Geoderma 303 (2017) 143–149

Contents lists available at ScienceDirect

Geoderma
journal homepage: www.elsevier.com/locate/geoderma

Relationship between magnetic properties and reddening of tropical soils as MARK

indicators of weathering
Holger Preetza,b, Jan Igela,⁎, Jacqueline A. Hannamc, Susanne Stadlerd
a
Leibniz Institute for Applied Geophysics, Stilleweg 2, D-30655 Hannover, Germany
b
Federal Competence Center for Soil and Groundwater Protection/UXO Clearance, Hannover, Germany
c
School of Water, Energy and Environment, Cranfield University, Cranfield, MK43 0AL, UK
d
Federal Institute for Geosciences and Natural Resources (BGR), Stilleweg 2, D-30655 Hannover, Germany

A R T I C L E I N F O A B S T R A C T

Keywords: This study evaluates the use of magnetic properties as an indicator of weathering of tropical soils. Soil samples
Tropical soil collected across the tropical belt were analysed for magnetic susceptibility and its frequency dependence.
Laterite Frequency dependence is caused by superparamagnetic (SP) ferrimagnetic nanoparticles, which are commonly
Soil development attributed to neoformation during soil forming processes. Magnetic properties are compared to the redness
Iron oxide
rating of soil colour, which is related to the hematite content and is an established proxy for soil weathering. The
Reddening
investigated samples comprise material of different weathering stages from unweathered and weathered rock to
Magnetic susceptibility
subsoil and strongly weathered topsoil. They cover a broad variety of parent materials: ultrabasic, basic,
intermediate, acid igneous rocks, clays and clay slates, phyllites and sandstones. The results show that soil
reddening, magnetic susceptibility and in particular frequency-dependent susceptibility generally increase with
proceeding weathering. However, there is a lithologic overprint and the parent material has to be taken into
consideration. Soils stemming from acid igneous rocks, clays, clay slates and phyllites show a positive correlation
between reddening and susceptibility or frequency dependent susceptibility, rendering these properties suitable
for indicating weathering. In contrast, soils stemming from ultrabasic, basic and intermediate igneous rocks and
sandstones show no significant correlation. The reason is the strong lithogenic overprint of ferrimagnetic
minerals including SP particles, which commonly occur in these rock types.

1. Introduction 1.2. Reddening

1.1. Lateritic soil formation Hematite is a main constituent of lateritic soils and is responsible for
their characteristic red colour (Torrent et al., 1983; Fontes and
Laterites, (ferralsols, plinthosols or oxisols FAO, 2006; NRCS, 2006) Carvalho, 2005), and is an indicator of weathering state
are one of the most widespread soil types in the humid tropics. They are (Schwertmann, 1993). The neoformation of hematite from the dehy-
the outcome of long-lasting and intense chemical weathering processes dration of ferrihydrite requires elevated soil temperatures and low
under humid conditions under variable soil moisture regimes. Laterites water activity (Schwertmann, 1993). Typical laterites are therefore
are generally characterised by the following features: strong weathering different to strongly weathered soils under permanent humid soil
of silicates; the release of Fe and Al ions forming new minerals such as moisture regimes which favour goethite over hematite formation
gibbsite and Fe oxides (predominantly goethite and hematite) (Tardy (Fontes and Carvalho, 2005), where in this case, hematite content is
and Roquin, 1992) and the prevalence of kaolinite. Laterites are formed not a suitable proxy for weathering. The verification of the paragenetic
from different geologies (e.g. silicate rocks and limestone) and their formation of maghemite and hematite was accomplished by Torrent
composition and properties are strongly controlled by the chemical et al. (2006, 2010a).
composition of the parent rock (Schellmann, 1981).
1.3. Soil magnetic susceptibility

Soil magnetic susceptibility is caused by the presence of ferrimag-

⁎
Corresponding author.
E-mail address: [email protected] (J. Igel).

http://dx.doi.org/10.1016/j.geoderma.2017.05.007
Received 2 November 2016; Received in revised form 4 April 2017; Accepted 3 May 2017
Available online 19 May 2017
0016-7061/ © 2017 Elsevier B.V. All rights reserved.
H. Preetz et al. Geoderma 303 (2017) 143–149

