вход по аккаунту


How effective is geophysical survey A regional review.

код для вставкиСкачать
Archaeological Prospection
Archaeol. Prospect. 16, 77–90 (2009)
Published online 3 March 2009 in Wiley InterScience
( DOI: 10.1002/arp.348
How Effective is Geophysical Survey?
A Regional Review
Archaeology Institute, University of Bern, Langasse Strasse 10, 3012 Bern, Switzerland
Geophysicsis such an acceptedpart of British archaeology that itseffectiveness seemsobvious.Yet if
there is no reason to doubt the benefits of geophysics why do some experienced archaeologists
use it so rarely and why is it little used in some countries which, in other ways, have highly developed
professional archaeology services? There are, often cheaper, alternatives for archaeological survey.
Yet since the performance of different survey methods has rarely been studied systematically there
is no objective basis on which to test which choices best meet archaeologists’ needs. Moreover the
geophysicists’ understandable desire to present successful rather than unsuccessful surveys, and
to discuss resultsin geophysicalrather than archaeologicalterms, makes such assessment more difficult. Thus although geophysical surveyors have strong grounds to claim that their work benefits
archaeology, those who pay for survey can reasonably ask that these benefits be clarified, quantified
where possible, and compared with alternatives, such as aerial photography or surface artefact
survey, so that they can make the best choices about its use.
This paper summarizes a study of all the geophysical surveys carried out in the northwest of England before 2006. The study assessed the performance of geophysical surveys in archaeological
terms and was centred on a detailed analysis of 35 sites for which there is good comparative
excavation data or whichhaveparticularlyillustrative casehistories.The studyconcludesthat, despite
the doubts in this area, geophysics serves archaeologists well and provides greater certainty in both
identifying where sites exist and where they do not exist than has been generallyassumed.It therefore
deserves more extensive and more rational use.Geophysics is, however, being underused because,
although abundant, surveys are formulaic and commercial surveyors are rarely able to fit methodologies to sites by a programme of reflective project development.Thus, although currently effective,
geophysics might be even more so if surveyors had the time and resources to do this and to answer
more complexand specific questions.The paperconsidershow these findingsrelate to the use ofgeophysical survey in other countries. Copyright # 2009 John Wiley & Sons, Ltd.
Key words: review; geophysical survey; England; planning
We have half a century of evidence that
archaeological geophysics ‘works’. Thousands
of surveys, carried out all over the world and in a
vast range of conditions, have mapped buried
remains. The wide use of geophysics in the
* Correspondence to: D. Jordan, Archaeology Institute, University of Bern, Langasse Strasse 10, 3012 Bern, Switzerland.
Contract/grant sponsor: Aggregates Levy under the management of English Heritage.
Copyright # 2009 John Wiley & Sons, Ltd.
cut-throat environment of British commercial
archaeology shows that some archaeologists
think it provides what they need against tough
competition from other methods of prospection
and site evaluation. Yet the relative rarity of
geophysical survey in many countries with welldeveloped archaeological services, and the infrequency of its use by some major British archaeological consulting organizations, suggest that
others remain sceptical.
It is striking that professional archaeologists
can disagree so much about the value of
geophysics. This disagreement is a result of the
Received 24 October 2008
Accepted 17 December 2008
complexity of the judgements archaeologists
must make about the methods they use and of
the persistent lack of objective measures of
survey effectiveness available to support those
judgements. Moreover, surveyors tend to consider their results in geophysical terms whereas
the users of survey results, overwhelmingly
archaeologists, are concerned with relative cost
and confidence in results as much as with the
simple detection of remains. Surveyors must
therefore answer some hard questions. Could
the same results have been better achieved, or
achieved more cheaply with other approaches –
aerial photography for example? What does
geophysics not tell us about the buried remains
of a site? What will it fail to find? The issues at
stake can be significant. For example an archaeologist might forbid the construction of a factory
on the eastern half of a field, where geophysics
has found a Roman villa, but permit it on the
western half where geophysics has failed to
find the ephemeral – but archaeologically vital –
traces of a Neolithic settlement. The key matter
here is not what geophysics can find but what
it cannot, because, in many such matters of
planning, it is the risk of failure to find remains
that is more significant.
This paper is the result of a review of
archaeological geophysics in the northwest of
England commissioned by English Heritage. It
was prompted by the concerns of some professional archaeologists, consultants and curators
(regional archaeology managers usually working
for local government), that geophysics did not
give them what they needed – in particular that it
too often failed to find remains for them to have
confidence when making complex planning
decisions. As a result some had doubts about
whether it should be used at all and many were,
in fact, making a conscious decision not to use it
despite its abundant use by colleagues working
nearby. These doubts seem widespread, both
nationally and internationally. Some countries
make abundant use of geophysics and others
very little.
The project therefore set out to identify all of
the surveys that had been carried out in the
region and to analyse those with sufficient
evidence to see how effective they had been. In
addition, discussions were held with almost all of
Copyright # 2009 John Wiley & Sons, Ltd.
D. Jordan
the senior professional archaeologists – surveyors, curators and consultants – working in
the region to find out how they view the use of
geophysics and its future potential. On the
understanding of anonymity some strong views
were expressed and it became clear that very few
are comfortable with some key aspects of the
professional practice of geophysics in England.
This is of wider significance because geophysics
is deeply integrated into routine archaeological
services in England and English survey practice
is influential elsewhere, not least through the
English Heritage Guidelines for geophysical
survey (English Heritage, 2008). This paper
distils both the study of survey outcomes and
the subsequent discussions to consider whether
archaeological geophysics is doing its job, at least
in this region, and how it might be better used.
The effectiveness of archaeological
Concerned archaeologists in the north west of
England refer to key cases, at Welsh Row,
Nantwich for example (Connelly and Power,
2005), where geophysics did not find the substantial remains that were later proved in
excavation. Discussions with archaeologists elsewhere in England, and the wide variations in the
use of geophysics nationally, suggest that these
doubts are widespread. Yet the use of geophysics
in English archaeology is well established.
