-
Catena 78 (2009) 234–249
Contents lists available at ScienceDirect
Catena
j ourna l homepage: www.e lsev ie r.com/ locate /catena
Linking cultural and environmental change in Peruvian
prehistory:Geomorphological survey of the Samaca Basin, Lower Ica
Valley, Peru
David Beresford-Jones a,⁎, Helen Lewis b,1, Steve Boreham
c,2
a Department of Archaeology, University of Cambridge, Downing
Street, Cambridge, CB2 3DZ, UKb School of Archaeology, University
College Dublin, Newman Building, UCD Belfield, Dublin 4, Irelandc
Department of Geography, University of Cambridge, Downing Place,
Cambridge CB2 3EN, UK
⁎ Corresponding author. Fax: +44 1223 333503.E-mail addresses:
[email protected] (D. Beresford-Jo
(H. Lewis), [email protected] (S. Boreham).1 Fax: +353 1 716
1184.2 Fax: +44 1223 333392.
0341-8162/$ – see front matter © 2008 Elsevier B.V.
Adoi:10.1016/j.catena.2008.12.010
a b s t r a c t
a r t i c l e i n f o
Article history:
The lower Ica Valley on th
Received 25 September 2007Received in revised form 23 October
2008Accepted 19 December 2008
Keywords:South coast
PeruDesertPalaeoenvironmentArchaeologyRiparian
dry-forestProsopis
e hyperarid south coast of Peru is today largely depopulated and
bereft ofcultivation. Yet its extensive archaeological remains
attest to substantial pre-Hispanic populations. Weprovide a
case-study of Pre-Hispanic culturally induced environmental change
through combining fieldarchaeological and geomorphological survey
with archaeobotanical, sedimentary and soil
micromorphologyapproaches. Our investigations reveal that, although
major El Niño climatic perturbations occurring aroundthe end of the
Early Intermediate Period are part of the explanation, more
gradual, human-induced reductionin riparian dry-forest vegetation
also lie behind major landscape change, which culminated during
theMiddle Horizon.
© 2008 Elsevier B.V. All rights reserved.
1. Introduction and area of study
The lower Ica Valley on the south coast of Peru is today
largelydepopulated and bereft of cultivation. Yet its extensive
archaeologicalremains attest to substantial pre-Hispanic
populations and present aprima facie case for changed ecological
and landscape conditions.This paper describes geomorphological
survey conducted in theSamaca Basin, as part of an archaeological
investigation exploringthis landscape change.
The Peruvian south coast is one of the oldest and driest deserts
inthe world. Its topography is typical of ‘basin-range’ deserts,
char-acterised by enclosed drainage systems (Cooke et al., 1993).
Samaca isone of several basins constituting the course of the lower
Ica River, cutinto the Tablazo de Ica sedimentary rock plateau
(Fig. 1). It is a well-defined and convenient landscape unit within
which to consider keygeomorphological agents affecting a lowland
valley system and inparticular human–environment interactions
(French, 2003). Theclimate is hyperarid with an average annual
precipitation of only0.3 mm (ONERN, 1971). In this environment
there are only two
nes), [email protected]
ll rights reserved.
natural geomorphological agents acting, in opposite directions,
withinthe Samaca Basin: the wind and the river.
The Río Ica exhibits a different configuration to the broad
fan-shaped delta complexes of most major Peruvian
westward-flowingcoastal rivers. Blocked by the uplifted formations
of the Tablazo de Ica,it is diverted south, parallel to the coast,
for 150 km. Alluvial sedimentsare deposited in a series of wide
basins that alternate with narrowsand canyons. Life and human
settlement since at least the EarlyHorizon (Cook, 1999, Fig. 2)
have been restricted to these riparianoases. The hydrology is
erratic and seasonal, arising from summerrainfall in the distant
Andean highlands; flow lasts for around threemonths per year. Total
surface flow since 1922 has oscillated widelyabout a mean of 257
million m3 per annumwith a standard deviationof over 150, but with
annual flows of over 800 million m3, related toperturbations in the
Southern Oscillation Index or El Niño (Beresford-Jones, 2005), and
many years with practically no flow (SENAMHI-Ica,1997, 2002).
The wind regime of the lower Ica Valley is extraordinarily
strongand unimodal, from the south (Parker Gay, 1999).
Mega-yardangshundreds of metres high etched upon the surface of the
Tablazo de Icatestify to the stability of this wind regime over
great time depths(McCauley et al., 1977; Beresford-Jones, 2005).
The average monthlywind velocities measured continuously by Davis
cabled weatherstation over seven months, varied between 32.3 (for
October 2004)and 27.1 km/h (for December 2004). Maximum gusts
recorded variedbetween 115 and 82.1 km/h (Personal communication —
Ing. AlbertoBenavides).
mailto:[email protected]:[email protected]:[email protected]://dx.doi.org/10.1016/j.catena.2008.12.010http://www.sciencedirect.com/science/journal/03418162
-
Fig. 1. Lower Ica Valley (Landsat 7 ETM+ 2000).
235D. Beresford-Jones et al. / Catena 78 (2009) 234–249
Detailed geomorphological survey concentrated upon a c. 50
haarea of the upper Samaca Basin, known as ‘H-13’ (Cook, 1994),
whichexhibits evidence of environmental degradation and diminished
land
Fig. 2. Simplified Ica Valley Pr
use over time. The purpose of the survey was to describe
andcharacterise sediment and soil sequences to gain insight into
thedifferential processes of deposition and erosion though alluvial
andaeolian action in the basin over time. Understanding these
processes isrequisite to any attempt to reconstruct human impact on
vegetationand land use change, interpret extant archaeological and
landscapefeatures identified by archaeological survey, and identify
prospectiveareas for archaeological excavation.
This article describes the methods used in carrying out
thegeomorphological survey and its results and interpretations,
whichinclude the definition of the basic geomorphic units across
theupper Samaca Basin, the nature of the relict river terrace
underlyingH-13, the extent of landscape and ecological change
across H-13 asrevealed by buried soils, and evaluation of contexts
from whicharchaeological and archaeobotanical data were extracted
and subse-quently analysed.
2. Methods
Geomorphological study was carried out along a
systematictransect of 31 survey pits located 25 m (occasionally 50
m) apart,proceeding due west between the sedimentary base rock
edges of theSamaca Basin (Fig. 3). Field descriptions were made of
the exposedsoil and sediment profiles following Hodgson (1974) and
usingMunsell (2000).
Bulk samples were collected from selected horizons for a series
ofanalyses in theUniversity of Cambridge Physical Geography
Laboratory(Beresford-Jones, 2005): 1) magnetic susceptibility (‘Mag
Sus’),following Bennett et al. (1990); 2) Loss-on-ignition (‘LOI’)
followingBengtsson and Enell (1986) at 105, 550 and 950 °C, a
widely usedmethod to estimate moisture, organic matter and
carbonate contentsrespectively because it is inexpensive, rapid and
involves no hazardouschemicals (Dearing, 1986; Konen et al., 2002);
3) electrical conductiv-ity (‘EC’); and 4) grain size analysis
using a Malvern Mastersizer 2000laser particle sizer.
Problems and limitations with LOI analysis have been
reported(Heiri et al., 2001). In addition to organic carbon for
exampleincreasingly higher temperatures up to 550 °C can drive off
structuralwater from clays and other inorganic constituents,
decompose somecarbonates and hydrated salts and oxidise Fe2+
(Schulte and Hopkins,1996). However, sediments from the hyperarid
Samaca Basin contextare generally clay-poor and calcareous, for
which LOI produces results
e-Columbian Chronology.
-
Fig. 3. Samaca Basin geomorphological survey transect and
profile.
236 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
-
237D. Beresford-Jones et al. / Catena 78 (2009) 234–249
with accuracy comparable to more sophisticated geochemical
meth-ods (Dean, 1974). The meanings and limitations of the LOI
deriveddata are discussed in its interpretation below.
Grain size analysis sampleswere run twiceusing twodifferent
lenses(45 and 1000 µm) to get a full range of grain sizes
(2000–0.02 µm).10 gsampleswere treatedwith 40ml of 4.4%
sodiumpyrophosphate at 90 °Cfor 3 h to aid disaggregationprior to
analysis. Treated sampleswere thencentrifuged at 3500 rpm for 13
min, decanted and re-mobilised beforeintroduction to the laser
particle sizer.
