SSSAJ Grow Your Career with SSSA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.
Soil Science Society of America Journal 64:1108-1117 (2000)
© 2000 Soil Science Society of America

DIVISION S-9-SOIL MINERALOGY

Changes in Soil Mineralogy and Texture Caused by Slash-and-Burn Fires in Sumatra, Indonesia

Quirine M. Ketteringsa, Jerry M. Bighamb and Valérie Lapercheb

a Environmental Science Graduate Program, Ohio State Univ., 2021 Coffey Road, Columbus, OH 43210 USA
b School of Natural Resources, Ohio State Univ., 2021 Coffey Road, Columbus, OH 43210 USA

bigham.1{at}osu.edu


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 
We investigated the effect of fire intensity from slash-and-burn agriculture on the mineralogy of Oxisols in the Sepunggur area, Jambi Province, Sumatra, Indonesia, in both field and laboratory experiments. Samples were collected from two depths (0–5 and 5–15 cm) at locations exposed to 100, 300, 600, and >600°C surface temperatures during the burns. Soils under forest and slashed vegetation were collected as controls. The pre-burn soil mineralogy was dominated by kaolinite, gibbsite, anatase, and goethite. Changes in soil properties with burning were most pronounced in the 0- to 5-cm layer. Burning the topsoil led to coarser textures, especially at temperatures exceeding 600°C. Heat reduced the gibbsite and kaolinite concentrations and converted goethite into ultra-fine maghemite, thus increasing the magnetic susceptibility of the samples. The conversion of goethite did not take place until water in the samples had vaporized. Addition of organic matter to soil with a low organic C content before heating increased the magnetic susceptibility, indicating that organic matter was necessary (and limiting) for the complete conversion of goethite. Coarse-grained magnetite particles were present prior to and after the burning and, therefore, were not pyrogenic. Magnetic susceptibility measurements were highly discriminatory among heat treatments, whereas x-ray diffraction (XRD) was much less sensitive to fire-induced changes in mineralogy. Our research showed that severe burning had drastic effects on soil mineralogy, but changes should also be expected at lower fire intensities. Further research is needed to determine how important these changes in soil mineralogy are for nutrient availability in the growing season after the burn.

Abbreviations: CBD, citrate-bicarbonate-dithionite • DTA, differential thermal analysis • FC%mass, fractional conversion factor for magnetic susceptibility • TG, thermogravimetric analysis • XRD, x-ray diffraction • {kappa}fd, frequency-dependent magnetic susceptibility • {chi}lf, low frequency mass specific magnetic susceptibility


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 

Whenever he saw the sea, or a fire, he fell silent, impressed by their elemental force.

Paulo Coelho, The Alchemist

SMALL-SCALE RUBBER FARMERS in Sepunggur, Sumatra, Indonesia, indicated in a recent survey that fire is widely used to clear forests because it provides an easy and economical means of increasing access to the fields (Ketterings et al., 1999). Forest vegetation is slashed in the beginning of the dry season (March–April) and burned from June through August. If a first-time broadcast burn does not remove enough aboveground biomass, the remaining wood is piled in heaps of 200 to 400 kg and burned a second time.

Peak temperatures measured at the soil surface during broadcast and subsequent secondary burns vary widely. Intensities depend not only on the fuel load and the water content of the slashed biomass (Brown, 1988) but also on the stacking pattern, climatic conditions during the burn, and the size of the area being burned. Hartford and Frandsen (1992) measured maximum surface temperatures of 400°C in a ground (litter and duff) fire, and Sertsu and Sanchez (1978) reported temperatures >500°C in shifting cultivation fires. Similar intensities have been recorded by Wells et al. (1979) and Chandler et al. (1983), whereas burns in Senegal reportedly reached levels >700°C (Masson, 1948).

Heat transfer into the soil depends not only on the surface temperature but also on the duration of exposure, on the water content of the soil, and the soil pore distribution (Steward et al., 1990). Ghuman and Lal (1989) measured temperatures of 218°C at 1 cm below the soil surface in windrows burned after clearing a tropical rainforest. Temperatures dropped to 150, 104, and 70°C at 5-, 10-, and 20-cm depths, respectively. DeBano et al. (1979), as reported in Brown (1988), found that temperatures at the 5-cm depth did not change significantly even when surface temperatures reached 700°C.

Numerous papers have shown that low-intensity burns (<250°C) affect soil biological and chemical properties (e.g., Nye and Greenland, 1960), but the impact of burning on soil mineralogy has not been widely evaluated even though moderate (250–500°C) and severe fires (>500°C) have the potential to cause mineral transformations. Kaolinite, for example, has been shown to decompose at temperatures between 500 and 700°C (Richardson, 1972). Gibbsite may be completely destroyed by heating in air at 200°C (Rooksby, 1972), and goethite is altered to hematite at {approx}300°C (Cornell and Schwertmann, 1996). Synthetic magnetite transforms to maghemite at temperatures as low as 220°C (Sidhu, 1988), and pure maghemite may revert to hematite at temperatures around 350°C (Mullins, 1977). However, natural maghemites are often stabilized by impurities giving rise to transformation temperatures exceeding 600°C (Mullins, 1977).