Indeed statistics in the AIP Database (Archaeological Investigations Project, 2007) and the
experience of those involved (A. David, personal
communication) show that it is increasing both in
frequency and area of surveys, which suggests
that it has earned its place despite these doubts.
On the other hand, although English curatorial
archaeologists may need to defend their choices
of survey methods before a legal enquiry as
reasonable, enforceable and fair, the large
differences in the choices they make, under
similar circumstances, suggests that this has
done little to consolidate a common view of what
methods should be chosen and why. This allows
wide variations in practice, among the effects of
which is to undermine the confidence of
observers in the decisions archaeologists make
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
about the methods they use. Thus a manager of
two gravel quarries in adjacent counties may be
required to pay for very different methods of
archaeological evaluation in each – one with and
one without geophysics, for example – because of
the different preferences of the county curators.
The review began by creating a list (Jordan,
2007) of all the geophysical surveys carried out
for archaeology in the northwest of England
between 1990 and 2006, and then analysing a
number in greater detail to assess and evaluate
the extent to which they proved effective. Survey
and excavation reports were obtained from all
the principal commercial archaeologists and
geophysical surveyors and from public archives,
including Historic Environment Records. Information on the method, instrument, date and
the state of the ground were obtained where
possible. Topographic, geological and soil maps,
digital elevation models and aerial photographs
were obtained for each site and land-use
identified. Survey and excavation data were then
compared to see how well features found in
excavation had been detected by geophysics.
Interviews were held with almost all of those
involved, who were asked about their experience
of the effectiveness of geophysical survey in the
region, and of the way it is applied.
Only 87 of the 188 geophysical surveys
identified had been previously recorded in
national databases. Thus the project greatly
improved our record of work in the region and
would be worth repeating elsewhere for this
purpose alone. Thirty-five sites were selected for
more detailed study, 17 of these because there
were good, comparable survey and excavation
records and a further 18 because they illustrated
particular issues of wider significance.
The use of geophysical survey
in the region
The northwest of England contains a wide range
of upland and lowland landscapes with varied
geology, hydrology and soils. The region is
dominated by two groups of hills, the Lake
District and the Pennines, which reach heights up
to 1000 m. Much of the south, and the far north,
are lowland and low hills. Precipitation over the
Copyright # 2009 John Wiley & Sons, Ltd.
hills is high, averaging around 2000 mm per year,
but the lowlands receive less than 1000 mm
and the climate is generally mild. The geology is
varied. Limestone and sandstone underlie much
of the region, but there are large areas of coal
measure rocks, saline and calcareous mudstones
in the south and a mix of metamorphic and
igneous rocks of widely varying chemistry and
physical properties in the north. Over most of the
region the bedrock is covered by Devensian till
and outwash deposits, the complex and varied
composition of which reflects the wide variety
of its source rocks and the complex history of
glacial reworking (Taylor, 1971; BGS, 1977, 2007;
Aitkenhead, 2002).
The soils are likewise diverse with large
areas of wet, acid upland heath and peat but
with extensive lowland gleys and well-drained,
productive soils on the upland fringe and across
the lowlands (SSEW, 1984a,b).
This vast diversity of environments imposes
very variable background geophysical properties
that contrast with those of southern England.
Set within these diverse landscapes are a wide
range of archaeological remains, from ephemeral
prehistoric ditches to stone-built Roman forts and
Medieval abbeys.
Geophysical surveys have been relatively
sparse in the northwest compared with the
rest of England (Table 1, Figures 1 and 2; Darvill
and Fulton, 1995; Archaeological Investigations
Project, 2007). The new records show that the
majority have been carried out by magnetic
fluxgate gradiometry (67%) at a reading spacing
of 1 0.5 m. At the time that the records were
gathered (2006) only two surveys had been
carried out using a caesium vapour magnetometer, one of which was a small, experimental
survey, atypical of commercial practice. The
proportional area surveyed by magnetometer
is even greater (86%) because of a relatively small
number of very large unrecorded magnetometer
‘scan’ surveys (7% by numbers of surveys
but 43% by area surveyed). The average standard
(recorded) magnetometer survey covered 3.1 ha,
whereas the average magnetometer scan
covers 23.8 ha. In practice, unrecorded ‘scan’
surveys have been accompanied by smaller –
often very much smaller – areas of recorded
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
D. Jordan
Figure 2. The distribution of geophysical surveys in the northwest of England recorded by this project.
Figure 1. The distribution ofallgeophysicalsurveyssince 2000
(Archaeological Investigations Project data for England only).
Although incomplete, this map indicates the relative density of
recent surveys.
Electrical resistivity surveys have been common (27%) and much smaller, averaging only
1.3 ha. Almost all have been carried out using the
Geoscan RM15 resistance meter and a twinelectrode array at a fixed electrode spacing of
0.5 m and a reading spacing of 1 1 m. Finally
there had been only five ground-penetrating
radar (GPR) and two magnetic susceptibility
surveys up to 2006.
Thus the cheapest and fastest techniques, the
most widely available instruments and most
standard field methods dominate, as we might
expect. The proportion of resistivity surveys may
Copyright # 2009 John Wiley & Sons, Ltd.
be a little higher than elsewhere in England
because the northwest has benefited from more
research and conservation surveys over stonebuilt monuments, especially along Hadrian’s
Wall, where resistivity survey might be favoured.
The proportion of GPR and magnetic susceptibility surveys may be a little lower, perhaps
reflecting some conservatism in project design.
Survey success and survey design
The ‘success’ of survey is difficult to define and a
sensitive issue for commercial surveyors, whose
livelihoods depend on the degree to which they
are thought competent. Almost all of the surveys
examined for this project produced interpretable
images (although not necessarily of archaeological remains) and appear to have been
performed with professional care. Only a few
had noticeable striping, zig-zagging or other
artefacts resulting from poor field technique.