Targeted sampling for soil thin section analysis was also
under-taken; thin sections were produced in the McBurney
GeoarchaeologyLaboratory, University of Cambridge (after Murphy,
1986), andanalysed (following Bullock et al., 1985) in the Oxford
UniversityResearch Laboratory for Archaeology and the History of
Art using aNikon Optiphot-2 polarising microscope.
3. Results
3.1. Basic geomorphic units
The survey defined an asymmetrical distribution of four
basicgeomorphic units from east to west across the upper Samaca
Basin asfollows:
1. A short footslope covering 100m of the transect (Survey Pits
1/001to 1/005), with a sequence consisting of: 1) a colluvial
surface ofpoorly sorted coarse to medium sands and small rocks (to
5 cm)gradually diminishing westwards, overlying; 2) shallow, very
wellsorted, massive aeolian fine sand deposits with occasional
charcoallenses, overlying; 3) well sorted, massive, hard silty clay
loamlayers with varying degrees of calcium carbonate concretion
andiron staining.
2. The current Río Ica floodplain covering 300 m of the
transect(Survey Pits 1/006 to 1/012), and the only part today with
riparianflora, in largely degraded associations. The floodplain
consists of:newly stabilised dunes; massive, laminated silty clay
crusts offluvial and alluvial deposition (known locally as yapana);
incisedchannel walls at the river thalweg; and overbank
depositscomposed of interbedded fine layers of medium to fine sand
alongnumerous channel bedswithwell-rounded rocks. The
groundwatertable in the floodplain was encountered at c. 209 m
asl.
Fig. 4. Deflated duracrete surface of terrace H-13 w
The highly disturbed deposits across the floodplain are
indicative ofrecent high-energy flood events, including that of the
1997/98 ElNiño, which left overbankflood deposits up to 213.3masl.
Typical ofdryland compound rivers with highly erratic discharges,
the Río Icais laterally unstable and experiences a rapid transition
frommeandering to braided channel flow during these events
(Schumm,2003). During braided flow, river channels fill the whole
availablespace between low terraces (Graf, 2002). ‘Floodplain’ here
thusincludes the entire area influenced by lateral instability of
the riverduring these high-energy events. The complex first-order
topo-graphy of the floodplain has come about through vertical
accretionwith overbank flow depositing fine-grained materials in
horizontalbeds, with lateral accretion in cross-bedded forms as
coarsematerials are deposited in lag deposits from exceptional
flows.After flooding, as the river gradually returns to meandering
flow,the many minor channels of the braided system become filled
withfine sediments. These processes are highly influenced by
theoccurrence and nature of riparian floodplain vegetation,
particu-larly of phreatophytes such as Prosopis, the keystone
arborealspecies in the area (Beresford-Jones et al., in press).
Unsurprisingly,archaeological remains are scarce within the
floodplain area.
3. A relict river terrace standing around 5 m higher than the
floodplainand 10 m higher than the current Río Ica thalweg, with a
width ofabout 550m along the transect (Survey Pits no. 1/013 to
1/029). Thesurface of this terrace contains frequent archaeological
featuresdefined as area H-13, discussed in detail below. In general
its sedi-ments comprise: even, parallel beds of well consolidated,
blocky siltyclay loams with numerous ferric features and calcite
nodules;structureless, well sorted, fine sandy loams with few
calciumcarbonate concretions; and massive, well sorted, very fine
sandyloams. In certain horizons calcrete cementation attains very
hard,petrocalcic desert duracretes. Theseduracretes have
aprofoundeffecton the land surfaceofH-13 as discussed below. The
high, unvegetatedrelict terrace is exposed to the full force of the
very strong, unimodalwind regime of the region. The effects of
aeolian deflation of theH-13landscape are evident in frequent
palimpsests of multiperiod surfacescatters of ceramic sherds, and
complex second-order landformscreated by a classic inversion of
relief, whereby features such ascanals, which were once cut into
the land surface, are preserved byduracrete enrichment as
upstanding features above the deflatedsurrounding landscape
(Maizels, 1988; Cooke et al., 1993; Fig. 4).
ith relict canal Feature No. 40 in foreground.
-
Table 1Section 1/013 field descriptions.
Depth (cm) Colour Munsell colour Description
0–42 Very pale brown 10YR 7/3 Silty clay; massive to crumb;
hard, consolidated; ferric mottles (2–5 mm) common42–48 Brown 7.5
YR 4/3 Silt loam; sub-angular blocky (1–2 cm), moderately well
developed, hard; numerous ferric mottles; CaCO3 nodules48–53 Light
olive brown 2.5Y 5/3 Fine sandy loam; structureless; moderately
well sorted; CaCO3 concretions (2–4 mm)53–65 Light yellowish brown
2.5Y 6/3 Fine sandy loam; massive; moderately well sorted65–73
Brown 7.5YR 4/3 Silt loam; crumb to blocky; hard, consolidated;
ferric mottling; CaCO3 concretions (1–3 mm); roots73–75 Light olive
brown 2.5Y 5/3 Very fine sandy loam; structureless75–83 Light olive
brown 2.5Y 5/3 Fine sandy silt; massive; moderately well
sorted83–90 Brown 7.5YR 4/4 Silt loam; crumb to blocky, poorly
developed; CaCO3 concretions; same as 65–73 cm90–110 Light
yellowish brown 2.5Y 6/3 Same as 53–65 cm110–125 Brown 7.5YR 4/4
Same as 83–90 cm125–130 Light olive brown 2.5Y 5/3 Fine sandy loam;
structureless130–145 Brown 7.5YR 4/4 Silt loam; massive; well
consolidated; CaCO3 concretions; roots145–155 Brown 7.5YR 4/4 Silt
loam; sub-angular blocky, moderately well developed; CaCO3
concretions; rootlets to large roots155–165 Brown 7.5YR 4/4 Same as
130–145 cm165–182 Brown 7.5YR 4/4 Loam; blocky to crumb, weakly
developed; poorly sorted; CaCO3 concretions common; small
sub-angular gravels; ferric mottling; roots.182–205 Light olive
brown 2.5Y 5/3 Very fine sandy silt loam; massive205–226 Light
olive brown 2.5Y 5/3 Medium sandy silt; massive226–240 Light olive
brown 2.5Y 5/3 Very fine sandy silt; massive240–245 White 2.5Y 8/1
White calcrete horizon; hard245–255 Brown 7.5YR 4/4 Sandy loam;
massive, moderately well consolidated255–265 Brown 7.5YR 4/4 Silt
loam; massive; hard265+ Light brownish grey 10YR 6/2 Medium to
coarse sand; structureless
238 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
4. The western end of the transect (Survey Pits, 1/030 and
1/031) is alarge climbing dune, anchored on the sedimentary rock
slope onthe western flank of the basin. This part of the transect
isparticularly important because climbing dune deposits seal
buriedland surfaces, elsewhere lost across the surface of the H-13
terracebecause of wind deflation. These surfaces are discussed in
moredetail below. In general the profiles here comprise:
poorlyconsolidated coarse to fine sand in graded bedding, at the
base ofwhich are frequent cultural materials; overlying massive,
moder-ately well consolidated, moderately sorted layers of fine
sandy siltloam, with evidence of fining upward sequences, i.e. fine
sand atthe bottom grading upward to silt, with frequent ferric
staining andfew cultural materials; overlying hard, consolidated,
brown siltyclay with well developed blocky structure (1–5 cm).
These basic geomorphologic units are illustrated in Fig. 3,
whichshows the profile of the entire transect, as well as the
probable ancientland surface level of H-13 determined from the
height of the preservedrelict canal (Feature 40) and other
preserved archaeological features, asdiscussed below. Also shown
are the April 2002 groundwater levelduring survey, themaximum level
of the 1997/98 El Niño event, and thelevel of a huge flood event
that deposited fluvial layer Stratigraphic Unit19 (‘S.U. 19’),
recorded during excavation (Beresford-Jones, 2005).
Table 2Section 1/013 analyses results.
