close

Вход

Забыли?

вход по аккаунту

?

Geophysical evidence for fires in antiquitypreliminary results from an experimental study. Paper given at the EGS XXIV General Assembly in The Hague April 1999

код для вставкиСкачать
Archaeological Prospection
Archaeol. Prospect. 8, 211–225 (2001)
DOI: 10.1002/arp.170
Geophysical Evidence for Fires in
Antiquity: Preliminary Results from an
Experimental Study
Paper Given at the EGS XXIV General
Assembly in The Hague, April 1999†
N. T. LINFORD1,2 * AND M. G. CANTI1
1
Ancient Monuments Laboratory, English Heritage, Fort Cumberland, Eastney,
Portsmouth PO4 9LD, UK
2
Palaeomagnetic Laboratory, Department of Geological Sciences, University College
London, Gower Street, London WCIE 6BT, UK
ABSTRACT
The ubiquitous use of fire in antiquity has been identified as an important mechanism for the
magnetic enhancement of archaeological sediments and their subsequent identification through
the use of surface magnetometer or topsoil susceptibility surveys. However, despite a number
of specific studies there remains little quantitative evidence to examine the effect of domestic
fire on the magnetic properties of soils and estimate the degree of magnetic enhancement such
processes may create. This study reports results from a series of experimental fires lit over a
range of differing substrates to examine the geophysical response created by the fire and the
magnetic enhancement of the underlying soil. In all cases, the temperature profile both in the
air and soil was measured using an array of thermocouples and measurements of magnetic
susceptibility demonstrate a varying degree of enhancement dependent upon the depth from the
surface and underlying substrate. Initial results demonstrate the ‘softer’ (low coercivity) magnetic
nature of the thermally enhanced soils owing to an increased concentration of very fine-grained
superparamagnetic material. It is hoped that more detailed magnetic measurements on both burnt
and control samples collected before the fire was lit will reveal the critical temperatures required
for significant magnetic enhancement to occur. Copyright  Crown Copyright 2001. Recorded
with the permission of Her Majesty’s Stationery Office. Published by John Wiley & Sons, Ltd.
Key words: fire; magnetometer; magnetic susceptibility; rock magnetic measurements; thermal
alteration
Introduction
The deliberate use of fire in all its forms provides
evidence for anthropogenic activity throughout
antiquity and has long been recognized as an
important source for the creation of magnetic
anomalies detected through geophysical survey.
*Correspondence to: N. T. Linford, Ancient Monuments
Laboratory, English Heritage, Fort Cumberland, Eastney,
Portsmouth PO4 9LD, UK.
† Neither Her Majesty’s Stationery Office of the Department
of Archaeometry accept any responsibility for the accuracy of
any recipe, formula or instruction published in this article or
in the journal.
Indeed, the first application of a magnetometer
survey to archaeology arose from the expectation
of a considerable magnetic anomaly from Roman
pottery kilns (Aitken et al., 1958). Following the
success of this initial survey it was soon realized
that the action of fire played a much wider
role in the magnetic enhancement of topsoil and
subsequent site formation processes, leading to
the production of discernible magnetic anomalies
from a variety of buried features (e.g. Clark, 1990;
Scollar et al., 1990).
Experimental studies based on either the
controlled heating of soil in the laboratory or
actualistic measurements made on real fires have
Copyright  Crown Copyright 2001. Recorded with the permission of
Her Majesty’s Stationery Office. Published by John Wiley & Sons, Ltd.
Received 20 July 2000
Accepted 26 January 2001
N. T. Linford and M. G. Canti
212
also provided much useful information to determine the magnetic response of burnt features. To
date, these studies have investigated the enhancement of soil magnetic susceptibility through the
action of fire (e.g. LeBorgne, 1960; Tite and
Mullins, 1971; Mullins, 1974), the production of
magnetic anomalies from experimental hearths
(e.g. Gibson, 1986; Bellomo, 1993; Marshall, 1998,
Morinaga et al., 1999) and the contribution of
wood fuel ash to the magnetism of the resulting
sediments (McClean and Kean, 1993). The aim of
the present study was to enhance the available
literature by conducting a series of experimental
fires examining the influence of substrate type
and length of burn on the magnetic properties
of the underlying sediments and resultant magnetometer anomalies. In particular, it was hoped
that the combination of detailed soil temperature
data with the geophysical measurements would
allow the following points to be clarified:
(i) the degree of burning necessary to create
discernible magnetic anomalies;
(ii) the degree of enhancement experienced by
soils under typical short-duration fires and
the volume of enhanced material produced;
(iii) the identification of specific thermal histories
through magnetic measurements related to
both the length of exposure and maximum
temperature attained;
(iv) the ratio between induced and (thermo)
remanent components of the resulting magnetic anomaly to be determined for short,
camp-fire-type burning episodes.