Where remains existed but were not detected by
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
the geophysical survey, excavation subsequently
showed that there was some property of the site
that made it unlikely to have been detectable
given the technique, instrument and reading
spacing used. In most cases buried remains were
not detected either because the anomalies they
generated were too weak, were detectable over
too small an area at the ground surface or were
in some way obscured by material above or
close by. Thus failure to detect remains appears
to be largely the result of the nature of the sites
themselves and the choice of methods, sample
density, instruments and area covered, not of the
surveyor’s field or processing skills.
The project found that the range of methods,
instruments and spacings being used is narrow,
implying that surveys are not being optimised for
the conditions of the site or the aims of the
archaeologist, a situation with which many feel
uncomfortable. Surveyors’ skills are compromised by commercial pressures, which means that
they often have little time and background
information with which to design surveys. There
is rarely the opportunity, in particular, to
combine knowledge of site conditions with a
description of the remains an archaeologist needs
to be able to detect in order to model explicitly the
methods, instruments and reading spacings this
will require. Surveyors said that, all too often,
those who commission surveys expect them to
apply a standard approach prescribed by common practice and published guidelines, even
where these might not be optimal. Likewise the
date when survey takes place tends to be set
by the requirements of the client, which may,
for electrical resistance and GPR survey, be
Some of this conservatism in survey design is
self-imposed. Several surveyors said that do not
wish to challenge survey briefs they think
inadequate because they fear that the work
would simply be given to a competitor. This
might suggest that there is a vicious circle in
which curators (principally the County Archaeologists and their teams) do not adequately
challenge the designs that surveyors and consultants produce in response to the archaeological need the curators themselves define.
Assessments of survey ‘success’ are complicated by the fact that survey results, in creating
Copyright # 2009 John Wiley & Sons, Ltd.
new archaeological knowledge about areas
where we may initially know very little, change
the questions that might be asked. Thus a
single cycle of prospective survey might often
be usefully followed by a second cycle which
clarifies detail. Few of the surveys in the northwest were repeated in this way, although some
had been extended, suggesting once again that
geophysics is being used without sufficient
design or resources, and as a formula to be
applied rather than as an analytical tool.
Geophysical survey is often part of a package
of data from many sources and its effectiveness
must therefore be considered in terms of the
efficiency with which it adds new information.
Even where we can be confident that geophysical
survey will produce clear images of buried
remains there will be cases where other techniques will be just as effective. Thus, for example,
both magnetometry and aerial photography are
effective at finding new sites on the gravel soils
around Oxford, in central southern England, as at
Radley Barrows (English Heritage, 1983, 1985;
Barclay and Halpin, 1997). Since it is much
cheaper to trace crop-marks from aerial photograph archives it is reasonable for archaeologists
to ask why geophysics is needed. In practice
geophysical surveys in the northwest are often
supported by very little other data and such
cheaper options are not so often available.
Archaeologists said that, as a result, it is crucial
that geophysical survey is effective in the northwest because there are large areas where the only
alternative to geophysics, in looking for buried
remains, is excavation. The abundance of pasture, in particular, often makes aerial photography and field-walking ineffective. Moreover one
of the County Council Curators said that the
relative economic weakness of some parts of the
northwest can make the cost of evaluation
excavation unacceptable, and thus that geophysics is often the only survey method available.
Thus the confidence that users have in geophysical survey results is crucial to the protection of
remains because there may be no practical
alternative. Surveyors and curators confided
that geophysics is sometimes commissioned
not because they think it will allow them to
successfully evaluate a site, but simply because
nothing else is economically possible.
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
One might therefore expect that the lack of
alternatives would mean that geophysical surveys would be both more extensive and better
resourced than elsewhere in England, taking
place in stages with gradual refinements, extensions and repetitions at higher density and with
different instruments. This does not, however,
seem to be the case.
D. Jordan
exposures showed a high topsoil magnetic
susceptibility, decreasing down the soil profile.
Fluxgate magnetometer survey was carried out at
1 0.25 m. The clear and comprehensive survey
images produced show the fine structure of the
site as strong magnetic anomalies and form an
excellent basis for future conservation.
Unsuccessful or partly successful surveys
Case histories
To learn more about the reasons why specific
surveys produced good, or poor, results the
study looked at some surveys in further detail.
Those summarized here illustrate key influences
on survey outcomes seen repeatedly during the
Typical, successful surveys that accurately
identified the presence or absence of remains
Middlewich Eastern Bypass
This site lies in undulating pasture on poorly
drained soils formed in till over saline marl.
Excavation demonstrated that magnetometer
survey had comprehensively located ditches
beneath 25–30 cm of plough-soil on the top of
an interfluve.
Ulgill, Cumbria
This site is on wet, valley-side pasture,
overlying till derived from Coal Measures.
Caesium magnetometry at 1 0.25 m and twinelectrode resistivity survey at 1 0.5 m identified
nothing archaeologically significant. Excavation
confirmed agricultural and natural soil anomalies found by the geophysics and also showed
that there were no archaeological structures that
the geophysics had missed. Other surveys of
road routes in the region, along sections of the
A66 for example, likewise found large areas
apparently without archaeological remains,
the absence of which was then confirmed in
Maryport Roman fort and Vicus
This gently sloping grassland site overlies till
formed principally from the underlying iron-rich
St Bees Sandstone. Measurements on nearby
Copyright # 2009 John Wiley & Sons, Ltd.
Barker House Farm
This site lies in pasture, on poorly drained
lowland gley soils formed in stony till that
produces a background magnetic field with
many, very localized natural dipole anomalies.