Layers Loss on ignition (by mass) Mag Sus(LF Sus)
EC(mS/cm)
Grain
Depths(cm)
% Loss at105 °C
% Loss at550 °C
% Loss at950 °C
% Silicateresidue
% Siltclay (
0–42 5.0 4.3 3.1 91.2 88 7.33 9342–48 8.4 4.8 3.4 90.2 290 11.34
8948–53 1.5 1.3 10.6 83.4 210 2.67 4453–65 2.1 1.8 3.8 92.7 345
3.41 5265–73 5.6 3.4 2.8 92.6 459 7.78 7173–75 1.0 1.5 1.8 96.0 15
2.37 4775–83 3.2 2.4 2.6 93.8 345 4.7 6583–90 5.7 4.2 4.0 90.0 382
7.45 81
Notes:(1) grain size b63 µm.
(2) grain size N63 µm.
(3) grain size 63–2 µm.
(4) grain size N2 µm.
3.2. Section 1/013 — relict terrace H-13
A steep 5 m embankment incised by the river at some point in
thepast marks the edge of the H-13 relict terrace. Some 3 m of this
profilewas cleaned, described and subsequently analysed in detail
as Section1/013 (Table 1). The top of the embankment and sectionwas
surveyedat 218.79 m asl (Fig. 3). This profile represents the
history of theterrace prior to the deposition of archaeological
materials dating fromthe Early through to the Middle Horizon
scattered upon the surface ofthe H-13 relict terrace.
The 1/013 section is characterised by distinctive surface and
subsur-face horizons in various partly repeated sequences, and fits
descriptionsof aridisols according to USDA nomenclature (Cooke et
al., 1993).
But episodicity in alluvial and aeolian sedimentation may
producebedding that closely resemble pedogenic horizons, and as
Cooke et al.(1993: 51) note “processes of sedimentation and
pedogenesis arefrequently superimposed”. Key features include:
sub-surface enrich-ment from inputs of fine aeolian dust;
differential solubility leading todifferential distribution of
salts; clay illuviation; and processesassociated with fluctuations
in ground water levels (Cooke et al.1993: 51). All of these
processes are interrelated and may beassociated with both natural
and cultural activities. Flooding forexample can be related to
human manipulation of watercourses;
size analysis summary
and1)
% Sand(2)
% Silt(3)
% Clay(4)
Sortingcoefficient
Mediandiameter(µm)
Meandiameter(µm)
Inclusivegraphicskewness
8 80 12 1.97 12.16 43.2 0.14111 60 28 2.50 5.70 37.8 −0.12956 35
9 2.57 82.86 107.4 0.55048 43 9 2.40 57.63 70.7 0.51329 49 23 2.85
9.50 64.6 −0.10453 38 8 2.61 75.36 124.8 0.49835 54 11 2.41 37.83
54.4 0.41119 57 24 2.61 7.88 31.3 −0.085
-
Fig. 5. Grain size profiles for Section 1/013 profile.
Fig. 6. Variations with depth of sele
239D. Beresford-Jones et al. / Catena 78 (2009) 234–249
changes in vegetation cover can alter aeolian deposition, as
well asleading to carbonate mobilisation in the soil (Goudie,
1973). Togetherthey give rise to complex profile sequences whose
underlying causalprocesses are difficult to disentangle.
Secondary calcium carbonate enrichment is a common feature
ofarid zone soils (Cooke et al., 1993), and is the defining
characteristic ofthe H-13 relict terrace. Very tough enriched CaCO3
duracrete horizonsare known as calcretes or calíche on the south
coast of Peru. Mostcalcrete horizons in arid area soils are
pedogenic; their formation anddevelopment is influenced by several
factors including texture of thehost material, partial pressure of
CO2 in the soil air, rate of supply ofcalcium, amount of
infiltrating water, profile age, and the climatichistory of the
site (Cooke et al. 1993: 56–60, Goudie, 1973, 1983).
Estimates of the age of duracretes vary widely. Most are thought
tobe very ancient: from 10,000 BP to as far back as the Miocene
(Cookeet al., 1993). However duracretes formed from floodwaters
maydevelop quite quickly (Machette, 1985), and the preponderance
ofcalcrete deposited to over 3mdepth observed in the base of relict
canalfeature No. 40 during excavation (Beresford-Jones et al., in
press),suggests that at least some of the duracrete of terrace H-13
may relateto anthropogenic floodwater manipulation during the
archaeologicaltime periods. The Samaca climate is hyperarid and has
likely persistedfor millions of years (Petrov, 1976; Alpers and
Brimhall, 1988; Cookeet al., 1993; Böhlke et al., 1997; McKay et
al., 2003). Whilst climatechange in the immediate vicinity does not
likely explain duracreteformation during the time period discussed
here, any long-termvariations in seasonal rainfall on the distant
Andean foothills(Eitel et al., 2005; Mächtle et al., 2006) would
have affected ground-water levels. In Samaca, as in many deserts of
the USA (Machette,1985; Schlesinger, 1985), the primary source of
calcium carbonateenrichment is likely to be dust input fromwinds
blowing for millenniafrom the Pacific Ocean, over the highly
calcareous sedimentaryformations of the Tablazo de Ica.
Data from laboratory analyses (Table 2), field descriptions
(Table1), and grain size profiles (Fig. 5), show a series of
discernible episodic
cted Section 1/013 profile data.
-
240 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
sequences down the 1/013 profile (B–B′ on Fig. 6), below its
upper-most stratigraphic unit (0–42 cm). Although data for only the
upperpart of the profile have been determined, the observed
sequencesappear in field description to repeat to depth throughout
of the entireSection 1/013 (Table 1). They consist of:
1. Thin, brown (7.5YR 4/3) silty clay loams (42–48 cm; 65–73 cm
and83–90 cm): with sub-angular blocky structures, numerous
ferricmottling and calcite nodules; with relatively high LOI
readings at550 and 105 °C and relatively high electrical
conductivity andmagnetic susceptibility levels. They have bimodal
grain size profileswith negative inclusive graphic skewness towards
silts and clays.
Fig. 7. Samaca Basin, curren
2. Underlying, relatively thin, light olive brown (2.5Y 5/3),
structure-less sandy loams (48–53 cm and 73–75 cm); with very low
LOIreadings at 550 and 105 °C and low electrical conductivity
andmagnetic susceptibility levels. They have grain size profiles
withpositive inclusive graphic skewness towardsmoderatelywell
sortedfine sands.
3. Underlying, relatively thick light yellowish brown (2.5Y 6/3)
orlight olive brown (2.5Y 5/3), massive sandy loams (53–65 cm
and75–83 cm); with intermediate LOI readings at 550 and 105 °C
andelectrical conductivity and magnetic susceptibility levels.
Theyhave grain size profiles that are still well sorted but have
lesspositive inclusive graphic skewness towards coarser
materials.
t floodplain sequence.
-
241D. Beresford-Jones et al. / Catena 78 (2009) 234–249
We interpret these sequences as representing the history of
theH-13 terrace when it was part of the active river floodplain in
thedistant past. Episodic seasonal flood deposition of alluvial
silts andclays (1 above and B′ in Fig. 6) was followed by
subsequent dry periodswith aeolian deposition of well-sorted fine
sands in varying energies(2 and 3 above andB in Fig. 6) as seasonal
riparian vegetation graduallybecame established upon damp alluvium
(Fig. 7).
LOI at 105 °C represents a measure of pore water held within
thesample and is used here as a crude proxy for moisture content.
Sharpdiscontinuities in moisture readings down the profile
emphasise thecurrent standstill condition of the H-13 terrace
system in this hyper-arid environment. The existence of thin, dry
layers (e.g. 73–75 cm)that are more than 50% sand with large
amounts of pore space butwith only about 1% moisture would seem to
eliminate the likelihoodof ongoing hydrological continuity or
capillary action from the basinwater table up the profile.
LOI at 550 °C, following Dean (1974), is here taken as a
relativelyaccurate measure of organic carbon matter content in the
particularcontext of the lower Ica Valley. The measure does not
distinguishbetween elemental carbon and other organic matter,
althoughcharcoals were not noted in 1/013 contexts, either in field
descriptionsor archaeological flotation using 500 µm mesh. These
proxy organicmattermeasures are very low for these contexts, as
would be expectedin sediments of largely clastic originwith only
incipient soil formationin an arid environment. Organic matter
content is significantly higherin putative alluvial deposits (B′ in
Fig. 6) at around 4.5%. Indeed, thesevalues are fairly typical for
alluvial silts, whose fertility when annuallyreplenished, has of
course little to do with their organic contents.