netic Fe and Fe/Ti Oxides such as magnetite, titanomagnetite and soils of temperate or subtropical regions, only very few analyses are
maghemite that are stable in soils and can accumulate due to their available for tropical soils. It is generally assumed that an increase in
resistance to weathering. The absolute value of susceptibility is soil magnetic susceptibility occurs with an increasing age of the soil and
proportional to the amount of ferrimagnetic minerals and is also with climate being warmer and wetter (cf. Singer et al., 1996; Torrent
influenced by the mineral particle size. Grain sizes depend on the et al., 2010a). Therefore deeply weathered lateritic soils are suitable
provenance and conditions of formation. Macro-sized multidomain environments to verify the relationship between soil formation and soil
(MD) ferrimagnetic minerals tend to be of lithogenic origin (formed magnetic properties.
during the crystallisation of igneous rocks), derived from the weath- The aim of this study is to i) investigate the potential of ferrimag-
ering of ultrabasic and basic igneous rocks. In contrast, ultrafine netic compounds as a proxy for soil formation and weathering in
grained ferrimagnetic minerals are thought to form mainly during tropical environments, ii) identify if magnetic susceptibility is a result
pedogenesis (Maher and Taylor, 1988). These minerals are super- of pedogenic neoformation or if there is a significant lithogenic over-
paramagnetic (SP) and show frequency-dependence of magnetic sus- print and iii) investigate the relationship between soil reddening and
ceptibility. This behaviour is based on the relaxation of the SP minerals magnetic susceptibility according to different parent rock types.
which gives rise to a time decay of magnetisation termed magnetic
viscosity (Néel, 1949). Thus, quantifying this property can potentially
provide a proxy for soil formation and weathering in certain environ- 2. Materials and methods
ments.
2.1. Soil samples
1.4. Ferrimagnetic mineral formation and co-formation of hematite
The soil samples originate from the geoscientific collection of the
Ferrimagnetic minerals of pedogenic origin form during weathering Federal Institute for Geosciences and Natural Resources (BGR) in
and soil genesis by geochemical or bacterial processes. There are Hanover, Germany. The sample collection (n = 1475 samples) includes
several theories describing these processes. The first describes ferrihy- lateritic soils and their siliciclastic parent material from tropical
drite transforming to magnetite via partial dehydration and oxidation regions. Samples were collected from the field at the beginning of the
in the presence of excess Fe2 + ions in solution (Schwertmann, 1988; 1970s. These laterites are classified as Ferralsols and Plinthosols (FAO,
Dearing et al., 1996). Subsequently, magnetite may oxidise to maghe- 2006) or Oxisols (NRCS, 2006). The samples were originally collected
mite. This pathway requires the sufficient release of Fe from the parent to investigate mineral alterations under intense tropical weathering and
material through hydrolysis and subsequent dissolution by Fe-reducing the genesis of mineral deposits (Schellmann, 1974). At the time of
bacteria. When a critical concentration of Fe2 + is reached, it will be collection the samples were air-dried, mechanically crushed to < 2 mm
oxidised. These transformations are facilitated by wetting and drying and homogenised. Since preparation they have been stored in an air-
cycles in soils with corresponding changes in water activity, pH and dried state in the sample archive of the BGR in plastic boxes under
redox conditions. constant indoor climate.
A second theory assumes pedogenic enhancement of ferrimagnetic A sub-set of 506 samples was selected and includes topsoil and
minerals in tropical and subtropical soils also occurs in conjunction subsoil from various depths, in addition to weathered and unweathered
with pedogenic hematite formation (Torrent et al., 2006). Warm to hot parent rocks (Table 1). Table 1 also lists the countries of origin of the
and sub-humid to humid climates favour the dehydration that is samples and parent rock type based on the original field description.
necessary to transform ferrihydrite to hematite (Schwertmann, 1988). The classification of the igneous rocks was made by consulting available
The model suggests transformation occurs with intermediate steps geochemical data of the samples which resulted in basic gabbro being
producing ferrimagnetic minerals: ferrihydrite – SP (superparamag- classified as ultrabasic.
netic) maghemite (i.e. nanosized ferrimagnetic minerals) – SD (ferri-
magnetic single domain) maghemite – hematite (Torrent et al., 2006).
The intensity of each step of the pathway varies according to the
Table 1
pedoclimate and degree of weathering. This means that rates of Overview of the investigated material, including number of samples, country of origin
ferrihydrite and maghemite formation are relatively fast compared to and parent rock material. Ultrabasic igneous rocks: gabbro, phonolite, serpentinite; Basic
hematite formation in permanently moist but not saturated soils. In and intermediate igneous rocks: amphibolite, andesite, basalt, olivine-feldspar-basalt,
contrast, tropical soils undergoing seasonal drying tend to exhibit a charnockite, diabase, dolerite, gabbro, gneiss, phonolite; Acid igneous rocks: charnockite,
dolerite, gneiss, biotite-gneiss, granitic gneiss, granite; Clays/clay slates: pisolite, slate, shale,
faster hematite formation. However, Torrent et al. (2010b) stress that quartzitic slate, clay slate, carbonaceous clay, tertiary sediments; Phyllites: phyllite;
such processes are highly complex and should be carefully evaluated Sandstones: sandstone, quartzite.
and compared only when formed on similar parent materials.
The transformation pathways for ferrimagnetic minerals and hema- Origin Ultrabasic Basic/ Acid Clay/ Phyllites Sand- Σ
Inter- Clay stones
tite are in agreement with the general conditions that favour the
mediate slates
formation of laterites where various mechanisms of hydration and
dehydration create sequences of minerals in different hydration states Australia 9 39 26 10 84
(Tardy and Roquin, 1992). Thus laterites provide a key environment to Brazil 21 30 9 11 14 85
El Salvador 10 10
investigate the use of magnetic measurements as indicators of soil
Ghana 17 10 7 9 43
formation and development. Guatemala 8 8
While it is common to regard the visible Fe oxides such as hematite Hawaii 19 19
and goethite as products and indicators of soil development it is still India 9 36 10 55
rare to use the ferrimagnetic iron oxides for that purpose. It has been Madagascar 15 16 16 47
Mexico 4 4
demonstrated that the presence of the magnetic iron oxide fraction of New Caledonia 20 20
soils is also a widespread phenomenon that responds to and reflects soil Philippines 8 8
forming processes (Mullins (1977); Maher (1986). The advantage of Puerto Rico 21 4 25
using magnetic susceptibility as an indicator for soil development is Sri Lanka 4 20 24
Uganda 8 22 16 46
that this property provides a quantitative measure of the amount of
Venezuela 13 8 7 28
ferrimagnets and is relatively easy to measure. Σ 106 117 170 73 21 19 506
Many investigations of magnetic properties have been conducted in