Low magnetic susceptibility enhancement in the
archaeological strata coincided with a variable
natural background field to obscure small, but
archaeologically important remains. Thus the
fluxgate magnetometer survey failed to locate
a narrow ring-ditch known from previous
excavation but did find a second, broader ring
ditch to the southeast, confirmed in further
excavation. Simple, approximate modelling
shows that at the standard reading spacing of
1 0.5 m, approximately 280 magnetic gradient
readings will have been recorded above the
broader enclosure ditch and only about 36 within
any likely narrow ring ditch anomaly – nearly
five times fewer per linear metre of ditch (4.5 m 1
as against 0.96 m 1). This illustrates how the
density of survey determines, in any given
geophysical environment, not just whether
remains will be detected but what kind of
Buckleys Field, Middlewich
This site lies in pasture on wet soils derived
from mudstones. Fluxgate magnetometry was
carried out at 1 0.25 m, and resistivity survey
using a 0.5 m twin-electrode array at 1 1 m
intervals. Neither survey identified the abundant
remains found in excavation, but, instead, were
dominated by large anomalies due to recent
debris in the surface soil and to metal structures
around the site.
Welsh Row, Nantwich
This site lies in an area of flat grassland, on
poorly-drained soils derived from mudstone till.
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
Unrecorded magnetic ‘scanning’ was followed
by 6 ha of recorded fluxgate magnetometry
at 1 0.5 m intervals. Nothing archaeologically
significant was found. Excavation, however,
revealed abundant remains, some very large,
especially of Roman salt-making industry. This
was, however, buried beneath at least 50 cm of
topsoil and alluvium. The largest remains,
moreover, were wooden brine tanks containing
deposits very similar to the surrounding alluvia.
There was probably, therefore, little magnetic
susceptibility contrast between the tanks and the
soil to each side, and a larger contrast with
the soil above, which is likely to have been of
higher susceptibility. The result would therefore
be a weak anomaly due to the tanks – perhaps
detectable in itself but very difficult to distinguish within the more varying magnetic field
background produced by variations within the
topsoil above.
New Cowper Farm, Cumbria
Fluxgate and caesium magnetometry, at reading densities up to 0.5 0.25 m, were carried out
on a quartz-sand ridge, and revealed few of the
pits and ditches later found in excavation.
Analysis showed that the relatively high and
variable magnetic susceptibility of the deep
topsoils masked the much weaker contrasts
between the archaeological remains and the
quartz-dominated sand parent material in which
they lay.
Manchester Airport
Fluxgate magnetometer surveys, in pasture
overlying till and Holocene alluvia, found
nothing definitely archaeological. Anomalies in
one area, however, that might have had an
archaeological origin, were excavated and
found to be natural former stream courses,
potentially valuable sources of palaeoenvironmental evidence.
Survey outcomes
Of the 35 sites for which detailed records were
gathered, 24 produced results that fully met the
need for which they were commissioned. Among
these were surveys of major, protected sites that
Copyright # 2009 John Wiley & Sons, Ltd.
produced clear, comprehensive geophysical
images on which conservation and management
decisions could be based. The remaining 11 either
identified remains, or the absence of remains,
which were subsequently confirmed in excavation. Six of these 11 surveys produced equivocal
results. Five of these six missed some archaeologically important remains, whereas one identified apparently archaeological remains that
excavation showed to be shallow geological
structures. The remaining five of these 11 surveys
were demonstrably unsuccessful. They failed to
locate extensive archaeological remains later
proved in excavation despite the surveys having
been carried out to standards and with methods
in line with currently accepted practice.
Thus a high proportion of surveys achieve
their aims. The few outright failures can be
attributed to a limited range of circumstances.
Further analysis suggests that these often can be
identified in advance and thus might be avoided.
There is one particularly significant finding.
Some surveys identified large areas where
nothing, or nearly nothing, archaeological could
be found. In almost all cases excavation showed
that this correctly reflected an absence or a
very low density of remains. For planners,
identifying where development will not impinge
on archaeology is often a key goal. It is widely
assumed that geophysics cannot be used for this
– providing reliable evidence only for the presence
of remains. Thus, in contrast to Hey and Lacey
(2001), this study suggests that archaeologists
may be able to put more weight on negative
evidence in this region than has been thought.
Hey and Lacy (2001) argue that geophysical
survey should be verified by excavation yet the
need for this must depend on the reliability of
geophysics as a means to achieve a particular
purpose. It is clearly unwise to rely on any one
survey approach, where alternatives are available, particularly where commercial pressure
to reduce costs encourages archaeologists to cut
corners. Yet where, as is frequently the case, the
primary purpose of survey is to identify if there is
a risk of archaeological remains being present,
the possibility that geophysics alone may provide
reliable evidence for the absence of remains is
significant, even if this is only true in specific
circumstances. Although a combination of data
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
D. Jordan
sources is clearly best – and necessary for critical
planning decisions – there are many occasions,
especially early in the planning process, where
this is not feasible and thus where the reliability
of geophysics alone is crucial. Archaeologists
need not restrict their choices artificially to
geophysics or to some alternative, yet where
this restriction is forced upon them by circumstances, such as soil type, land-use or economics,
their confidence in geophysics becomes key.
The crucial factor in the reliability of negative
evidence on the sites studied here is the absence
of reasons why geophysics will not work at a
particular site, and thus for results to be
unreliable. This project identified a limited range
of ‘risk factors’ due to which surveys had not
found remains and which usually can be
predicted by a little preliminary map and field
study and avoided by choosing an appropriate
survey approach, which may not include geophysics. It might be expected that the reliability
of such negative evidence will vary considerably
with the regional geomorphology, parent
material types and density of past and present
occupation. The sample of surveys gathered by
this project is too small to draw firm conclusions
and it would be very interesting to conduct a
wider study to find out if this is more widely true.
The results show that geophysical surveys in the
northwest of England largely fulfil their purpose,
despite the complex geophysical environment.
Important remains have been missed in only a
few cases (five out of the 35 reviewed). This may
give us confidence that geophysics is also
proving effective in less complex environments
across much of the rest of Britain. The study
found that most failures result from a combination of a very few, well-understood circumstances. Thus survey might usefully be preceded
by a simple map and field study to identify if
these circumstances are likely to be present.