Fig. 8. Prosopis calcium carbonate pseud
When LOI at 550 °C across all the 1/013 data is plotted as a
functionof grain size for each grain size class, whilst a strong
positivecorrelation is noted for finer grained sediments (e.g.
R2=0.775 forvery fine silt), with increasing grain size this
correlation steadilybecomes weak (e.g. R2=0.024 for very coarse
silt), and then morestrongly negative (e.g. R2=−0.899 for fine
sands). This observationtogether with those above may indicate that
organic matter in thesecontexts is mainly represented by finely
divided organic matterbrought in with alluvium.
Magnetic susceptibility readings show wide fluctuation down
the1/013 profile from a high of 383 in silt layers of proposed
flooddeposition to a low of only 15 (S.I. units x 10−8 m3 kg−1) in
putativeaeolian deposit layers (Fig. 6). However it is not
correlatedsignificantly with % silt and clay (R2=0.031). This may
be becausevery highmagnetic susceptibilities observed in putative
flood deposits(e.g. 42–48 cm) are the result of their post-flood
exposure, prior togradual burial by aeolian deposits and
establishment of vegetation.
LOI at 950 °C in this context is taken as a fairly accurate
measure ofcalcium carbonate content since there are no magnesium
carbonatecontaining rocks in the Río Ica catchment (ONERN, 1971).
Althoughcalcite contents generally vary in accordance with the
observedepisodic sequence there are notable exceptions (e.g. the
high contentof layer 48–53 cm and the white petrocalcic K horizon
of almost purecalcrete at 240–245 cm). Calcite enrichment is thus
largely theproduct of processes of secondary enrichment (Goudie,
1973;Raghavan et al., 1991) and suggests much higher groundwater
levelsfrom those persisting today some 10 m beneath the
standstillcondition terrace.
omorphs in H-13 duracrete layers.
-
242 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
Although EC is strongly positively correlated with percentage
siltsand clays (R2=0.804), calcite contents are not
(R2=−0.153).Calcrete formation here therefore does not appear to
conformstraightforwardly to the model proposed by Gile (1966), Gile
et al.(1997), in which laminations form as water through-flow
ischannelled over a plugged, fine textured hardpan. However
anotherfactor in H-13 calcite enrichment may be previously extant
riparian
Fig. 9. Context and details
vegetation. Dominant calcite rhizoliths and Prosopis leaf
litterpseudomorphs are observed in calcrete horizons across the
barrensurface of H-13 (Fig. 8), and nodules within in and below the
proposedalluvial layers of Section 1/013 (Table 1). These may
indicate drawingof groundwater up-profile by roots (Goldberg and
Macphail, 2006) orchanges in CO2 partial pressures in the
rhizosphere (Bhojvaid andTimmer, 1998; Garg, 1999; Mishra and
Sharma, 2003).
of Survey Pit 1/030.
-
243D. Beresford-Jones et al. / Catena 78 (2009) 234–249
Moreover, the uppermost part of 1/013 stratigraphy shows anumber
of striking contrasts to the underlying sequence thus fardiscussed.
This thick (0–42 cm), upper unit of pale brown, massivesilty clay
with frequent ferric mottles (Table 1) exhibits significantlyless
sorting and a symmetrical grain size profile about a median,medium
silt, quite distinct from underlying layers (Table 2 and Fig. 5).In
view of the relict canal system across the surface of H-13 it
istempting to interpret this distinction as the result of the
imposition ofhuman control, agriculture and pedogenesis upon the
energies of theepisodic clastic sediment regime of the underlying
river floodplain.
Although of considerable thickness, this is interpreted as a
truncatedlayer since it is apparent from the extant form and
distribution ofarchaeological evidence, exemplified by relict canal
fragment Feature 40(Fig. 4), that the relict H-13 terrace surface
has undergone severedeflation over time. The characteristics of
canal Feature 40 establishedthrough excavation and survey
(Beresford-Jones, 2005), including itsdirection parallel to the
river, substantial dimensions, long duration ofuse, and unlined
channel form, all fit Moseley et al.'s (1983: 316) Class 1canal: a
principal supply canal, supporting a sub-critical flow regime,with
a gentle down-course gradient. Based upon survey of this
canalfragment (average extant surface of 220.06 m asl; maximum
surfacealtitude of 220.88masl) the original land surface of theH-13
terrace canbe estimated to have lain at around 220.5 m asl. Survey
of the deflatedflanks of the canal (average 218.55m asl, and
coincident with the top ofSection 1/013), implies some 2 m of
deflation since its abandonment.Thus the 0–42 cm layer observed
today in Section 1/013 is likelytruncated by some 1.5–2 m of wind
deflation, and represents only thebottom part of the original
surface profile of the terrace when canal 40was in operation.
The amount of entrainment of grains by wind and thus of
aeoliandeflation experienced in a particular environment is a
complexfunction of wind speed; surface conditions including
vegetationcover; and grain size and cohesion, which in turn depend
uponproperties such as moisture and salt enrichment (Fryberger and
Dean,1979; Maizels, 1988; Cooke et al., 1993). But since the date
by whichthe H-13 relict terrace was completely abandoned by
vegetation canbe estimated through various lines of archaeological
data, discussedbelow, to around the end of the Middle Horizon we
can estimate theaverage rate at which the stable, unimodal Samaca
wind regime isdeflating the terrace. This is useful because, as
Cooke et al. observe,‘there are few reliable measurements of rates
of transport in the field’(1993: 263). Assuming a deflation depth
as determined above, thetotal volume of material removed from the
surface of the H-13terrace, much of it deposited in the climbing
dune anchored on thewestern edge of the basin, can be estimated by
GIS to total some377,000 m3. The flank of the terrace perpendicular
to the unimodalSamaca wind direction is around 1100 m in length,
which implies anaverage deflation rate of around 0.31 m3 (m
width)−1 yr−1; or, byaround 2 mm depth per year.
3.3. Survey Pit 1/030 — buried land surface and buried soil
In this deflated landscape only the western end of the
transectsealed beneath the climbing dune seemed to offer
possibilities of
Table 3Survey Pit 1/030 field descriptions.
Depths (cm) Colour Munsell colour Description
0–33 Light brownish grey 10YR 6/2 Fine to coarse sand with
graded be33–46 Dark greyish brown 10YR 4/2 Sandy loam; massive;
moderately w
marine mollusc shell frequent. At b46–60 Brown 10YR 4/3 Silt
loam; massive; well consolidate60–124 Brown 10YR 4/3 Sandy silt
loam; well developed blo
moved down fissures; charcoal (1 m124+ Greyish brown 10YR 5/2
Silt loam; sub-angular blocky to cru
calcite nodules (1–5 mm), (same a
buried land surfaces. Survey Pit 1/030 was located at the very
base ofthe climbing dune at a surveyed altitude of 219.89 m asl
(Fig. 3). Thealtitude of the base of this pit was therefore
approximately coincidentwith the top of Section 1/013, some 560 m
distant across the relictterrace H-13 (Fig. 3). Survey Pit 1/030
was later expanded to becomethe first quadrant (1A) of 3×1 m Test
Pit No. 1 (Fig. 9; Table 3).
The profile for Survey Pit 1/030 was unique among all
thoserecorded on the survey transect because it appeared to
includeevidence of a buried soil horizon. In summary this profile
comprised:1) a surface layer (0–33 cm) of unconsolidated, graded
sandsconstituting the aeolian deposit of the climbing dune,
overlying2) a sequence of moderately well consolidated and
moderatelysorted sandy and silt loams (33–46 cm and 46–60 cm), over
3) abrown, well developed blocky, sandy silt loam (60–124 cm).
Thislatter element was tentatively identified in the field as a
buried soilhorizon.
Three samples for soil micromorphological analysis weretherefore
taken from the interfaces across 46–60 cm and 60–124 cm to further
refine the characterisation and interpretation ofthis profile:
Samaca 5 (46–54 cm); Samaca 2 (55–67 cm) andSamaca 4 (68–80 cm),
(Fig. 9C).