Method
Experimental fires
A total of five experimental fires were conducted
during the summer of 1998 at a site near
Yarnton, Oxfordshire, UK, where the opportunity
arose to examine two extreme substrates, a dry,
bare, gravelly sand and a waterlogged clay soil,
together with an intermediate sandy loam soil all
within a modest distance of each other. Fires
were kindled on each of the three substrates
and fed with a mixture of wood fuel for a
period of approximately 2 h before being left
to cool. Additional longer term fires rekindled
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
every morning over a period of 4 days were also
conducted on the gravel and clay substrates.
Temperature measurements
The temperature of the underlying soil was
determined through a linear array of K-type
thermocouples buried in a narrow north–south
trench across the centre of each experimental
fire. Measurements were made at five sample
stations running across the mid-point of the fire
0.5 m apart at depths of 1 cm and 4 cm beneath
the ground surface. An additional array of
temperature measurements was collected above
the fire from the ground surface to a height of
30 cm at 10 cm vertical intervals. These latter
measurements were made along the same line as
the buried thermocouple array at five horizontal
sample stations separated by 0.25 m aligned on
the centre of each fire.
Magnetic survey
Prior to the introduction of the thermocouple
arrays to each of the experimental fire sites a
detailed magnetometer survey was conducted
with a Geoscan FM36 fluxgate gradiometer.
The survey was then repeated following the
completion of the fire experiment and the careful
excavation of the ferrous thermocouple array. For
all surveys, readings were collected at a 0.25 m
interval from a 5 mð 5 m grid along north–south
orientated parallel traverses separated by 0.5 m.
To enhance low-magnitude magnetic anomalies
a mechanical modification to the carrying handle
of the magnetometer was made to permit the
bottom sensor to be lowered to within 0.05 m of
the ground surface.
Extreme care was taken during data acquisition
to ensure that sample stations were spatially
identical in both the before and after surveys
conducted over each site. This allowed the
subtraction of the initial from the final data set
to provide an estimation of the residual anomaly
created by each of the experimental fires.
Soil samples and magnetic susceptibility
Individual ca. 10 g soil samples were collected
adjacent to the five thermocouple measurement
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
stations at depths of 1 cm, 4 cm and 7 cm both
before and after each of the experimental fires.
The post-burn samples were offset by 0.25 m
from the original collection point to ensure that
soil undisturbed by the initial sampling was
recovered. Additional samples from the ash layer
were also collected where these had developed
over the original ground surface. Measurements
of magnetic susceptibility were then made on
all samples in the laboratory at two frequencies
(470 Hz and 4700 Hz) with a Bartington MS2
AC susceptibility meter and MS2B sensor. The
dry mass of each sample was subsequently
determined after air-drying at room temperature.
Following the measurement of natural remanent magnetization (NRM) all samples were
exposed to laboratory fields for the determination of anhysteretic and isothermal remanent
magnetisations (ARM and IRM). A Dtech D2000
alternating field demagnetizer was used to impart
an ARM to the samples through a peak alternating field of 199 mT superimposed over a
0.05 mT steady field. The resulting ARM199 mT
was subsequently demagnetized by exposure to
a peak 40 mT alternating field generated by the
same instrumentation ARM40mT/199mT . Samples were then exposed to a maximum direct
field of 2.5 T IRM2.5T generated by an ASC
Ranger pulse magnetizer and then progressive reverse fields of 30, 100, 300 and 1000 mT
IRM30mT/100mT/300mT/1000mT . Resulting magnetizations were all measured in a Geofysika
JR5A spinner magnetometer.
Magnetic anomaly models
The magnetic susceptibility of the underlying soil
and ash layers were used to construct numerical
models to estimate the expected magnitude of
gradiometer response for comparison with the
field data. Models were based on a series of thin
rectangular prisms representing each of the soil
samples from the three measured depths. The
prisms were extended laterally to produce an
approximately circular distribution of heated soil
and ash similar to the results of the experimental
fires recorded in the field. Magnetic anomalies
of Zs ZsC0.5m (where S is the height of the
lower sensor above the ground surface) were
calculated following Linington (1972), assuming
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
213
the local magnitude and inclination of the Earth’s
magnetic field are 50 000 nT and 68° respectively.
Results
Soil and air temperatures
Figure 1 provides a typical illustration of the air
temperature recorded above the fires taken from
results of the 4-day fire on the clay soil and
demonstrates that the fuel burns at temperatures
well in excess of 600 ° C. Note also the rapid and
prolonged heating of the first 1 cm of soil in the
centre of the fire to 200–400 ° C. A more detailed
discussion of the results from the temperature
variations recorded during the five experimental
fires is given in Canti and Linford (2000).