Most studies of the heat stability of minerals have been performed in the laboratory with either synthetic or purified natural minerals. Under field conditions, fire has been observed to collapse some 2:1 phyllosilicates and destroy kaolinite (Ulery et al., 1996). The presence of maghemite in soils has often been attributed to the dehydroxylation of goethite or lepidocrocite by burning in the presence of organic matter (Van der Marel, 1951; Taylor and Schwertmann, 1974; Anand and Gilkes, 1987; Stanjek, 1987). Maghemite formation is of interest because of its ferrimagnetic character and the potential for using magnetic techniques to evaluate the conversion process. Magnetic susceptibility (the total magnetic force in a material divided by the strength of the magnetic field inducing magnetization) is one such method.

Besides aiding in the identification of Fe oxide transformations due to fire, magnetic susceptibility measurements may give an indication of fire intensity when results are expressed as frequency-dependent magnetic susceptibility ({kappa}fd) or as a fractional conversion factor (FC%mass). Dearing (1994) has suggested that soils with {kappa}fd between 10 and 15% are dominated by ultrafine ferrimagnetic particles that are often products of high-intensity burning. The FC%mass is calculated as the ratio of the magnetic susceptibility of a field soil to the magnetic susceptibility of the same soil after being exposed to 550°C for 8 h in an oven. High values of the fractional conversion factor indicate that Fe oxide conversion took place prior to reheating. Lower values are obtained where prior burning was either absent or of low intensity. In situations where laboratory facilities for XRD are not available, Fe concentrations are very low, field measurements are needed, or the amount of sample is limited, magnetic susceptibility may be a good tool for measuring fire intensities and fire-induced changes in Fe oxides.

We investigated the effect of heat intensity on the mineralogy of Oxisols in the Sepunggur area, Jambi Province, Sumatra, Indonesia, where slash-and-burn agriculture is still commonly used (Ketterings et al., 1999). Our objectives were to evaluate (i) the effects of burn intensity on soil mineralogy, (ii) the use of magnetic susceptibility as a tool for measuring burn intensity, and (iii) factors influencing changes in magnetic susceptibility.


   Materials and methods
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
Results and discussion
 Conclusions
 REFERENCES
 
Two sites in the Sepunggur area (102°14' E, 1°29' S) were selected for Field Experiments I and II, respectively. The soils at both sites were classified as Hapludox and Kandiudox according to U.S. soil taxonomy (Soil Survey Staff, 1999).

Field Experiment I
Field Experiment I was a 12- to 15-yr-old secondary forest that was burned in July 1997. The entire field (0.75 ha) was subjected to a low-intensity broadcast burn by the farmer. Following the broadcast burn, three burn piles were established (400 kg of wood with stem diameters between 3.2 and 6.4 cm) on an area of 3 by 3 m. The piles were ignited, and temperatures were measured using heat-sensitive crayons (Cole Parmer, Vernon Hills, IL) situated directly on top of the mineral soil and beneath any existing vegetative litter. Crayons were placed in the center, at the edge, and outside each pile. A control sample was taken in a 12- to 15-yr-old secondary forest near the burned field (Table 1) . The farmer indicated that the age of the forest surrounding the burned plot was identical to the forest that was slashed, burned, and sampled during this experiment. Therefore, the samples taken in the forest reflect the situation prior to slash-and-burn, the samples taken outside the piles reflect the situation after a low-intensity broadcast burn, and the samples taken underneath the piles reflect the effect of high-intensity secondary fires at two temperatures. All locations were sampled at depths of 0 to 5 cm (hereafter referred to as topsoil) and 5 to 15 cm (hereafter referred to as subsoil). Ash, charcoal, and unburned wood remaining on the field after the secondary burn were carefully removed prior to soil sampling. Each location was sampled before the burn and 1 d and 1, 2, 4, 8, and 12 wk after the burn. Soil water contents (105°C) were determined at each sampling time. Samples were returned to the laboratory the same day, sieved through a 2-mm sieve, and air-dried for further analysis.


View this table:
[in this window]
[in a new window]
  		Table 1 Characterization of unburned forest soil from Field Experiments I and II

 
Field Experiment II
For Field Experiment II, a 20-yr-old secondary forest was selected (Table 1). Bulk soil samples (4–5 kg) in this 5-ha field were taken after slashing and after a broadcast burn at locations based on the temperature (100, 300, and 600°C) reached during the burn as measured with heat-sensitive crayons. Samples were taken at depths of 0 to 5 and 5 to 15 cm. Subsamples of completely combusted, brick-like topsoil were taken in locations where high fuel loads had produced temperatures that exceeded those measurable with the crayons (>600°C). In addition, control samples were taken in the surrounding unslashed and unburned forest. All samples were taken 2 wk after the burn, and sample processing was identical to that for the samples from Field Experiment I.