Thin sections Samaca 5 and Samaca 2 have
well-accommodatedsubangular blocky microstructure with short planar
voids and illuvial‘dusty’ clay inclusions within the thin cracks
and voids (Fig. 10). Thefine groundmass of the deposits is weakly
to moderately striated, atypical characteristic of clay-rich
sediments undergoing shrink–swellprocesses associated with periodic
wetting and drying. Iron mottlesand manganese nodules are also
associated with these cracks andchannels, and the groundmass is
stained with oxidised iron. Samaca 2also shows polyconcave voids,
suggesting soil collapse related towatersaturation. These layers
appear to be alluvial, representing low-energyflooding, possibly
with some aeolian input. Later illuviation andshrink–swell/wet–dry
processes have created many of the maincharacteristics of the
horizon. Sands do not seem to contribute muchuntil after primary
deposition.
Thin section Samaca 4 appears to confirm its identification as
anold soil horizon. Some aggregates similar to the fabrics
describedabove appear near the top of the thin section. However
themain fabricis light brown under plane polarised light, and is
more stained withamorphous organic matter than the iron-stained
fabrics of theoverlying sections (Fig. 10). Coalesced crumb peds
and microaggre-gates suggest reworking by soil fauna for some
time.
Although organic materials within the thin section
appeardepleted, they are nevertheless still a major component of
the sectionat between 15 and 20% (visual estimate of area). Some
‘dusty’ claycoatings occur, but there are relatively few clay
infillings and thegroundmass is mainly stipple specked. The sandy
loam texture ismuch more mixed than that of the layers above. The
generally moreorganic and mixed nature of this horizon suggests
that it was a soilhorizon proper at some time, which was
subsequently buried by lowenergy alluvial deposits. This sequence
was then overlain by theaeolian deposits of the climbing dune.
Because of deflation of H-13and the removal of its once extant
stratigraphy, only isolated frag-ments of this soil layer are
observed outside the area sealed by the
dding, alternating layers of well sorted grain sizes; poorly
consolidated in lower 15 cmell consolidated, moderately sorted;
charcoal (1 mm–3 cm) common;oundary with level below — very fragile
wood, (same as 1/028 4–32 cm)d; moderately sortedcky (1–5 cm); well
consolidated, hard; material from above hasm–2 cm) commonmb (1–2
cm), moderately well developed; hard, well consolidated,s 1/026
25–45 cm)
-
Fig. 10. Micrographs of Survey Pit 1/030 thin sections.
244 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
-
245D. Beresford-Jones et al. / Catena 78 (2009) 234–249
dune, most notably along parts of the eastern flank of relict
canalFeature No. 40, which are protected from the effects of
unimodal winderosion by the canal itself.
It is worth reemphasising the significance of the identification
ofalluvial deposits, overlying a once organic buried soil horizon,
in themidst of a today's barren landscape, over half a kilometre
west, andabout 5 m above, the edge of the current riparian Río Ica
floodplain(Figs. 3 and 9). As Survey Pit 1/030 was expanded to form
Test Pit 1,and its unconsolidated aeolian climbing dune deposits
were brushedaway, a remarkably preserved print of an individual's
left foot wasfound on the surface of the hard, underlying silt–loam
(33–46 cm,S.U. 5 in Test Pit 1), (Fig. 9). This footprint is
further, rather poignant,evidence that S.U. 5 represents an ancient
surface, buried and protectedby the climbing dune. Indeed, when the
individual stepped upon thissurface, it was moist.
3.4. Survey Pit 1/031 — gradual aeolian deposition
Survey Pit 1/031marked thewestern limit of the transect, 25 m
upthe ascending face of the climbing dune from 1/030 at a
surveyedaltitude of 222.43 m asl, some 2.5 m higher than the dune
surface at1/030. The base of Survey Pit 1/031 therefore lay around
0.5 m higherthan the surface at 1/030 (Fig. 3).
The upper 0–64 cm of Survey Pit 1/031 comprised light
brownishgrey, unconsolidated fine to coarse sands in graded bedding
(Table 4;Fig. 11), constituting the aeolian deposits of the
climbing dune.The rest of its profile exhibited frequent evidence
of archaeologicaldeposition including midden materials, significant
amounts ofcharcoal and an in situ hearth. In addition its sequence
includedsignificant quantities of desiccated Prosopis tree litter
(known locallyas poña).
Data derived in subsequent laboratory analyses (Table 5 andFig.
11) show a steady increase in the proportion of fine materialsdown
the profile, and no clearly discernible repeating sequences,
likethose noted in the data of Section 1/013.
Interspersed among the upper parts of the sequence is
frequentevidence of anthropogenic activity. The identification of
Nasca 2/3ceramic fragments in the 76–80 cm unit, near the top of
the sequence,dates these deposits to the Early Intermediate Period.
Layer 92–98 cmshows high readings for carbonates, electrical
conductivity andmagnetic susceptibility, reflecting its
interpretation as an in situhearth. Magnetic susceptibility
readings are particularly high indeposits with anthropogenic
input.
Organic matter readings increase down the profile and show
veryhigh levels for arid area sediments (e.g. 10.4% for layer
98–102 cm).These reflect the significant quantities of desiccated
plant macro-fossils dominated by Prosopis tree litter that increase
down the
Table 4Survey Pit 1/031 field descriptions.
Depths (cm) Colour Munsellcolour
Description
0–30 Light brownish grey 10YR 6/2 Coarse to fine sand in graded
beddin30–64 Light brownish grey 10YR 6/2 Medium sand,
unconsolidated, weak64–76 Dark yellowish brown 10YR 4/4 Sandy loam;
poorly consolidated, st76–80 Dark yellowish brown 10YR 4/4 Frequent
organic inclusions in samem80–92 Brown 10YR 4/3 Sandy silt loam;
sub-angular to sub-
organic inclusions: Prosopis sp. leave92–98 Very dark grey 7.5YR
3/1 Charcoal (to 8 mm) and ash (in-situ98–102 Black 7.5YR 2.5/1
Charcoal, ash, charred poña leaf litte102–120 Strong brown 7.5YR
4/6 50% organic matter/50% silty sand m
of massive, moderately well consoli120–140 Yellowish brown 10YR
5/4 Same as above but with 33% poña le140–190 Pale brown 10YR 6/3
Sandy silt loam with c. 10% organic
hard, sandy silt loam (similar to 1/0190+ Yellowish brown 10YR
5/4 Silty sand; well consolidated, massiv
in layers above. Capped by hard clay
profile. The excellent preservation of both plant macro fossils
andpollen extracted from Survey Pit 1/031 contexts
(Beresford-Jones,2005; Beresford-Jones et al., in press)
demonstrate that they wereabove the sequences of high pH floodwater
wetting and drying thatcharacterise the layers of 1/030, some 2.5 m
lower down on thesurface of H-13.
In a unidirectional wind regime, the western flank of the basin
andthe deposits of Survey Pit 1/031 are interpreted as the surfaces
uponwhich progressive accumulation of aeolian materials have
becomeanchored and deposited over time. The energy of this
depositionappears to have increased over time, with finally the
graded sandaeolian deposits of the climbing dune. The preponderance
of fossiltree litter observed at depths of up to 140 cm is evidence
of treesgrowing close to this location.
4. Discussion and conclusions
We beganwith the observation that first stimulated this research
–the extensive basins of the lower Ica Valley have abundant
archae-ological remains and yet today support few people and
littleproductive agriculture – thereby presenting a prima facie
case fordramatic landscape change. The geomorphological survey that
wehave described is the basis of research that seeks to
combinearchaeological survey and archaeobotanical analysis to
answer thequestions of when andwhy this change occurred. The story
it tells is asfollows.
Today, the floodplain of the Río Ica in the Samaca Basin is
narrow. Ithas a highly disturbed geomorphology –with overbank
deposits, siltyclay alluvial crusts and newly stabilised dunes of
remobilised riverbeddeposits – indicative of laterally unstable,
high-energy flood events. Atsome time in the past the river has
incised into its floodplain, exposingthe 5 m embankment described
as Section 1/013 and creating theH-13 relict terrace.
