Chapter 11

From Terraces to Trees: Ancient and Historical

Landscape Changes in Southern Peru
Gregory Zaro

Indigenous agriculture across American landscapes has traditionally

been diverse and complex (Denevan 2001; Doolittle 2000; Whitmore
and Turner 2001). In some cases, diversity of practices, including the
creation of landesque capital, is seen as a hallmark of indigenous farm-
ing, and when coupled with wider subsistence strategies, this diversity
may be interpreted to help manage socioecological risk in a variety of
cultural and environmental settings (Homburg and Sandor 2011; Mar-
ston 2011; Zaro 2007; Zaro and Umire Alvarez 2005). In the prehispanic
Andean region of South America, the focus of many landscape studies
can generally be included under the rubric of landesque capital (for an
extensive review of related works, see Contreras 2010), but the concept
itself is rarely made explicit, and even less attention is paid to the complex
processes that surround its creation or devaluation. Consequently, many
studies focus on either the impetus for systematic investment in landesque
capital or its total abandonment, with much less attention given to the
historically contingent processes that lead to more complex outcomes like
incremental landscape change or cases in which landesque capital may
be only partially devalued without total abandonment.
As a research program, historical ecology treats landscapes as the
primary analytical unit with which to evaluate the interconnectedness
of humans and the environment in the context of dynamic change (Balée
2006). Landesque capital constitutes an important focus of landscape
studies because it represents a form of human agency that often leads to
both intentional and unintentional outcomes, which others have referred
to as “land use legacies” (Foster, Frederick, Aber, Burke, Brokaw, Tilman,
and Knapp 2003). However, given its temporal and spatial components,
the value of landesque capital is subject to change according to both the
cultural importance placed on such landscapes and the socioenviron-
mental conditions surrounding it.
This chapter takes an historical-ecological approach to evaluate the
complex processes surrounding the creation and devaluation of landesque
capital along the Peruvian south coast. The period of interest includes

232

Landesque Capital: The Historical Ecology of Enduring Landscape Modifications, by N. Thomas

Hakansson & Mats Widgren (eds.), 232-250. 2014 Left Coast Press, Inc.
 From Terraces to Trees  ▪  233 

significant socioenvironmental change, with European colonization rep-

resenting the most pronounced cultural and demographic shift, and the
generally arid environment experiencing both punctuated and prolonged
changes in local parameters. As part of a more broadly diverse subsis-
tence strategy, late prehispanic villagers invested in the construction and
maintenance of stone-faced agricultural terraces and canals, whereas the
most enduring feature on the coastal desert landscape today is the olive
tree, first introduced to the region by European colonizers in the six-
teenth century. The results of this study suggest that the degree to which
landesque capital actually stockpiles labor or replaces/improves land for
future use will vary in historically contingent ways. It also provides an
example of the complex manners in which landesque capital can become
devalued over time.
Creation, Management, and Devaluation
of Landesque Capital
Here, landesque capital is broadly conceived, but it generally follows
Blaikie and Brookfield’s (1987, 9) concept of “any investment in land
with an anticipated life well beyond that of the present crop, or crop
cycle. The creation of landesque capital involves substantial ‘saving’
of labour and other inputs for future production.” Terraces, irrigation
canals, water impoundment tanks, and drainage systems all represent
forms of “banked” labor, but less obvious types of investments can also
include cultivated or managed trees and anthropogenic soils, both of
which retain productive characteristics beyond the current harvest cycle
(Brookfield 2001, 55).
Although a historical overview of the concept of landesque capital is
not within the scope of this chapter (see Widgren 2007), several factors
deserve mention here. First, it is fixed in space, and although some forms
might increase resilience or mitigate certain kinds of historically contin-
gent risks associated with farming, they may also be more susceptible to
other kinds of risks, given their immobility within a dynamic socionatural
environment. In some cases, available labor may be redirected toward
other activities with greater flexibility in space, like herd management
or fishing. Such economic shifts historically may signal periods of stress
or risk mitigation (Marston 2011), or perhaps socioeconomic changes
more broadly. Any critical study of landesque capital thus requires that
it be contextualized within the total landscape.
Second, there is an explicit temporal component to landesque capital
because it is inextricably linked to the notion that current labor appli-
cations are “stored” in landscape features for future use. Labor can be
applied systematically in the construction of stone-faced terraces or in
the planting of olive trees, with the expectation that such creations will
require only minimal labor to maintain them once established. Such in-
vestments reflect long-term planning, because the intent of the practice
234  ▪  Gr egory Za ro