Further examples of the variation in subsoil
temperatures are given in Figures 2A and 2B,
which show results from the first day of the 4day fires conducted over the clay and gravelly
sand substrates respectively. Both demonstrate
maximum subsoil temperatures at a depth of
1 cm of approximately 500 ° C and substantial
areas of the near-surface heated to ca. 250 ° C for
considerable periods of time. The insulating effect
of the rapidly formed fuel ash layer is also visible
in Figure 2A, suppressing the peak temperature
of thermocouple 8. Table 1 provides a summary
of the maximum temperatures attained by the
subsoil thermocouples. The extreme temperature
attained at 1 cm depth on the sandy soil appears
anomalous and probably results from disruption
of the surface soil during fuelling, causing direct
contact between the thermocouple and a burning
ember.
Magnetic survey
The residual magnetic anomalies recorded over
each of the experimental fires are illustrated
Table 1. Summary of maximum recorded subsurface temperatures (° C)
Bare gravel
1-day fire
4-day fire
Sandy soil
Clay
1 cm
4 cm
1 cm
4 cm
1 cm
4 cm
433
436
276
289
570
N/A
318
N/A
456
443
232
199
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
214
0 mins
15
30
1000
60
650
deg C
75
90
105
120
350
10
135
150
165
180
Position of thermocouples
195
210
225
240
Figure 1. Cartoon illustrating the variation in temperature recorded during the first day of the 4-day fire on clay in both the
body of the fire and from the subsurface thermocouples.
in Figure 3 in both XY trace-plot and greytone
image form. For comparison, the vertical scale
of the trace plots and the tone mapping of the
greytone images is identical for each plot. In
all cases a discernible magnetic anomaly has
arisen following the experimental fire, with a
peak (positive) magnitude in the range from
50 to 100 nT. The magnitude of the 4-day
fires on the sand and clay soils is slightly
greater than the equivalent 1-day burn and
is further distinguished by the development
of a more pronounced annulus of negative
readings.
The anomaly developed following the 1-day
fire on the sandy loam soil is considerably
more subtle than the response from similar
duration burn on either the sand or clay. In
part, this may be due to the higher background
magnetic susceptibility (Figure 4) of the sandy
loam although the subtraction of the pre-burn
survey results should have reduced the response
to that of the residual anomaly only.
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
Soil and ash samples
The variation of for the soil samples beneath
each of the fire experiments is shown in Figure 4,
where solid symbols represent the pre-burn soil
samples. The figure illustrates both differing
degrees of enhancement between the soil types
and also the influence of a greater exposure to
the heat source during the 4-day fires. In particular, note the 4-day fire over the gravelly sand,
which produced a discernible magnetic enhancement to a depth of 7 cm. Results from the 1-day
fire over clay are of interest owing to the modest increase in demonstrated by the post-burn
samples. Despite this apparent lack of enhancement the same experimental fire produced a
strong (75 nT) residual magnetometer anomaly
(Figure 3), suggesting either a strong thermoremanent component to the magnetization or the
influence of a considerably enhanced ash layer.
From empirical observations, Dearing et al.
(1996) suggest that the normalized parameter
Archaeol. Prospect. 8, 211–225 (2001)
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
0
60
180
240
Duration (Minutes)
120
300
16
360
S
200cm
0
Section
•10•9 •8 •7 •6
•5 •4 •3 •2 •1
Plan
>100°C
>250°C
4cm
Location of Thermocouples
1cm
Temperature °C
Temperature °C
30
10
30
10
0
100
200
300
400
0
100
0
60
180
Duration (Minutes)
120
insulating ash layer
240
8
16
27
3
4
9
300
5 10
Wind directions
0
S
200cm
1cm
Section
•10•9 •8 •7 •6
•5 •4 •3 •2 •1
Plan
>100°C
>250°C
4cm
Location of Thermocouples
N
Figure 2. Detailed graphs of the variation of subsurface soil temperature during the first day of the 4-day fires conducted on the gravelly sand (A) and clay (B) substrates.
Inset diagrams show the positions of the thermocouples and the prevailing wind direction during the two experiments.