To summarize, geophysical survey only failed
to meet the archaeologists needs where:
(i) Remains lay too deeply buried to be
distinguished – more than 50 cm at Welsh
Copyright # 2009 John Wiley & Sons, Ltd.
Row, Nantwich and at New Cowper Farm,
Cumbria for example. More deeply buried
remains can be detected geophysically but
the weaker and more dispersed anomalies
detected are increasingly obscured by other
variations in the soil around them.
The remains are overlain by, or embedded
in, soil containing natural or man-made
debris with a strong geophysical contrast to
its surroundings, such as brick fragments or
igneous stones in finer till.
The geophysical anomalies due to remains
are masked by anomalies caused by
nearby surface structures such as (for
magnetometry) metal-clad buildings, fences
and vehicles.
The remains and natural soil do not contrast
geophysically. This appears rare. Despite the
wide range of geologies, mineralogies and
soil types in which surveys have taken place
in the northwest, only on the clean quartz
sands of the New Cowper ridge and in the
large brine-tanks of Welsh Row, Nantwich
was magnetic susceptibility contrast too
small to allow remains to be found and,
even here, alternative methods proved, or
might prove, effective.
The remains are too small, in spatial extent,
to be detected, given the spacing of geophysical readings and the sensitivity of the
instrument used.
In practice, as the above examples show,
remains are often obscured by combinations of
these effects. Thus, although the more deeply
buried remains at Welsh Row and New Cowper
Farm might be distinguished in a geophysically
quiet environment, the presence of stronger
anomalies, forming complex distributions in
the soil above, can make the pattern they form
indistinguishable and thus uninterpretable as
archaeological remains.
Discussions with surveyors suggest ways in
which we might reduce the risks of such failures,
and improve survey practice, with a few simple
Most of the risks above could be anticipated.
Those sites found in this study to be too deep to
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
be detected by standard survey methods either
lay within the flood range of a river, which had
buried them in alluvium, or in coarse, easily
eroded soils, where they were masked by
colluvium or deep cultivation. Sites with such
risks can be identified from existing maps
including the new BGS Parent Material map
(Lawley and Smith, 2008) and brief preliminary
fieldwork, given a few geomorphological and
soil criteria and a simple check-list of risk
factors. Likewise a little research on site history,
using historical and modern maps, would have
identified those sites in this study where remains
were obscured by surface debris or nearby
Sites that contain natural debris (such
as igneous cobbles), which creates a disturbed
geophysical background, are more difficult to
identify in advance from maps alone. Most of
those in northwest England are likely to lie on
gravels or till (as at Barker House Farm) with a
proportion of larger stones that have a raised
magnetic susceptibility and electrical resistivity.
Areas of gravel are restricted and may be
identified from existing geological and aggregate
maps. Stony till, however, is common and
undistinguished on British geological maps. Such
sites can, however, be identified by an initial
walk-over survey, digging a few spade holes and
filling in a standard check-list on the way. Such
preliminary surveys have much else to recommend them, in identifying the wide range of
factors that affect how a survey might best be
designed. It is not clear, however, who should
perform them. It might be best if they were
performed by an experienced geophysical surveyor employed, by the project consultant, to
design an appropriate brief. On the other hand,
such an additional step in the process of survey
commissioning might prove too costly or cause
too much delay to be acceptable in the context of
British commercial archaeology.
Survey briefs stipulating ‘standard’ methods are
too limiting. A magnetometer reading spacing
of 1 0.25 m is now recommended (English
Heritage, 2008), which is a considerable improvement over the reading spacing of 1 0.5 m
Copyright # 2009 John Wiley & Sons, Ltd.
commonly used in surveys in the northwest
before 2006. This was adequate, for example,
for detecting typical 1-m-wide ditches against
simple geophysical backgrounds but inadequate
where remains are small, anomalies weak
and the background complex. The question of
survey orientation in relation to the orientation
of remains has become more critical as the new
standard for narrower spacings of readings
along the survey traverse direction means that
survey sampling patterns are yet more strongly
orientated. Some of these remains are likely to be
archaeologically important but undetectable at
current sampling densities. Thus it might be wise
to abandon the use of ‘standard’ survey reading
spacings and, instead, for survey briefs to state
explicitly what scale of structures surveyors are
expected to be able to find and thus the spacing
this implies in a particular soil environment, with
a given level of background variation against
which remains must be resolved. Where we have
prior knowledge of the archaeology at a site this
should already have been made explicit since it is
current practice, in England, that all relevant
information on known remains should be
included in the survey brief (ACAO, 1993). Thus
the required survey design could be based on
what more the survey needs to tell us. Where we
do not have prior knowledge of remains (as in
most cases of commercial site evaluation) survey
designs could be based on local research
requirements. In England these could be drafted
from regional research frameworks which provide an agreed list of research priorities. Detecting sufficiently small remains to meet a particular
research need then requires that the surveyor
responds by setting an appropriate survey
density and uses a sufficiently sensitive survey
method and instrument identified through an
additional stage of numerical forward anomaly
modelling. There is no need for this to be
complex, lengthy or costly if it is made routine
– although its precision is bound to be limited by
prior knowledge of the site. Indeed, the wider use
of modelling has much to recommend it, at the
design and interpretative stages, particularly if
the testing of anomaly interpretations against
forward models were, as a common part of the
reporting process, able to show how interpretations had been arrived at. It remains unclear,
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
D. Jordan
however, how the geophysical expertise of the
surveyor can be used in this way to design briefs
that are then issued by the curator, who normally
lacks the expertise required. Although modelling
is bound to be imprecise it may be of particular
benefit if it provides a means by which geophysicists and archaeologists can interact, helping
archaeologists to see their sites in geophysics
terms – as the geophysicist must – by making
explicit how archaeological stratigraphy is ‘translated’ into distributions of geophysical properties. This might avoid both a misunderstanding
of the limits of geophysics and an insufficient
appreciation of its potential as a tool for the
analysis of remains, not just their bare depiction
as grey shades on a plan.