is to create a landscape for current and future use. However, agricultural

landscapes may also change incrementally over time, and in some cases
resulting in the slow creation of landesque capital, simply from repeated
cultivation practices (Doolittle 1984). Such investments may not neces-
sarily reflect long-term intent, but the outcomes might nevertheless be
carried over beyond the current cropping cycle. Although it generally
receives much less attention, the devaluation of such investments can
also be systematic (abandonment, total loss) or incremental, in which
specific components of landesque capital may continue to hold value
while others are lost given changes in the socioenvironmental conditions
surrounding its use. This underscores the importance of examining the
complex and historically contingent processes that underlie the creation,
management, and devaluation of landesque capital (see also Morrison
1994), and necessarily calls for greater attention to local and regional
sequences of landscape change (Zaro et al. 2008).
The Peruvian South Coast
Though often overshadowed by the neighboring highlands, the Andean
desert coast is quite ecologically diverse and has historically offered a
plethora of opportunities under the right socioecological conditions.
Along the Tambo-Ilo intervalley coast of southern Peru, a diverse strategy
that incorporated marine, freshwater, and lomas (inland fog-fed forest)
resource bases seems to have been productive and sustainable for periods
of time (Zaro 2007). Today, however, the area is littered with abandoned
field systems, with only a few scattered farmsteads and historic olive
groves surviving into the present. Many of the inland hills remain desic-
cated as well, but remains of woody and herbaceous plants and trees tes-
tify to a more vegetated landscape in the past. Historic records mention a
thriving lomas four to five centuries ago in the Ilo region, noting that they
were once some of the most productive of coastal Peru (Rostworowski
1981; Vázquez de Espinosa 1942[1618]). However, lomas contraction was
notable by the early eighteenth century, with historic accounts attributing
much of their decline to deforestation (Frezier 1982[1713]). Continued
collection of woody material in the present era has contributed to the
barren landscape observed today, which has been, in many respects,
dominated by an industrial scale copper smelter since the 1960s.
Water is a limiting factor in any desert setting, but along the Andean
coast its presence on the surface is very patchy and usually results from
either runoff from highland precipitation, excavation down to ground
water aquifers, or freshwater springs. Between the Tambo and Ilo rivers,
springs have historically constituted the primary source of fresh water, but
they are characterized by low flow and often confined to incised channels.
Rain falls only rarely and typically results from recurring, though infre-
quent, El Niño episodes (Maasch 2008). Particularly strong events can be
catastrophic (Moseley and Keefer 2008), but the resultant precipitation
 From Terraces to Trees  ▪  235 

also helps to spawn growth among inland stands of lomas and improve
pastureland for domestic and wild herbivores (Oficina de Información
Agraria 2000). Aquifer recharge in the study area has not been intensively
investigated, but it may relate to fluctuations in highland runoff (Clement
and Moseley 1991), variation in the strength and frequency of El Niño
events (Magilligan et al. 2008), or a combination of the two.
Analyses of ice (Thompson et al. 1994) and lake sediment (Abbott et
al. 1997) cores from the high Andes suggest considerable variation in
highland precipitation over the past 1,500 years, including drier than
average conditions from the tenth through fifteenth centuries, followed
by relatively wetter conditions from the late fifteenth through the sev-
enteenth centuries. Along the coast, the geomorphic record provides
evidence for at least two potentially catastrophic El Niño events in the
fourteenth and seventeenth centuries, respectively (Moseley and Keefer
2008), but historic accounts also suggest recurrences of El Niño over the
past four centuries (Quinn and Neal 1992). Unfortunately, the spatial
extent of coastal rainfall varies considerably, and the local impact, if any,
of specific events is difficult to gauge.
Although there is evidence for farming in the area early in the first mil-
lennium A.D., (Owen 2009), the ephemeral nature of the archaeological
record immediately following this early period suggests that population
levels subsequently declined and remained relatively low or absent for
several centuries (Owen 1993). Beginning around A.D. 900 or slightly
earlier, the region began to support a number of farming and fishing
villages that formed part of the Chiribaya señorío, a culture group of
moderate socioeconomic and political complexity (Lozada and Buikstra
2002; Umire Alvarez and Miranda 2001). Chiribaya farmers constructed
major canals and terraces within the Ilo river valley, whereas villages
closer to the river’s mouth focused predominantly on fishing. Chiribaya
material expression persisted for about 5 centuries, and based on the total
size of habitation areas, the lower river valley appears to have experienced
population growth, possibly substantial, from early to late Chiribaya
periods (Owen 1993, 526).
Along the intervalley coast, the most pronounced archaeological
component dates to the late Chiribaya phase (A.D. 1200–1400) and
consists of habitation, mortuary, and agricultural sectors organized
around localized springs (Zaro et al. 2010). Individual complexes were
noticeably smaller than those of the Ilo valley, but combined, they formed
a significant component of the overall Chiribaya cultural landscape. In-
tervalley villages pursued an explicitly diverse set of economic activities
that included agriculture, fishing, and herding; but the relative strength
of these may have varied from one location to another. The convergence
of a massive late fourteenth century El Niño with prolonged drought
may have catalyzed the disappearance of Chiribaya cultural expression
(Satterlee et al. 2000), but farming in intervalley contexts seems to have
236  ▪  Gr egory Za ro