144
H. Preetz et al. Geoderma 303 (2017) 143–149

20 10 ml plastic boxes and measured at two frequencies (low frequency

(LF): 465 Hz and high frequency (HF): 4.65 kHz) at a magnetic field
strength of approximately H = 80 A/m. Each sample was measured ten
times to ensure high reproducibility and the mean value was used for
further analysis. In addition, the filling levels of the boxes were
15 measured and the susceptibility results were adjusted according to
the sample volume (Preetz et al., 2008). Volume-specific susceptibility
RR, person 2

κ was transformed into mass-specific susceptibility χ by dividing by the

r = 0.90
sample bulk density.
10
2.3. Determination of soil colour

Most soils containing hematite have Munsell hues between 5YR and
10R. The occurrence of these colours usually relate to the state of
5
weathering due to the pigmenting power of hematite (Schwertmann,
1993).
The colours of the dry and homogenised soil samples were visually
determined using the “Munsell Soil Color Charts” (Munsell Color
0 Company, 1975). The determination was carried out by two indepen-
0 5 10 15 20 dent observers under controlled lighting conditions by the use of a
natural light lamp with a colour temperature of 6500 K and a spectral
RR, person 1
composition equal to natural light. The resulting Munsell values were
Fig. 1. Correlation of colour determination by two individual observers: linear regression transformed to a redness rating (RR) originally developed by Torrent
(solid) and unity line (dashed). et al. (1980) as a measure for the intensity of reddening. It is expressed
as follows: RR = (10 − H)·C/V, with H = hue, C = chroma and
2.2. Determination of magnetic susceptibility V = value. While this method is subjective it was selected so that
redness could be determined by any observer without the need for a
Measurements of volume-specific magnetic susceptibility κ were spectrophotometer and can also be applied to standard field description
conducted using a MS2B sensor (Bartington Instruments, Witney, UK; of soils where a Munsell colour has been recorded in the absence of
Dearing, 1999). The dry homogenised soil samples were filled into actual samples.

100000 10000

10000 1000
m3/kg]

m /kg]

1000 100
3
−8

100 10
−8
χLF [10

Δχ [10

10 1

1 0.1

0.1 0.01
ub b/int. acid clay phy. sand. ub b/int. acid clay phy. sand.

25 100

20 80
Fe O [%]

15 60
RR [−]

2 3

10 40

5 20

0 0
ub b/int. acid clay phy. sand. ub b/int. acid clay phy. sand.
Fig. 2. Low frequency magnetic susceptibility χLF, frequency-dependent susceptibility Δχ, redness rating and Fe content of the investigated samples. The samples are grouped according
to their parent material and the groups are further subdivided into subgroups of unweathered rock, weathered rock, subsoil and topsoil (from left to right). The black dots inside the white
circles represent the medians, while the boxes represent the lower and upper quartiles. The whiskers denote the most extreme values within the interval of the first quartile minus, and the
third quartile plus 1.5 times the interquartile range, respectively. These correspond to a roughly 99% coverage of the data. Outliers with values beyond the ends of the whiskers are
displayed as circles.

145
H. Preetz et al. Geoderma 303 (2017) 143–149

Table 2
Median values of magnetic susceptibility χLF, frequency dependence Δχ, redness RR and iron content and correlation coefficients r. The level of significance of r, the probability of error α,
has been determined by a Student's t-test. Accumulation factors (AF) represent the ratio between the median values of the topsoil samples and the median values of the unweathered
parent rocks.

