF, x-ray fluorescence
spectroscopy; N, volume-specific magnetic susceptibility; F, mass-specific magnetic susceptibility; Ffd%,
frequency-dependent magnetic susceptibility; Flf, low-frequency mass-specific magnetic susceptibility;
FT, temperature-dependent magnetic susceptibility.
Ryan E. North
U.S. Army Corps of Engineers
Engineer Research and Development Center
3909 Halls Ferry Rd.
Vicksburg, MS 39180
M
any techniques have been developed for processing electromagnetic and magnetic data to discriminate unexploded
ordnance (UXO) from non-UXO, but their performance is negatively affected at sites with magnetic soils and rocks. In an environment with highly magnetic soil, magnetic and electromagnetic
sensors often detect large anomalies that are of geologic, rather
than metallic, origin (Butler, 2003). Therefore, it is crucial to
understand the variability of the magnetic characteristics in a survey area. Improved detection and discrimination performance will
Soil Sci. Soc. Am. J. 72:244-257
doi:10.2136/sssaj2006.0281
Received 14 Aug. 2006.
*Corresponding authhor ([email protected]).
© Soil Science Society of America
677 S. Segoe Rd. Madison WI 53711 USA
All rights reserved. No part of this periodical may be reproduced or
transmitted in any form or by any means, electronic or mechanical,
including photocopying, recording, or any information storage and
retrieval system, without permission in writing from the publisher.
Permission for printing and for reprinting the material contained
herein has been obtained by the publisher.
244
only be realized when the interactions of magnetic soil material
with geophysical sensors are understood and efficient data reduction techniques have been developed.
The magnetic properties of soils result from the presence
of Fe oxides in different forms and quantities (Mullins, 1977).
Paramagnetic and antiferromagnetic Fe oxides such as goethite
and hematite are the most abundant Fe oxide minerals in soils
but play a minor role in determining the magnetic character of
a soil. Ferrimagnetic minerals such as magnetite, maghemite,
and pyrrothite are the most magnetic of the Fe oxides, and are
of primary importance in their effects on geophysical sensors.
Iron oxide minerals can be both pedogenic (i.e., a product of soil
formation) and lithogenic (i.e., unweathered minerals from the
parent material) in origin. Two types of magnetic behavior associated with magnetic soil minerals can have a significant impact
on the magnetic and electromagnetic characteristics of the subsurface. Both magnetic susceptibility and viscous remanent magnetization (VRM) can have a negative effect on the performance
of geophysical sensors (Billings et al., 2003).
Currently, the knowledge of the spatial distribution and
temporal evolution of soil magnetic properties in tropical and
SSSAJ: Volume 72: Number 1 • January–February 2008
other soils is limited (Das et al., 2002; Van Dam et al., 2005b).
In Hawaii, soils are commonly highly magnetic, the cause of
which is Fe-bearing minerals in the parent material. In Hawaii,
the basaltic rocks contain significant amounts of magnetite
and other Fe-containing minerals (Wright and Clague, 1989).
In soils developed on the relatively uniform parent material of
Hawaii, it has been shown, however, that the magnetic susceptibility varies significantly across relatively short distances and
depends strongly on the degree of soil development (Van Dam
et al., 2004). We conducted a field and laboratory investigation
for soils on Kaho’olawe and the Big Island of Hawaii. The objective of the study was to understand the magnetic mineralogy
of soils by characterizing the origin, frequency dependence, and
variability of the magnetic signature.
SOIL MAGNETICS
Table 1 shows the magnetic class and magnetic susceptibilities
for several Fe and Fe–Ti oxides, Fe sulfides, and other soil constituents. Water and quartz are diamagnetic and have a small negative
magnetic susceptibility. Hydrated Fe oxides like goethite, which is the
most abundant Fe oxide in soils around the world, ferrihydrite, and
lepidocrocite play a minor role in determining the magnetic character of soils. Also, hematite, which is the most abundant Fe oxide in
tropical soils, pyrite, and ilmenite hardly affect the magnetic soil characteristics. As follows from Table 1, the magnetic character of a soil
is dominated by the combined amount of lithogenic and pedogenic
ferrimagnetic magnetite (Fe3O4) and maghemite (J-Fe2O3) minerals in a soil (Ward, 1990). Soils formed on basaltic parent material
typically contain significant amounts of magnetite or maghemite.
These ferrimagnetic minerals originate from primary lithogenic Fe
oxides in magmatic rocks such as magnetite (Fe3O4), titanomagnetite (Fe3xTixO4), and ilmenite (FeTiO3). The average magnetite and
ilmenite content of tholeiitic and alkali olivine basalts, characteristic
for Kaho’olawe (Clague and Dalrymple, 1987), lies between 37 to 46
and 24 to 50 g kg1 (Wedepohl, 1969).
The presence of Fe oxides in different forms and quantities is the
predominant cause of the magnetic properties of soils. Iron oxide minerals can be both pedogenic (i.e., a product of soil formation) and lithogenic
(i.e., unweathered minerals from the parent material) in origin. Although
MATERIALS AND METHODS
pure Fe can occur naturally in rocks and soil, it is very rare. Specific types
of Fe oxides, Fe–Ti oxides, and Fe sulfides are the predominant causes of
Kaho’olawe Field Site
magnetic soil characteristics. Iron is the fourth most common element in
Kaho’olawe is the smallest of the eight major islands in the state of
the crust of the earth. Iron-containing minerals can be found in igneous
Hawaii and lies approximately 11 km southwest of Maui (Fig. 1). It is one
rock such as basalt, gabbro, and granite, but also in metamorphic and
of the smallest single-shield volcanoes of the Hawaiian Islands and is 1 to
sedimentary rocks. The concentration of (magnetic) Fe oxides in soils is
2 million yr old (Wood and Kienle, 1990). The rocks consist of tholeiitic
affected by the parent material, soil age, soil-forming processes, biologiand alkaline basalt of shield and capping stages (Clague and Dalrymple,
cal activity, and soil temperature (Singer et al., 1996; Fabris et al., 1998).
1987). Kaho’olawe has not erupted in historic time and is considered to be
Magnetite and ilmenite are among the most stable lithogenic Fe oxides.
extinct. Due to its position on the leeward side of Haleakala Volcano on
Overall, the Fe oxide content in tropical soils may be as high as several
Maui and its limited elevation (maximum 450 m), Kaho’olawe receives
hundred grams per kilogram (Goulart et al., 1998).
very little rainfall throughout the year. The annual rainfall is estimated to
Many books and review papers have addressed the physical backrange from 250 to 600 mm (Nakamura and Smith, 1995). Overgrazing
ground of magnetic minerals in general (Lindsley, 1991; Thompson
by cattle (Bos taurus) that were first introduced in 1864 has caused large
and Oldfield, 1986) and the magnetic mineralogy of soils in particular
areas of the island to be stripped of vegetation. The bare surface and the
(Cornell and Schwertmann, 2003; Mullins, 1977; Stucki et al., 1988).
strong trade winds from the east have led to severe soil erosion. The sparse
Three types of magnetization are commonly used to describe the magvegetation of the central plateau consists of piligrass [Heteropogon connetic behavior of a material:
tortus (L.) P. Beauv. ex Roem. & Schult.] and kiawe trees [Prosopis pal1. Remanent magnetization occurs within ferromagnetic and ferrilida (Humb. & Bonpl. ex Willd.) Kunth]. The island of Kaho’olawe was
magnetic minerals and exists in the absence of an applied field.
used for military training and target practice from 1941 to 1993. Despite
2. Magnetic susceptibility—when a low-intensity magnetic field
extensive cleanup operations by the U.S. Navy, large parts of the island
is applied to a material, the net magnetic moment (magnetization, M) is proportional to the applied field strength Table 1. Magnetic susceptibilities (F) for several lithogenic and pedo(H). Therefore, the low-field magnetic susceptibility, which
genic Fe oxides and soil constituents. Data from Thompson and
Oldfield (1986) and Cornell and Schwertmann (2003).
is defined as M/H and expressed per unit volume (N) or per
unit mass (F), is a material-specific property. The magnetic
Material
Chemical formula
Magnetic class
F
susceptibility affects the performance of both electromag108 m3 kg1
netic and magnetic sensors (Billings et al., 2003).