persisted for decades longer before widespread abandonment in the

middle of the fifteenth century (Zaro et al. 2013). The region remained
largely depopulated during the subsequent Andean Late Horizon (~A.D.
1476–1532), the period preceding Spanish colonization.
In the sixteenth century, Spanish colonists initially organized wealth in
the Andes through the encomienda, an institution used to extract surplus
labor and resources from indigenous populations (Andrien 2001). As part
of an encomienda in 1540, the small coastal village of Ilo consisted of
a mere 20 native Andeans, a testament to continued population depres-
sion in the early colonial period (Adriazola Flores 1998, 40–41). By the
seventeenth century, commercial agriculture on the Peruvian coast was
anchored to wheat, wine, and sugar (Keith 1976, 65), but along the far
south coast of Peru, historic era farmers managed olive groves in addi-
tion to small-scale subsistence agriculture, probably driven in part by the
desire for European products (olive oil) and the growth of an economic
trunk line across the Central Andes (Andrien 2001). Population in the
area remained low until the latter half of the twentieth century, when
Southern Peru Copper Corporation established its smelter along the in-
tervalley coast, catalyzing significant growth in Ilo as a company town
(Hall 1992). Only a few olive groves survive along the intervalley coast
today, significantly reduced in number and capacity from the historic and
more ancient agricultural landscapes.
Landesque Capital and the Intervalley Coast
Significant creation of landesque capital began during the late Chiribaya
period, and many of those modifications endure, either as abandoned
archaeological relicts or, in many fewer cases, as living fields (Zaro et
al. 2010). Archaeological survey identified a total of 23 independent
Chiribaya field systems, with an average size of only about 9 ha per
system (Figure 11.1). Despite the modest size of each system, canals and
terraces constitute the majority of physical remains of prehispanic farm-
ing in the area, although various furrowing strategies and other forms of
soil management are also preserved under the right conditions. Specific
components of these capital investments are described here, followed
by a discussion in the context of socioenvironmental change outlined
previously.
Terracing and Irrigation, Twelfth to Fifteenth Centuries
Terraces
Terrace construction creates planting surfaces, improves water reten-
tion, augments soil thickness, and reduces soil erosion (Denevan 2001;
Homburg and Sandor 2011). Although soil fertility may not necessarily
improve immediately following terrace construction (e.g., Posthumus and
Stroosnijder 2010), it is often enhanced and maintained when combined
 From Terraces to Trees  ▪  237 

Figure 11.1  Map of the study region showing locations of Chiribaya agricultural spring complexes.

with other management strategies that include crop rotation, multicrop-

ping, irrigation, fertilization, or fallowing (Holliday 1992).
Following Treacy and Denevan (1994), several terrace types initially
defined in highland Andean contexts are notable along the Tambo-Ilo
intervalley coast, including valley bottom terracing (referred to here as
channel margin terraces), bench or contour terracing, and grid terrac-
ing. Most are faced with small to medium blocks of pink granite and
constructed to heights of about 0.5 to 1.0 meter. Chiribaya terraces most
often exhibit a fairly homogeneous fill of silt, sand, and small gravels with
very gradational changes between strata, suggesting that regardless of
type, individual terrace complexes seem to have been constructed rela-
tively systematically over a short period of time rather than incrementally
over long periods (Zaro and Umire Alvarez 2005).
Channel margin terraces were commonly used to create cultivable
terrain along the moderately sloping margins of coastal ravines (Figure
11.2), whereas contour terracing is generally found on steeper terrain away
from primary channels, or in isolated discontinuous blocks among hills
or desert plains between ravines. Their primary purpose was likely to
generate a level planting surface and create a deeper soil horizon. Spacing
of terraces appears to be related to slope: the steeper the slope, the more
238  ▪  Gr egory Za ro

closely spaced the terrace walls. In some cases, layers of small cobbles
had been placed at the base of terrace fill, possibly to facilitate drainage.
Many constructions also contain some element of grid terracing, though
one site in particular includes an entirely gridded field of roughly 35 x
60 meters. Its general purpose may have been simply to accommodate
underlying geomorphology: As the terrain undulated, grid terraces were
able to maintain relatively small, quadrangular planting surfaces.
In most locations, terrace fill is generally darker than apparently uncul-
tivated lands, and moderate amounts of carbon, marine shell, and other
cultural refuse were also commonly observed. Though not a ubiquitous
practice, indigenous Andean farmers traditionally integrated domestic
refuse into field systems or practiced a form of mixed farming, in which
herd management (camelids and, later, livestock) and manuring were
important components of cultivation practices. Both strategies introduce
high amounts of organic waste into cultivated terrain, contributing to the
general health of soil maintenance (Sandor and Eash 1995).
Under conditions of optimal lighting and good preservation, ancient
furrows are also visible. In most cases, specific patterns cannot be distin-
guished, but several examples on abandoned terraces and pampa surfaces
around Cola de Zorro may reflect box furrows morphologically similar
to waffle gardens (Doolittle and Neely 2004), or sinuous caracoles that
are common elsewhere in the Andes (Denevan 2001). Pollen and mac-
robotanical remains from Wawakiki and Cola de Zorro suggest that
maize, beans, and squash/pumpkin were probably common crops among
Chiribaya field systems (Zaro 2007; Zaro et al. 2013).
Irrigation
Irrigation complexes along the intervalley coast are much smaller and
more localized than those of principal river valleys of coastal Peru. The
longest irrigation system among intervalley complexes stretched only
about 2 kilometers, and most operated on an even smaller scale of a few
hundred meters. All terraced fields were irrigated by spring water, either
directly through canals, or by first storing water in impoundment tanks
and subsequently redirecting it to cultivated surfaces. The few surviving
farmsteads today rely solely on impoundment tanks because low spring
discharge precludes its transport for any great distance. Chiribaya farm-
ers also occasionally drew from multiple spring sources to irrigate single
units of land, a strategy that would have been more resilient in the face
of dynamic spring flow (Zaro and Umire Alvarez 2005).
Abandonment, Fifteenth to Sixteenth Centuries
The paucity of Late Horizon artifacts coupled with available radiocarbon
dates are broadly suggestive of widespread abandonment of the interval-
ley coastal region sometime in the middle of the fifteenth century and
 From Terraces to Trees  ▪  239 

Figure 11.2  Abandoned channel margin terraces at Cola de Zorro.