Routine Soil Analyses
Whole soil samples were analyzed for organic C according to the Walkley–Black method (Walkley, 1947). The pH was determined electrometrically in a 1:5 mass ratio of soil/water after shaking the samples on a reciprocal shaker for 2 h. Particle-size analyses were performed using standard sieve and pipette methods (Kilmer and Alexander, 1949) following removal of organic matter with 30% (w/w) H2O2. The samples were extracted with citrate-bicarbonate-dithionite (CBD) (Mehra and Jackson, 1960) to dissolve crystalline Fe oxides, except magnetite. Final extracts were analyzed for Fe by atomic absorption spectrophotometry.

Soil Fractionation
Soil samples taken from Field Experiment II and from Field Experiment I before the second burn were treated with 1 M NaOAc-HOAc buffer (pH 5) and 30% (w/w) H2O2 to remove organic matter (Jackson, 1975). The residual materials were washed free of excess salts, dispersed in water, and separated into sand (2–0.05 mm), silt (0.05–0.002 mm), and clay (<0.002 mm) fractions using standard sieve and gravity sedimentation techniques (Jackson, 1975). The sand and silt fractions were dried at 60°C, whereas the clays were Na-saturated, quick frozen, and freeze-dried.

Magnetic Susceptibility
All whole soil samples and particle-size fractions were analyzed for magnetic susceptibility with a Bartington MS2 Susceptibility System (ASC Scientific, Carlsbad, CA). This system was equipped with a dual-frequency sensor that accepted 10-cm3 samples in standard plastic pots. Mass-specific magnetic susceptibility was calculated as the volume-specific susceptibility ({kappa}) divided by the density of the sample in the 10-cm3 pots. Low-frequency (0.46 kHz) magnetic susceptibility ({chi}lf, units of m3 kg-1) was measured for all samples prior to exposing them to a high-frequency magnetic field (4.6 kHz). The dual-frequency measurements permitted the calculation of frequency-dependent magnetic susceptibility as:

(1)
where {kappa}fd is the volume-specific frequency-dependent magnetic susceptibility in percentage, {kappa}lf is the low-frequency (0.46 kHz) volume-specific magnetic susceptibility x 10-5, and {kappa}hf is the high-frequency (4.6 kHz) volume-specific magnetic susceptibility x 10-5. The mean diamagnetic volume specific magnetic susceptibility of the standard 10-cm3 pot was -0.4 x 10-5. All volume-specific magnetic susceptibility measurements were adjusted for this value prior to calculating mass-specific susceptibility. Soils with {kappa}fd between 10 and 15% were considered to contain ultrafine (<0.035 µm) ferrimagnetic minerals as reported by Dearing (1994).

A second set of dried (60°C) samples was exposed to 550°C for 8 h and analyzed for {chi}lf after cooling to room temperature. Fractional conversion values, which may serve as indicators of prior burning (Dearing, 1994), were calculated as:

(2)
where FC%mass is the fractional conversion value in percentage, {chi}lf is the mass-specific low-frequency (0.46 kHz) magnetic susceptibility x 10-8 m3 kg-1, and {chi}lf @ 550°C is the {chi}lf of 550°C exposed soil x 10-8 m3 kg-1.

X-Ray Diffraction
The sand and silt samples were ground in a ceramic mortar and pestle to pass a 250-µm sieve and backfilled to produce randomly oriented sample mounts. Two-hundred milligrams of each clay were mixed with 15 mL of deionized water, and one drop of 0.5 M NaOH was added to enhance dispersion. The suspensions were then ultrasonically dispersed for 30 s, and 5 mL of the suspension were filtered through a 0.22-µm Millipore filter (Millipore Corp., Bedford, MA). Filtration was stopped when the deposits were almost dry. The oriented aggregates were then transferred onto glass slides for XRD analyses. The oriented clays were scanned after air drying and after heating to 550°C for 4 h. All samples were analyzed with a Philips diffractometer (Philips Elec. Inst., Mahwah, NJ) using CuK{alpha} radiation (35 kV, 20 mA) and a graphite monochromater. Measurements were made by step scanning with a fixed time of 4 s per 0.05°2{theta} for routine work and a fixed time of 6 s per 0.01°2{theta} for detailed analyses of Fe oxides. In most cases, the XRD patterns were obtained from 2 to 70°2{theta}. Minerals occurring in the sand, silt, and clay fractions were identified on the basis of their diagnostic XRD peaks and their response to heating. Hydroxy-Al vermiculite was distinguished from vermiculite and chlorite by the partial collapse of the 1.4-nm peak with heating to 550°C.