Water
H2O
diamagnetic
0.9
3. Viscous remanent magnetization refers to the delay of the sec- Quartz
diamagnetic
0.6
SiO2
ondary magnetic field relative to the primary magnetic field
Pyrite
FeS2
paramagnetic
30
and has been linked to the presence of superparamagnetic
paramagnetic
40
Ferrihydrite
5Fe2O3·9H2O†
(SP) grains (Dearing, 1994). It causes frequency depen- Lepidocrocite J-FeO·OH
paramagnetic
70
dence in magnetic susceptibility, which has important Ilmenite
FeTiO3
superparamagnetic
200
implications for time- and frequency-domain electromag- Hematite
antiferromagnetic
60
D-Fe2O3
netic sensors (Billings et al., 2003).
Goethite
D-FeO·OH
antiferromagnetic
70
The formation of SP grains is generally considered to be associated with
ferrimagnetic
Pyrrhotite
Fe7/8/9S8/9/10
~5,000
processes occurring under normal pedogenic oxidation–reduction cycles
ferrimagnetic
40,000
Maghemite
J-Fe2O3
(Mullins, 1977; Thompson and Oldfield, 1986), but the exact processes
ferrimagnetic
50,000
Magnetite
Fe3O4
leading to recrystallization of these SP grains are poorly understood.
† The amount of OH and H2O is variable (Cornell and Schwertmann, 2003).
SSSAJ: Volume 72: Number 1 • January–February 2008
245
Fig. 1. Location and setting of the field areas: (a) the eight main Hawaiian islands and detailed satellite images (Landsat) of Kaho’olawe
and the northern part of the Big Island of Hawaii with the approximate locations of the field sites given by the circles; (b) outline
of the Navy QA Grid on Kaho’olawe and locations of the four soil pits (the gray squares in the inset show the locations of the five
surface samples studied in detail); and (c) aerial photograph of the field site with the locations of Test Grid 2E and the soil pits.
are still not cleared. The magnetic properties of the Kaho’olawe geology
and soils have been responsible for a large number of false positives in
previous UXO detection surveys.
The Navy QA Grid that was selected for the geophysical measurements lies in the Kunaka/Na’alapa province and is dominated by soils
of the Kaneloa and Puu Moiwi series (Nakamura and Smith, 1995).
The parent material of these soils is strongly weathered volcanic ash over
strongly weathered basic igneous rock. The Kaneloa soils reach saprolite
(weathered bedrock) within 40 to 75 cm from the surface, while the
Puu Moiwi soils have a deeper soil cover. The soil colors range from
reddish brown to brown for the Puu Moiwi soils and from brown to yellowish brown for the Kaneloa soils. For both soils the texture is mostly
silty clay loam and both contain detectable amounts of hematite and
magnetite at all depths (Nakamura and Smith, 1995).
Big Island Field Sites
Hawaii is the youngest and largest of the islands in the Hawaiian
archipelago and consists of five subaerial and one submerged volcanoes. The ages of the volcanoes range from 0.43 million yr for Kohala
Volcano in the north to active in the south. Potassium–argon dates of
shield-stage tholeiitic basalts indicate an age of 0.375 million yr for
Mauna Kea (Clague and Dalrymple, 1989). This age should be treated
with caution, as the individual lava flows at the surface show a wide
range of ages (Wolfe and Morris, 1996).
Both the Waimea Geophysical Prove-out near the town of Waimea
and the Geophysical Prove-out at Waikoloa Village are part of the former
246
Waikoloa Maneuver Area and lie on the northwestern flanks of Mauna
Kea Volcano in lava flows of the Hamakua Volcanics (Fig. 1a). Lava flows
originated from vents widely distributed across the flanks of Mauna Kea
and have both alkaline and tholeiitic characteristics (Wolfe and Morris,
1996). The rocks contain variable amounts of olivine, plagioclase, and
clinopyroxene phenocrysts. Rocks of the Hamakua Volcanics have been
K–Ar dated between 250,000 and 65,000 yr (Pleistocene).
The parent material of the soils at both sites is young (relatively
unweathered) basalt, covered by local tephra falls and windblown material
of variable thickness. Mean annual rainfall ranges from to 150 to 500 mm
near Waikoloa Village to 1500 mm near Waimea. As a result, the soils at
the Waimea Geophysical Prove-out are better developed than those at the
Geophysical Prove-out at Waikoloa Village, where soils are mostly stony
and are characterized by surface rock outcropping. The Waimea soils are
characterized by a shallow, organic-rich soil over shallow bedrock. Rocks
are common at the surface.
Field Sampling and Measurements
Four soil pits located in areas that represent the variation in magnetic properties of the site were excavated in and around Test Grid 2E
on Kaho’olawe (Fig. 1b). The locations were selected based on previously
collected time-domain electromagnetic (TEM) data, TEM data collected
in this field campaign (Walker et al., 2005), and geomorphological characteristics of the terrain. Soil Pit KH-A was located in the area with the
lowest magnetic background, just outside Grid 2E, while Soil Pits KH-B
and KH-C were dug in areas with intermediate to high magnetic backSSSAJ: Volume 72: Number 1 • January–February 2008
ground readings. Soil Pit KH-D was dug at a distance of approximately 75
m south of Grid 2E in a topographic low. All soil profiles were described
using standard methods and samples were collected for laboratory analyses
of soil texture, mineralogy, and frequency-dependent magnetic properties.
A total number of 50 soil and two saprolite samples were collected from the
soil pits. The magnetic susceptibility was measured at vertical increments of
5 cm using both Bartington MS2D (0.958 kHz) and MS2F (0.580 kHz)
sensors (Bartington Instruments, Witney, UK).
The 30- by 30-m Grid 2E that is located within the Navy QA area
was subdivided into regular 5- by 5-m cells. A random number generator was used to select one sampling location in each cell. This stratified
random sampling strategy ensured a good coverage of the entire plot
(Webster and Oliver, 1990). At each of the 36 locations, the surface
magnetic susceptibility was measured using a Bartington system with an
MS2D sensor at a frequency of 0.958 kHz. Also, a frequency-domain
Geonics EM-38 system (Geonics Ltd., Mississauga, ON, Canada) was
used to calculate the magnetic susceptibility of the soil from the inphase component of a 14.6-kHz signal measured at 1.5 m above the
ground and at the ground surface. Finally, at each of the 36 locations,
surface samples were collected for laboratory analysis of soil mineralogy
and frequency-dependent magnetic properties. We will focus on five
surface samples, indicated by the gray polygons in Fig. 1b.
The former Waikoloa Maneuver Area field sites are designed as geophysical test locations. Both sites consisted of three 30- by 30-m grids.