0
30
10
27
8
9
10
Wind directions
•° •° •° •° •°
100
3
4
5
200
30
10
(B) Clay Substrate first day of 4-Day fire
•° •° •° •° •°
200
0
100
0
100
200
0
100
200
300
400
N
(A) Sand Substrate first day of 4-Day fire
Geophysical Evidence for Fires
215
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
216
SAND 4-DAY
N
SAND 1-DAY
CLAY 1-DAY
CLAY 4-DAY
SANDY SOIL 1-DAY
50 nT
−12.35
−0.48
11.38
23.25
nT
0
6m
Figure 3. Residual magnetometer anomalies recorded over the site of the experimental fires.
of FD % (determined from the measurement
of magnetic susceptibility at two differing frequencies) provides an indication of the average
grain size for superparamagnetic (SP) material
within a sample, whereas the absolute parameter of FD is related to the concentration of
SP material present. However, such determinations will hold only where the ratio of frequency
dependent SP material is high compared with
single/multidomain particles and paramagnetic
components. Analysis of a FD % versus FD biplot
(Figure 5) demonstrates an approximately linear
relationship between these two parameters for
the pre-burn samples, where the low values of the
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
normalized term (FD %) most probably reflect the
high concentration of non-frequency dependent
material within the sample rather than the effective grain-size. The linear relationship between
FD % and FD begins to fail for the (near-surface)
post-burn samples that demonstrate a more consistent increase in FD than FD %.
Results from the ash layer are of interest as
samples collected from the 1-day fire on the
clay are removed from the other substrates
on the FD % versus FD biplot, suggesting a
relationship between the substrate type and
magnetic properties of the resultant ash. Further
comparison between all the ash samples and the
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
217
SAND 4-DAY
SAND 1-DAY
10000
χ 10−8 m3kg−1
χ 10−8 m3kg−1
1000
100
10
1000
100
10
1
1
0
100
200
0
Position [cm]
CLAY 1-DAY
200
CLAY 4-DAY
1000
1000
100
100
χ 10−8 m3kg−1
χ 10−8 m3kg−1
100
Position [cm]
10
10
1
1
0
100
Position [cm]
200
0
100
200
Position [cm]
SANDY SOIL 1-DAY
χ 10−8 m3kg−1
1000
1cm PRE
1cm POST
4cm PRE
4cm POST
7cm PRE
7cm POST
Ash
100
10
0
100
200
Position [cm]
Figure 4. Magnetic susceptibility data from five fire experiments. Solid, open and shaded symbols show values for pre-burn,
post-burn and ash layer samples respectively.
corresponding post-burn soil samples shows a
similar range for FD % but enhanced values of
FD . As a similar fuel source was used for all
the fire experiments it would appear that the
highly enhanced magnetic properties of the ash
layer are derived from ‘roasting’ soil in contact
with the fuel as opposed to the enhancement of
iron minerals within the wood itself (cf. McClean
and Kean, 1993). This soil may adhere to the
wood during its preparation or be derived from
the immediate ground surface on which the fire
was lit.
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
Natural remanent magnetization (NRM)
Figure 6 shows the NRM of all the samples
collected from the 4-day fires on the gravel
and clay substrates. The results demonstrate a
distinguishable increase in NRM for the majority of the post-burn samples, with particularly
enhanced values from the ash layer. In comparison to the susceptibility data (Figure 4) the NRM
results provide a more diagnostic identification
of the post-burn samples and demonstrate that an
enhanced NRM is also produced in the deeper soil
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
218
SAND 4-DAY
SAND 1-DAY
16
16
14
14
12
10
1cm Post
10
4cm PRE
8
4cm POST
6
7cm PRE
4
7cm POST
χFD%
1cm PRE
χFD%
12
6
4
2
0
0.01
8
2
1
0.1
0
10
0.01
0.1
χFD 10−8m3kg−1
1
10
χFD 10−8m3kg−1
CLAY 4-DAY
CLAY 1-DAY
14
8
12
6
10
1cm PRE
4cm PRE
4
4cm POST
χFD%
χFD%
1cm Post
8
6
7cm PRE
4
7cm POST
2
2
0
0
0
0.5
1
0.1
1.5
1
χFD 10−8m3kg−1
SANDY SOIL 1-DAY
100
ASH LAYER
16
16
14
14
10
1cm PRE
12
1cm Post
10
4cm PRE
8
4cm POST
χFD%
12
χFD%
10
χFD 10−8m3kg−1
Clay 1-Day
Clay 4-Day
Burnt Soil
Sand 1-Day
Sandy Soil 1-Day
8
6
7cm PRE
6
4
7cm POST
4
2
Sand 4-Day
2
0
0
0
10
20
χFD 10−8m3kg−1
0.1
1
10
100
1000
χFD 10−8m3kg−1
Figure 5. Biplot of FD versus FD % for the soil samples from the five fire experiments together with the combined results from
the ash layer. Solid and open symbols indicate pre-burn and post-burn soil samples respectively. The absolute parameter
FD represents the loss of mass specific magnetic susceptibility between the two measurement frequencies.
samples (7 cm), where the maximum temperature
attained was apparently less than 50 ° C.