Thermal Analysis
Thermal analyses of clay and silt samples were conducted to quantify kaolinite and gibbsite concentrations using a Seiko SSC 5020 instrument (Haake Inst., Paramus, NJ) that provided simultaneous thermogravimetric (TG) and differential thermal analysis (DTA). Approximately 50 mg of sample were heated from 25 to 1000°C at a rate of 20°C min-1 under a continuous flow (200 mL min-1) of dry N2 gas. Calibration of the temperature signal was achieved using the melting points of In and Sn. Calibration of the thermal balance was performed using a reference weight provided by the instrument manufacturer. Thermal events associated with dehydroxylation of kaolinite and gibbsite were assigned based on published literature (e.g., Dixon, 1966) and analyses of standard mineral samples.

Oven Experiment I: Effect of Added Carbon
Duplicate, 10-g samples of forest soil (0–5 and 5–15 cm) from the control area of Field Experiment II were mixed with 0, 20, 40, 80, 100, 150, and 200 g kg-1 air-dried, processed Indian tea (Camellia sinensis L.) or 50 and 100 g kg-1 granulated sugar. Samples were exposed to 600°C for 930 min to ensure the complete conversion of antiferromagnetic to ferrimagnetic minerals. Weight loss, water content, and magnetic susceptibility were measured after heating. The tea used in this experiment was combusted at 600°C for 8 h to determine the ash content by weight loss.

Oven Experiment II: Kinetics
Duplicate 25-g soil samples (0–5 and 5–15 cm) from the control area of Field Experiment II were placed in 2-cm-diam. glass vials and heated in a muffle furnace. Samples were exposed to 300 or 600°C for 1, 2, 4, 6, 10, 15, 30, 45, 90, 120, 240, 360, and 660 min. Weight loss, water content, and magnetic susceptibility were measured after heating. Samples were, in addition, analyzed for CBD-extractable Fe oxides. The magnetic susceptibility of the material remaining after CBD extraction was measured in order to determine the contribution of the CBD-extractable Fe oxides to the total magnetic susceptibility.

Statistical Analyses
All results of Field Experiment I were initially analyzed in a complete randomized block design in three replicates with three factors: treatment (forest, first burn estimated as 100°C exposure, 300°C exposure during second burn, and 600°C exposure during second burn), depth (0–5 and 5–15 cm), and time (0, 1, 7, 14, 28, 56, and 84 d after burning). Parameters were considered time dependent when differences over time were larger than the standard deviations of the individual treatments, the uncertainty of the means when sampling in triplicate, or the standard deviations of the forest treatment. Analyses of the 60 grid samples (0–5 cm depth) in the forest neighboring that of Field Experiment I showed that taking three samples to calculate a mean would result in an uncertainty of the mean of 63.7 x 10-8 m3 kg-1 for {chi}lf and 2.6% for {kappa}fd. Uncertainties for the subsoil (5–15 cm) means were less: 26.3 x 10-8 m3 kg-1 for {chi}lf and 1.9% for {kappa}fd. Although several magnetic susceptibility parameters showed significant differences among sampling times, those differences did not follow a recognizable pattern. Nor were they larger than the standard deviations of the parameters in the forest treatment or the uncertainty of the mean estimates when taking three replicates. The time series samples of Field Experiment I were thus combined to calculate the average for each replicate per treatment per layer for further heat exposure comparisons. Because the samplings over time cannot be considered completely independent, we calculated the averages of the 18 measurements and analyzed Field Experiment I as a two-factorial experiment (fire intensity at four levels and depth of sampling at two levels) in three replicates. Analyses of variance were performed using the software package Genstat 5 for Windows, Release 3.2 (Genstat 5, 1987).

Field Experiment II did not have independent replicates due to the fact that bulk samples were taken. Each bulk sample was separated into two subsamples that were analyzed separately and in duplicate in order to assess the precision of the measurements. Results were expressed as means and standard deviations.


   Results and discussion
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
Conclusions
 REFERENCES
 
Texture Changes
The soil samples from Field Experiment II showed some natural variability in texture but, in general, were high in clay (>55%) and low in sand (<5%) prior to burning. Exposure of the soil to extreme temperatures (>600°C) resulted in a sharp increase in the amount of sand and a decrease in silt and, especially, clay (Table 2) . Increases in sand were also observed at lower temperatures but were mostly limited to the 0- to 5-cm depth. These results are similar to those obtained in laboratory studies by Dyrness and Youngberg (1957), Sertsu and Sanchez (1978), and Duriscoe and Wells (1982), and in field studies by Sreenivasan and Aurangabadkar (1940) and Ulery and Graham (1993).


View this table:
[in this window]
[in a new window]
  		Table 2 Effect of heat on soil texture in Field Experiment II.{dagger}

 
Mineralogical Changes
The sand fraction of the forest soil from Field Experiment II (both 0–5 and 5–15 cm) contained mainly quartz and gibbsite with traces of kaolinite. Diagnostic XRD peaks for gibbsite in the topsoil sand fractions gradually decreased in intensity with increasing burn temperature and disappeared completely in the combusted (>600°C) samples (data not shown). Weakened intensities in the subsoil samples were observed only at 600°C. Quartz XRD peak intensities were unaffected except for a decrease in the topsoil sands exposed to temperatures of 600°C or greater. Because quartz should not vitrify at this temperature, the diminished intensities must be caused by encapsulation of quartz particles in a glassy matrix.