The deep-time history of H-13, recorded in this section
showsrepeated episodes of flooding – with the deposition of
alluvial silts –followed by dry periods –with increased aeolian
inputs of well-sortedfine sands. During this time the terrace was
part of the active riverfloodplain. However, the upper part of the
Section 1/013 sequencereflects a very different environment: one
under a controlled energyregime, perhaps indicative of human
influence and control over theprocess of episodic sedimentation and
of pedogenesis. This is quiteunsurprising in view of the ancient
canal systems preserved across thesurface of the H-13 terrace.
Pollen data from the 1/031 contexts and plant macrofossil
datafrom various excavated contexts across H-13 provide evidence
that inthe Early Intermediate Period the terrace sustained
productivefloodplain agriculture of maize, cotton and other
domesticated plants
g; poorly consolidatedgrading
ructureless; organic matter: poña, wood, root, Prosopis sp.
leaves, charcoal (1–4 cm)atrix as above: twigs, beans, seeds,
insect remains, wood, Nasca 2/3 phase ceramic fragmentsrounded
gravels (1 cm); poorly consolidated, poorly sorted,
structureless;s, poña leaf litter, charcoal (few), wood fragments,
ceramic fragmentshearth)r.atrix. Organic matter exclusively poña
leaf litter. Aggregates (2–20 cm)
dated silty sand; random orientationaf litter/66% matrixmatter.
Aggregates as above but also aggregates (1–8 cm) of well
consolidated,30 46–124 cm)e, moderately sorted; ferric mottling.
Material for some aggregateslayer (8–10 cm).
-
Fig. 11. Survey Pit 1/031 grain size profiles.
246 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
within a landscape still dominated by Prosopis woodland
(Beresford-Jones, 2005; Beresford-Jones et al., in press). Over 60
relict Prosopistree stumps preserved in this hyperarid climate have
been surveyedand excavated across today's barren H-13 landscape
(Fig. 3).
When did the river incise into its floodplain to expose the
H-13terrace? Various evidence link this river entrenchment event to
afluvial layer labelled S.U. 19 recorded during the excavation of
Test Pit4 on a secondary terrace just below H-13 (Fig. 3).
Preserved fragmentsof S.U. 19, up to 60 cm deep, were surveyed at
scattered points in thelandscape (Fig. 12). GIS simulation of the
flood event that depositedS.U.19 indicates that it was huge: far
greater in its spatial extent thanthe so-called ‘Super El Niño’ of
1997/8 (Bendix et al., 2002). Thisancient flood event can be dated
approximately to the end of theEarly Intermediate: firstly S.U. 19
caps directly midden contexts inTest Pit 4 that date to Nasca 2/3
(Beresford-Jones et al., in press); and
secondly, our simulation shows marked coincidence with the
extantforms of relict canal feature no. 40/34 and other Early
Intermediatearchaeological features (Fig. 12).
Once river entrenchment had occurred, productive
agriculturewasabandoned and the H-13 terrace eventually became
exposed toaeolian deflation. The presence of different degrees of
secondary cal-cium carbonates in the H-13 terrace, in part due to
human floodplainmanagement, and the subsequent exposure of these
deposits bydeflation, has created a complex second-order
landformwith invertedrelief. Cemented features such as canals
survive today standing abovethe deflated relict terrace, whilst
artefacts spanning some 2000 yearsfrom the Early through to the
Middle Horizon, have been left as mixedassemblages scattered across
it. Calcite cementation also attests tohow groundwater levels in
the basin have fallen: the result of the riverentrenchment
described, but perhaps also indicative of long-term
-
Table 5Survey Pit No. 1/031 analyses results.
Layers Loss on ignition (by mass) Mag Sus(LF Sus)
EC(mS/cm)
Grain size analysis summary
Depths(cm)
% Loss at105 °C
% Loss at550 °C
% Loss at950 °C
% SilicateResidue
% Silt andclay (1)
% Sand(2)
% Silt(3)
% Clay(4)
Sortingcoefficient
Mediandiameter(µm)
Meandiameter(µm)
Inclusivegraphicskewness
64–76 1.7 1.8 2.0 95.4 355 5.73 43 58 31 11 3.27 98 195
0.41376–80 3.9 – – – 119 7.24 39 61 31 9 3.73 806 694 0.80780–92
2.8 3.9 1.8 93.5 317 5.62 44 56 35 9 2.83 89 132 0.49092–98 4.0 8.7
8.6 78.9 312 7.43 52 48 43 9 2.69 59 104 0.37898–102 4.7 10.4 3.8
84.1 259 6.37 58 42 47 11 2.87 40 102 0.242102–120 5.6 4.8 2.7 91.3
238 7.54 63 38 47 15 3.17 18 101 −0.022120–140 6.0 4.6 1.8 92.7 119
9.45 72 29 54 18 2.85 10 59 −0.156140–190 5.3 3.1 3.2 92.2 254 7.53
78 22 62 16 2.64 10 60 −0.163190+ 2.3 1.5 0.9 97.2 106 3.62 – – – –
– – – –
Notes:(1) grain size b63 µm.
(2) grain size N63 µm.
(3) grain size 63–2 µm.
(4) grain size N2 µm.
247D. Beresford-Jones et al. / Catena 78 (2009) 234–249
rainfall changes in the distant upper Río Ica catchment (Eitel
et al.,2005; Mächtle et al., 2006).
These archaeological data thus approximately date the processes
ofgeomorphological change. Archaeological survey of the H-13
relictterrace revealed almost no materials that post-date the
MiddleHorizon, despite the presence of two large Late Intermediate
to LateHorizon habitation sites in the lower Samaca Basin (Cook,
1994;Beresford-Jones, 2005). We adduce therefore that sometime
towardsthe end of the Middle Horizon this terrace had become an
entirelydegraded landscape without any vegetation cover and
completelyexposed to the effects of the region's extraordinary wind
regime. Thedifference in heights between the indurated extant canal
surface andthe surrounding deflated landscape can therefore be used
to estimatethe ongoing rate of deflation of the H-13 terrace.
Our research thus records further evidence of the catastrophic
eventsthat so characterise Andean archaeological interpretations:El
Niñofloodsand/or La Niña droughts (viz., perturbations in the
Southern OscillationIndex, ‘ENSO’). Ice-core records from the
Quelccaya ice cap (Thompsonet al.,1985)havebeenused topostulate
thatmajor climatic perturbationscorrelate with so-called
‘punctuations’ evident in the archaeologicalrecord of the Peruvian
coast (see for example Shimada et al., 1991;Moseley, 1992;
Silverman and Proulx, 2002). However, we also haveevidence that
this punctuated equilibrium model may be simplistic.
Firstly, excavation of Middle Horizon midden contexts on
relictProsopis nabkhas, or phytogenic mounds (e.g. Mound 23, Fig.
3),indicates that although the H-13 terrace was by then exposed
toserious wind erosion, some large trees and human occupation
stillpersisted there (Beresford-Jones et al., in press). This
indicates moregradual landscape change.
Moreover, the products of that deflation have been deposited by
aunidirectional wind regime as a climbing dune, anchored against
thewestern flank of the basin. This deposition has buried a portion
of theancient landscape along that western edge and thereby
preserved it,as evidenced by Survey Pits 1/030 and 1/031. Survey
Pit 1/030provides evidence about the extent of alluvial deposition
and organicrich soil horizons once present on H-13, and thus of the
dramatic scaleof landscape and ecological change. But the
contextual evidence fromSurvey Pit 1/031, in conjunction with well
preserved plant macro-fossil and pollen data, also provides
evidence that geomorphologicalchange here followed the gradual
removal of the keystone arborealgenus of the area, Prosopis, as
well as other riparian tree species(Beresford-Jones et al., in
press).
The proximate cause of the abandonment of the H-13 relict
terracewas indeed river entrenchment resulting from the S.U. 19 El
Niño eventat the end of the Early Intermediate Period. Similarly
devastating river
entrenchments have been noted in the archaeology of the
Peruviannorth coast and explained through tectonic uplift (Moseley
et al., 1983).Yet, there is little geomorphological evidence of
significant tectonicuplift on the Peruvian coast during
archaeological time periods (Cookeet al., 1993;Wells and
Noller,1999). Rather, the phenomenon of ‘arroyo-cutting’ has been
known since the work of the geomorphologist KirkBryan in the 1920s
in the Gila River Basin of the south-western USA, andhas been
intimately linked there to the removal of riparian vegetation(Cooke
and Reeves,1976; Hackenberg,1983; Nabhan,1986; Cooke et al.,1993;
Graf, 2002).