Values of the Königsberger ratio Q D MR /H,
the ratio between remanent MR and induced
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
magnetisation MI D H in the Earth’s magnetic
field (H ' 50 000 nT), vary from insignificant
levels in the pre-burn samples <0.1 to >1 in
the post-burn samples with the maximum values
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
219
Clay 4-Day
1000
100
100
NRM [µAm2Kg−1]
NRM [µAm2Kg−1]
Sand 4-Day
1000
10
1
0.1
0.01
0.001
0
50
100
150
200
250
Location [cm]
10
1
0.1
0.01
1cm Pre
0
1cm Post
50
100
150
200
250
200
250
Location [cm]
4cm Pre
4cm Post
7cm Pre
7cm Post
Clay 4-Day
Ash
1000
NRM(AF−200mT) [µAm2Kg−1]
NRM(AF−200mT) [µAm2Kg−1]
Sand 4-Day
100
10
1
0.1
0.01
0.001
1000
100
10
1
0.1
0.01
0.001
0
50
100
150
200
250
Location [cm]
0
50
100
150
Location [cm]
Figure 6. Natural remanent magnetization (NRM) data together with the magnetization remaining after demagnetization in a
peak 200 mT AF field (NRMAF200mT ) for the 4-day fires conducted over the sand and clay substrates.
recorded in the ash layer. There also appears to be
little variation in the range of Q values produced
by the gravelly sand and clay substrates.
Demagnetization of the samples in a peak
200 mT alternating field (Figure 6) suppresses
the NRM by an order of magnitude, with only
the more intensely magnetized soil and ash
samples remaining distinguishable from the preburn values. This suggests that low to medium
coercivity magnetic minerals form the principal
NRM carriers.
Laboratory induced remanence
To investigate further the origin of the NRM carriers, an ARM was induced in each of the samples
in a steady field with magnitude (0.05 mT) similar to the Earth’s magnetic field (Figure 7). Again
the results demonstrate enhanced ARM values in
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
the post-burn samples, suggesting that thermal
alteration has occurred producing an increased
concentration of fine-grain ferrimagnetic material (Banerjee et al., 1981). This is particularly
evident for the 4-day sand experiment, where
the majority of post-burn samples towards the
centre of the fire are enhanced in contrast to the
more ambiguous susceptibility results (Figure 4).
Although the enhancement of ARM in the sand to
a greater depth than in the clay substrate may represent a greater penetration of heat, an alternative
explanation, such as the translocation of very
fine-grained enhanced material from the surface
through the wider pore spaces of the sand cannot
be discounted.
Demagnetization of the ARM in a peak alternating field of 40 mT (Figure 7) demonstrates
that the majority of the post-burn magnetization is held by low coercivity minerals. Figure 7
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
220
Sand 4-Day
Clay 4-Day
10000
ARM(0.05mT) [µAm2Kg−1]
ARM(0.05mT) [µAm2Kg−1]
10000
1000
100
10
1000
1
100
10
1
0
50
100
150
200
250
0
50
Location [cm]
100
150
200
250
200
250
200
250
Location [cm]
Clay 4-Day
Sand 4-Day
0.5
0.5
1cm Post
0.3
4cm Pre
4cm Post
0.2
7cm Pre
0.1
7cm Post
ARM(−40mT)/ARM
ARM(−40mT)/ARM
1cm Pre
0.4
0.4
0.3
0.2
0.1
Ash
0
0
0
100
200
0
50
Location [cm]
Sand 4-Day
150
Clay 4-Day
100
100
(s)IRM(2.5T) [m Am2Kg−1]
(s)IRM(2.5T) [m Am2Kg−1]
100
Location [cm]
10
1
0.1
10
1
0.1
0
50
100
150
200
250
Location [cm]
0
50
100
150
Location [cm]
Figure 7. Values of anhysteretic remanent magnetization (ARM0.05mT ) for samples from the 4-day fires conducted over the
sand and clay substrates together with the residual magnetization after demagnetization in a peak 40 mT AF field (ARM40mT ).
The lower two figures show the same samples after exposure to a 2.5 T pulse laboratory field ((s)IRM2.5T .
also illustrates the higher initial coercivity of
the pre-burn sand samples through their greater
resistance to demagnetization in comparison with
the clay soil.
Application of a large 2.5 T pulse magnetic
field produces comparable results IRM2.5T to
the ARM and again distinguishes the post-burn
samples (Figure 7). Although IRM2.5T generally is
considered to be indicative of the overall concentration of ferro/ferrimagnetic material within a
sample its value is strongly influenced by both
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
the type and domain size of the minerals present.
These results support the interpretation of thermal alteration having occurred in the post-burn
samples, producing a fine-grained strongly magnetic mineral.