Rock group Material Number χLF AF Δχ AF(Δχ) RR [−] AF Fe2O3 [%] AF r (RR α [%] r (RR α [%]
of [10− 8 m3/ (χLF) [10− 8 m3/ [10− 8 m3/ (RR) (Fe2O3) vs. lg vs. lg
samples kg] [−] kg] kg] [−] [−] (χLF)) (Δχ))
[−] [−]

Ultrabasic All 106 1589 15 11.8 58.9 0.14 < 20 0.14 < 20
Topsoil 17 1906 5.6 47 23.5 12 3.5 63.3 6.4 0.13 > 50 − 0.1 > 50
Subsoil 53 2514 41 12.5 63.4 −0.41 <1 − 0.56 <1
Weath. rock 28 384 2.7 9.3 20.3 0.05 > 50 0.21 > 20
Unw. rock 8 341 2.0 3.4 9.9 −0.07 > 50 − 0.27 > 50
Basic/intermediate All 117 820 7.0 9 23.7 −0.06 > 50 0.15 < 20
Topsoil 23 793 1.4 15 15.0 10 3.7 27.4 1.9 −0.39 < 10 − 0.21 > 20
Subsoil 38 584 7.1 8.9 27.5 −0.14 > 20 − 0.25 < 20
Weath. rock 41 972 7.0 9.3 22.9 −0.15 > 20 0.05 > 50
Unw. rock 15 547 1.0 2.7 14.5 −0.2 > 20 0.25 > 20
Acid All 170 34 1.6 7.2 8.45 0.30 <1 0.51 <1
Topsoil 36 94 5.5 3.8 19 8.3 2.2 13.1 4.9 0.28 < 10 0.30 < 10
Subsoil 74 41 2.4 7.5 13.8 0.40 <1 0.38 <1
Weath. rock 47 15 1.2 5 4.4 0.20 < 20 0.19 < 20
Unw. rock 13 17 0.2 3.8 2.7 −0.5 < 10 0.0 > 50
Clay/clay slate All 73 32 1.7 6 15.9 0.79 <1 0.74 <1
Topsoil 13 189 67.5 8.4 21 7.5 3.1 24.2 26.9 0.86 <1 0.66 <2
Subsoil 39 26 1.3 6 23.5 0.53 <1 0.59 <1
Weath. rock 15 15 0.8 4.4 7.7 0.91 <1 0.85 <1
Unw. rock 6 2.8 0.4 2.4 0.9 0.75 < 10 0.82 <5
Phyllite All 21 58 3.5 8 10.9 0.6 <1 0.61 <1
Topsoil 2 1264 66.5 107 357 10 3.2 20.7 1.9 – – – –
Subsoil 14 112 7.9 8 10 0.40 < 20 0.29 > 20
Weath. rock 4 8.2 0.6 2.9 6.5 −0.09 – − 0.5 > 20
Unw. rock 1 19 0.3 3.1 10.9 – – – –
Sandstone All 19 25 1.2 8 14.9 0.26 > 20 0.03 > 50
Topsoil 7 171 155 7.1 36 9 2.4 20 33.3 −0.4 > 20 − 0.34 > 20
Subsoil 7 23 1.2 8 15.1 0.34 > 20 0.35 > 20
Weath. rock 4 9.5 0.7 6.3 5.7 −0.07 > 50 − 0.29 > 50
Unw. rock 1 1.1 0.2 3.8 0.6 – > 20 – –

A comparison of the colour determination by the two individual rocks generally show higher susceptibilities than the other samples
observers is given in Fig. 1. Both results correspond well and demon- with median values > 800 × 10− 8 m3/kg (Fig. 2, Table 2) and there is
strate the reproducibility of the method. For further analysis, the mean no significant correlation between LF susceptibility and RR (Fig. 3) or
of both individual results is used. between FD susceptibility and RR (Fig. 4). These soils also show low
enrichment factors in magnetic susceptibility (Table 2), particularly for
2.4. Geochemical analysis the basic/intermediate rocks where there is little difference between
the soil and parent material. These samples also show frequency
Geochemical analysis of the samples was accomplished using X-ray dependence of susceptibility in the soil and rock samples. The
fluorescence (XRF) with a Philips spectrometer PW 1220. The samples unweathered parent materials also have high susceptibilities (median
were prepared by melting with a flux (1 part of sample and 5 parts of Li- 341 and 547 × 10− 8 m3/kg for ultrabasic and basic/intermediate
metaborate) at 1250 °C, and international standards were used for rocks, respectively) indicating a significant ferrimagnetic mineral
calibration (Schellmann, 1986). concentration in the parent rock.
Magnetite and titanomagnetite are ubiquitous minerals in these
3. Results and discussion rock types and are typically concentrated in the stable heavy metal
fraction of tropical soils (Kanig, 1990). This results in enhancement of
The samples were grouped according to their parent material and frequency-dependent magnetic susceptibility and in the case of soils
each group was further subdivided according to the degree of weath- from ultrabasic rock also the low-frequency susceptibility from residual
ering into unweathered parent material, weathered parent material, enrichment of weathering-resistant magnetite/titanomagnetite of all
subsoil and topsoil (Table 1). Generally there is a trend for increasing grain sizes (MD, SSD and SP). The general stability of magnetite, similar
low-frequency magnetic susceptibility (χLF), frequency-dependent mag- to that of hematite, has also been identified in Mediterranean soils (Di
netic susceptibility (Δχ) and redness rating (RR) from the parent rock to Figlia et al., 2007). SP minerals are typically thought to be an indicator
the soil (i.e. as weathering increases) on all rock types (Fig. 2). of pedogenic processes but they can appear in volcanic rock itself,
Although the magnitude of the enhancement varies between the formed from a short crystallisation phase during very rapidly cooling of
different parent material types. Fe contents are highest for ultrabasic magma. Hence, they can preferably occur in volcanic rocks such as
and basic parent rock material (Fig. 2) and increase with ongoing glasses and basalts (Zhou et al., 2000; Knudsen et al., 2005). This
weathering, reaching up to 63% for soils stemming from ultrabasic questions the use of FD susceptibility as suitable proxy for tropical soil
rocks. development on these parent material types because the SP minerals
can also be of lithogenic origin.
The lithogenic input of coarser ferrimagnetic particles can further
3.1. Ultrabasic and basic/intermediate parent rocks
contribute to the increase of frequency-dependent susceptibility via the
disintegration of MD and SSD particles rather than neoformation of SP
Material derived from ultrabasic and basic/intermediate parent