Due to security restrictions, our samples were collected from unseeded
grids. We had no permission to dig soil pits and were restricted to surface
sampling of both the Waikoloa and Waimea sites. The format of the surface soil sampling was similar to that on Kaho’olawe. At both sites, a 5- by
5-m grid was laid out inside the unseeded grid. From each 5- by 5-m grid,
one random sample was taken. Thirty-six samples were collected from
both sites. In addition, two samples of unweathered basalt were collected
from a road cut from Kohala Volcano on the Big Island of Hawaii.
Laboratory Measurements
To understand the origin and mineralogy of the soils, and to characterize the frequency- and temperature-dependent magnetic properties, a range of laboratory measurements was performed. Samples
were dried for a minimum of 48 h at a low temperature (<50°C) so
as not to cause any chemical reactions. The soil and rock material was
crushed (manually) to silt- and sand-size grain sizes where necessary.
Magnetic Susceptibility
To measure magnetic susceptibility, 10-cm3 pots were filled for
measurement of the magnetic susceptibilities and their weight was
measured on an A&D GR-120 balance with 0.1-mg accuracy. Next,
the volume-specific magnetic susceptibility (N) was measured at two
frequencies (0.46 and 4.6 kHz) using a Bartington MS2B dual-frequency sensor. All measurements were conducted at the 1.0 sensitivity
setting. Each sample was measured three times, with an air reading
before and after each series for correction of drift. The mass-specific
susceptibility (F) was calculated from the volume-specific susceptibility using F = 10N/m, where m is the sample mass in grams (Dearing,
1994). The (percentage) difference between readings at the two frequencies can be used to estimate the VRM of the samples. The frequency-dependent susceptibility was calculated using Ffd% = [(Flf Fhf )/Flf ]100, where Flf is the mass-specific susceptibility measured at
a frequency of 0.46 kHz and Fhf is the mass-specific susceptibility
measured at 4.6 kHz (Dearing, 1994).
SSSAJ: Volume 72: Number 1 • January–February 2008
Chemical Dissolution Techniques
The content and composition of pedogenic Fe oxides was determined by selective chemical dissolution techniques. The total amount
of secondary Fe oxyhydroxides (dithionite extractable, Fed) was determined using the concentrated citrate–dithionite method (Holmgren,
1967). Poorly crystalline and amorphous Fe oxyhydroxides (oxalate
extractable, Feo) such as ferrihydrite were extracted using the hydroxylamine procedure (Chao and Zhou, 1983). The Fe content was measured using an Instrumentation Laboratory (Wilmington, MA) video
12 flame atomic absorption spectrometer. An increase in Fed and Feo are
both indicative of weathering of mafic minerals from ferrous (Fe2+) to
ferric (Fe3+). Amorphous Fe oxides (Feo) are considered unstable under
most conditions and are expected to be present only during initial soil
formation, or when the presence of silicate and organics impede formation of crystalline Fe oxides (Cornell and Schwertmann, 2003).
X-Ray Techniques
X-ray techniques were used to perform chemical, mineralogical,
and crystallographic analyses of fine-powdered soil material. X-ray diffraction specroscopy (XRD) allows the identification of polymorphic
Fe oxides (for example D-Fe2O3, hematite, and J-Fe2O3, maghemite).
Because of overlapping peaks in the resulting spectrum, however, it is
very difficult to make the distinction between lithogenic magnetite
(Fe3O4) and pedogenic maghemite. Also, XRD cannot be used to identify poorly crystalline minerals such as ferrihydrite (Fe5HO8·4H2O).
We used a Rigaku D/MAX II x-ray diffractometer (Rigaku Americas,
The Woodlands, TX) with a Cu anode to scan across a 2T range of 2 to
70° with a step width of 0.03° and a count time of 0.5 s per step. The
Cu KD2 peak was removed using JADE software before peak search
with the International Center for Diffraction Data (ICDD, Newtown
Square, PA) database and its minerals subset.
X-ray fluorescence spectroscopy (XRF) was used to quantitatively
analyze the chemical composition of selected samples. To ensure homogeneity, mixtures of 1 g of finely ground soil material, 9 g of lithium tetraborate (flux), and 0.25 g of NH4NO3 (oxidizer), were fused at 1000°C
in a Pt crucible in an oxidizing flame for at least 30 min, and then poured
into Pt molds to produce glass disks. These disks were used for all analyses using a Bruker AXS (Madison, WI) S-4 Pioneer, a 4-kW, wavelength
dispersive instrument. The XRF element analyses were determined by
reducing the x-ray intensities using SPECTRAplus software (from Bruker
AXS), which uses alpha factors to determine the fundamental parameters.
Twelve well-known standards were used for calibration. Accuracy was
checked by analyzing two known standards.
Thermal Analysis
Thermal analysis refers to the study of the behavior of materials as
a function of temperature change and gives information on sample composition, thermal characteristics of the sample, and the products formed
during heating (Speyer, 1994). While thermogravimetric analysis (TGA)
shows the change in weight with time or temperature, differential thermal
analysis (DTA) shows whether specific events are endothermic or exothermic. Several factors such as the heating rate, amount of sample, and particle size play a role in the shape of the curves and the magnitude of events
(Speyer, 1994). We used a Linseis Thermal Balance L81/1550 (Linseis Inc.,
Princeton Junction, NJ), which is sensitive to the nearest microgram, for
approximately 250-mg starting weights. The instrument has a temperature
range of 20 to 1550°C and a variable heating rate of 0.01 to 50°C min1.
Changes in magnetic susceptibility as a result of chemical transformations that occur due to temperature variations can be measured
247
with a furnace and susceptibility meter combination. The effect of
temperature on magnetic susceptibility (FT) was measured with
a CS-2 furnace and a Kappabridge KLY-2 AC susceptibility bridge
(Advanced Geoscience Instruments Co., Brno, Czech Republic) using
a 300 A m1 induced magnetic field at a single frequency of 920
Hz in an Ar (reducing) atmosphere. A small glass test tube is used
to hold approximately 0.5 cm3 of sample material during the susceptibility measurements. The mass is measured to normalize the
recorded susceptibility values to the sample density. The test tube is
then placed into a water-cooled jacket integrated in the Kappabridge.
A thermocouple device attached to the furnace rests directly on the
sample. Susceptibility measurements began at room temperature and
were taken approximately every 18 s while the temperature increased
at about 10qC min1. Volume-specific instrument readings were normalized to mass-specific susceptibility.
RESULTS
Soil Descriptions
Soil Pit A (KH-A) is located in an area of low magnetic background (Walker et al., 2005) and magnetic susceptibility readings.
Fig. 2. Pictures and description of (a) Soil KH-A and (b) Soil KH-B on Kaho’olawe, Hawaii; see Fig. 1 for location.
248
SSSAJ: Volume 72: Number 1 • January–February 2008
The area has been significantly eroded. Weathered bedrock is found
within 40 cm of the surface (Fig. 2a). A very thin remnant of a lower
B horizon lies directly over the Co horizon. The surface color at this
site is yellowish to dark gray to green (Fig. 1c), similar to what is
seen at depth in the bedrock. The entire soil profile is dominated
by partly weathered basalt or saprolite. The Co1 and Co2 horizons
below the Bw/Co horizon are characterized by increasing amounts
of saprolite. This soil, at the highest elevation of the four soil pits,
has been subject to more severe erosion than the other soils. Before
the erosion started, the Bw/Co horizon that is exposed here at the
surface was located below A and B horizons in the profile.
Soil Pit B (KH-B), the only one located within Grid 2E, has
the highest magnetic background readings of any site (Walker
et al., 2005). In appearance, the soil is very similar to KH-C
and KH-D, although the upper horizon appears more purple,
perhaps indicating a higher hematite content. Six horizons were
identified, all of them showing significant weathering (Fig. 2b).