Analysis of backfield IRM ratios (Figure 8)
confirms the higher coercivity of the pre-burn
sand samples and the low coercivity of the
post-burn samples from both substrates. Again
it would appear from the backfield IRM ratios
that the pre/post-burn gravelly sand samples are
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
221
Clay 4-Day
Sand 4-Day
1
IRM(−30mT)/(s)IRM
IRM(−30mT)/(s)IRM
1
0.5
0
−0.5
−1
0.5
0
−0.5
−1
0
50
100
150
200
250
0
50
Location [cm]
Sand 4-Day
150
200
250
200
250
200
250
200
250
Clay 4-Day
1
1
IRM(−100mT)/(s)IRM
IRM(−100mT)/(s)IRM
100
Location [cm]
0.5
0
−0.5
−1
50
100
150
200
250
Location [cm]
0
−0.5
−1
1cm Pre
0
0.5
0
1cm Post
50
4cm Pre
100
150
Location [cm]
4cm Post
7cm Pre
Sand 4-Day
Clay 4-Day
7cm Post
1
1
IRM(−300mT)/(s)IRM
IRM(−300mT)/(s)IRM
Ash
0.5
0
−0.5
−1
0.5
0
−0.5
−1
0
50
100
150
200
250
0
50
Location [cm]
Sand 4-Day
Clay 4-Day
IRM(−1000mT)/(s)IRM
IRM(−1000mT)/(s)IRM
150
0
0
−0.5
−1
100
Location [cm]
0
50
100
150
200
250
Location [cm]
−0.5
−1
0
50
100
150
Location [cm]
Figure 8. Backfield magnetization ratios for the samples imparted with an (s)IRM2.5T in incremental reverse fields of 30, 100,
300 and 1000 mT.
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
222
more readily distinguished than those from the
clay soil.
Magnetic anomaly models
Figure 9 shows the results of surface magnetometer anomaly models calculated from the and Q
values of recovered samples from the 4-day fires
over the sand and clay substrates. The models
were calculated initially using only the subsurface samples and then modified to include data
from the ash layer (Figure 9 upper and lower
parts respectively). Comparison with Figure 3
illustrates that inclusion of the ash layer is necessary in order to provide a good agreement
between the magnitude of response predicted
by the models and the corresponding field data.
A more accurate description of the field anomaly
could be obtained by increasing the density of soil
samples recovered for inclusion within the model.
This is particularly evident for the sand model,
where a single highly enhanced ash sample from
the centre of the fire has led to the overestimation of the magnetometer response. Note that
the response from the enhanced and thermoremanence induced within the subsoil beneath the
CLAY 4-DAY MODEL
SAND 4-DAY MODEL
N
Soil Only
50 nT
−16.35
−5.07
6.22
17.50
nT
Soil + Ash
0
6m
Figure 9. Model magnetometer anomalies for the 4-day fires conducted on the gravelly sand and clay substrates estimated
from the measured soil samples presented with the same plotting parameters as the field data shown in Figure 3. The
models are calculated both without and including the presence of the surface ash layer (upper and lower parts of the figure
respectively).
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
fires is visually suppressed in Figure 9 in order
to accommodate both the intense magnetization
of the ash layer and provide a direct comparison,
with identical plotting parameters, to the field
data shown in Figure 3.
Discussion
Actualistic fire experiments will always remain
highly subjective owing to the differing temperature regimes that may be generated through the
design of the fire and prevailing ambient conditions. Many factors, including the length of burn,
fuel source, wind direction and strength will vary
between different experiments and often are difficult to control or evaluate accurately in the
field. However, such experiments are warranted
as they provide results from within the envelope
of temperature regimes that would have been
generated by the use of fire in antiquity and are,
perhaps, more realistic than extrapolating results
based upon laboratory determinations.
Of particular interest to this study was the
generation of magnetic anomalies from shortterm camp fires and the possibility that the
response from such features would survive over
a significant period of time. In this regard the
significance of the surface ash to the magnitude
of the resultant magnetometer anomaly should be
considered, owing to the physical mobility of this
layer and the effects of both deliberate clearance
and weathering. The survival of the highly
enhanced ash-layer in situ will depend upon
the nature of the original feature, for example
whether the fire has been kindled in a more
protected recessed hollow or directly upon the
ground surface. The translocation of the ash layer
may well be of greater significance as it is likely
to be weathered from the site of the original fire
and accumulate with topsoil in contemporary cut
features, producing an enhanced susceptibility
fill. Deliberate removal of the ash layer from
domestic hearths is also common, where, for
example, it may enhance the susceptibility of
midden features or be added to cess pits.
Other studies (e.g. Hamilton et al., 1986;
McClean and Kean, 1993; Peters et al., 1999) confirm the enhanced magnetic properties found in
ash deposits but it remains difficult to identify
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
223
the source of the original iron minerals. Magnetic measurements made on various species of
wood ash, prepared by burning washed wood
in a laboratory furnace (Cole and Canti, personal
communication), demonstrate the presence of ferrimagnetic minerals within the original organic
material. The level of magnetic enhancement
reported during these experiments from a given
quantity of wood, however, could not account for
the properties of the ash layers produced during
the present study. In this case it seems likely that
much of the source of enhanced magnetic material in the ash layer originated from either roasted
soil from the ground surface directly beneath the
fire or material adhering to the wood fuel as it
was burnt.