146
H. Preetz et al. Geoderma 303 (2017) 143–149

ultrabasic (n = 106) basic (n = 117) acid (n = 170)

25 25 25
r = 0.14 r = −0.06 r = 0.30
20 20 20

15 15 15
RR [−]

10 10 10

5 5 5

0 0 0
−1 0 1 2 3 4 5 −1 0 1 2 3 4 5 −1 0 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

clay (n = 73) phyllite (n = 21) sandstone (n = 19)

25 25 25
r = 0.79 topsoil r = 0.26
20 20 subsoil 20
weath. rock
unw. rock
15 15 15
RR [−]

r = 0.60
10 10 10

5 5 5

0 0 0
−1 0 1 2 3 4 5 −1 0 1 2 3 4 5 −1 0 1 2 3 4 5
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
−8 3 −8 3 −8 3
χLF [10 m /kg] χLF [10 m /kg] χLF [10 m /kg]

Fig. 3. Correlation between the logarithm of LF susceptibility and redness rating. Grouping is according to parent rock types and the number of samples is given in parentheses. The
correlation coefficient r and the regression line is given for the entire groups without differentiating between different degree of weathering and the dashed line delineates 95%
confidence bounds.

particles. Comminution of lithogenic MD particles by weathering between the magnetic susceptibility of the parent material and soil
(Maher, 1998) may be intensified in laterites due to the long timescales (Fig. 2) also indicated by an enhanced accumulation factor (Table 2).
of weathering. During the process of comminution, less hematite seems There is a strong correlation between RR and magnetic properties in
to be formed than during the inorganic neoformation of SP minerals, soils developed from clays and clay slates. Compared to soils on basic
which would explain the low correlation of χLF or Δχ with RR for the and ultrabasic rocks there is a greater differentiation between the soil
group of ultrabasic, basic and intermediate rocks. and parent material magnetic susceptibility and a moderate enhance-
ment of Δχ that is also positively correlated with RR.
3.2. Acid igneous rocks Although the soils developed from clays were subject to similar
chemical weathering to the other parent material groups the increase of
Topsoils developed on acid igneous rocks have moderate suscept- frequency-dependent susceptibility and RR in weathered claystones and
ibility (median χLF 94 × 10− 8 m3/kg) and low concentrations of SP slates suggests that neoformation of SP minerals already occurs in early
minerals (median Δχ 3.8 × 10− 8 m3/kg) (Table 2). There is a weak stages of the laterisation process. This indicates the dominant mechan-
positive correlation between magnetic parameters and RR (Figs. 3 and ism is neoformation of ferrimagnetic minerals for magnetic enhance-
4). This weak positive correlation between RR and χLF and Δχ indicate ment in these soils.
a relationship between the neoformation of ferrimagnets in the SP grain The model of neoformation of Dearing et al. (1996) might be
size and the formation of hematite. Since acid rocks commonly only assumed since preconditions such as a cyclic change of redox condi-
contain small proportions of magnetite a residual enrichment mechan- tions, which provide sufficient Fe supply and dehydration and oxidation
ism for the enhanced magnetic susceptibility can be excluded. There is cycles, prevail for laterites. It could be argued that these highly
also a higher AF for Δχ susceptibility than for χLF (Table 2), supporting weathered soils do not provide sufficient primary Fe-bearing minerals
the assumption of neoformation of ferrimagnetic minerals over a (as the source material of Fe2 +) required for the formation of
lithogenic input. ferrihydrite, the precursor for magnetite. In soils formed in the Tertiary
red clay in the Chinese Loess plateau a decline of ferrimagnetic minerals
3.3. Clay and clay slates with corresponding enhancement of hematite was found with increas-
ing soil development (Hu et al., 2009). This was attributed to the
Topsoils developed on clays have moderate susceptibility (median ferrimagnetic phase transforming into hematite during aging of Fe
χLF 189 × 10− 8 m3/kg) and moderate concentrations of SP minerals oxides (cf. Torrent et al., 2006). However, laterites are developed under
(median Δχ 8.4 × 10− 8 m3/kg) (Table 2). There is a clear separation different conditions to the soils of the Chinese Loess Plateau, and these