The sand fraction contributed <3% of the total soil magnetic susceptibility after all heat treatments except for the combusted topsoil, where it was responsible for 84% of the total (Table 3) . The glassy, sand-sized particles showed high spontaneous magnetization, with {chi}lf values of 2584 x 10-8 m3 kg-1 (Table 3). The magnetic susceptibility of the subsoil sand fraction was one-half that of the overlying topsoil in areas exposed to 600°C surface temperatures. Silt in the forest soil contained quartz, gibbsite, kaolinite, goethite, anatase, and traces of hydroxy-Al-vermiculite. Once again, burning the topsoil resulted in a decrease in quartz peak intensities after exposure to 600°C (Fig. 1) . A decrease in gibbsite concentration occurred at 300°C and was evident in both XRD (Fig. 1) and TG analyses (Table 4) . No further reductions with heating were noted from thermal analysis, but XRD peaks for gibbsite decreased in intensity at 600°C and were absent in the combusted (>600°C) topsoil (Fig. 1). Kaolinite concentrations in the silt fraction of the topsoil were mostly unaltered by heating. In the subsoil, quartz peak intensities remained unchanged through 300°C but were lower after exposure to 600°C (data not shown). Kaolinite, gibbsite, and goethite concentrations in the subsoil silt fractions were unaltered by the surface fires. Heat exposure increased the magnetic susceptibility of silt fractions from both the topsoil and subsoil (Table 3).


View this table:
[in this window]
[in a new window]
  		Table 3 Low frequency (0.46 kHz) mass–specific magnetic susceptibility ({chi}lf) of sand, silt, and clay fractions of heat-exposed soil from Field Experiment II{dagger}

 


View larger version (21K):
[in this window]
[in a new window]
  		Fig. 1 X-ray diffraction patterns, showing anatase (A), gibbsite (Gb), goethite (Go), hematite (H), kaolinite (K), and quartz (Q), of the silt fraction from topsoil exposed to different intensities of burns in Field Experiment II

 

View this table:
[in this window]
[in a new window]
  		Table 4 Kaolinite and gibbsite contents of silt and clay fractions from Field Experiment II as determined by thermogravimetric analysis

 
The clay fraction at both depths contained some hydroxy-Al-vermiculite and anatase but was dominated by kaolinite, gibbsite, and goethite. Losses of kaolinite with heating and the formation of a glassy phase were pronounced only in the combusted topsoil where kaolinite concentrations were reduced from 750 g kg-1 (100°C) to 270 g kg-1 (>600°C) (Fig. 2 , Table 4). Losses of gibbsite occurred at lower heat intensities. Gibbsite concentrations in the topsoil clay decreased from 100 to 50 g kg-1 at 600°C, and clay-sized gibbsite disappeared in the combusted topsoil (>600°C). An increase in magnetic susceptibility of the clays upon heating (slashed to 600°C) indicated that ferrimagnetic minerals were gradually being formed by conversion of goethite to maghemite or magnetite (Table 3). Hematite was also a combustion product. The appearance of these phases was difficult to confirm through conventional XRD analysis of the clays except in the combusted topsoil. The persistence of ferrimagnetic phases at surface temperatures exceeding 600°C indicates that heat transfer into the soil is a fairly inefficient process. Moreover, trace impurities, such as Al, tend to inhibit thermal transformations of these minerals (Sidhu, 1988).



View larger version (25K):
[in this window]
[in a new window]
  		Fig. 2 X-ray diffraction patterns, showing anatase (A), gibbsite (Gb), goethite (Go), hematite (H), kaolinite (K), maghemite and/or magnetite (M), and quartz (Q), of the clay fraction prior to burning (slashed soil) and following combustion at >600°C in Field Experiment II

 
Source of Increased Magnetic Susceptibility: Maghemite or Magnetite?
The percentage of the magnetic susceptibility lost after CBD extraction of the samples from Field Experiment II increased from 64% in the slashed topsoil to 93% in the combusted topsoil (Fig. 3) . Hence, heating causes the CBD-extractable Fe oxides to become increasingly susceptible to spontaneous magnetization. Because magnetite is resistant to CBD extraction (McKeague et al., 1971; Gamble and Daniels, 1972; Walker, 1983; Fine and Singer, 1989), our results suggest that maghemite is formed during the burning process.