Although the results of the study indicate
the importance of the ash layer, the magnetic
anomaly models calculated from the subsoil samples alone also produce a distinguishable magnetic response. It is difficult to predict how long
such anomalies would survive, particularly as
much of the resultant magnetization apparently is
the result of low-temperature thermoremanence,
which may be easily disturbed by bioturbation
or subsequent land use (cf. Marshall, 1998). In
addition, the sample methodology used during
the field magnetometer survey is, perhaps, more
rigorous than that usually recommended for the
evaluation of large areas (David, 1995) and thus
the ephemeral response from a degraded camp
fire may easily be missed.
Mineral magnetic identification of burnt soil
Results from the rock magnetic study demonstrate the sensitivity of iron minerals in the soil
to surface burning, even at the relatively modest temperatures attained by the deeper samples.
Although values of are certainly enhanced after
burning, in this case at lower temperatures other
magnetic parameters, such as ARM and IRM,
appear to provide a more diagnostic means of
identifying the heated soil. This suggests that the
low-temperature thermal alteration favours the
production of fine-grained ferrimagnetic material, possibly through either the dehydration of
lepidocrocite to maghemite (e.g. Özdemir and
Banerjee, 1984) or the conversion of paramagnetic
clay minerals. Alternatively, the effect of heat
Archaeol. Prospect. 8, 211–225 (2001)
N. T. Linford and M. G. Canti
224
Conclusion
This study demonstrates that short-term camp
fires provide sufficient heat to produce a distinguishable magnetometer anomaly through the
production of both a magnetically enhanced ash
layer and the thermal alteration of the underlying soil. With respect to the magnetometer
anomaly, the thermoremanent component of the
ash layer provides the most significant source of
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
1
0.1
χ / χMax
may produce a physical as opposed to chemical
alteration of the constituent magnetic minerals,
perhaps either reducing the size of multidomain
grains or reducing the separation of very fine
SP material to the point that grain–grain interactions occur. Both of the latter mechanisms might
well enhance the remanent magnetization of the
sample with little change to .
Further laboratory measurements are required
to identify the mineral alterations occurring
within the three soil types and confirm the necessary conditions (e.g. temperature, length of
exposure, presence of organic matter) required to
produce significant magnetic enhancement. Perhaps of greater interest is the longevity of these
thermally altered minerals, as many of the subsurface samples contain no indication of burning
(such as the presence of charcoal or an obvious
colour change) other than their diagnostic magnetic behaviour. Should this magnetic signature
survive over significant periods of time, mineral
magnetic measurements may well provide a useful means of reconstructing the thermal history
of a particular soil (e.g. Marmet et al., 1999).
In this case it would appear that the enhanced
magnetic properties of the sand and clay samples
accurately reflect the differing depth of heat penetration into the soil recorded by the thermocouple
arrays during the fire experiments. The degree of
enhancement with depth demonstrated by the
clay samples, however, may also be the result
of greater sensitivity to thermal alteration of this
substrate. For example, Figure 10 illustrates the
variation in for the three substrates after heating
fresh samples in a laboratory furnace to a maximum temperature of 600 ° C, and demonstrates
that the clay undergoes a much greater degree of
enhancement at low temperatures (<300 ° C).
0.01
Sand
Sandy Soil
Clay
0.001
0
100
200
300
400
Temperature °C
500
600
Figure 10. Variation of magnetic susceptibility with temperature for samples of the three substrates heated under
controlled conditions in a laboratory furnace.
magnetization, arising through the high temperature ‘roasting’ of soil both from the immediate
ground surface and that adhering to the wood
fuel itself. As a result of exposure of the ash
layer to weathering it seems likely that it will
be dissipated relatively quickly, leaving only the
thermally altered soil beneath the ground surface
resulting in a considerably reduced magnetometer anomaly. Results from the repeated 4-day
fires on the gravelly sand and clay soils confirmed that the increased exposure time resulted
in greater degree of magnetic enhancement. As
the maximum subsurface temperatures did not
increase significantly between the 1 and 4-day
experiments it would appear that the degree of
magnetic enhancement is also highly dependent
upon the length of exposure to this temperature.
From the recorded thermocouple measurements it would appear that the thermal alterations
of the soil can occur at relatively modest maximum temperatures (ca. 150 ° C), although the
degree of enhancement in terms of increased
magnetic susceptibility is limited. In this study
other rock magnetic parameters, particularly
ARM, IRM2.5T and IRM backfield measurements,
proved more sensitive to the detection of thermally altered soils and appear to correlate with
the depth of heat penetration from the surface fire.