147
H. Preetz et al. Geoderma 303 (2017) 143–149

ultrabasic (n = 106) basic (n = 117) acid (n = 170)

25 25 25
r = 0.14 r = 0.15 r = 0.51
20 20 20

15 15 15
RR [−]

10 10 10

5 5 5

0 0 0
−2 −1 0 1 2 3 4 −2 −1 0 1 2 3 4 −2 −1 0 1 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

clay (n = 73) phyllite (n = 21) sandstone (n = 19)

25 25 25
r = 0.74 topsoil r = 0.03
20 20 subsoil 20
weath. rock
unw. rock
15 15 15
RR [−]

r = 0.61
10 10 10

5 5 5

0 0 0
−2 −1 0 1 2 3 4 −2 −1 0 1 2 3 4 −2 −1 0 1 2 3 4
10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10
Δχ [10−8 m3/kg] Δχ [10−8 m3/kg] Δχ [10−8 m3/kg]
Fig. 4. Correlation between the logarithm of frequency-dependent susceptibility and redness rating. Grouping is according to parent rock types and the number of samples is given in
parentheses. The correlation coefficient r and the regression line are given for the entire groups without differentiating between different degree of weathering and the dashed line
delineates 95% confidence bounds.

conditions provide enough opportunities for a repeated Fe release from pedogenic neoformation. Lithogenic ferrimagnets inherited from parent
weathering of primary and secondary Fe minerals. Thus the continued material may play a minor role.
production of neoformed ferrimagnetic minerals could occur.
The correlation of RR with low frequency and frequency-dependent 3.5. Sandstone
susceptibility corresponds to the formation model of Torrent et al.
(2006) where ferrihydrite is transformed by dehydration to maghemite The number of laterite samples derived from sandstone is low
that subsequently oxidises to hematite. The first step is faster than the (n = 19) but the results provide an indication of increased χLF and Δχ
second. It is also coupled to soil conditions remaining moist, but not with weathering intensity (Fig. 2). However there is little correlation
saturated most of the year and warm enough for substantial mineral between the magnetic properties and RR (Figs. 3 and 4). This lack of
transformation. These conditions are typical for the investigated reddening and increase in SP minerals during soil genesis suggests
lateritic soil types. either a dominant lithogenic input for the origin of the magnetic
minerals or soil conditions are less conducive to the formation of
3.4. Phyllite rocks ferrimagentic minerals. In soils derived from sandstones ferrimagnetic
particles can be detrital, stemming from fresh rocks of the source area
Results for soils from phyllites should be interpreted with caution of the sediments and/or from pedogenic processes that take place
due to small sample numbers (n = 21). Phyllites are formed during before the material was eroded and deposited (Hounslow, 1996; Maher
low-grade regional metamorphism of predominantly clay slates and and Taylor, 1988; Von Suchodoletz et al., 2009). Moreover, magnetite
other clayey sediments so they provide comparable conditions for is often concentrated as a placer deposit along the border of shorelines
tropical soil formation as the clay/clay slate group. They also show a which is conserved in the resulting sediments. Although our data set is
trend towards increasing χLF, Δχ and RR with weathering intensity too small for general conclusions in this context, it provides evidence
(Fig. 2). The topsoil samples (n = 2) have high χLF (median that lithogenic inputs into magnetic compounds of tropical soils could
1264 × 10− 8 m3/kg), subsoil samples also show enhancement of also occur from acid clastic sediments.
susceptibility (Table 2). Samples also show a positive but slightly
weaker correlation between RR and χLF and Δχ than the clay/clay slate 4. Summary and conclusions
sediments (Figs. 3 and 4). Based on the discussion above we conclude
that enhanced magnetic susceptibility in combination with increased The relationship between reddening (as redness rating RR) and
reddening in laterites developed from phyllites is mainly due to magnetic properties (low frequency magnetic susceptibility χLF and