View larger version (29K):
[in this window]
[in a new window]
  		Fig. 3 Effect of fire intensity on the contribution of citrate-bicarbonate-dithionite (CBD)-extractable Fe oxides to the total bulk soil magnetic susceptibility (Field Experiment II). Error bars indicate standard deviations of the means

 
Fractionation of ground subsamples of the field combusted sand fraction using a hand magnet produced two different types of magnetic particles, some that were black and others that were orange-brown. X-ray diffraction analyses showed that the black particles contained maghemite and hematite in addition to magnetite (Fig. 4a) . Microprobe analyses yielded Ti/Fe ratios ranging from 0 to 1. These particles were also present in the soil prior to burning and, therefore, were not pyrogenic. The orange-brown fraction could be dissolved in CBD, whereas the black particles were mostly unaffected by CBD treatment. X-ray diffraction analysis of the orange-brown particles (Fig. 4b) indicated the presence of both hematite and maghemite. The approximate {chi}lf of the maghemite can be calculated as 53870 x 10-8 m3 kg-1 if we assume that these particles are mostly maghemite. This value corresponds well with the {chi}lf of 54000 x 10-8 m3 kg-1 found by Graham and Scollar (1976) and is slightly higher than those previously reported for pure maghemites (Mullins, 1977; Dearing, 1994).



View larger version (22K):
[in this window]
[in a new window]
  		Fig. 4 X-ray diffraction patterns showing hematite (H), magnetite (Ma), and maghemite (Mh) of the black (a) and orange-brown (b) ferrimagnetic fractions obtained from completely combusted topsoil in Field Experiment II. X-ray diffraction d spacings (nm) are given for an unidentified phase

 
Variability of Fire Induced Mineralogical Changes
In Field Experiment I, the {chi}lf of forest topsoil increased upon exposure to fire (600°C) by a factor of 1.2, whereas in Field Experiment II the increase was by a factor of 15.5 (compare {chi}lf for forest with {chi}lf at 600°C in Table 5) . Furthermore, the magnetic susceptibility of unburned forest soil in Field Experiment I was six times (topsoil) and four times (subsoil) larger than the magnetic susceptibility of the forest soil in Field Experiment II. The FC%mass values for topsoil amounted to 20.8 and 2.3% for Field Experiments I and II, respectively. For subsoil, the FC%mass values were 56.2 and 1.7% for Field Experiment I and II, respectively. These observations suggest that the field in Field Experiment II had not been seriously impacted by prior burns; whereas, the forest soil in Field Experiment I had probably been burned previously. The topography of the field in Field Experiment I was flat, and the area was surrounded by forest. We therefore can assume that erosion did not play a major role in changing the magnetic susceptibility of the soil so that the effects of burning may be detectable for long periods of time.


View this table:
[in this window]
[in a new window]
  		Table 5 Effect of burning on magnetic susceptibility and derived parameters from <2-mm soil material (Field Experiments I and II).{dagger}

 
Increasing fire intensity in the 1997 burns resulted in an increase in both {chi}lf and {kappa}fd in the topsoil (0–5 cm) at both field sites (Table 5); however, the increases were more dramatic in Field Experiment II. Subsoil (5–15 cm) effects were absent in Field Experiment I and present but less extreme than in the 0- to 5-cm depth at Field Experiment II. These results suggest that heat transfer did not reach below 5 cm in Field Experiment I but did at Field Experiment II. The location of Field Experiment II (on top of a small hill), favorable climatic conditions, a longer period of drying after the slash, and a larger burn area might have contributed to higher fire intensities and deeper heat penetration into the soil at this site relative to Field Experiment I. An increase in {kappa}fd indicates the formation of ultrafine superparamagnetic grains upon heating (Dearing, 1994). In the combusted (>600°C) topsoil from Field Experiment II, {kappa}fd approached the upper limit of 15% (Dearing, 1994), indicating that ultrafine maghemite particles dominated the magnetic fraction.

Fractional conversion factors (FC%mass) also increased with heat intensity. Combusted topsoil (>600°C) from Field Experiment II gave a FC%mass value of {approx}100%, indicating that maximum conversion of goethite to maghemite had taken place in the field (Table 5). The FC%mass values appear to be good indicators of fire intensity for the 0- to 5-cm depth samples. In this regard, it is surprising that the 5- to 15-cm depth samples showed higher FC%mass values for all heat intensities in Field Experiment I and for the 100 and 300°C treatments in Field Experiment II. It might be expected that the subsoil FC%mass values would be lower because there was less exposure to the high surface temperatures. However, the total organic matter contents of the 5- to 15-cm depth samples were, on average, half those of the 0- to 5-cm depth materials and may have limited the conversion of goethite to maghemite. We investigated the contribution of soil organic matter and Fe oxide concentration, as well as the kinetics of change in magnetic susceptibility at two temperatures (300 and 600°C), in two oven experiments using the forest soil from Field Experiment II (Table 1).

Contribution of Soil Organic Matter to Increases in Magnetic Susceptibility
Adding additional organic material to the unburned topsoil of Field Experiment II prior to exposing it to 600°C for 930 min caused a linear decrease in magnetic susceptibility (Fig. 5) according to the relationship:

(3)
where {chi}lf is the mass-specific magnetic susceptibility at low frequency in 10-8 m3 kg-1, and masstea is the number of grams of air-dry tea kg-1 soil. Standard errors amounted to 7.1 x 10-8 m3 kg-1 and 0.067 x 10-8 m3 g-1 for the intercept and the slope, respectively. The tea contained 61 g kg-1 ash, and this paramagnetic ash was apparently responsible for the decrease in magnetic susceptibility due to a dilution effect.