There is no strong indication of clay translocation in the soil.
Originally, these horizons may have been formed under an A
horizon, characterized by the accumulation of organic matter, as
found in most of the other soil types on Kaho’olawe (Nakamura
and Smith, 1995). The absence of an A horizon demonstrates
the erosive history of the soil. Starting from a depth of around
0.4 m, the soil color gradually changes from red (10R) to yellowish red (5YR). Below 1.1 m, the amount of unweathered parent
material in the soil increases, and the physical properties start to
resemble the material in Soil Pit KH-A.
The profiles and horizon characteristics of Soils KH-C and
KH-D are very similar to that of Soil KH-B and are not discussed in detail. One notable difference between Soil KH-D
and the other soils is the 20-cm-thick layer of eolian or fluvially
deposited material on top of the original profile. This soil is in
a topographic low and has accumulated eroded material from
the soils upslope.
Magnetic Susceptibility
A contour plot of the surface susceptibility readings for
Kaho’olawe indicates a change from high values in the west and
north of the grid to lower values in the east to southeast (Fig. 3a).
The region of large susceptibility corresponds to the area with a
high TEM signal strength (Walker et al., 2005). The values are
comparatively much higher than on Waimea and Waikoloa (Fig.
3b and 3c). At the Waimea site, the ground surface was covered
by grass and a shallow litter layer. The low bulk density and the
high organic matter content reduced the susceptibility readings.
The laboratory Bartington measurements with the MS2B
sensor system indicate a very large magnetic susceptibility in the
top horizons of Soils KH-B and KH-C (Fig. 4). Below a depth of
0.7 m, the magnetic susceptibility decreases significantly to values
similar to those found in Soil KH-A. The high magnetic susceptibility values in the Bw1 to Bw4 horizons (Fig. 2b) are probably
associated with the increase in concentration of Fe oxides, which
are the most stable end product for low-pH soils in this climate.
In Soil KH-A, the magnetic susceptibility values are much lower
than those in Soil KH-B, and similar to values found for unweathered, fresh basalt from the Big Island of Hawaii (Fig. 4b–4d). The
magnetic susceptibility values (measured with the MS2D sensor)
for the top horizons (Bw) in Pits KH-A and KH-B (1000 and
SSSAJ: Volume 72: Number 1 • January–February 2008
Fig. 3. Surface soil magnetic susceptibilities (corrected for air drift)
measured in the field using a Bartington MS2 sensor with Dloop at a frequency of 0.958 kHz at (a) Kaho’olawe, (b) Waikoloa, and (c) Waimea field sites. The filled circles indicate the
exact locations of the measurements, distributed according to
a stratified random sampling scheme in the 30- by 30-m grids.
Sample material from the numbered locations (0,5; 0,20; 5,25;
15,15; and 20,0) has been used for additional analysis (e.g.,
Table 3). The open squares give locations of the soil pits.
3000 u 105, respectively) are consistent with those observed in
the surface measurements (Fig. 3a).
249
Fig. 4. Results of magnetic susceptibility measurements for Soils A, B, and C from Kaho’olawe and fresh basalt from the Big Island, measured using the Bartington MS2 system with B sensor. The fresh basalt value is an average of measurements on two separate rock
samples (Table 3): (a) low-frequency (0.46-kHz) volume-specific susceptibility (Nlf), (b) low-frequency mass-specific susceptibility
(Flf), (c) the difference between volume-specific susceptibility measured at 0.46 and 4.6 kHz ('N), which is a good measure for the
amount of viscous remanent magnetization, and (d) frequency dependence percentage (Ffd), measured at 0.46 and 4.6 kHz, which
is a measure of the amount of neoformed ultra-fine ferrimagnetic minerals.
The frequency dependence is highest in the B horizon of
Soil KH-B (Fig. 4). This trend is in agreement with observations in earlier studies and is usually associated with neo-formation of superparamagnetic ultra-fine maghemite (Mullins,
1977; Singer et al., 1996).
Mineralogy
X-Ray Analysis Results
Four soil samples from different depths in Soil Pits KH-A
and KH-B were analyzed using XRD (Table 2). Comparison of
the diffractogram peaks with the ICDD database has allowed us
to qualitatively group the data into three categories: (i) present,
for which the peaks in the diffractogram showed a good overlap
with the database; (ii) indication of presence, for which there was
overlap of most peaks; and (iii) not present, for which there was no
or very poor overlap between peaks in the diffractograms and the
database. Soil KH-A is characterized by the abundance of numerous minerals at all depths in the profile. Kaolinite is only present in
the top two samples. Gibbsite is present in all samples, except for
Sample KH-A3, where an indication of the mineral’s presence was
found. Soil KH-B is dominated by the Fe oxide minerals goethite,
hematite, and magnetite or maghemite in the top three samples.
The absence of ilmenite, one of the most stable lithogenic Fe minerals (Cornell and Schwertmann, 2003), shows that the soil material in KH-B is highly weathered. The Bw/Co horizon sample has
characteristics similar to Soil KH-A: both gibbsite and muscovite
are present, and an indication of the presence of hematite and
magnetite or maghemite was observed. The yellowish color of Soil
KH-A and the Bw/Co horizon in Soil KH-B can be caused by several factors. Among the Fe oxides, goethite is usually thought to be
responsible for a yellowish soil color (Cornell and Schwertmann,
2003). In this case, while there is no indication for a change in the
amount of goethite, it is possible that a reduction in red-coloring
hematite (Table 2) has resulted in more influence on the soil color
from goethite. Alternatively, the color change may be explained by
the presence of gibbsite (Nakamura and Smith, 1995).
X-ray fluorescence spectroscopy analysis was performed on
five surface samples from the Kaho’olawe, Waimea, and Waikoloa
Table 2. Results of x-ray diffraction analysis of selected samples from Soils KH-A and KH-B on Kaho’olawe.
Sample
Depth
Horizon
Goethite
D-FeOOH
Hematite
D-Fe2O3
Magnetite Fe3O4/
Maghemite J-Fe2O3
Ilmenite
FeTiO3
Gibbsite
Al(OH)3
Muscovite
(Mica’s)
x
x
x
x
x
o
x
x
x
x
x
x
x
x
Kaolinite
m
KH-A1
KH-A3
KH-A5
KH-A7
0–0.05
0.1–0.25
0.4–0.6
0.8–1.1
Bw/Co
Co1
Co2
Co2
x†
x
x
x
x
o
o
o
x
x
x
x
KH-B1
KH-B5
KH-B8
KH-B15
0–0.05
0.2–0.3
0.5–0.6
1.2–1.3
Bw1
Bw3
Bw4
Bw/Co
x
x
x
x
x
x
x
o
x
x
x
o
x
o
† x = clear presence of the particular mineral (i.e., diffractogram peaks show a good overlap with the International Center for Diffraction Data database), o = indication of the mineral’s presence, i.e., a few of the mineral-specific peaks in the database do not overlap with the peaks in the
x-ray spectrum for the sample.
250
SSSAJ: Volume 72: Number 1 • January–February 2008
Table 3. Results of x-ray fluorescence measurements for soil pit and surface samples from Kaho’olawe (Kaho) and surface samples from the Waikoloa and Waimea sites on the Big Island of Hawaii. Fresh basalt samples (BI-S3) were collected on Kohala Volcano on the Big Island of Hawaii (Soil 3 in Van Dam et al., 2005a). All data have been corrected for loss on ignition.