240  ▪  Gr egory Za ro

well before Spanish colonization (Zaro et al. 2013). Without continued

maintenance, the intervalley agricultural landscape was subject to broader
geophysical processes of decay; predominantly wind, gravity, and oc-
casionally water.

Olive Groves and the Changing Landscapes of

the Seventeenth to Twentieth Centuries
Terraces and Olives
Although many terraces of more gently sloping terrain remained relatively
intact at the time of Spanish arrival, the retention walls of those on steeper
hillsides had been breached in various locations, leading to considerable
erosion (Zaro 2005). Spanish colonial land use was organized around
existing infrastructure, but in most cases, a degree of contraction is evi-
dent, leaving many distal terraced fields to continue the process of decay.
However, there is little indication that Spanish colonial farmers utilized
even more intact terraces in the same manner as indigenous farmers. On
the promontory at Wawakiki, for example, geologic profiles identified
late Chiribaya terrace walls and fill buried by thin stratified deposits of
sand and gravels with very sharp contacts between layers, suggesting
that punctuated and recurrent alluvial episodes led to their incremental
accumulation over time. A thin lens of Huaynaputina tephra dated to
A.D. 1600 is also visible in profile, providing good chronological control
between strategies.
Olive trees became a primary economic focus of many coastal spring
systems, with smaller subsistence crops cultivated at distal ends of these
groves. Semicircular stone terraces were sometimes constructed around
individual trees, no doubt drawing on materials, in part, from available
prehispanic terraces (Figure 11.3). Deliberate deconstruction of prehis-
panic agricultural terraces in the historic era was most evident at Cola
de Zorro, in which a number of retention walls were stripped of their
stones and redirected toward the construction of a large corral around
the former fields (Figure 11.4). Dung and other remains of only European
introduced herbivores (caprines, equines) were identified in excavation and
surface inspection of this corral, although smaller enclosures elsewhere
at the site were used to pen camelids in the early fifteenth century (Zaro
et al. 2013). Limited farming after A.D. 1600 is apparent, but there is
no evidence of olive cultivation at the site.
Spanish colonial fields often buried late Chiribaya terraces, which in
some cases makes possible a stratigraphic comparison of cultivated soil.
As noted previously, abandoned prehispanic soils visibly contain carbon,
marine shell, and other cultural refuse, but post-A.D. 1600 soil horizons
are noticeably lacking in these additives (Zaro 2005).
Furrowing strategies can rarely be identified specifically for the Spanish
colonial period, but several fields that were likely abandoned in the late
 From Terraces to Trees  ▪  241 

Figure 11.3  Olive grove at Pocoma.

Figure 11.4  Historic era corral constructed around prehispanic agricultural terraces. Stones used in
corral construction were apparently stripped from terrace walls.
242  ▪  Gr egory Za ro

nineteenth or early twentieth century exhibit elements of more traditional

Andean farming strategies, and they may represent reintroduction of those
practices in the Spanish colonial or modern era. Furrowing patterns tend
to be visible at the distal ends of olive groves, with the most pronounced
patterns reflecting E-type furrows or sinuous caracoles (Figure 11.5),
both of which have a long history in the Andes and have been identified
in both prehispanic and historic contexts (Denevan 2001).
Irrigation
Because water is the limiting factor in desert agriculture, it is not surpris-
ing that historic and more recent fields have been tethered to agricultural
infrastructure of greater antiquity. Unfortunately, few excavations within
irrigation canals have been performed, but one at Wawakiki suggests
that historic era farmers renovated and reused at least one primary
canal that had been abandoned in the late fourteenth or early fifteenth
century (Zaro 2005, 85–87). It is also clear that although specific prac-
tices in the Spanish colonial period are difficult to glean from available
information, farming into the present era continues to utilize strategies
that include water impoundment tanks and multiple spring sources to
irrigate units of land. In some cases, impoundment tanks were reduced
in size, probably to accommodate decreases in spring discharge (Zaro
2005). Similar modifications can also be observed in more recent tanks
at Alastaya grove.
Discussion
It is unclear what precipitated Chiribaya expansion into the intervalley
coastal region in the twelfth or thirteenth century, but a growing popu-
lation in the context of an intensifying drought may have placed added
stress on the specialized, exchange-based economy of Chiribaya farmers
and fisherfolk settled along the Ilo River. What is clearer, however, is
that Chiribaya villagers expanded into a relatively depopulated region
and explicitly pursued a diverse strategy that integrated farming, fish-
ing, and herding. Although a diverse economic strategy alone may be
evidence of risk reduction, the specific kinds of agricultural investments
and the diversity with which they were implemented also permitted in-
tervalley farmers to manage risks associated with this hyper-arid region,
particularly as they relate to water security and soil. The employment
of multiple terracing strategies in unique geomorphic contexts along the
Tambo-Ilo coast permitted local farmers to increase the diversity of con-
texts in which farming could be implemented. Ultimately, the importance
of terrace construction, canal engineering, and soil management lay in
their integration: Indigenous farmers implemented these investments in
concert to manage an array of diverse land forms along the coast, creating
an assorted agricultural landscape that remained productive for several
centuries as part of a more broadly diverse economic strategy.
 From Terraces to Trees  ▪  243 

Figure 11.5  Plan of late nineteenth or early twentieth century caracoles at Wawakiki.
244  ▪  Gr egory Za ro