Indeed, in arid riparian environments phreatophyte vegetation
–exemplified here by Prosopis, which has one of the deepest and
mostlaterally extensive root systems of any tree (Stone and Kalisz,
1991;Díaz Celis, 1995) – can ‘significantly alter channel geometry
byincreasing bank resistance to erosion, inducing deposition
andincreasing roughness’, as Cooke et al. (1993: 153) observe. The
deepProsopis root architecture ‘underpins’ floodplain and river
meanders,providing the edaphic conditions for other riparian
species, whichtogether make for a robust, erosion-resistant system
(Galan de Mera,1996, 1999). In the lower Ica Valley context,
humans, graduallyclearing riparian woodland for agriculture, are
the only plausibleagents of deforestation. And, this process
culminated in the MiddleHorizon, a period of great social change
across the Andes (Beresford-Jones et al., in press).
So, whilst our investigations confirm that major El Niño
eventsaround the end of the Early Intermediate Period likely offer
part of theexplanation for marked landscape change in the Samaca
Basin, wealso demonstrate the significance of more gradual,
human-induceddestruction of Prosopis-dominated riparian dry-forest.
Prosopis pallida(known as huarango on the south coast) is a
remarkable leguminoushardwood that lives for over a millennium and
provides forage, fuel,and food. Moreover, it plays a crucial role
in integrating fragile desertecosystems, enhancing soil fertility
and moisture, and accomplishingdesalination and microclimatic
amelioration (Beresford-Jones, 2005).Most obviously in an
environment that frequently experiences windsof well over 100 km/h
and extremely erratic river flows, trees withextensive root systems
physically maintain soil stability (Pasieczniket al., 2001).
Successful agriculture is simply not possible herewithoutthe
protection afforded by trees.
Thus, our point here is not to deny the significance of
chaoticfluctuations in the biophysical environment such as El Niño
floods, butto observe that their effects would only be precipitated
by ongoingprocesses of gradual, human induced change, eventually
causing themto breach critical desert geomorphic thresholds. The
limitation of thepunctuated equilibriummodel of Andean
archaeological interpretation
-
Fig. 12. S.U. 19 ancient flood event.
248 D. Beresford-Jones et al. / Catena 78 (2009) 234–249
is precisely that it relegates pre-Hispanic human culture to
ecologicalpassivity, living within an effectively static landscape,
periodicallyimpacted by catastrophic climatic events (and see
Butzer, 1996; Wellsand Noller, 1999).
Although our archaeological focus here is avowedly environmental
–both geomorphological and archaeobotanical – this does not
neces-sarily make our interpretations environmentally
deterministic. On thecontrary, it allows for more informed
discrimination between humanagency and environmental response. Many
of the changes recorded inthe extant landscape of the Samaca Basin
are the products of anthro-pogenic geomorphology.
Acknowledgements
We would like to thank Julie Boreham (McBurney
Geoarchae-ological Laboratory, University of Cambridge) for her
meticulous
preparation of thin sections; Susana Arce (Museo Regional de
Ica),Brian Pitman, Sandy Pullen, and Oliver Whaley (RBG Kew) for
theirfieldwork contributions; Fraser Sturt (Dept. of Archaeology,
Uni-versity of Southampton), Kevin Lane (Dept. of Archaeology,
Uni-versity of Manchester) and Ing. Cesar Patroni for their survey
andmapping contributions; Chris Rolfe (Dept. of Geography,
University ofCambridge) for his help with laboratory work; Charly
French,Elizabeth DeMarrais and Martin Jones (Dept. of
Archaeology,University of Cambridge) for their advice; Mark Pollard
and ChrisDoherty at Oxford University Research Laboratory for
Archaeologyand the History of Art; and most importantly the people
of theEscuela Libre de Puerto Huamaní and Don Alberto Benavides G.
forbeing the father of it all. Funding was provided by the
NaturalEnvironmental Research Council (NERC), UK, the British
Academyand the McDonald Institute of Archaeological Research,
Universityof Cambridge.
-
249D. Beresford-Jones et al. / Catena 78 (2009) 234–249
References
Alpers, C.N., Brimhall, G.H., 1988. Middle Miocene climatic
change in the AtacamaDesert, northern Chile: evidence from
supergene mineralisation at La Escondida.Geological Society of
America Bulletin 100, 1640–1656.
Beresford-Jones, D.G., 2005. Pre-Hispanic Prosopis—human
relationships on the SouthCoast of Peru: riparian forests in the
context of environmental and culturaltrajectories of the Lower Ica
Valley (http://www.arch.cam.ac.uk/dgb27/). Ph.D.Thesis, University
of Cambridge, UK.
Beresford-Jones, D.G., Arce Torres, S., Whaley, O.Q.,
Chepstow-Lusty, A.J., in press. Therole of Prosopis in ecological
and landscape change in the Samaca Basin, Lower IcaValley, South
Coast Peru from the Early Horizon to the Late Intermediate
Period.Latin American Antiquity.
Bendix, A., Bendix, J., Gämmerier, S., Reudenbach, C., Weise,
S., 2002. The El Niño 1997/98 as seen from space — rainfall
retrieval and investigation of rainfall dynamicswith GOES-8 and
TRMM data. Proceedings of the 2002 Eumetsat MeteorologicalSatellite
Conference, Dublin, Ireland, 3rd – 4th September, pp. 647–652.
Bengtsson, L., Enell, M., 1986. Chemical analysis. In: Berglund,
B. (Ed.), Handbook ofHolocene Palaeoecology and Palaeohydrology.
JohnWiley, Chichester, pp. 423–451.
Bennett, K.D., Simonson, W.D., Pelgar, S.M., 1990. Fire and man
in postglacial woodlandsof eastern England. Journal of
Archaeological Science 17, 635–642.
Bhojvaid, P.P., Timmer, V.R., 1998. Soil dynamics in an age
sequence of Prosopis julifloraplanted for sodic soil restoration in
India. Forest Ecology and Management 106,181–193.
Böhlke, J.K., Ericksen, G.E., Revesz, K., 1997. Stable isotope
evidence for an atmosphericorigin of desert nitrate deposits in
northern Chile and southern California, USA.Chemical Geologist 136,
135–152.
Bullock, P., Fedoroff, N., Jongerius, A., Stoops, G., Tursina,
T., Babel, U.,1985. Handbook forSoil Thin Section Description.
Waine Research Publications, Wolverhampton.
Butzer, K.W., 1996. Ecology in the long view: settlement
histories, agrosystemicstrategies and ecological performance.
Journal of Field Archaeology 23, 141–150.
Cook, A.G., 1994. Invesitigaciones de Reconocimiento
Arqueologico en la Parte Baja delValle de Ica. Informe Final
(1988–1990) al Instituto Nacional de Cultura del Perú, Lima.
Cook, A.G., 1999. Asentamientos Paracas en el Valle Bajo de Ica,
Perú. GacetaArqueologica Andina 25. INDEA, Lima, Peru, pp.
61–91.
Cooke, R.U., Reeves, R.W., 1976. Arroyos and Environmental
Change in the AmericanSouthwest. Oxford University Press,
Oxford.
Cooke, R.U., Warren, A., Goudie, A.S., 1993. Desert
Geomorphology. University CollegeLondon Press, London.
Dean, W.E., 1974. Determination of carbonate and organic matter
in calcareoussediments and sedimentary rocks by loss on ignition:
comparison with othermethods. Journal of Sedimentary Petrology 44,
242–248.
Dearing, J., 1986. Core correlation and total sediment influx.
In: Berglund, B. (Ed.),Handbook of Holocene Palaeoecology and
Palaeohydrology. JohnWiley, Chichester,pp. 423–451.
Díaz Celis, A., 1995. Los Algarrobos. CONCYTEC, Lima.Eitel, B.,
Hecht, S., Mächtle, B., Schukraft, G., Kadereit, A., Wagner, G.A.,
Kromer, B., Unkel,
I., Reindel, M., 2005. Geoarchaeological evidence from desert
loess in the Nazca-Palpa Region, Southern Peru: palaeoenvironmental
changes and their impact onPre-Columbian cultures. Archaeometry 47
(1), 137–158.