Location
Waikoloa, surface
Waikoloa, surface
Waikoloa, surface
Waikoloa, surface
Waikoloa, surface
Waimea, surface
Waimea, surface
Waimea, surface
Waimea, surface
Waimea, surface
Kaho, surface
Kaho, surface
Kaho, surface
Kaho, surface
Kaho, surface
Kaho, Pit B
Kaho, Pit B
Kaho, Pit B
Kaho, Pit B
Kaho, Pit A
Kaho, Pit A
Kaho, Pit A
Kaho, Pit A
Kaho, Pit A
Big Island
Big Island
Sample name
WKL-20–00
WKL-15–15
WKL-05–25
WKL-00–20
WKL-00–05
WM-20–00
WM-15–15
WM-05–25
WM-00–20
WM-00–05
KH2E-20–00
KH2E-15–15
KH2E-05–25
KH2E-00–20
KH2E-00–05
KH-B1
KH-B5
KH-B8
KH-B15
KH-A1
KH-A3
KH-A5
KH-A7
KH-A9-saprolite
BI-S3-basalt1
BI-S3-basalt2
SiO2
TiO2
Al2O3 Fe2O3 MnO MgO CaO Na2O K2O P2O5
—————————————— % (w/w) ———————————————
42.1
3.9
26.0
17.9
0.3
3.1
4.0
0.9
1.0
0.8
44.2
3.6
23.9
16.7
0.3
3.4
4.0
1.8
1.5
0.6
49.8
2.9
19.7
11.9
0.2
3.2
6.6
3.0
1.7
0.9
48.7
3.1
20.7
12.9
0.3
2.9
6.2
2.7
1.5
0.9
42.0
4.2
26.1
19.3
0.4
2.7
3.2
0.4
0.7
0.9
46.9
3.5
23.5
16.9
0.6
1.7
2.8
1.0
1.5
1.6
47.7
3.6
23.5
17.0
0.5
1.4
2.8
1.0
1.5
1.1
47.5
3.6
23.7
17.2
0.5
1.4
2.8
0.6
1.5
1.3
46.3
3.7
24.3
18.1
0.4
1.3
2.7
0.5
1.4
1.4
46.3
3.7
24.0
18.2
0.5
1.4
2.5
0.6
1.3
1.4
20.4
10.3
19.0
48.6
0.3
0.4
0.1
0.0
0.5
0.4
19.4
10.2
20.8
48.1
0.1
0.4
0.1
0.0
0.5
0.4
27.4
6.7
32.0
32.9
0.1
0.2
0.1
0.0
0.3
0.3
29.5
6.5
34.4
28.7
0.1
0.3
0.0
0.0
0.2
0.2
22.6
8.5
26.6
41.0
0.1
0.3
0.1
0.0
0.4
0.4
19.3
10.6
17.9
50.3
0.1
0.5
0.3
0.0
0.6
0.4
16.6
12.0
11.9
57.0
0.1
0.9
0.5
0.0
0.5
0.4
16.8
11.9
12.2
55.8
0.3
1.1
1.0
0.0
0.5
0.5
29.1
6.3
29.6
32.0
0.0
0.9
1.5
0.0
0.3
0.2
32.9
6.7
32.4
27.1
0.1
0.3
0.0
0.0
0.2
0.3
33.2
7.0
31.5
27.5
0.1
0.2
0.1
0.0
0.2
0.3
30.8
7.5
32.8
28.1
0.1
0.2
0.1
0.0
0.1
0.3
29.9
7.3
30.1
31.2
0.1
0.6
0.2
0.0
0.2
0.4
39.5
5.4
31.6
22.3
0.1
0.4
0.2
0.1
0.2
0.3
46.4
2.8
13.6
14.2
0.2 10.8
9.3
1.8
0.6
0.3
46.5
2.9
13.6
14.1
0.2 10.9
9.2
1.7
0.6
0.3
Rb
Sr
kg1
—— mg
22
407
25
434
38
1031
38
901
14
235
54
319
52
320
52
299
39
218
36
236
31
46
28
54
17
43
10
39
24
42
32
87
31
141
25
136
8
98
5
52
4
36
0
35
0
32
5
40
14
398
13
401
Zr
——
506
449
499
539
289
291
266
246
197
240
546
564
408
435
543
610
681
628
380
494
506
499
524
431
196
197
Thermal Analysis
field sites, as well as two samples of unweathered basalt
from Kohala Volcano on the Big Island of Hawaii (Table
3). The samples can be grouped according to weathering stage. High Ca and silica contents show that the
Big Island basalt samples are the least weathered (Fig.
5a and 5b), followed by the material from Waikoloa
and Waimea. The Kaho’olawe soil samples are the most
intensely weathered, but the surface samples show that
the mineralogy varies widely, depending on the location
in the grid. The degree of weathering increases from the
saprolite in Soil KH-A (A9 in Fig. 5b), through the
lower weathered horizons (Bw/Co horizon in Soil KHB; B15 in Fig. 5d), to the material in the B horizon of
Soil KH-B (B5 in Fig. 5c).
Analysis of the XRF results indicates that initial
weathering of the basaltic rocks caused Ca to disappear
(Fig. 5a) and silica contents to decrease (Fig. 5b), leading to a relative enrichment of Al and Fe. Continued
weathering leached out Al and increased the Fe content (Fig. 5c). The Fe content is linearly correlated
(R2 = 0.97) to Ti (Fig. 5d), which is considered to be
very stable and least subject to transformations and
phase changes. The Ti/Fe ratio agrees well with val- Fig. 5. Results of x-ray fluorescence spectroscopy measurements for different data
ues reported in the literature for basic extrusive rocks
sets from Kaho’olawe and the Big Island of Hawaii (see also Table 3). All val(Piper, 1987).
ues have been adjusted for loss on ignition. Samples A9, B5, and B15, which
are discussed in the text, are identified for clarity.
SSSAJ: Volume 72: Number 1 • January–February 2008
251
In the plots of thermal analysis, five distinct events can be recognized. In Table 4, the observed
events are correlated with literature
data on the thermal behavior of soil
material (Blazek, 1973; Földvari,
1991; Van Dam et al., 2002). Event
1 (100–120°C) is associated with a
loss of free and capillary water. The
peak shifted to higher temperatures
for more weathered, and thus more
clay-rich, material. The endothermic Event 2 (300°C) may be related
to the dehydroxylation of gibbsite
(Table 4), but the XRD analysis
suggests that gibbsite is present
only in Sample KH-A3 (Table 2).