148
H. Preetz et al. Geoderma 303 (2017) 143–149

frequency dependent magnetic susceptibility Δχ) was investigated for during lateritic soil formation: a case study in the middle Amazon basin. Clay Clay
Miner. 55, 253–259.
its potential as a proxy for characterising weathering and environ- Boyle, J.F., Dearing, J.A., Blundell, A., Hannam, J.A., 2010. Testing competing
mental conditions of tropical soil formation. All of the investigated hypotheses for soil magnetic susceptibility using a new chemical kinetic model.
samples belong to the same general soil type and were therefore subject Geology 38, 1059–1062.
Dearing, J.A., 1999. Environmental Magnetic Susceptibility - Using the Bartington MS2
to comparable environmental and soil forming conditions such as System. Chi publishing, Kenilworth.
changes in wetting and drying, and related redox potential changes. Dearing, J.A., Hay, K.L., Baban, S.M.J., Huddleston, A.S., Wellington, E.M.H., Loveland,
The data indicates that given sufficient Fe supply magnetic suscept- P.J., 1996. Magnetic susceptibility of soil: an evaluation of conflicting theories using
a national data set. Geophys. J. Int. 127, 728–734.
ibility increases with increasing soil age (Boyle et al., 2010) and with Di Figlia, M.G., Bellanca, A., Neri, R., Stefansson, A., 2007. Chemical weathering of
weathering regimes with increasing temperature and moisture (Singer volcanic rocks at the island of Pantelleria, Italy: information from soil profile and soil
et al., 1996). However, this relationship has limitations because the solution investigations. Chem. Geol. 246, 1–18.
FAO (Food and Agriculture Organization of the United Nations), 2006. World reference
underlying mechanisms of magnetic enhancement show differences
base for soil resources 2006 - a framework for international classification, correlation
depending on the nature of the parent rock. The dataset does not and communication. World soil resources reports. 103. Rome. Available at. www.fao.
support a universal pathway of magnetic enhancement for all laterite org/ag/Agl/agll/wrb/doc/wrb2006final.pdf.
soils. Parent rock is shown to have a strong influence on the chemical Fontes, M.P.F., Carvalho, I.A., 2005. Color attributes and mineralogical characteristics,
evaluated by radiometry, of highly weathered tropical soils. Soil Sci. Soc. Am. J. 69,
composition of lateritic soils (Schellmann, 1986), causing differentia- 1162–1172.
tion between lateritic soils on different patent materials. Hounslow, M.W., 1996. Ferrimagnetic Cr and Mn spinels in sediments: residual magnetic
The following conclusions can be drawn regarding our results of RR minerals after diagenetic dissolution. Geophys. Res. Lett. 23 (20), 2823–2826.
Hu, X.-F., Xu, L.-F., Pan, Y., Shen, M.-N., 2009. Influence of the aging of Fe oxides on the
and magnetic properties being used as proxies for soil development decline of magnetic susceptibility of the Tertiary red clay in the Chinese Loess
stages in lateritic soils: Plateau. Quat. Int. 209, 22–30.
Kanig, M., 1990. Lateritische Verwitterungsdecken im Bereich des Alkalikomplexes von
Jacupiranga, Sao Paulo, Brasilien.- Mitteilungen zur Mineralogie und
i) The enhancement of ferrimagnetic minerals in tropical soils deriving Lagerstättenkunde. 35 RWTH Aachen University 260 pp., PhD thesis.
from ultrabasic and basic/intermediate rocks is due to a strong Knudsen, M.F., Abrahamsen, N., Riisager, P., 2005. A rock- and palaeomagnetic study of
lithogenic overprint. Thus magnetic measurements in these soils are geologically recent lavas and 1995 volcanic glass on Fogo (Cape Verde Islands). Bull.
Geol. Soc. Den. 52, 39–52.
not useful proxies for neoformed pedogenic ferrimagnetic minerals
Maher, B.A., 1986. Characterisation of soils by mineral magnetic measurements. Phys.
and corresponding soil development. The enrichment of weathering- Earth Planet. Inter. 42, 76–92.
resistant ferrimagnetic iron oxides is concordant with lateritic Maher, B.A., 1998. Magnetic properties of modern soils and Quaternary loessic paleosols:
paleoclimatic implications. Palaeogeogr. Palaeoclimatol. Palaeoecol. 137, 25–54.
processes and enrichment of certain elements such as Al
Maher, B.A., Taylor, R.M., 1988. Formation of ultrafine-grained magnetite in soils. Nature
(Schellmann, 1981). 336, 368–370.
ii) Soils stemming from clay and clay slate, phyllite and acid igneous Mullins, C.E., 1977. Magnetic susceptibility of the soil and its significance in soil science –
rocks show a strong positive correlation between RR and magnetic a review. J. Soil Sci. 28, 223–246.
Munsell Color Company, 1975. Munsell Soil Color Charts. Macbeth Division of
susceptibility and indicate neoformation of SP ferrimagnetic miner- Kollmorgen, Baltimore, MD.
als, thus making these measurements a suitable proxy for soil Néel, L., 1949. Théorie du traînage magnétique des ferromagnétiques en grains fins avec