The narrow range of survey methods and
instruments used in northwest England makes it
difficult to determine how much alternatives
might have contributed to better survey outcomes. Such a calculation is complex. A recent
paper (Linford et al., 2007) indicated that the
ability of caesium magnetometers to resolve
smaller variations in magnetic field increases
the range of environments in which magnetometry can be successful and the range of remains
detected. Moreover the total field measurements
made by such instruments offer considerable
advantages in understanding the origins of
the anomalies detected. The advantages of
resolution might be even greater with the new
generation of extremely sensitive instruments
such as portable SQUID magnetic gradiometers
(Schultze et al., 2008). Although this high degree
of resolution will give benefits in some areas of
the northwest of England, the abundance of
sites formed in magnetically variable substrates
(such as stony till) suggests that the limiting
factor in detection will often be the spatial and
amplitude scales of the magnetic ‘noise’ this
generates, in comparison with the density of
measurements taken. The difference between an
instrument able to resolve 0.001 nT and one
only capable of 0.1 nT is clearly less significant
where background variations are one or two
orders of magnitude higher at the spatial scale of
normal survey sampling. Increasing the density
of survey, to better depict the distribution of
soil ‘noise’ (which is actually unassimilated
information), may be fruitful but will not always
Copyright # 2009 John Wiley & Sons, Ltd.
help to distinguish the archaeological anomalies
with which such ‘noise’ is melded. These
limitations may be more widely true. The
spatially complex geology, varied mineralogy
and dense pattern of glacial remains impose a
much more varied geophysical background on
much of the northwest than on the southeast of
England. Thus the same issues of redundancy in
instrument sensitivity may well apply to other
geophysical methods and might, likewise, need
to be overcome by greater sampling density
rather than more sensitive instruments. This is
not, however, to say that the greater geophysical
resolution of newer instruments is not valuable
in geophysically quieter environments where
it can be used effectively. Despite its demonstrable benefits, the time-costs of increasing
sample density are considerable, especially if
this requires the surveyor to cover more ground
by reducing the spacing between survey traverses. Thus one of the most interesting challenges facing archaeological geophysicists is to
design survey strategies that optimise the
recovery of archaeological anomalies by varying
the spatial density of readings. This is analogous
to the optimization schemes that can, for
example, be applied to electrical resistance
tomography data gathering (Stumme et al.,
2004; Wilkinson et al., 2006), where an initial
sample of measurements is used to determine
what further measurements will add most
information. The analysis required in archaeological survey is rather different, however,
because the time-costs of covering ground more
than once are high. Therefore, one of the future
roles of rapid, low-resolution survey may be to
map the distribution of geophysical property
variance in order to provide a basis for adaptive
but detailed survey designs.
Many surveyors complained that they are not
routinely given reports on excavations that
followed their surveys, and therefore they do
not have the opportunity to find out how well
their survey performed, or how it might have
been better carried out. It seems clear that if
surveys are to be evaluated and improved the
best people to do it are surveyors. Obviously it
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
therefore would be worthwhile for a summary of
excavation results to be sent to surveyors and
that surveys be routinely evaluated against
remains found in excavation as a required part
of archiving the excavation records. This feedback is particularly important where further
survey in adjacent areas might be designed and
interpreted with reference to this previous
comparison of datasets.
Most other professions accumulate evidence
about what methods to apply in different
circumstances – which is why we trust surgeons
and airline pilots – using formal methods that are
integrated into routine practice and feed back,
directly, into that practice. Replication and self
analysis is, after all, a defining characteristic of
science. By contrast, very little archaeological
fieldwork, geophysics included, gathers objective
evidence of its own effectiveness. Equally
surprising, perhaps, is that geophysical surveyors themselves rarely include in their surveys, or describe in their publications, a phase
of measurement replication at different scales
and with different controls, in order to gather
evidence from which the characteristics of the
data itself can be assessed. Thus at present
most surveys give us no comparable means to
measure data quality and analyse the sources of
variability. Where archaeologists have tried to
evaluate methods they have usually done so as a
special exercise and the result, therefore, tends to
involve few cases – as with the study published
by Hey and Lacy (2001). Yet the great diversity of
archaeological remains, and the environments
in which we find them, requires that we must
formally analyse very many cases in order to
extract trends. This suggests that we need to do
as other professions and make the analysis of the
effectiveness of our methods, and the sources of
variance within our data, a routine part of our
practice. This, in turn, requires that we adopt
survey standards that include a formal stage of
replication and analysis.
Mapping the background
Another good reason to make analysis a routine
part of practice is that we have a lot more to
learn about the archaeological and geophysical
landscape. Some of the soil-physical mechanisms
Copyright # 2009 John Wiley & Sons, Ltd.
behind archaeological geophysical anomalies
are well understood but our knowledge of the
distribution of the soil properties that determine
these physical behaviours across the landscape is
usually too coarse to confidently predict the best
geophysical survey approach at any particular
site. In looking to improve both the spatial
scale and physical specificity of the soil and
geological mapping this requires, archaeologists
hold a key advantage; there is perhaps no other
profession that gets to see so many sites, and so
much soil, at such close quarters. Soil scientists
and geologists are interested in objective observations that archaeologists might make about the
soil horizons and rocks they see because this may
allow them to improve their maps and analyses,
and to do so on sites and in areas to which they
have no access. Moreover geologists are increasingly recognizing the importance of Quaternary
geological features that archaeologists often
reveal in excavation. Archaeological geophysical
survey data itself has a value beyond archaeology, which gives archaeological surveyors a
further resource to exchange. On the one hand,
geologists and soil scientists themselves need to
understand soil geophysical properties, as do
archaeological surveyors, because this is of much
wider relevance to their clients – in the electricity
distribution industry, for example. On the
other hand much archaeological survey data
contains evidence of soil property distributions
of an extent and at spatial resolutions that soil
surveyors are rarely able to achieve. Thus there
are good reasons why other earth scientists
want archaeological survey data. There is
the basis for a symbiotic relationship in which
archaeologists provide soil scientists and geologists with all the data they can (simply but
correctly gathered during routine operations).