Therefore, this effect more likely represents dehydroxylation of FeO˜OH
polymorphs (probably goethite)
and transformation to hematite
(2FeO·OH o Fe2O3 + H2O) (Van
Dam et al., 2002). This reaction,
and thus the presence of goethite, is
absent in unweathered basalt from
the Big Island (Fig. 8). Strongly
weathered Sample KH-B5 contains
about 3.5 times as much goethite
as Sample KH-A9. Events 3 and 4
are associated with the same mineral. The weight loss Event 3 around
500°C is probably caused by the loss
of structural water from kaolin or
mica type clay minerals, such as muscovite. Muscovite-type clay would be
most prominent in young or relatively undeveloped material such as
Soil KH-A and Sample KH-B15. In
more weathered soil samples, such
as KH-B5 and KH-B8, muscovite
has deteriorated and the secondary
clays are more stable. This observation is supported by the difference in
clay minerals found using the XRD
Fig. 6. (a) Time–temperature curves of thermogravimetric analyses and their first derivatives
for selected samples from Soil KH-A on Kaho’olawe ('w/'T is the first derivative of timeanalysis (Table 2). The exothermic
dependent weight loss); and (b) differential thermal analyses (DTA) curves for the same
Event 4 at around 950°C is caused
sample set. In the DTA curves, negative peaks denote endothermic reactions (i.e., consum- by a chemical or structural change
ing heat) while positive peaks indicate exothermic reactions (i.e., giving off heat). Insets in
in muscovite- or kaolinite-type clay
(b) show specific events that are discussed in the text.
minerals. Event 5 is associated with
the melting of hematite (Blazek,
Eight samples, representing the variation in magnetic proper1973). Possibly, variations in crystallinity or cation substitutions
ties in Soils KH-A and KH-B, were selected for TGA. In addition,
caused the observed irregular pattern.
one sample of saprolite (KH-A9) and a sample of fresh (unweathMaghemite is expected to produce an exothermic reaction
ered) basalt from a road cut on Kohala Volcano on the Big Island
between 510 and 570°C (Blazek, 1973), while magnetite is
of Hawaii were analyzed. The measurements were performed
expected to produce an exothermic reaction between 590 and
from room temperature up to 1515°C at a constant heating rate
650°C. Possibly, the broad exothermic trend between 500 and
of 10°C min1. The results are presented in Fig. 6 through 8. The
850°C in the DTA curves is related to this effect.
plots of TGA give the normalized weight loss percentage and the
first derivative.
252
SSSAJ: Volume 72: Number 1 • January–February 2008
Chemical Dissolution Techniques
for Iron Mineralogy
Figure 9 shows the variation of Fed
and Feo concentration with depth for
Soils A through D. In all soil profiles, the
amount of Fed is larger than that of Feo
because dithionite extracts both amorphous and crystalline forms of Fe oxides.
Furthermore, the Feo/Fed ratios are very
small, which is indicative of the general
maturity of the soils (Simón et al., 2000).
The lower values of Fed and Feo in soil
KH-A, however, demonstrate that it is
less weathered than the other three soils.
The pattern of distribution of Fed and Feo
with increasing soil depth varies between
the two Fe species. While the Feo values
remain relatively constant below 100cm depth for Soils KH-B and KH-C
(Fig. 9a), the values for Fed show a clear
decline. This pattern is also observed in
the measurements for Ffd% (Fig. 4d) and
Flf (Fig. 4b), respectively.
Temperature-Dependent Magnetic
Susceptibility Analysis
Using a Kappabridge KLY-2 AC susceptibility bridge and CS-2 furnace, the
magnetic susceptibility was measured
across a broad temperature range of 30 to
650°C for eight samples from Soils KH-A
and KH-B and three surface samples from
the three field locations. The susceptibilities at the start of the measurements (indicated by closed symbols, Fig. 10) agree
reasonably well with the values found
with the MS2 system (Fig. 4). Small differences may be explained by intersample
variations in mineralogy and by equipment effects (magnetic field strength and
frequency). The large discrepancy between Fig. 7. (a) Time–temperature curves of thermogravimetric analyses and their first derivatives
starting susceptibility and that measured
for selected samples from Soil KH-B on Kaho’olawe ('w/'T is the first derivative of
time-dependent weight loss); and (b) curves of differential thermal analyses (DTA) for
with the MS2 system for Sample KH-B15
the same sample set. For reference, the results of a fresh basalt sample (BI-S3) have
is not understood at this point.
been included. In the DTA curves, negative peaks denote endothermic reactions while
The results reveal the presence of a
positive peaks indicate exothermic reactions. The drop in the DTA curve between 700
number of processes. The first process is
and 800°C for the Big Island basalt sample (BI-S3) is due to an automatic equipment
the general increase in susceptibility from
scale adjustment. Insets in (b) show specific events that are discussed in the text.
room temperature to around 300°C. This
is possibly the result of the production of
tite exists in any of the samples; lithogenic magnetite would retain
small amounts of magnetite or maghemite under the influence of
its high susceptibility across the entire temperature range.
organic matter (Tite and Linington, 1975). Another process is the
A number of differences were observed between the
broad peak in the susceptibility around 300°C, which is probably
curves from Soils KH-A and KH-B (Fig. 10a and 10b). The
associated with the dehydroxylation of goethite, which was also
decrease in susceptibility associated with the transformation
observed in the TGA and DTA diagrams (Fig. 6 and 7). Between
of maghemite to hematite started at a lower temperature for
around 450 and 520°C, the susceptibility rapidly decreased and
the samples from Soil KH-B than those from Soil KH-A. The
completely disappeared. This effect is the result of the transformadifferent return curves differ markedly between samples from
tion of maghemite to hematite—a process that, under reducing
Soils KH-A and KH-B; samples from Soil KH-A show a gradconditions, is absent in magnetite (Cornell and Schwertmann,
ual increase in F with cooling, while Sample KH-B5 is char2003). This result shows that no lithogenic or pedogenic magneacterized by a rapid increase, followed by a decrease in F. The
SSSAJ: Volume 72: Number 1 • January–February 2008
253
Fig. 8. Thermogravimetric analysis curves and first derivatives
for a sample of unweathered basalt (BI-S3-basalt2) from
Kohala Volcano on the Big Island of Hawaii, a saprolite
sample from Kaho’olawe Soil KH-A (KH-A9, see Table 3),
and a strongly weathered sample (KH-B5) from Soil KH-B
on Kaho’olawe ('w/'T is the first derivative of time-dependent weight loss).
observed differences in FT curves between samples from Soils
KH-A and KH-B may be used as an estimate for the amount
of hyperfine maghemite present, which would be an important
new way to characterize the presence of superparamagnetic
minerals. Because the processes and rates of thermal interconversions between maghemite and hematite are dependent on
crystal size (Cornell and Schwertmann, 2003), however, more
research is required to make quantitative estimates of the proposed relationship between FT and hyperfine maghemite.
DISCUSSION AND CONCLUSIONS
In Fig. 11 to 13, the different measures of mineralogy and
magnetic properties are compared for selected samples. Not all
analyses have been performed for all samples so that, in this discussion, there is a focus on the samples from Kaho’olawe and the
fresh basalt from the Big Island of Hawaii. The XRF analysis of a
large range of soils on Kaho’olawe and the Big Island showed that
Fig. 9. Results of selective chemical dissolution of pedogenic Fe
oxides using a flame atomic absorption spectrometer for
Soils A, B, C, and D: (a) total amount of secondary Fe oxyhydroxides (Fed); and (b) poorly crystalline and amorphous
Fe oxyhydroxides (Feo).
an increase in the degree of soil development is accompanied by a
gradual increase in the Fe content of the soils (Table 3, Fig. 5). For
the samples from Kaho’olawe, this increase
in Fe content, in turn, is positively correTable 4. Comparison of thermogravimetric and differential thermal analyses (DTA)
with events described in the literature.
lated (R2 = 0.87) with the Flf (Fig. 11a). The
fresh basalt samples and soil material from
DTA peak
Reaction
Probable effect
Temperature
Waimea, on the Big Island of Hawaii, seem
temperature
type
to fit with this general trend. The samples
°C
°C
from Waikoloa, however, exhibit anomaloss of free water
110
endothermic
~100
transformation of ferrihydrite to hematite
lously high magnetic susceptibility values
dehydration of gibbsite and formation of böhmite†
250–330
(Fig. 11a), which are as of yet unexplained.