Additional laboratory experiments are currently
being undertaken to assess whether the rock magnetic measurements may yield a more detailed
Archaeol. Prospect. 8, 211–225 (2001)
Geophysical Evidence for Fires
thermal history indicating the approximate maximum temperature and exposure time to which a
soil sample was subjected.
Acknowledgements
The authors wish to express their thanks to
Hanson Aggregates for allowing the fire experiments to be conducted at the Yarnton–Cassington
extraction site and for their assistance in providing clearance wood for fuel. Thanks also is due
to colleagues at the Oxford Archaeological Unit,
particularly Gill Hey, Chris Bell and Robin Bashford and to Tim Horsley of Bradford University
who provided invaluable non-flammable assistance in the field. Louise Martin of the Ancient
Monuments Laboratory greatly assisted with the
measurement of the samples to whom thanks is
also duly proffered.
References
Aitken MJ, Webster G, Rees A. 1958. Magnetic prospecting. Antiquity 32: 270–271.
Banerjee SK, King JW, Marvin JA. 1981. A rapid
method for magnetic granulometry with applications to environmental studies. Geophysical Research
Letters 8: 333–336.
Bellomo R. 1993. A methodological approach for identifying archaeological evidence of fire resulting from
human activities. Journal of Archaeological Science 20:
525–553.
Canti MG, Liuford N. 2000. The Effects of Fire on
Archaeological Soils and Sediments: Temperature
and Colour Relationships, Proceedings of the Prehistoric Society 66: pp 385–395.
Clark AJC. 1990. Seeing Beneath the Soil. Batsford:
London.
David A. 1995. Geophysical Survey in Archaeological Field
Evaluation. English Heritage: London.
Dearing JA, Hay KL, Baban SMJ, Huddleston AS, Wellington EMH, Loveland PJ. 1996. Magnetic susceptibility of soil: an evaluation of conflicting theories
Copyright  Crown Copyright 2001. Recorded with the
permission of Her Majesty’s Stationery Office.
Published by John Wiley & Sons, Ltd.
225
using a national data set. Geophysical Journal International 127: 728–734.
Gibson TH. 1986. Magnetic prospection on prehistoric
sites in Western Canada. Geophysics 51: 553–560.
Hamilton AC, Magowan W, Taylor D. 1986. Use of the
Bartington meter to determine the magnetic susceptibility of organic-rich sediments from western
Uganda. Physics of the Earth and Planetary Interiors 42:
5–9.
LeBorgne E. 1960. Influence de feu sur les propriétés
magnétique anormale du sol et du granite. Annales
de Géophysique 16: 159–195.
Linington RE. 1972. A summary of simple theory
applicable to magnetic prospecting in archaeology.
Prospezioni Archeologiche 7: 9–60.
Marmet E, Bina M, Fedoroff N, Tabbagh A. 1999. Relationships between human activity and the magnetic
properties of soils: a case study in the medieval
site of Roissy-en-France. Archaeological Prospection 6:
161–170.
Marshall A. 1998. Visualizing burnt areas: patterns of
magnetic susceptibility at Guiting Power 1 Round
Barrow (Glos., UK). Archaeological Prospection 5:
159–177.
McClean RG, Kean WF. 1993. Contributions of wood
ash magnetism to archaeomagnetic properties of fire
pits and hearths. Earth and Planetary Science Letters
119: 387–394.
Morinaga H, Inokuchi H, Yamashita H, Ono A, Inada
T. 1999. Magnetic detection of heated soils at palaeolithic sites in Japan. Geoarchaeology 14: 377–399.
Mullins CE. 1974. The magnetic properties of the soil
and their application to archaeological prospecting.
Archaeo-Physika 5: 143–347.
Özdemir O, Banerjee SK. 1984. High temperature stability of maghemite. Geophysical Research Letters 11:
161–164.
Peters C, Church MJ, Coles G. 1999. Mineral magnetism and archaeology at Galson on the Isle of
Lewis, Scotland. Poster presented at the ‘Magnetic
Moments’ Conference, Oxford, 1999.
Scollar I, Tabbagh A, Hesse A, Herzog I. 1990. Archaeological Prospecting and Remote Sensing. Cambridge
University Press: Cambridge.
Tite MS, Mullins CE. 1971. Enhancement of the magnetic susceptibility of soils on archaeological sites.
Archaeometry 13(2): 209–219.
Archaeol. Prospect. 8, 211–225 (2001)
Документ
Категория
Без категории
Просмотров
3
Размер файла
244 Кб
Теги
paper, april, hague, general, xxiv, geophysics, fire, antiquitypreliminary, experimentov, 1999, assembly, stud, evidence, give, results, egs
1/--страниц
Пожаловаться на содержимое документа