These earth scientists share the maps and
underlying interpretations they produce, with
this help, to guide the planning and interpretation of archaeological studies – geophysics
and aerial photography in particular. The
creation of the British National Parent Material
map, now in progress (Lawley and Smith,
2008), could form a good focus on which such
symbiosis might develop in Britain and form a
template for such mapping collaboration in other
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
In Britain at least there is little research in
progress to improve our key archaeological
survey methods – geophysics, aerial photography and fieldwalking. Stagnant funding
reflects neither the needs, nor potential, of
professional geophysical survey. Although the
growing output of commercial surveys and the
success of journals such as Archaeological Prospection give the impression of feverish activity,
most commercial reports remain unpublished
and many publications of surveys are descriptive
rather than self-analytical and thus do less than
they might to advance the discipline. This is a
pity because the abundance of surveys in Britain
means that we could learn a lot, and learn
quickly, if even modest research funding became
The wider picture
In the absence of similar reviews it is not possible
to say how relevant this study is beyond the
northwest of England. Many of the planning
issues apply nationwide, indeed Europe-wide
and worldwide under international conventions.
The diversity both of the archaeological remains
and of the natural environment in the northwest
suggest that the practical lessons are widely
applicable. In particular the five factors limiting
survey effectiveness are common worldwide,
although their frequency and degree of influence
will vary with local geology, soils, archaeological
remains and planning context. Thus there are
reasons to think that much of the foregoing
discussion is relevant more widely. There can
hardly be a region or a type of monument
anywhere in the world where archaeological
geophysical survey has not been successful, yet
the difference in the degree to which it is routinely
used, and thus its impact on professional
archaeology, in different countries is striking.
Some of this is the understandable product of
differences in survey effectiveness and practicality in different geologies and landscapes.
Magnetometry is less able to resolve small
anomalies in ‘noisy’ Scottish gravel soils full of
igneous cobbles as on the A74 survey (Banks,
Copyright # 2009 John Wiley & Sons, Ltd.
D. Jordan
2006) than in the uniform loess plains of
Germany, as at Xanten (Scollar, 1971) and other
central European sites (such as examples in
Neubauer, 2001), for example. It is less practical
to survey in jungle than in an English pasture.
Needs likewise vary. In regions such as
some parts of Spain – Jaen, for example (Zafra
de la Torre, 2006), archaeological remains and
recent development largely coincide in
town centres. Here those geophysical techniques
dominant in Britain, which excel in physically
uncomplicated rural sites, are less useful, producing ambiguous images in circumstances where
clarity is particularly necessary. By contrast, in
countries going through periods of rapid infrastructure development – such as motorways in
Ireland or gas pipeline networks in Britain to take
two recent examples – there is a period during
which rapid evaluation survey by geophysics
over large areas can adopt a central role in
circumstances to which it is well suited. Yet the
planning and commercial context of archaeological control is clearly vital in determining where
geophysics is actually applied. In France and
Switzerland, for example, the rapid development
of the motorway network over recent decades
has not been accompanied by the widespread use
of geophysics, at least in part because it has been
argued there that the total excavation of such
routes makes geophysics redundant. In Switzerland geophysics may now find a new role as, with
the completion of the motorway network, the
focus of professional services swings back
towards enhancing the wider archaeological
record and focuses less on specific, high-impact
A further constraint may be a lack of surveyors
and, in less developed regions, of equipment.
This may be compounded by an associated lack
of university departments providing training
and support. In Britain this is rarely an issue
because equipment is available and geophysics is
familiar to almost all professional archaeologists,
through university training, personal experience
and television programmes. Elsewhere, despite
the abundance of geophysical expertise in
universities and industry, archaeologists have
much less personal experience of its use and
access to the equipment and skills is more
difficult – circumstances that are self-perpetuat-
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
How Effective is Geophysical Survey?
Table 1. Summary of geophysical surveys in the northwest
of England to 2006 ^ types of surveys and areas covered
(values rounded to whole numbers of hectares)
Magnetic scanning
Magnetic susceptibility
Ground-penetrating radar
Number of
surveyed (ha)
ing. Thus, if archaeological geophysics becomes
more widely taught in university and equipment
is made more widely available, the growing
familiarity of archaeologists with its potential,
worldwide, may lead to its wider use.
Yet, as in the northwest of England, there
remain significant doubts as to the value of
geophysics even where archaeological services
are sophisticated and well funded, in Britain and
beyond. What lessons may we therefore draw
from this study of relevance to archaeological
geophysics elsewhere? The first, perhaps, is that
these widespread doubts about the value of
geophysics may originate, at least in part, from
the same lack of appraisal – both of geophysical
survey itself and of its performance in comparison with alternatives – in other countries as well
as in Britain. An objective appraisal of past
geophysical surveys in France, for example, and
the great depth of relevant geophysical expertise,
breadth of research and innovation in related
fields among French geophysicists, might persuade an independent observer that geophysics
deserves a much larger role in French archaeology. Might it be, therefore, that the wider use of
geophysics will benefit from a similar appraisal
of its potential, extending that undertaken by this
study? Might not sceptical archaeologists be
more willing to make selective use of geophysics
if surveyors could assess previous surveys to
identify where it will and will not produce
what archaeologists need – and do so, at least
semi-quantitatively and with relative costs, in
comparison to alternatives such as excavation,
fieldwalking and aerial photography?