280
endothermic dehydroxilation of gibbsite‡
270–380
The positive correlation between Fe
dehydroxilation of goethite§
280
content
and the 300°C weight loss event
decomposition of böhmite (J-AlO·OH)†
470
endothermic
(estimated using the 260–320°C weight
loss of structural water from kaolin or micas†
loss percentage in Fig. 11b [R2 = 0.95])
decomposition of böhmite (J-AlO·OH)†
510
endothermic
loss of structural water from kaolin or micas†
indicates an increase in goethite content
chemical or structural change in kaolinite or mica
during initial soil development. For the
900
exothermic
(muscovite)†
most strongly developed soil material in
~1380
endothermic hematite melting†
1360–1440
Soil KH-B (thus excluding Sample KH† From Blazek (1973).
B15), however, the 300°C weight loss
‡ From Földvari (1991).
(~goethite content) stabilized at around
§ From Van Dam et al. (2002).
3.5 to 4%. The weight loss event at around
254
SSSAJ: Volume 72: Number 1 • January–February 2008
Fig. 10. Plots of magnetic susceptibility (F) vs. temperature for
samples from (a) Kaho’olawe Soil A, (b) Kaho’olawe Soil
B, and (c) surface samples from the Kaho’olawe, Waimea,
and Waikoloa sites. Sample codes refer to those in Table 3.
The start of each curve is marked with a closed symbol. The
return curves that show the change in magnetic susceptibility for decreasing temperatures are, where present, marked
with an open symbol.
500°C (estimated using the 480–550°C weight loss percentage), which can be related to loss of structural water from clay
minerals, is inversely proportional (R2 = 0.99 for Kaho’olawe
samples) to the Fe content of the soil material (Fig. 11c). This
shows that muscovite-type clays, which are most prominent
in relatively weakly developed soil material such as Soil KH-A
and Sample KH-B15, are replaced by different clay types in
more mature soil material, such as KH-B5 and KH-B8 (see
also Table 2).
The plots in Fig. 12, which directly correlate the mineralogical data with the magnetic properties for Kaho’olawe Soils
KH-A and KH-B and the Big Island fresh basalt, show two clusters of data points and one individual sample (the fresh basalt).
These clusters may represent individual uncorrelated groups
or indicate a lack of data for intermediate conditions. Since in
most previous analyses the Kaho’olawe surface samples occupy
intermediate positions between Soils KH-A and KH-B (see Fig.
SSSAJ: Volume 72: Number 1 • January–February 2008
Fig. 11. Cross plots showing correlations between different measures of mineralogy and magnetic properties for soil samples
from Kaho’olawe and the Big Island of Hawaii: (a) low-frequency magnetic susceptibility (Flf) vs. x-ray fluorescence
(XRF) Fe (Fe2O3), (b) the 300°C weight loss event vs. XRF Fe,
and (c) the 500°C weight loss event vs. XRF Fe. Pearson correlation coefficients are given for subsets of the data.
5a–5d and 11a), we believe that it is likely that future analysis
of the Kaho’olawe surface samples using thermal analysis and
255
Fig. 13. Cross plot showing amorphous Fe oxyhydroxides (Feo) percentage vs. frequency-dependent magnetic susceptibility (Ffd).
Fig. 12. Cross plots showing correlations between different
measures of mineralogy and magnetic properties for soil
samples from Kaho’olawe and the Big Island of Hawaii: (a)
the 300°C weight loss event vs. low-frequency magnetic
susceptibility (Flf), (b) the 500°C weight loss event vs. Flf,
and (c) the 1400°C weight loss event vs. Flf. The legend in
(b) applies to all panes. Samples A9, B5, and B15, which
are discussed in the text, are identified for clarity.
256
chemical dissolution techniques may fill the gap between the
two data clusters. Figure 12a, which correlates the Flf with the
300°C weight loss event (~goethite content), suggests that Flf
does not increase significantly until the apparent maximum goethite content (Fig. 11b) is reached. For Kaho’olawe soils, the Flf
correlates negatively with the 500°C weight loss event, showing
that a reduction in muscovite-type clays, which are more prominent in weakly developed soils, is associated with an increase in
magnetic susceptibility. Figure 12c shows a positive correlation
between Flf and the 1400°C weight loss event (estimated using
the 1350–1500°C weight loss percentage). This weight loss event
is positively correlated with XRF estimated amounts of Fe2O3,
SiO2, and TiO2, and is probably an indication of the sample’s
hematite content (Table 4).
Figure 13 shows the correlation between the Feo content
and the Ffd% (~superparamagnetic minerals), suggesting that
the amount of superparamagnetic minerals in a soil can be estimated not only through its Ffd%, but also using hydroxylamine
extraction of amorphous Fe oxyhydroxides.
This study dealt with the mineralogical causes of the magnetic behavior of soils on Kaho’olawe and the Big Island, Hawaii.
While the primary origin of the magnetic minerals in the soils
is the basaltic parent material, the variation in mineralogy and
magnetic properties of the soils reflects a range of processes as
well as the duration and intensity of weathering. For example,
the relatively weak magnetic susceptibility of the Big Island sites
is a reflection of the low degree of weathering of the soils. On
Kaho’olawe, which has an exposed surface, the variation in soil
magnetic properties is related to the amount of erosion that the
soils have received. Of the two soil series at our field site, the
Kaneloa soils, which are found in higher topographical locations,
have been stripped of their B horizon. Magnetic measurements
and mineralogical composition indicate that Soils KH-B and
KH-A are separate segments of one continuous soil profile, so
that the Kaneloa and Puu Moiwi series (Nakamura and Smith,
1995) could be classified as one soil type.
The results from this study show the mineralogical and magnetic effects associated with soil development on basaltic parent
SSSAJ: Volume 72: Number 1 • January–February 2008
material. The XRD analyses show that some primary-rock minerals
still persist in young soils (e.g., ilmenite in Soil KH-A), while newly
formed Fe oxides are present in large amounts in the B horizon of
Soil KH-B. The XRF analysis shows that initial soil development
led to rapid leaching of Ca and a gradual decrease in silica, which
resulted in a relative enrichment of Al and Fe. On Kaho’olawe, with
a surface age of at least 1 million yr, the conditions have been favorable for leaching of Al and a further development toward an Oxisol
with large amounts of Fe. Thermal analysis allowed estimation of
the relative amounts of goethite and muscovite, both of which vary
systematically with position in the soil profile and correlate with the
observed variations in magnetic properties. The results from the FT
analysis shows that the high magnetic susceptibility and the large
VRM observed in geophysical surveys in Kaho’olawe are completely
due to secondary, or neoformed, minerals.
ACKNOWLEDGMENTS
We thank the Kaho’olawe Island Reserve Commission (KIRC) for
permission to collect data on Kaho’olawe. We thank KIRC staff for
logistical support, Len Pasion and Sean Walker for organization of the field
campaign, Chris McKee (New Mexico Bureau of Geology and Mineral
Resources’ XRD Lab), Scott Morton, and Shinku Sky for field and lab
assistance, and Mark Dekkers, José Fabris, Russell Harmon, Jan Hendrickx,
Michael Velbel, and Tom Vogel for discussions. The research was funded by
grants from the U.S. Army Research Office (STIR W911NF-04-1-0433
to RvD) and SERDP (UX-1414 to R. Van Dam, J.B.J. Harrison, and Y. Li,
and UX-1355 to Y. Li). R.E. North was funded by the U.S. Army ERDC
Basic Research Program and granted permission to publish by the director
of the Geotechnical & Structures Lab.