Terrace and canal construction served to physically transform desert

regions into agricultural farmland. Canalization of water meant that
farming would not be restricted solely to the drainage channels them-
selves (most of which are deeply incised and offer no arable land), but
rather were engineered along the channel margins or across pampa sur-
faces away from the channels. In effect, the combination of canals and
terraces permitted the systematic creation of an agricultural landscape
(capital) that would otherwise have been impossible during most years.
The construction of such infrastructure may have required considerable
labor investments initially, but there is no indication that it was done so
under the direction of a larger central authority. Most intervalley spring
complexes were small in scale and highly localized. In contrast, lower
Ilo Valley agriculture was predominantly linked to a single 9-kilometer-
long canal and likely required more coordinated effort to construct or
repair (Reycraft 2000). The archaeological record suggests a sense of
permanence with respect to Chiribaya villagers along the intervalley
coast because most agricultural complexes were also associated with
domestic/residential sectors and formally bounded cemeteries, both of
which may signal a degree of territoriality (Zaro et al. 2010). This also
implies sufficient onsite availability of labor to construct and subsequently
manage investments in an engineered agricultural landscape, including
terrace maintenance, canal cleaning, and the regular introduction of
organic refuse into cultivated soils. The systematic construction of such
a landscape certainly implies confidence in long-term stability and the
intentional stockpiling of labor for future use.
The archaeological record is consistent with widespread abandonment
along the intervalley coast sometime around the middle of the fifteenth
century. The reason for this is unclear, but an intensifying highland
drought may have adversely impacted coastal aquifers and limited
freshwater availability beyond the resilience of multispring irrigation
systems and water impoundment tanks. Expansion of massive terrace
and irrigation field systems under Inca management in the adjacent up-
per Moquegua/Ilo drainage at sites like Camata (Dayton 2008) during
the Late Horizon may have also contributed to declining runoff and
subsequent aquifer recharge along the coast. The presence of water does
not necessarily precipitate farming in a desert environment, but the lack
of water would most certainly guarantee its absence.
A wetter period from about A.D. 1500–1720 (Thompson et al. 1994)
may have fostered conditions once again favorable for small-scale farm-
ing, but along the intervalley coast land use appears to have remained
relatively inactive until the late sixteenth or early seventeenth centuries.
Much like the situation encountered by late phase Chiribaya farmers,
Spanish colonists, albeit few in numbers, also encountered a landscape
that had been relatively abandoned for a century or more. Most farmers
took advantage of a landscape already naturally equipped with canals and
 From Terraces to Trees  ▪  245 

terraced fields, though with a much-reduced supply of labor and more ex-
pedient mentality toward land use. Farming was often diminished in scale
and apparently with little labor applied to existing terrace infrastructure.
The absence of even moderate organic contents in historic fields also sug-
gests that Spanish colonists did not engage in similar fertilizing strategies
as indigenous Andean farmers, which might be explained by the general
difference in settlement between indigenous Andean communities and
Spanish colonizers. Indigenous Andeans were generally more dispersed
about the landscape and engaged in rural lifeways. In contrast, Spanish
colonial populations were much more urban focused and tended to reside
in nucleated centers (Keith 1976). The Chiribaya domestic, fishing, and
herding components that accompanied many intervalley agricultural
complexes meant that there would have been a ready supply of organic
waste and other refuse for field systems. This has not been the case since
before late sixteenth century recolonization of the region, where resident
populations along the intervalley coast have remained sparse, with the
majority centered in principal river valleys. Thus, a change in population
and settlement distribution may partially explain the lack of significant
organic inputs and labor applications to post-1600 fields.
Temporal studies of landscape change at Wawakiki (Zaro 2005) and
Carrizal (Clement and Moseley 1991) suggest that since the seventeenth
century, the extent of land use at these springs has declined incremen-
tally into the twentieth century, and possibly an indication that spring
discharge has been slowing. Furthermore, a comparison of sixteenth
century ethnohistoric records and historic era descriptions with the cur-
rent situation among the inland hills suggest that the lomas also may
have incrementally declined into the present day. Farther north along the
Andean coast, Canziani (1998) argued that deforestation and overgrazing
(European-introduced herbivores) of the Lomas de Atiquipa since the
Spanish colonial period were primary factors in the desertification of that
landscape. It is unclear what role European-introduced herbivores have
played in lomas desiccation along the Tambo-Ilo coast, but historic era
corrals are found at many locations among current and former agricul-
tural complexes and among the inland hills.
Immovable vs. Mobile Capital and the Intervalley Coast
In general, the infrequency of coastal precipitation coupled with the
dynamic nature of low discharge springs has necessarily made farming
along the intervalley coast a particularly risky endeavor. The spatially
fixed nature of landesque capital means that agroecological systems
are either resilient to such risks in their current location or subject to
abandonment. Chiribaya farmers mitigated risks in several ways: They
utilized multiple spring sources to irrigate single units of land, captured
low-discharge spring water in impoundment tanks for its subsequent
distribution, and created an assorted agricultural landscape in various
246  ▪  Gr egory Za ro

geomorphic settings. Importantly, they maintained these investments

in the context of a diverse subsistence economy that included camelid
herding and fishing. Indeed, the very construction of an agricultural
landscape expanded the limits of vegetation and may have made herding
more tenable along the desert coast. In the Mediterranean, the inclusion
of livestock in a broader subsistence strategy is argued to buffer subsis-
tence stress precisely because of its mobility (Butzer 1996, 143). Hence,
with the appropriate labor availability, Chiribaya combined significant
investments in immovable capital (canals and terraces) with those of
greater mobility (herding and fishing) to navigate the socioenvironmental
dynamics of the Peruvian south coast.
The importance of herbivores in the prehispanic, Spanish colonial,
or even more contemporary Tambo-Ilo cultural landscapes is difficult
to quantify, but the material record seems to reflect a greater number
of historic era corrals than prehispanic camelid pens (Zaro and Umire
Alvarez 2007). This may signal an increase in the relative importance
of domestic animals in the local economy since Spanish colonization,
particularly in the context of agricultural contraction and significantly
lower population levels. Gade (1992) has noted that significant popula-
tion decline more broadly across the Central Andes may have catalyzed
rapid adoption of European livestock, citing their more manageable labor
requirements than those for maintaining terraced field systems. Along
with cultural values that were tied to an emerging European economy
in Peru, the relatively diminished labor availability along the Tambo-Ilo
coast may not only have led to more expedient methods of farming and
the devaluation of terrace constructions but also to a greater importance
of livestock in the local and regional economy.
Landesque Capital and a Process of Devaluation
Agricultural landscapes are constantly in a state of development, and in
general they are created through a mix of both systematic and incremen-
tal additions. Chiribaya farmers in the twelfth through early fifteenth
centuries systematically constructed networks of canals and terraces, but
they also regularly integrated organic refuse into fields that incrementally
contributed to the general health of soils. On the other hand, the most
enduring agricultural landscape feature of the Spanish colonial era is the
olive tree. Olive trees, which can live for centuries, may be viewed as sys-
tematic investments, particularly because they require several years before
producing a first crop of fruit, but annual maintenance in pruning and
irrigation literally cultivates capital incrementally over the lifespan of a tree.
Similarly, the devaluation of landscape modifications may also be rapid
or incremental, where components of constructed landscapes may retain
value long after others have fallen out of favor. The devaluation of canal
and terrace systems in the fifteenth century was rapid because the entire
intervalley coast appears to have experienced a century or more period
 From Terraces to Trees  ▪  247 

of abandonment. As noted previously, the reasons for this are unclear,

but the combination of precipitation variability and upper valley irriga-
tion practices may have created considerable water shortages for even
the most resilient coastal spring systems to manage. Spanish colonial
and historic era farming was also organized around coastal springs,
and although some prehispanic canals were renovated and reused in the
colonial period (continued appreciation), most terraced fields were not
valued in the same manner. In an extreme case, retention walls from
prehispanic terraces were stripped and redirected toward the construction
of a corral. In many cases, however, portions of prehispanic agricultural
fields were utilized for farming, but apparently without any investment in
terrace maintenance. Consequently, older terraces deteriorated and were
simply buried from sight after years of cultivation. Limited labor avail-
ability may be partly responsible for the continued devaluation of terrace
systems throughout the colonial and historic era, but because water is
the limiting factor for desert farming, foregoing canal construction or
renovation was not an option.
Conclusion
In a historical ecological framework, the integration of humans and envi-
ronment are manifested in landscapes. Consequently, a focus on landesque
capital necessarily reveals the complexity of relationships between people,
environment, and climate. In the context of non-equilibrium dynamics,
change is historically contingent and therefore demands that landscapes
be studied on local and regional scales, in multiple cultural and envi-
ronmental settings. Archaeology can provide a unique perspective on
this process—indeed, the physical nature of landesque capital gener-
ally leaves physical markers—but the coarse chronological resolution
normally available to the archaeologist limits the discussion to broader,
centuries-to-millennia-scale, overviews or trends. However, in doing so,
such studies can point to potential relationships among key variables that
surround landesque capital.
The approach taken here combines the archaeological and historic
records with the contemporary landscape to evaluate the historical ecol-
ogy of enduring modifications. Indeed, landesque capital may be viewed
as “banking labor,” yet it does not do so indefinitely and may undergo
a complex process of devaluation as socioecological conditions vary. In
an extreme desert environment like that of southern coastal Peru, water
availability has been a key variable in the spatial distribution and extent
of farming in the region. However, the archaeological record suggests that
the process of socioenvironmental change over the past eight centuries
was complex, and that relationships between settlement, labor avail-
ability, and culturally defined practices of resource management were
also of great importance in the continued appreciation or devaluation
of landesque capital.
248  ▪  Gr egory Za ro

Acknowledgments
This work has been supported, in part, by the National Science Foun-
dation (BCS #0222040), the Heinz Foundation for Latin American
Archaeology, the University of New Mexico, and the University of
Maine. Permits to conduct archaeological research along the Tambo-Ilo
coast were granted by the National Institute of Culture in Lima, Peru. I
wish to thank the editors of this volume, Thomas Håkansson and Mats
Widgren, for their invitation to participate in such a stimulating project,
and for their thoughtful comments and editorial work. I also thank Bill
Doolittle for his observations on an earlier draft of this paper. Although
editorial and reviewer comments have helped to strengthen this paper,
any interpretive errors or oversights are strictly my own.

References
Abbott, Mark B., Michael W. Binford, Mark Brenner, and Kerry R. Kelts. “A 3500 C-14 yr High-Resolution
Record of Water-Level Changes in Lake Titicaca, Bolivia/Peru.” Quaternary Research 47, no. 2
(1997): 169–180.
Adriazola Flores, Edwin. 1998. Apuntes para una Historia General de Ilo. Ilo: Archivo Municipal
Andrien, Kenneth J. 2001. Andean Worlds: Indigenous History, Culture, and Consciousness under Spanish
Rule, 1532–1825. Albuquerque: University of New Mexico Press.
Balée, William. “The Research Program of Historical Ecology.” Annual Review of Anthropology 35 (2006):
75–98.
Blaikie, Piers. and Harold Brookfield. 1987. Land Degradation and Society. New York: Methuen.
Brookfield, Harold. 2001. Exploring Agrodiversity. New York: Columbia University Press.
Butzer, Karl. “Ecology in the Long View: Settlement Histories, Agrosystemic Strategies, and Ecological
Performance.” Journal of Field Archaeology 23 (1996): 141–150.
Canziani, Jose. 1998. “Las Lomas de Atiquipa: Un Caso de Paisaje Cultural en la Costa Desértica del
Sur del Perú.” In Paisajes Culturales en los Andes, edited by E. Mujica Barreda. Arequipa, Peru:
Patrimonio Mundial.
Clement, Christopher O., and Michael E. Moseley. “The Spring-Fed Irrigation System of Carrizal, Peru: A
Case-Study of the Hypothesis of Agrarian Collapse.” Journal of Field Archaeology 18 (1991): 425–443.
Contreras, Daniel A. 2010. “Landscape and Environment: Insights from the Prehispanic Central Andes.”
Journal of Archaeological Research 18 (2010): 241–288.
Dayton, Christopher L. “Late Prehistoric and Modern Irrigation Agriculture in Torata, Peru.” Unpublished
Ph.D. diss., Department of Archaeology, Boston University, 2008.
Denevan, William M. 2001. Cultivated Landscapes of Native Amazonia and the Andes: Triumph over the
Soil. Oxford: Oxford University Press.
Doolittle, William E. 1984. “Agricultural Change as an Incremental Process.” Annals of the Association of
American Geographers 74 (1984): 124–137.
Doolittle, William E. 2000. Cultivated Landscapes of Native North America. Oxford: Oxford University Press.
Doolittle, William E., and James A. Neely, eds. 2004. The Safford Valley Grids: Prehistoric Cultivation in
the Southern Arizona Desert. Anthropological Papers of the University of Arizona 70. Tucson: The
University of Arizona Press.
Foster, David, Frederick Swanson, John Aber, Ingrid Burke, Nicholas Brokaw, David Tilman, and Alan
Knapp. 2003. “The Importance of Land-Use Legacies to Ecology and Conservation.” Bioscience
53, no. 1 (2003): 77–88.
Frezier, Amadeo. 1982[1713]. Relación del Viaje por el Mar del Sur. Ayacucho, Peru: Biblioteca Ayacucho.
Gade, Daniel W. 1992. “Landscape, System, and Identity in the Post-Conquest Andes.” Annals of the
Association of American Geographers 82 (1992): 460–477.
Hall, Arthur C. 1992. Nothing Down. Tucson: Sundance Press.
 From Terraces to Trees  ▪  249 

Holliday, Vance T., ed. 1992. Soils in Archaeology: Landscape Evolution and Human Occupation. Washington,
DC: Smithsonian Institution Press.
Homburg, Jeffrey A., and Jonathan A. Sandor. “Anthropogenic Effects on Soil Quality of Ancient Agri-
cultural Systems of the American Southwest.” Catena 85 (2011): 144–154.
Keith, Robert G. 1976. Conquest and Agrarian Change. Cambridge: Harvard University Press.
Lozada, Maria Cecilia, and Jane E. Buikstra. 2002. El Señorío de Chiribaya en la Costa Sur del Perú. Lima:
Instituto de Estudios Peruanos.
Maasch, Kirk A. 2008. “El Niño and Interannual Variability of Climate in the Western Hemisphere.” In
El Niño, Catastrophism, and Culture Change in Ancient America, edited by Daniel H. Sandweiss and
Jeffrey Quilter. Washington, DC: Dumbarton Oaks Research Library and Collection, Harvard
University Press.
Magilligan, Francis J., Paul S. Goldstein, G. Burch Fisher, Benjamin C. Bostick, and Rebecca B. Man-
ners. “Late Quaternary Hydroclimatology of a Hyper-Arid Andean Watershed: Climate Change,
Floods, and Hydrologic Responses to the El Niño-Southern Oscillation in the Atacama Desert.”
Geomorphology 101 (2008): 14–32.
Marston, John. M. “Archaeological Markers of Agricultural Risk Management.” Journal of Anthropological
Archaeology 30 (2011): 190–205.
Morrison, Kathleen D. “The Intensification of Production: Archaeological Approaches.” Journal of Ar-
chaeological Method and Theory 1, no. 2 (1994): 111–159.
Moseley, Michael E., and David K. Keefer. 2008. “Deadly Deluges in the Southern Desert: Modern and
Ancient El Niños in the Osmore Region of Peru.” In El Niño, Catastrophism, and Culture Change in
Ancient America, edited by Daniel H. Sandweiss and Jeffrey Quilter. Washington DC: Dumbarton
Oaks Research Library and Collection, Harvard University Press.
Oficina de Información Agraria. 2000. Areas Perdidas y Afectadas durante el Fenómeno “El Niño.” Lima:
Ministerio de Agricultura.
Owen, Bruce D. “A Model of Multiethnicity: State Collapse, Competition, and Social Complexity from
Tiwanaku to Chiribaya in the Osmore Valley, Peru.” Unpublished Ph.D. diss., Department of An-
thropology, University of California, Los Angeles, 1993.
Owen, Bruce D. 2009. “Early Agriculture in the Coastal Osmore Valley, Peru: Synchronous Events and
Macroregional Processes in the Formation of Andean Civilization.” In Andean Civilization: A Tribute
to Michael E. Moseley, edited by J. Marcus and P. R. Williams. Monograph 63. Los Angeles: UCLA
Cotsen Institute of Archaeology.
Posthumus, Helena, and Leo Stroosnijder. “To Terrace or Not: The Short-Term Impact of Bench Ter-
races on Soil Properties and Crop Response in the Peruvian Andes.” Environment, Deelopment and
Sustainability. 12 (2010): 263–276.
Quinn, William H., and Victor T. Neal. 1992. “The Historical Record of El Niño events.” In Climate Since
A.D. 1500, edited by Raymond S. Bradley and Philip Jones. New York: Routledge.
Reycraft, Richard M. 2000. “Long-term Human Response to El Niño in South Coastal Perú, circa A.D.
1400.” In Environmental Disaster and the Archaeology of Human Response, edited by Garth Bawden
and Richard M. Reycraft. Anthropological Papers No. 7. Albuquerque: Maxwell Museum of An-
thropology.
Rostworowski, Maria de Diaz Canseco. 1981. Recursos Naturales Renovables y Pesca, Siglos XVI y XVII.
Lima: Instituto de Estudios Peruanos.
Sandor, Jonathan A., and Neal S. Eash. “Ancient Agricultural Soils in the Andes of Southern Peru.” Soil
Science Society of America Journal 59, no. 1 (1995): 170–179.
Satterlee, Dennis R., Michael E. Moseley, David K. Keefer, and Jorge E. Tapia A. “The Miraflores El Niño
Disaster: Convergent Catastrophes and Prehistoric Agrarian Change in Southern Peru.” Andean
Past 6 (2000): 95–116.
Thompson, Lonnie G., Mary E. Davis, and Ellen Mosley-Thompson (1994). “Glacial Records of Global
Climate: A 1500-Year Tropical Ice Core Record of Climate.” Human Ecology 22, no. 1 (1994): 83–95.
Treacy, John M., and William M. Denevan. 1994. “The Creation of Cultivable Land through Terracing.”
In The Archaeology of Garden and Field, edited by N. F. Miller and K. L. Gleason. Philadelphia:
University of Pennsylvania Press.
Umire Alvarez, Adan, and Ana Miranda. 2001. Chiribaya de Ilo. Arequipa, Peru: Eskann.
250  ▪  Gr egory Za ro

Vázquez de Espinosa, Antonio. 1942 [1618]. Compendium and Description of the West Indies. Smithsonian
Miscellaneous Collections 102. Washington, DC: Smithsonian Institution.
Whitmore, Thomas M.. and Billie L. Turner. 2001. Cultivated Landscapes of Middle America on the Eve of
Conquest. Oxford: Oxford University Press.
Widgren, Mats. 2007. “Pre-Colonial Landesque Capital: A Global Perspective.” In Rethinking Environ-
mental History: World-System History and Global Environmental Change, edited by Alf Hornborg,
John Robert McNeill, and Juan Martínez Alier. Lanham, MD: Altamira Press.
Zaro, Gregory. “Human Ecology and Ancient Agrarian Land Use at Wawakiki Spring, South Coast
Peru, A.D. 1000–2000.” Unpublished Ph.D. diss., Department of Anthropology, University of
New Mexico, 2005.
Zaro, Gregory. “Diversity Specialists: Coastal Resource Management and Historical Contingency in the
Osmore Desert of Southern Peru.” Latin American Antiquity 18 (2007): 161–179.
Zaro, Gregory, Heather Builth, Claudia Rivera, Jimena Roldán, and Graciela Suvires. “Landscape Evolu-
tion and Human Agency: Archaeological Case Studies from Drylands in Western South America
and Australia.” Chungará, Revista de Antropología Chilena 40, Número Especial (2008): 261–271.
Zaro, Gregory, Kenneth C. Nystrom, Alfredo Bar, Adan Umire-Alvarez, and Ana Miranda. “Tierras
Olvidadas: “Chiribaya Landscape Engineering and Marginality in Southern Peru.” Latin American
Antiquity 21 (2010): 355–374.
Zaro, Gregory, Kenneth C. Nystrom, and David K. Keefer. “Environmental Catastrophe and the Ar-
chaeological Record: Complexities of Volcanism, Floods, and Farming in South Coastal Peru, A.D.
1200–1700.” Andean Past 11 (2013): 233-262.
Zaro, Gregory, and Adán Umire Alvarez. “Late Chiribaya Agriculture and Risk Management Along the
Arid Andean Coast of Southern Peru, A.D. 1200–1400.” Geoarchaeology: An International Journal
20 (2005): 717–737.
Zaro, Gregory, and Adán Umire Alvarez. 2007. “Informe Técnico: Proyecto Arqueológico Tambo-Ilo
2006.” Field report submitted to the Instituto Nacional de Cultura, Lima, Peru.