Beyond this, the same constraints on the
objective use of geophysics that this study
identifies in the northwest of England appear
Copyright # 2009 John Wiley & Sons, Ltd.
to be widespread. Thus the same benefits may
accrue from the wider use of design based on
forward models, of replication to assess sources
of variance in data and of a formal evaluation
of the effectiveness of survey and the causes
of anomalies. Close collaboration between geophysicists and geoarchaeologists is therefore
This project was funded by the Aggregates Levy
under the management of English Heritage
which is gratefully acknowledged. Particular
thanks are due to the English Heritage geophysics and geoarchaeology staff and the many
Curators, Consultants, Survey Contractors and
Developers who gave very generously of their
time: Matt Canti, Neil Linford, Andrew David,
David Batchelor, Sue Stallibrass, Jim Williams,
Ingrid Ward, Jennie Stopford, Sarah Cole, Tony
Wilmott, Mark Leah, Jill Collens, Sarah-Jane Farr,
Mark Adams, Ron Cowell, Norman Redhead,
Mike Nevell, Rob Iles, Peter McCrone, Jeremy
Parsons, Mike Morris, David Mason, Richard
Newmann, Jeremy Parsons, Jo Mackintosh, Jamie
Quartermaine, Alan Lupton, Emily Mercer, Karl
Taylor, Brian Grimsditch, Phillip Day, Mark
Noel, Anne and Martin Roseveare, Peter Barker,
Alister Webb, Alan Biggins, Ian Brooks, Russell
Colman, Peter Stephenson, Nick Edwards, Bob
Bewley, Alister Bartlett, Frank Harkness, Frank
Geicco, Tim Malim, Ben Stephenson, Mark
Fletcher, Chris Gaffney, John Gater, John Dearing, Frances Tattum, Patricia Crompton, Armin
Schmidt, Charles Parry, Chris Jones, Roy Canham, David Knight and Simon Buteux. Apologies
to any that have been missed.
Aitkenhead N. 2002. The Pennines and Adjacent Areas.
Regional Geology Guide 8, British: Geological
Survey Keyworth.
Archaeological Investigations Project. 2007. http://
ACAO. 1993. Model Briefs and Specifications for
Archaeological Assessments and Field Evaluations. Association of County Archaeological Officers.
Banks I. 2006. Going over old ground. Perspectives
on archaeological geophysical and geochemical
survey in Scotland. Proceedings of a conference
held at the Department of Archaeology, University of Glasgow, Scotland, August 2003, Jones RE,
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
Sharpe L (eds). British Archaeological Report 416:
Barclay A, Halpin C. 1997. Excavations at Barrow
Hills, Radley, Oxfordshire. Volume 1. The Neolithic
and Bronze Age Monument Complex. Oxford
University Committee for Archaeology.
BGS. 1977. Bedrock Geology of the UK: North.
British Geological Survey Map at 1:625000 scale.
BGS. 2007. Quaternary Geology of the UK: North.
British Geological Survey Map at 1:625000 scale.
Connelly P, Power D. 2005. Salt making in Roman
Nantwich. Recent discoveries at Kingsley Fields,
Welsh Row. In Brine in Britannia: Recent Archaeological Work on the Roman Salt Industry in Cheshire,
Nevell M, Fielding AP (eds). Archaeology North
West 7 (17); 31–40.
Darvill T, Fulton A. 1995. The Monuments at Risk
Survey of England 1995. English Heritage Geophysical Survey Database.
English Heritage. 1983. Unpublished Geophysical
Survey Report 822, Radley Barrows.
English Heritage. 1985. Unpublished Geophysical
Survey Report 823, Radley Barrows.
English Heritage. 2008. Geophysical Survey in
Archaeological Field Evaluation. English Heritage.
Hey G, Lacy M. 2001. Evaluation of Archaeological
Decision Making Processes and Sampling Strategies. Oxford Archaeology Unit.
Jordan D. 2007. Evaluating Aggregate in North West
England: the Effectiveness of Geophysical Survey. Report on English Heritage ALSF project
number 3835.
Lawley R, Smith B. 2008. Digital soil mapping at a
national scale: a knowledge and GIS based
approach to improving parent material and property information. In Digital Soil Mapping with
Limited Data, Hartemink AE, McBratney A,
Maria de Lourdes Mendonça-Santos M (eds).
Springer; 173–182.
Copyright # 2009 John Wiley & Sons, Ltd.
D. Jordan
Linford N, Linford P, Martin L, Payne A. 2007.
Recent results from the English Heritage caesium
magnetometer system in comparison with recent
fluxgate gradiometers. Archaeological Prospection
14(3): 151–166.
Neubauer W. 2001. Magnetische Prospektion in der
Archäologie. Verlag der Österreichischen Akademie der Wissenschaften.
Scollar I. 1971. A magnetometer survey of the Colonia Ulpia Trajana near Xanten, West Germany.
Prospezioni Archeologiche 6: 83–92.
Schultze V, Linzen S, Schüler T, Chwala A, Stolz R,
Schulz M, Meyer H-G. 2008. Rapid and sensitive
magnetometer surveys of large areas using
SQUIDs – the measurement system and its application to the Niederzimmern Neolithic doublering ditch exploration. Archaeological Prospection
15(2): 113–131.
SSEW. 1984a. Soils and their Use in Northern England. Bulletin 10, Soil Survey of England and
SSEW. 1984b. Soils and their Use in Midland and
Western England. Bulletin 12, Soil Survey of England and Wales.
Stumme P, Maurer H, Green AG. 2004. Experimental design: electrical resistivity data sets that provide optimum subsurface information. Geophysics
69: 120–139.
Taylor BJ. 2001. Northern England. Regional Geology
Guide 7. British Geological Survey: Keyworth.
Wilkinson PB, Meldrum PI, Chambers JE, Kuras O,
Ogilvy RD. 2006. Improved strategies for the
automatic selection of optimised sets of electrical
resistivity tomography measurement configurations. Geophysical Journal International 167:
Zafra de la Torre N. 2006. De los campamentos
nómadas a las aldeas campesinas. La provincia
de Jaén en la Prehistoria. Jaén: Universidad de
Archaeol. Prospect. 16, 77–90 (2009)
DOI: 10.1002/arp
Без категории
Размер файла
189 Кб
survey, regional, review, effective, geophysics
Пожаловаться на содержимое документа