French, C., 2003. Geoarchaeology in Action. Routledge,
London.Fryberger, S.G., Dean, G., 1979. Dune forms and wind regime.
In: McKee, E.D. (Ed.),
A Study of Global Sand Seas. USGS Professional Paper 1052. USGS,
Washington,pp. 305–397.
Galan de Mera, A., 1996. Relación entre los suelos y la
vegetación del Perú. Arnaldoa 4,87–94.
Galan de Mera, A., 1999. La Clases Fitosociologicas De la
Vegetación del Perú. Boletín deLima 117, 84–94.
Garg, V.K., 1999. Leguminous trees for the rehabilitation of
sodic wasteland in northernIndia. Restoration Ecology 7 (3),
281–287.
Gile, L.H., 1966. Coppice dunes and the Rotura soil. Soil
Science Society of AmericaProceedings 30, 657–660.
Gile, L.H., Gibbens, R.P., Lenz, J.M., 1997. The near-ubiquitous
pedogenic world ofmesquite roots in an arid basin floor. Journal of
Arid Environments 35, 39–58.
Goldberg, P., Macphail, R.I., 2006. Practical and Theoretical
Archaeology. Blackwell,Malden.
Goudie, A.S., 1973. Duricrusts in Tropical and Sub-Tropical
Landscapes. Clarendon Press,Oxford.
Goudie, A.S., 1983. Calcrete. In: Goudie, A.S., Pye, K. (Eds.),
Chemical Sediments andGeomorphology. Academic Press, London, pp.
93–131.
Graf, W.L., 2002. Fluvial Processes in Dryland Rivers. Blackburn
Press, New Jersey.Hackenberg, R.A., 1983. Pima and Papago
ecological adaptations. In: Sturtevant, W.C.
(Ed.), Handbook of North American Indians, vol. 10. Smithsonian
Institution,Washington D.C., pp. 161–177.
Heiri, O., Lotter, A.F., Lemcke, G., 2001. Loss on ignition as a
method for estimatingorganic and carbonate content in sediments:
reproducibility and comparability ofresults. Journal of
Paleolimnology 25, 101–110.
Hodgson, J.M., 1974. Soil survey field handbook. Soil Survey
Technical MonographsNo. 5. Soil Survey, Harpenden.
Konen, M.E., Jacobs, P.M., Burras, C.L., Talaga, B.J., Mason,
J.A., 2002. Equations forpredicting soil organic carbon using
loss-on-ignition for north central US soils. SoilSociety of America
Journal 66, 1878–1881.
Machette,M.N.,1985. Calcic soils of the southwesternUnited
States. In:Weide, D.L. (Ed.),Soils and Quaternary Geology of the
Southwestern United States. Geological Societyof America Special
Paper, vol. 203. Geological Society of America, Boulder, pp.
1–21.
Mächtle, B., Eitel, B., Kadereit, A., Unkel, I., 2006. Holocene
environmental changes in thenorthern Atacama Desert, southern Peru
(14°30'S) and their impact on the rise andfall of Pre-Columbian
Cultures. Zeitschrift fuÉr Geomorphologie N.F. Supplemen-tary 142,
47–62.
Maizels, J.K., 1988. Palaochannels: Plio-Pleistocene raised
channel systems of thewestern Sharqiyah. Journal of Oman Studies
Special Report 3, 95–112.
McCauley, J.F., Breed, C.S., Grolier, M.J., 1977. Yardangs. In:
Doehring, D.O. (Ed.),Geomorphology in Arid Regions. Allen and
Unwin, Boston, pp. 233–269.
McKay, C.P., Friedmann, E.I., Gómez-Silva, B.,
Cáceres-Villanueva, L., Andersen, D.T.,Landheim, R., 2003.
Temperature and moisture conditions for life in the extremearid
region of the Atacama Desert: four years of observations including
the El Niñoof 1997–1998. Astrobiology 3 (2), 393–406.
Mishra, A., Sharma, S.D., 2003. Leguminous trees for the
restoration of degraded sodicwasteland in eastern Uttar Pradesh,
India. Land Degradation and Development 14,245–261.
Moseley, M.E.,1992. The Incas and their Ancestors, the
Archaeology of Peru. Thames andHudson, London.
Moseley, M.E., Feldman, R.A., Ortloff, C.R., Narvaez, A., 1983.
Principles of agrariancollapse in the Cordillera Negra, Peru.
Annals of Carnegie Museum 52 (13),299–327.
Munsell, 2000. Munsell Soil Colour Charts. Gretag Macbeth, New
York.Murphy, C.P., 1986. Thin Section Preparation of Soils and
Sediments. AB Academic
Publishers, Berkhamsted.Nabhan, G.P., 1986. Papago Indian desert
agriculture and water control in the Sonoran
Desert, 1697–1934. Applied Geography 6, 43–59.ONERN,1971.
Inventario, Evalacíon y Uso Regional de los Recursos Naturales de
la Costa.
Cuenca del Rio Ica, vols. 1 & 2. Oficina Nacional de
Evaluacíon de Recursos Naturales(ONERN), Lima.
Pasiecznik, N.M., Felker, P., Harris, P.J.C., Harsh, L.N., Cruz,
G., Tewari, J.C., Cadoret, K.,Maldonado, L.J., 2001. The Prosopis
juliflora– Prosopis pallida Complex: aMonograph.HDRA, Coventry.
Parker Gay Jr., S., 1999. Observations regarding the movement of
barchan sand dunes inthe Nazca to Tanaca area of Southern Peru.
Geomorphology 77, 279–293.
Petrov, M.P., 1976. Deserts of the World. John Wiley & Sons,
New York.Raghavan, R., Gaillard, C., Rajaguru, S.N., 1991. Genesis
of calcretes from the calc-pan site
of Singi Talav near Didwana, Rajasthan, India: a
micromorphological approach.Geoarchaeology 6, 151–168.
Schlesinger, W.H., 1985. The formation of caliche in soils of
the Mojave Desert,California. Geochimica et Cosmochimica Acta 49,
57–66.
Schulte, E.E., Hopkins, B.G., 1996. Estimation of organic matter
by weight loss-on-ignition. In: Magdoff, F.R., Tabatabai, M.A.,
Hanlon, E.A. (Eds.), Soil Organic Matter:Analysis and
Interpretation. Soil Science Society of America Special
Publication,vol. 46. SSSA, Madison, pp. 21–31.
Schumm, S.A., 2003. The Fluvial System. Blackburn Press, New
Jersey.SENAMHI-Ica, 1997. Boletine Meteorológico. Servicio Nacional
de Meteorología e
Hidrología, Ica.SENAMHI-Ica, 2002. Boletine Meteorológico.
Servicio Nacional de Meteorología e
Hidrología, Ica.Shimada, I., Schaaf, C.B., Thompsonm, L.G.,
Mosely-Thompson, E., 1991. Cultural impacts
of severe droughts in the prehistoric Andes: application of a
1,500-year ice coreprecipitation record. World Archaeology 22,
247–265.
Silverman, H., Proulx, D.A., 2002. The Nazca. Peoples of the
America Series. BlackwellPublishers, Oxford.
Stone, E.L., Kalisz, P.J., 1991. On the maximum extent of tree
roots. Forest Ecology andManagement 46, 59–102.
Thompson, L.G., Mosely-Thompson, E., Bolzan, J.F., Koci, B.R.,
1985. A 1,500-year recordof tropical precipitation in ice cores
from the Quelccaya Ice Cap, Peru. Science 229,971–973.
Wells, L.E., Noller, J.S., 1999. Holocene coevolution of the
physical landscape and humansettlement in northern coastal Peru.
Geoarchaeology 14 (8), 755–789.
http://www.arch.cam.ac.uk/dgb27/
Linking cultural and environmental change in Peruvian
prehistory: Geomorphological survey of th.....Introduction and area
of studyMethodsResultsBasic geomorphic unitsSection 1/013 — relict
terrace H-13Survey Pit 1/030 — buried land surface and buried
soilSurvey Pit 1/031 — gradual aeolian deposition
Discussion and conclusionsAcknowledgementsReferences