View larger version (19K):
[in this window]
[in a new window]
  		Fig. 5 Effect of organic material addition (Indian tea) on maximum achievable mass-specific magnetic susceptibility for initially unburned topsoil (0–5 cm) and subsoil (5–15 cm) in Oven Experiment I. Samples were heated at 600°C for 930 min to ensure complete combustion. OC is organic C

 
By contrast, adding organic material to the subsoil increased the {chi}lf up to 50% (Fig. 5). The {chi}lf increased until 105 g kg-1 organic material were added and then reached a plateau of 2147 x10-8 m3 kg-1. Eighty-five percent of this increase was achieved with the addition of 80 g kg-1 organic material. The organic C content of the forest subsoil prior to burning was 20 g kg-1. Addition of 105 g kg-1 tea (with 61 g kg-1 ash) to the subsoil made the organic matter level similar to that of the topsoil. Apparently, 70 g kg-1 organic C was sufficient to keep this soil reduced during the burn and enabled complete conversion of goethite to maghemite.

Addition of 50 and 100 g kg-1 granulated sugar increased the {chi}lf of the topsoil to 654 and 696 x 10-8 m3 kg-1, respectively. For subsoil the results were 1517 and 1869 x 10-8 m3 kg-1 for 50 and 100 g kg-1 sugar addition, respectively. These values deviated only slightly from the values obtained with the tea experiment.

These observations support our earlier hypothesis that the amount of organic matter controlled the conversion of goethite to maghemite in the subsoil. Similar results were obtained by Brown (1988) who observed an increase in magnetic susceptibility when flour was added to a sample before igniting it at 550°C for 1 h. The relatively high FC%mass for subsoil as compared with topsoil (Table 5) could thus be explained by insufficient organic matter to keep the environment reduced during burning. The final {chi}lf of the subsoil material after addition of extra organic matter was higher than for topsoil with the same organic matter content, which is probably due to the higher CBD-extractable Fe oxide content of the subsoil (24 g kg-1 Fe in subsoil vs. 12 g kg-1 Fe in the forest topsoil).

Conversion Kinetics
The increase in {chi}lf during the first 30 min of heat exposure in Oven Experiment II is shown in Fig. 6 . Also indicated in the same figure are the final levels of {chi}lf reached after 660 min of exposure. Soil water determinations (results not shown) indicated that pore water was lost at 300 and 600°C after 10 and 1 min of exposure, respectively. The lag time in response to temperature for the 300°C curve can be explained by the initial vaporization of soil water; that is, conversion of goethite to maghemite did not take place until water in the samples had vaporized. After water vaporization, {chi}lf increased sharply to 80% of the maximum level within 220, 115, 300, and 75 min for topsoil at 600°C, subsoil at 600°C, topsoil at 300°C, and subsoil at 300°C, respectively.



View larger version (21K):
[in this window]
[in a new window]
  		Fig. 6 Mass-specific magnetic susceptibility ({chi}lf) as a function of time and temperature of exposure for forest topsoil (0–5 cm, 67 g kg-1 organic C, 12 g kg-1 Fe) and subsoil (5–15 cm, 20 g kg-1 organic C, 24 g kg-1 Fe) in Oven Experiment II. On the right for each treatment is the {chi}lf measured after 660 min exposure

 
The maximum {chi}lf obtained by heating the topsoil for an extended period of time at 600°C was {approx}50% of that observed for the subsoil. Part or all of this difference in obtainable {chi}lf was probably due to the fact that the forest subsoil contained 24 g kg-1 CBD-extractable Fe; whereas, the topsoil contained only 12 g kg-1 CBD-extractable Fe. The CBD-extractable Fe in the samples exposed to 600°C decreased rapidly with time of exposure. A similar decrease was observed in the combusted topsoil in Field Experiment II and was most likely due to the formation of glassy, sand-sized particles that encapsulated and protected the ultrafine maghemite particles.

Field vs. Laboratory Levels of Magnetic Susceptibility
In both field experiments, burning produced measurable changes in magnetic susceptibility, especially in the topsoil where fires lasted for a longer time (Field Experiment II). The extent of change differed significantly from those obtained in oven experiments, where temperature levels and duration of exposure were controlled. Maximum achievable {chi}lf levels with oven experiments using forest soil from Field Experiment II were 728 (topsoil) and 1333 x 10-8 m3 kg-1 (subsoil) at 600°C and 543 (topsoil) and 1188 x 10-8 m3 kg-1 (subsoil) at 300°C (Fig. 6). The values obtained at comparable temperatures in the field were much lower. Topsoil exposed to 300°C in the oven reached field-observed levels after 12 min of exposure. For subsoil material, 3 min of oven exposure was sufficient to reach field levels. For exposure to 600°C in the oven, field-observed levels were reached after 3 min (topsoil) and 15 s (subsoil). Because field samples had a water content of {approx}450 g kg-1 directly before the burn (much larger than the water content of the 60°C dried soil that was used in the oven experiments), an increase in time needed for evaporation of the remaining water probably accounts for the substantial increase in time of exposure required in the field.