REFERENCES
Billings, S.D., L.R. Pasion, D.W. Oldenburg, and J. Foley. 2003. The influence
of magnetic viscosity on electromagnetic sensors. p. 123–130. In EUDEMSCOT2, Int. Conf. on Requirements and Technologies for the Detection,
Removal and Neutralization of Landmines and UXO, Brussels. 15–18
Sept. 2003. Vrije Universiteit Brussels, Brussels, Belgium.
Blazek, A. 1973. Thermal analysis. Van Nostrand Reinhold Co., London.
Butler, D.K. 2003. Implications of magnetic backgrounds for unexploded
ordnance detection. J. Appl. Geophys. 54:111–125.
Chao, T.T., and L. Zhou. 1983. Extraction techniques for selective dissolution
of amorphous iron oxides from soils and sediments. Soil Sci. Soc. Am.
J. 47:225–232.
Clague, D.A., and G.B. Dalrymple. 1987. The Hawaiian-Emperor volcanic
chain. Part II. p. 5–54. In R.W. Decker et al. (ed.) Volcanism in Hawaii.
Vol. 1. U.S. Dep. of the Interior, Washington, DC.
Clague, D.A., and G.B. Dalrymple. 1989. Tectonics, geochronology, and
origin of the Hawaiian-Emperor Chain. p. 188–217. In E.L. Winterer
et al. (ed.) The Eastern Pacific Ocean and Hawaii: The geology of North
America. Geol. Soc. Am., Boulder, CO.
Cornell, R.M., and U. Schwertmann. 2003. The iron oxides: Structure, properties,
reactions, occurrences and uses. Wiley-VCH, Weinheim, Germany.
Das, Y., J.E. McFee, and G. Cross. 2002. Soil properties database for
humanitarian demining: A proposed initiative. In Proc. World Congr.
of Soil Sci., 17th, Bangkok [CD-ROM].14–21 Aug. 2002. Soil and
Fertilizer Soc. of Thailand, Bangkok.
Dearing, J.A. 1994. Environmental magnetic susceptibility: Using the
Bartington MS2 system. Chi Publ., Fairfield, CT.
Fabris, D., J.M.D. Coey, and W.D.N. Mussel. 1998. Magnetic soils from
mafic lithodomains in Brazil. Hyperfine Interact. 113:249–258.
SSSAJ: Volume 72: Number 1 • January–February 2008
Földvari, M. 1991. Measurement of different water species in minerals by means
of thermal derivatography. p. 84–101. In W. Smykatz-Kloss and S.S.J.
Warne (ed.) Thermal analysis in geosciences. Springer-Verlag, Berlin.
Goulart, A.T., J.D. Fabris, M.F. de Jesus Filho, J.M.D. Coey, G.M. da Costa,
and E. De Grave. 1998. Iron oxides in a soil developed from basalt.
Clays Clay Miner. 46:369–378.
Holmgren, G.G.S. 1967. A rapid citrate–dithionite extractable iron procedure.
Soil Sci. Soc. Am. Proc. 31:210–211.
Lindsley, D.H. (ed.). 1991. Oxide minerals: Petrologic and magnetic
significance. Rev. Mineral. Ser. 25. Mineral. Soc. Am., Chantilly, VA.
Mullins, C.E. 1977. Magnetic susceptibility of the soil and its significance in
soil science—A review. J. Soil Sci. 28:223–246.
Nakamura, S., and C.W. Smith. 1995. Soil survey of the Island of Kahoolawe,
Hawaii. U.S. Dep. of the Interior, Washington, DC.
Piper, J.D.A. 1987. Paleomagnetism and the continental crust. John Wiley &
Sons, New York.
Simón, M., S. Sánchez, and I. García. 2000. Soil landscape evolution on a
Mediterranean high mountain. Catena 39:211–231.
Singer, M.J., K.L. Verosub, P. Fine, and J. TenPas. 1996. A conceptual model
for the enhancement of magnetic susceptibility in soils. Quat. Int. 34–
36:243–248.
Speyer, R.F. 1994. Thermal analysis of materials. Marcel Dekker, New York.
Stucki, J.W., B.A. Goodman, and U. Schwertmann (ed.). 1988. Iron in soils
and clay minerals. D. Reidel Publ. Co., Dordrecht, the Netherlands.
Tite, M.S., and R.E. Linington. 1975. Effect of climate on the magnetic
susceptibility of soils. Nature 256:565–566.
Thompson, R., and F. Oldfield. 1986. Environmental magnetism. Allen &
Unwin, London.
Van Dam, R.L., J.B.J. Harrison, J.M.H. Hendrickx, B. Borchers, R.E. North, J.E.
Simms, C. Jasper, C.W. Smith, and Y. Li. 2005a. Variability of magnetic soil
properties in Hawaii. p. 157–164. In R.S. Harmon et al. (ed.) Proc. Detection
and Remediation Technologies for Mines and Minelike Targets, 10th, Orlando.
28 Mar.–1 Apr. 2005. Soc. for Optical Eng., Bellingham, WA.
Van Dam, R.L., J.M.H. Hendrickx, J.B.J. Harrison, and B. Borchers. 2005b.
Conceptual model for prediction of magnetic properties in tropical soils.
p. 177–187. In R.S. Harmon et al. (ed.) Proc. Detection and Remediation
Technologies for Mines and Minelike Targets, 10th, Orlando. 28 Mar.–1
Apr. 2005. Soc. for Optical Eng., Bellingham, WA.
Van Dam, R.L., J.M.H. Hendrickx, B. Harrison, B. Borchers, D.I. Norman,
S.A. Ndur, C. Jasper, P. Niemeyer, R. Nartey, D. Vega, L. Calvo, and
J.E. Simms. 2004. Spatial variability of magnetic soil properties. p.
665–676. In R.S. Harmon et al. (ed.) Proc. Detection and Remediation
Technologies for Mines and Minelike Targets, 9th, Orlando. 12–16 Apr.
2004. Soc. for Optical Eng., Bellingham, WA.
Van Dam, R.L., W. Schlager, M.J. Dekkers, and J.A. Huisman. 2002. Iron
oxides as a cause of GPR reflections. Geophysics 67:536–545.
Walker, S.E., L.R. Pasion, S.D. Billings, Y. Li, and D.W. Oldenburg. 2005. Examples
of the effect of magnetic soil environments on time domain electromagnetic
data. In Proc. Symp. Application of Geophysics to Eng. and Environ. Problems,
Atlanta. 3–7 Apr. 2005. Environ. Eng. Geophys. Soc., Denver.
Ward, S.H. (ed.). 1990. Geotechnical and environmental geophysics. Soc.
Explor. Geophys., Tulsa, OK.
Webster, R., and M.A. Oliver. 1990. Statistical methods in soil and land
resource survey. Oxford Univ. Press, Oxford, UK.
Wedepohl, K.H. 1969. Composition and abundance of common igneous
rocks. p. 227–249. In K.H. Wedepohl (ed.) Handbook of geochemistry.
Springer-Verlag, Berlin.
Wolfe, E.W., and J. Morris. 1996. Geologic map of the island of Hawaii
(1:100,000). U.S. Dep. of the Interior, Washington, DC.
Wood, C.A., and J. Kienle (ed.). 1990. Volcanoes of North America: United
States and Canada. Cambridge Univ. Press, New York.
Wright, T.L., and D.A. Clague. 1989. Petrology of Hawaiian lava. p. 218–237.
In E.L. Winterer et al. (ed.) The Eastern Pacific Ocean and Hawaii: The
geology of North America. Geol. Soc. Am., Boulder, CO.
257