development in these soils. Pedogenic hematite has been suggested applications aux terres cuites. Ann. Geophys. 5, 99–136.
NRCS (Natural Resources Conservation Service), 2006. Keys to Soil Taxonomy, 10th ed.
to derive from the transformation of ferrimagnetic minerals. If this NRCS, Washington, DC Available at: soils.usda.gov/technical/classification/
were the case as hematite content increases, magnetic properties taxonomy/.
should decrease, which has not been shown in these samples. The Preetz, H., Altfelder, S., Igel, J., 2008. Tropical soils and landmine detection—an
approach for a classification system. Soil Sci. Soc. Am. J. 72 (1), 151–159.
coexistence of hematite and ferrimagnetic minerals after such long Schellmann, W., 1974. Abschlussbericht zum Forschungsvorhaben Nr. 1630 “Laterit”,
weathering histories in laterites suggest that other pathways may Teil 1: Allgemeine Forschungsergebnisse. Archive Report. Federal Institute for
take place during tropical weathering and laterite formation. In Geosciences and Natural Resources (BGR), Hannover.
Schellmann, W., 1981. Considerations on the Definition and Classification of Laterites.
laterites derived from clay substrates two phases (detrital and Lateritisation Processes, Proceedings of the International Seminar on Lateritisation
secondary) of kaolinite have been observed (Balan et al., 2007). Processes, Trivandrum/India. pp. 1–10.
This suggests long-term active phases of dissolution and neoforma- Schellmann, W., 1986. On the geochemistry of laterites. Chem. Erde 45, 39–52.
Schwertmann, U., 1988. Occurrence and formation of iron oxides in various
tion are evident in laterites on these substrates and these processes
pedoenvironments. In: Stucki, J.W., Goodman, B.A., Schwertmann, U. (Eds.), Iron in
could also be relevant for the continued production of pedogenic Soils and Clay Minerals. Reidel, Dordrecht, pp. 267–308.
ferrimagnetic minerals over long weathering timescales. Schwertmann, U., 1993. Relations between iron oxides, soil color, and soil formation. In:
Bigham, J.M., Ciolkosz, E.J. (Eds.), Soil Color. 31. Soil Sci. Soc. Am. Spec. Publ., pp.
51–69.
Further investigations should address if the stability of SP minerals Singer, M.J., Verosub, K.L., Fine, P., TenPas, J., 1996. A conceptual model for the
that stem from neoformation are similar to lithogenic sources. This enhancement of magnetic susceptibility in soils. Quat. Int. 34-36, 243–248.
could provide fundamental information to identify the different origins Tardy, Y., Roquin, C., 1992. Geochemistry and evolution of lateritic landscapes. In:
Martini, I.P., Chesworth, W. (Eds.), Weathering, Soils and Paleosols – Developments
for magnetic enhancement in tropical soils and the subsequent relia- in Earth Surface Processes. 2. Elsevier, pp. 407–443.
bility as a soil development index in lateritic soils. Torrent, J., Schwertmann, U., Schulze, D.G., 1980. Iron oxide mineralogy of some soils of
two river terrace sequences in Spain. Geoderma 23, 191–208.
Torrent, J., Schwertmann, U., Fechter, H., 1983. Quantitative relationships between soil
Acknowledgements color and hematite content. Soil Sci. 136, 354–358.
Torrent, J., Barrón, V., Liu, Q., 2006. Magnetic enhancement is linked to and precedes
This work has in part been supported by the Federal Ministry of hematite formation in aerobic soil. Geophys. Res. Lett. 33 4 pp. http://dx.doi.org/10.
1029/2005GL024818.
Education and Research (BMBF) (Grant no. 01RX0310) and the Federal Torrent, J., Liu, Q.S., Barrón, V., 2010a. Magnetic susceptibility changes in relation to
Ministry of Defence (Grant no. E/KR2J/6A100/5F093), Germany. We pedogenesis in a Xeralf chronosequence in northwestern Spain. Eur. J. Soil Sci. 61,
thank T. Putzmann H. Schröder, D. Renk and K. Haase for carrying out 161–173.
Torrent, J., Liu, Q.S., Barrón, V., 2010b. Magnetic minerals in calcic luvisols (chromic)
the susceptibility measurements and determining the sample colours.
developed in a warm Mediterranean region of Spain: origin and paleoenvironmental
We are also indebted to the anonymous reviewers and the editor significance. Geoderma 154, 465–472.
Edward Nater for their constructive comments, which helped to Von Suchodoletz, H., Kühn, P., Hambach, U., Dietze, M., Zöller, L., Faust, D., 2009. Loess-
improve the manuscript. like and palaeosol sediments from Lanzarote (Canary Islands/Spain) – indicators of
pelaeoenvironmental change during the Late Quaternary. Palaeogeogr.
Palaeoclimatol. Palaeoecol. 278, 71–87.
References Zhou, W., Van der Voo, R., Peacor, D.R., Zhang, Y., 2000. Variable Ti-content and grain
size of titanomagnetite as a function of cooling rate in very young MORB. Earth
Planet. Sci. Lett. 179, 9–20.
Balan, E., Fritsch, E., Allard, T., Calas, G., 2007. Inheritance vs. neoformation of kaolinite

149