The highest magnetic susceptibility found in Field Experiment II (combusted topsoil) was 2467 x 10-8 m3 kg-1. The FC%mass approached 100% for the combusted topsoil, indicating that this was the maximum obtainable magnetic susceptibility. The magnetic susceptibility of combusted topsoil in Field Experiment II was higher than the maximum achievable level for forest topsoil (670 x 10-8 m3 kg-1) in Oven Experiment II and also of subsoil with the addition of extra organic matter (2147 x 10-8 m3 kg-1). Because topsoil from the originally unburned forest soil used in the oven experiment contained less than one-third of the CBD-extractable Fe oxides (12 g kg-1 Fe in forest topsoil, 41 g kg-1 Fe in slashed topsoil prior to burning), a higher magnetic susceptibility for the field-combusted topsoil should be expected.


   Conclusions
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
REFERENCES
 
Field observations showed that in intense burns, such as in Field Experiment II where average surface temperatures were between 470 and 550°C (data not presented), a total of 30 to 35% of the surface of the burned field was covered by combusted (reddened) soil material. At places where tree trunks were present, underground burning of roots led to the formation of holes that could extend up to 50 cm deep. This severe burning was observed more frequently in our research area and covered larger areas than reported by others.

In a recent survey in the Sepunggur area (Ketterings et al., 1999), farmers indicated that reddened soil was undesirable for crop growth, based upon their observation that it could not hold water and was not fertile. Our research showed that severe burning had drastic effects on soil texture and mineralogy. Changes in soil mineralogy also take place at lower fire intensities and are still likely to influence the soil fertility status. Further research is needed to determine how important these changes in soil mineralogy are for nutrient availability in the growing season after the burn. Magnetic susceptibility should be an excellent tool for determining the spatial variability of fire intensity for soil fertility evaluations.Genstat 5


   ACKNOWLEDGMENTS
 
This research was funded by a Scholarship from the Mervin G. Smith International Studies Fund, an Ohio State University Graduate School Alumni Research Award, and by two projects within the International Center for Research in Agroforestry Southeast Asia Regional Program: the Smallholder Rubber Agroforestry Project (CIRAD/GAPKINDO/ ICRAF) and the Alternatives-to-Slash-and-Burn project supported by the Global Environment Facility with United Nations Development Program sponsorship. Salary and research support for J.M.B. were provided by state and federal funds appropriated to the Ohio Agricultural Research and Development Center, The Ohio State University. We thank Sandy Jones for running the texture analyses and are grateful to Mr. Zulkifli (Field Experiment I) and the chief of the village of Sepunggur (Field Experiment II) for allowing us to take samples and to study their fields during the burning season of 1997. Constructive criticisms from Dr. D.W. Ming, Associate Editor, and three anonymous reviewers led to several improvements in the manuscript. And finally, terimah kasih to all the kids of Sepunggur. Their presence made digging soil so much more interesting.

Received for publication February 1, 1999.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 Materials and methods
 Results and discussion
 Conclusions
 REFERENCES
 




This article has been cited by other articles:


Home page
 Soil Sci.Home page
C. Guerrero, J. Mataix-Solera, V. Arcenegui, J. Mataix-Beneyto, and I. Gomez
Near-Infrared Spectroscopy to Estimate the Maximum Temperatures Reached on Burned Soils
Soil Sci. Soc. Am. J., May 16, 2007; 71(3): 1029 - 1037.
[Abstract] [Full Text] [PDF]

Home page
 Clays and Clay MineralsHome page
R. Zanelli, M. Egli, A. Mirabella, M. Abdelmoula, M. Plotze, and M. Notzli
'BLACK' SOILS IN THE SOUTHERN ALPS: CLAY MINERAL FORMATION AND TRANSFORMATION, X-RAY AMORPHOUS Al PHASES AND Fe FORMS
Clays and Clay Minerals, December 1, 2006; 54(6): 703 - 720.
[Abstract] [Full Text] [PDF]

Home page
 Soil Sci.Home page
Q. M. Ketterings and J. M. Bigham
Soil Color as an Indicator of Slash-and-Burn Fire Severity and Soil Fertility in Sumatra, Indonesia
Soil Sci. Soc. Am. J., September 1, 2000; 64(5): 1826 - 1833.
[Abstract] [Full Text]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (24)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Ketterings, Q. M.
Right arrow Articles by Laperche, V.

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Crop Science
Journal of Natural Resources
and Life Sciences Education		 Vadose Zone Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome