Social-Ecological Functional Types: Connecting People and Ecosystems in the Argentine Chaco

Abstract

Sustainability science recognizes the importance of the integrated assessment of the ecological and social systems in land-use planning. However, most studies so far have been conceptual rather than empirical. We developed a framework to characterize the social-ecological systems heterogeneity according to its functioning through the identification of social–ecological functional types (SEFT). The SEFT framework builds on the plant, ecosystem and agent functional type approaches, taking a step forward to integrate the dimensions of social–ecological systems into an operational product to characterize administrative units in a hierarchical way. To illustrate this novel framework, we described the heterogeneity of SEFT in the Argentine Chaco by clustering administrative entities. This area is a global deforestation hotspot and has diverse social actors that harness ecosystem services in multiple, and sometimes contrasting and conflictive, ways which determines an urgent need for land-use planning. We combined data from national census and remote sensing to identify SEFT by clustering census tracts based on 17 input variables that integrate key human, ecological and interaction processes across landscapes. We identified three classes and eight subclasses of SEFT. Ecological variables defined the first level of heterogeneity (classes), while human variables and the variables of interactions between the human and ecological components defined a second level of heterogeneity (subclasses). The degree of anthropization and mean annual productivity were important variables to explain the first two axes in the ordination (32% of the total variance). This framework offers a conceptually novel and comprehensive approach to understand the spatial heterogeneity of social–ecological systems functioning, which could play a pivotal role to support conservation or land-use planning in rural areas.

Full text

Social-Ecological Functional Types:
Connecting People and Ecosystems
in the Argentine Chaco
M. Vallejos,
1,2,3
* S. Aguiar,
1,4
G. Baldi,
5
M. E. Mastra
´ngelo,
6
F. Gallego,
7
M. Pacheco-Romero,
8,9
D. Alcaraz-Segura,
9,10,11
and J. M. Paruelo
1,2,3,7
1
Laboratorio de Ana
´lisis Regional y Teledeteccio
´n (LART), IFEVA, Facultad de Agronomı
´a, Universidad de Buenos Aires, CONICET,
Buenos Aires, Argentina;
2
Departamento de Me
´todos Cuantitativos y Sistemas de Informacio
´n, Facultad de Agronomı
´a, Universidad
de Buenos Aires, Buenos Aires, Argentina;
3
Instituto Nacional de Investigacio
´n Agropecuaria, INIA La Estanzuela, Ruta 50 km 11,
Colonia, Uruguay;
4
Ca
´tedra de Ecologı
´a, Facultad de Agronomı
´a, Universidad de Buenos Aires, Buenos Aires, Argentina;
5
Instituto de
Matema
´tica Aplicada San Luis, Universidad Nacional de San Luis and CONICET, San Luis, Argentina;
6
Grupo de Estudio de
Agroecosistemas y Paisajes Rurales, Universidad Nacional de Mar del Plata, CONICET, Balcarce, Argentina;
7
Grupo de Ecologı
´ade
Pastizales, Instituto de Ecologı
´a y Ciencias Ambientales, Universidad de la Repu
´blica, Montevideo, Uruguay;
8
Departamento de
Biologı
´a y Geologı
´a, Universidad de Almerı
´a, Almerı
´a, Spain;
9
Centro Andaluz para la Evaluacio
´n y Seguimiento del Cambio Global,
Universidad de Almerı
´a, Almerı
´a, Spain;
10
Departamen de Bota
´nica, Facultad de Ciencias, Universidad de Granada, Av. Fuente-
nueva s/n. 18071, Granada, Spain;
11
iecolab, Interuniversitary Institute for Earth System Research (IISTA), University of Granada, Av.
del Mediterra
´neo s/n. 18006, Granada, Spain
ABSTRACT
Sustainability science recognizes the importance of
the integrated assessment of the ecological and social
systems in land-use planning. However, most stud-
ies so far have been conceptual rather than empiri-
cal. We developed a framework to characterize the
social-ecological systems heterogeneity according to
its functioning through the identification of social–
ecological functional types (SEFT). The SEFT
framework builds on the plant, ecosystem and agent
functional type approaches, taking a step forward to
integrate the dimensions of social–ecological sys-
tems into an operational product to characterize
administrative units in a hierarchical way. To illus-
trate this novel framework, we described the
heterogeneity of SEFT in the Argentine Chaco by
clustering administrative entities. This area is a glo-
bal deforestation hotspot and has diverse social ac-
tors that harness ecosystem services in multiple, and
sometimes contrasting and conflictive, ways which
determines an urgent need for land-use planning.
We combined data from national census and remote
sensing to identify SEFT by clustering census tracts
based on 17 input variables that integrate key hu-
man, ecological and interaction processes across
landscapes. We identified three classes and eight
subclasses of SEFT. Ecological variables defined the
first level of heterogeneity (classes), while human
variables and the variables of interactions between
the human and ecological components defined a
second level of heterogeneity (subclasses). The de-
gree of anthropization and mean annual productiv-
ity were important variables to explain the first two
axes in the ordination (32% of the total variance).
This framework offers a conceptually novel and
comprehensive approach to understand the spatial
heterogeneity of social–ecological systems func-
tioning, which could play a pivotal role to support
conservation or land-use planning in rural areas.
Received 10 January 2019; accepted 21 June 2019
Electronic supplementary material: The online version of this article
(https://doi.org/10.1007/s10021-019-00415-4) contains supplementary
material, which is available to authorized users.
Author’s Contribution MV, SA, GB, MEM, FG and JMP conceived of
or designed the study; MPR and DAS contributed new methods or
models; MV analyzed data, performed research and wrote the paper and
SA, GB, MEM, FG, MPR, DAS and JMP participated in the wring process.
*Corresponding author; e-mail: vallejos@agro.uba.ar
Ecosystems
https://doi.org/10.1007/s10021-019-00415-4
Ó2019 Springer Science+Business Media, LLC, part of Springer Nature

Key words: social–ecological systems; land-use
planning; remote sensing; functional types; hier-
archical analysis.
HIGHLIGHTS
Social–ecological functional types (SEFT) are
administrative units that share a similar social–
ecological functioning
SEFT were defined by clustering administrative
units based on variables that integrate key
human ecological and interaction processes
across landscapes
The degree of anthropization and mean annual
productivity were the variables that captured
most of the spatial variability of SEFT
Ecological variables (for example primary pro-
ductivity) defined a first level of heterogeneity
and human variables a second one
INTRODUCTION
Solutions to environmental problems increasingly
require systemic perspectives that integrate key
aspects of the structure and function of ecological
and social systems (Holling 2001). Although the
full complexity of systems can never be character-
ized by maps (Hamann and others 2015), the def-
inition of relatively homogeneous units in terms of
its social–ecological functioning can contribute to
land-use planning by spatially characterizing social
and ecological dynamics together (Martı´n-Lo
´pez
and others 2017). Traditionally, rural planning has
characterized the spatial heterogeneity of land-
scapes through land zoning (that is, the delimita-
tion of homogeneous areas) mainly based on
biophysical properties (FAO 1996). Thus, human
dimensions were seldom considered and biophysi-
cal properties were generally related to ecosystems
structural attributes, whereas functional ones gen-
erally received less attention (Guerry and others
2015). The explicit spatial integration of these ne-
glected properties of social–ecological systems could
be an operational tool for understanding complex
systems and defining evidence-based policies for
sustainable development. This study develops a
conceptual framework to map social–ecological
systems according to their functioning. Using the
Argentine Chaco as a case study, we applied this
framework in order to identify social–ecological
functional types (SEFT) by clustering the smallest
possible administrative unit at two nested levels.
This area is a global deforestation hotspot and has
diverse social actors that harness ecosystem services
in multiple, and sometimes contrasting and con-
flictive, ways which determines an urgent need for
land-use planning.
FUNCTIONAL FRAMEWORK TO CLASSIFY
SOCIAL–ECOLOGICAL SYSTEMS
Functional characterizations are increasingly being
considered a central aspect for sustainable man-
agement (Oliver and others 2015). According to
Jax (2010), the functioning of a system refers to its
‘performance’ or ‘behavior,’ which is regulated by
different mechanisms and interactions at multiple
spatial and temporal scales. Previous studies have
characterized functional types for ecological and
human systems separately. For ecological systems,
plant functional types were defined as groups of
plants with a similar response to environmental
conditions or with a similar effect on ecosystems
(Walker 1997), so its use has spread, for example,
to predict the responses of vegetation to global
change (Bonan and others 2002). In a higher level
of organization, ecosystem functional types were
defined as ecosystems that regardless of their
structure and composition have similar matter and
energy dynamics (Paruelo and others 2001), and
were useful for capturing the spatial and temporal
heterogeneity of ecosystem functioning, as well as
for improving the performance of general atmo-
sphere circulation models (Mu
¨ller and others
2014). For human systems, agent functional types
were theoretically defined as groups of agents with
a similar role and behavior regarding decision
making related to land-use change (Arneth and
others 2014). Although the usefulness of functional
approaches for different levels of organization is
widely acknowledged, a framework for classifying
social–ecological systems according to its function-
ing, to our knowledge, has not been developed yet.
As plant species, ecosystems and agents can be
grouped according to common functional charac-
teristics, social–ecological systems should be too.
Social–ecological systems are understood as sys-
tems where ecological and social components are
strongly coupled through multiple mechanisms of
interaction (Ostrom 2009). The social component
benefits from the services provided by the ecosys-
tem and, in turn, human agency modifies—directly
or indirectly—the functioning and structure of
ecosystems (Berkes and others 2003). As social–
ecological systems cannot be identified or found in
M. Vallejos and others

nature (they are defined by the observer under a
certain conceptual framework or question), their
spatial limits are strongly dependent on the specific
perspectives of the observer. Tracing these bound-
aries can be done in many different ways depend-
ing on the problem to be addressed and the data
availability (Martı´n-Lo
´pez and others 2017). Al-
though natural scientists frequently use landscape
units or pixels as entities, social scientists focus on
the individuals, households, farms or communities
as units of analysis and demographers or decision
makers base their analyses on administrative units
to achieve better management efficiency.
Most studies on social–ecological systems in the
literature have been conceptual rather than
empirical (Herrero-Ja
´uregui and others 2018).
However, there have been some efforts to develop
approaches for mapping them by combining bio-
physical and social attributes. At the global scale,
Ellis and Ramankutty (2008) used data on human
population density, land use and land cover to
derive a classification of anthromes (that is,
anthropogenic biomes). More recently, Va
´clavı´k
and others (2013) integrated land-use intensity
indicators with underlying environmental and
socioeconomic factors to map global land system
archetypes. On a regional level, Alessa and others
(2008) identified social–ecological hotspots in the
Kenai Peninsula of Alaska, as areas where there
was a convergence between the social (human
perceptions of the biological value of the landscape)
and ecological (high net primary productivity)
space. Based on the ecosystem services framework,
Raudsepp-Hearne and others (2010) mapped the
supply of ecosystem service bundles (that is, groups
of services that appear together repeatedly) that
allowed to identify emerging social–ecological
dynamics across diverse landscapes in Quebec,
Canada. With a similar approach, Hamann and
others (2015) derived a classification based on the
direct use level of local ecosystem services by
households in South Africa, using national popu-
lation censuses. Although the spatial integration of
the human and ecological components has grown
in recent years, there is still a lack of operational
mapping approaches for characterizing the social–
ecological system from a functional perspective.
In this study, we adopted the social–ecological
system framework presented by the Resilience Al-
liance (2007, p. 8) as a basis to develop a more
detailed framework to characterize the functioning
of the social–ecological systems (see Figure 1). This
framework emphasizes the interdependence and
two-way feedbacks that exist between humans and
the environment. In this framework, the social–
ecological system is composed of three intercon-
nected components, which comprises two or three
key dimensions. Within the human component,
the population distribution is the way in which
people are spatially arranged in the territory,
whereas well-being and development integrate
aspects related to housing, work, income, educa-
tion, health status, governability and social con-
nections, among others (OECD 2015). Within the
ecological component, the natural capital refers to
the stock of natural resources, which can provide
people with free goods and services, while the
ecosystem functioning refers to the fluxes of energy
and matter that sustain the ecosystem over time
and space (Jax 2010). Within the interaction
component, the pressure on the environment re-
flects the impact of human activities on the
ecosystem; the ecosystem services are the benefits
people obtain from the natural environment (Mil-
lennium Ecosystem Assessment 2005), and the
territorial link reflects the degree of decoupling
between people and their surrounding environ-
ment (Liu and others 2007; Hamann and others
2015).
We define social–ecological functional types
(SEFT) as administrative units that share a similar
social–ecological functioning, that is, have similar
dynamics in terms of (a) the socioeconomic aspects
(human component), (b) biophysical aspects (eco-
logical component) and (c) the two-way interac-
tions between the previous subsystems
(interactions component). This approach can be
useful to explore complex systems, to understand
Figure 1. Conceptual diagram of the functional
framework for assessing social–ecological systems. Each
component (that is human, ecological and interactions) is
comprised of two or three dimensions (for example,
population distribution), which in turn should be
characterized by one or more variables (for example,
occupation of the territory) and its indicator (for
example, population density) [Adapted and modified
from Resilience Alliance (2007, p. 8)].
Social-Ecological Functional Types

the spatial heterogeneity of social–ecological sys-
tems, to provide better inputs for modeling and as a
tool to support conservation or land-use planning
(for example, guiding participatory processes to
regulate land use). The questions that guided this
research are: (1) How to characterize SEFT using
attributes of the social–ecological system function-
ing? (2) Which are the SEFT in the Argentine
Chaco at the census tract resolution? (3) How are
the attributes of social–ecological functioning cor-
related in the study area? (4) What are the spatial
patterns among the SEFT classes at different levels
of nesting?
METHODS
Study Area
The Humid and Dry Chaco ecoregions (hereafter,
Chaco, see location in Figure A.1—Appendix
A—Supplementary Material) (Olson and others
2001) of Argentina have a great biological and
productive diversity (Baldi and others 2015), cul-
tural richness (Leake 2008) and agricultural
potential (Murray and others 2016), as well as high
levels of poverty and inequality (Paolasso and
others 2012). The social actors that occupy this
region include large-scale capitalized farmers,
small-scale farmers (‘puesteros’), indigenous com-
munities and mennonite colonists, among others
(Baldi and others 2015). These actors interact with
the environment in multiple ways by capturing
ecosystem services and modifying the environment
(for example, deforestation for industrial agricul-
ture, silvopastures for livestock farming, selective
logging or hunting and gathering). The rapid con-
version of native forests to annual crops and pas-
tures makes the Chaco region a global deforestation
hotspot (Hansen and others 2013; Vallejos and
others 2015). This transformation brought an in-
crease in the production of commodities (for
example, crops, meat) for the foreign markets, but
at the same time is compromising the supply of
multiple ecosystem services (Grau and others 2005;
Paruelo and others 2011; Mastrangelo and Laterra
2015), creating social inequalities (Laterra and
others 2019) and increasing social conflicts (Redaf
2011;Ca
´ceres 2015; Aguiar and others 2016). De-
spite some legislation efforts (for example, Senado
yCa
´mara de Diputados de la Nacio
´n Argentina
2007, Law N°26,331), the territorial configuration
responds more to an economic rationality, than to
ecological, social and cultural criteria (Garcı´a Col-
lazo and others 2013; Aguiar and others 2018). The
rapid changes are affecting the configuration of the
territory, not only in the ecological aspects (natural
capital, ecosystem functioning), but also in the
human aspects (population distribution, well-being
and development) and in the interaction between
both (pressure on the environment, ecosystem
services and territorial link).
Data Sources and Preprocessing
The approach we followed to define SEFT consists
of integrating human, ecological and interaction
variables across rural landscapes to identify homo-
geneous administrative entities in terms of their
social–ecological functioning. The resolution of the
analysis was the ‘rural census tract,’ -that is, group
of farms—an administrative unit designed to carry
out the field surveys of agricultural censuses in
Argentina. These divisions are delimited according
to the way that population is distributed and settled
in rural areas (divisions contain an average of 300
households), and also by the accessibility and dis-
tance between households, achieving a mean area
of 180 km
2
. Census data at this resolution are the
finest and publicly available because the personal
data protection prohibits the distribution of census
data at finer resolutions in Argentina (Senado y
Ca
´mara de Diputados de la Nacio
´n Argentina 1968,
Law N°17,622). Because this study focused on
rural land-use planning, census tracts correspond-
ing to urban areas were eliminated. From 2188
rural census tracts, 22 were discarded due to
missing or incongruent information (for example,
area under irrigation larger than the total land), so
we retained 2166 units.
For the classification of SEFT, we selected 17
descriptive variables, following the functional
framework for assessing social–ecological systems
described above (adapted and modified from the
Resilience Alliance 2007, p. 8), that is, we selected
variables and indicators according to the available
data (Table 1) to characterize the dimensions of the
three components of the framework (human, eco-
logical and interactions). We gathered data from
multiple governmental open-access databases, the
‘National Census of Population, Households and
Housing’ (INDEC 2001) and the ‘National Agri-
cultural Census’ (INDEC 2002). To characterize the
ecosystem functioning, we used the 2002 seasonal
dynamics of the enhanced vegetation index images
derived from MODIS 005.MOD13Q1 product (16-
day and 232 m resolution). Mean annual EVI and
the intra-annual coefficient of variation were used
as indicators of the primary productivity and its
seasonality (difference in C gains between the
growing and non-growing seasons), respectively.
M. Vallejos and others

These two indicators are descriptors of ecosystem
functioning (Alcaraz-Segura and others 2006; Vo-
lante and others 2012) and have also been suc-
cessfully used to quantify the provision of
ecosystem services (Paruelo and others 2016). All
variables where aggregated at the census tract res-
olution. We omitted the use of data from last Na-
tional Agricultural Census of Argentina (2008), as
Table 1. Variables Used for the Definition of Socio-ecosystem Functional Types at the Census Tract (CT)
Resolution
Component Dimension Variable Indicator Metric Abbreviation Data source
Human Population
distribu-
tion
Occupation
of territory
Population den-
sity
#Inhabitants/km
2
Pop INDEC (2001)
Well-being
and devel-
opment
Quality of life Poverty Unsatisfied basic needs Pov INDEC (2001)
Employment Permanent labor #Permanent workers/
1000 ha
Lab INDEC (2002)
Educational
infrastruc-
ture
Number of
schools
#Schools/km
2
School Ministerio de
Educacio
´n,
Argentina
(2012)
Ecological Natural cap-
ital
Natural cover Remnant native
forest
Native forest area/CT
area (%)
Nf INDEC (2002)
Ecosystem
function-
ing
Productivity Primary produc-
tivity (mean)
Annual mean of EVI
(year 2002)
PPm MOD13Q1-
EVI - 2002
Seasonality Primary produc-
tivity (coeffi-
cient of
variation)
Annual CV of EVI
(year 2002)
PPs MOD13Q1-
EVI - 2002
Interactions Pressure on
the envi-
ronment
Cultural
practices
Surface with
irrigation
Irrigated area/km
2
Irrig INDEC (2002)
Machinery Number of trac-
tors
#Tractors/1000 ha Tract INDEC (2002)
Livestock
intensity
Stocking rate Livestock unit (cattle
equivalent)/live-
stock area (ha)
Cattle INDEC (2002)
Loss of natu-
ral cover-
age
Natural cover
transformation
Deforested area/CT
area (%)
Defor LART (2012)
Ecosystem
services
Agricultural
production
Annual crops
area
Annual crops area/CT
area (%)
Crops INDEC (2002)
Livestock
industrial
production
Forage crops
area
Forage crops area/CT
area (%)
Forage INDEC (2002)
Cattle breed-
ing activity
Cow pregnancy #Pregnant cows/Total
#cows (%)
Calf INDEC (2002)
Territorial
link
Access Transport net-
work connec-
tivity
Km of roads/km
2
Roads Instituto Geo-
gra
´fico Na-
cional (2015)
Type of man-
agement
Legal type of
farmer
Area with legal type of
farmer ‘Physical
Person’/CT area (%)
Farmer INDEC (2002)
Land tenure Tenure regime Area with land tenure
regime ‘Owner’/CT
area (%)
Tenure INDEC (2002)
Social-Ecological Functional Types

it turned out to be unreliable due to a non-ex-
haustive survey (Giarracca and others 2008).
Statistical Analyses
We explored the Pearson correlation between the
17 variables to analyze the sign and magnitude of
the relationships between them. We standardized
the variables and performed a hierarchical cluster-
ing analysis using Ward’s method (Ward 1963)to
identify SEFT classes and subclasses, in two nested
levels of detail. Hierarchical clustering is useful to
recognize discontinuities in the dataset of multiple
variables, where the units of inferior-ranking
clusters (subclasses) are members of larger and
higher ranking (classes) (Legendre and Legendre
1998). Then, we performed a principal component
analysis to understand the multivariate structure of
data and to find which of the variables are more
important to describe the heterogeneity of the data
between the census tracts (Hair and others 2010).
Once the SEFT subclasses were defined, we map-
ped them using QGIS development team (2016).
Finally, we described each SEFT class and subclass
in terms of the indicators conforming each class.
Data analysis was done with R Core Team (2015),
using the following packages: ‘cluster’ (Maechler
and others 2015), ‘ggplot2’ (Wickham 2009), ‘cor-
rplot’ (Taiyun Wei 2013), ‘rgdal’ (Bivand and oth-
ers 2015), ‘raster’ (Hijmans 2015) and ‘sp’
(Pebesma and Bivand 2005).
RESULTS
Correlation Between Variables
Regarding the correlation between variables, of the
136 possible pairwise combinations, 75% were
significant (p<0.01). Of these, 66% presented
positive relationships, while the remaining 34%
were negative. The most significant positive corre-
lation was the occupation of the territory (Pop) and
the machinery (Tractors)(r= 0.63). The most sig-
nificant negative correlations were between the
natural cover (Nf) and the agricultural production
(Crops), and between the productivity (PPm) and its
seasonality (PPs)(r=-0.40, both) (Figure 2).
SEFT Classification
Rural census tracts were classified into three classes
and eight subclasses of SEFT by using two cutting
levels in the hierarchical clustering analysis. Class
1, 2 and 3 (defined by the first level) are composed
by 550, 634 and 982 census tracts, respectively.
Subclasses 1a, 1b, 1c, 2a, 2b, 3a, 3b and 3c (defined
by the second nested level) are composed by 157,
136, 257, 332, 302, 95, 252 and 635 census tracts,
respectively (Figure 3). The principal component
(PC) 1, (21.6% of the total variance), corresponds
to an axis that varies from less-transformed census
tracts with greater remnant native forests to more
transformed tracts with higher population density.
The PC 2, (10.9% of the total variance), corre-
sponds to a gradient that goes from tracts with
greater livestock pressure and higher mean annual
productivity to tracts with greater agricultural
activity and greater seasonality (Figure 4). Principal
component analysis showed that seven axes were
needed to explain 75% of the spatial variability
between the census tracts (see complete analysis in
Table B.1—Appendix B—Supplemental Material).
SEFT Mapping
We mapped the social–ecological functional types
in the Argentine Chaco at the census tract resolu-
tion (Figure 5). Not all the census tracts are repre-
sented on the map because some tracts were
discarded from the analysis (lack or inconsistency
of data). We also described SEFT classes and sub-
classes in terms of the variables used for their
classification (see Figure C.1 and C.2—Appendix
B—Supplemental material).
Class 1 (Agricultural functioning systems) occupied
census tracts in the western and central area of the
study region, coinciding with foci of agricultural
advance in the margins of the Dry Chaco, where
rainfall is greater. This class covers an area of
62,000 km
2
, and it has, on average, high popula-
tion density (Pop), high levels of well-being and
development (Schools &Roads), high land-use
intensity (Crops,Forage,Defor,Irrig &Tract), high
employment (Lab), high seasonality in the pro-
ductivity (PPs) and low natural cover (Nf). Subclass
1a (Small-scale intensive agriculture) was located
mainly in small-sized census tracts located in the
province of Tucuma
´n, where the sugarcane is the
predominant crop. This subclass had the highest
population density (Pop), employment (Lab),
machinery (Tract) and agricultural production
(Crops). Subclass 1b (Traditional middle-scale agricul-
ture) was located mainly in the central southwest-
ern region of the province of Chaco, where
immigrant settlers deforested for cotton production
in the mid-twentieth century and were then con-
verted to the soybean production. This subclass has
the highest loss of natural cover in the region (De-
for) in the region. Subclass 1c, (Expanding agri-
business), was located in new foci of agricultural
expansion in the eastern and western Dry Chaco.
M. Vallejos and others

Figure 2. Correlation analyses between descriptive variables. The correlation coefficients are represented by ellipses (the
more pronounced the shape of the ellipse, the greater the correlation) and colors (degrees of blue for positive correlations
and degrees of red for negative correlations, the legend for correlation coefficients is at the left). The correlation
significance is also shown (*are not significant correlations, pvalue >0.01).
Figure 3. Hierarchical cluster analysis to identify social–ecological functional types. Three classes and eight subclasses
were identified in the Argentine Chaco at the census tract resolution.
Social-Ecological Functional Types

This was the subclass with the highest livestock
industrial production (Forage) and the greatest
irrigated area (Irrig) of the whole region.
Class 2 (Extensive stockbreeding functioning systems)
is located mainly in the eastern portion of the study
region, coinciding with the Humid Chaco, where
agricultural aptitude is low due to the water surplus
during the rainy season, and the predominance of
low and flooded land. This class covers an area of
124,000 km
2
, and has, on average, high produc-
tivity (PPm), low seasonality (PPs), low deforesta-
tion rates (Defor), high proportion of cattle breeding
activities (Calf) and high stocking rates (Cattle).
Subclass 2a (Cattle breeding dominance) was located
in the eastern and southern part of the Humid
Chaco, where the floods are mostly of fluvial origin
(Paraguay and Parana
´Rivers floodplains). This
subclass had the highest proportion of cattle
breeding activities (Calf) of the whole region. Sub-
class 2b (Livestock farming dominance) was located in
the northern and eastern part of the Humid Chaco.
This subclass had the highest stocking rate (Cattle)
of the whole region. Within Class 2, it is observed
that subclass 2a has lower primary productivity
mean (PPm) and higher seasonality (PPs) than
subclass 2b, because in 2a, there is naturally more
pasture than in 2b. The deforestation rates are
lower in part because the amount of native forests
is lower too.
Class 3 (Forest functioning systems) occupy most of
the western zone of the study region, coinciding
with the central and southern portion of the Dry
Chaco, where rainfalls are low and the lands are
marginal for agriculture. This class covers an area of
340,000 km
2
, and has, on average, high native
forest area (Nf), low productivity (PPm), low pres-
sure on the environment (Irrig, Tract, Cattle, Defor),
low population density (Pop), low levels of well-
being and development (Pov, Lab, Schools) and low
access (Roads). Subclass 3a (Subsistence activities
dominance) was located in census tracts with pre-
dominance of indigenous communities or small
Figure 4. Principal component analysis (PCA) for the eight subclasses of SEFT. The first axis explains 21.6% of the
variability of the data, whereas the second axis, 10.9%. Pop: population density; Pov: unsatisfied basic needs (poverty);
Lab: permanent labor; School: number of schools; Nf: native forest area; PPm: mean net annual primary productivity; PPs:
annual seasonality of net primary productivity; Irrig: irrigated area; Tract: number of tractors per area; Cattle: stocking rate;
Defor: natural cover transformation; Crops: annual crops area; Forage: forage crops area; Calf: % cow pregnancy; Roads:
roads density; Farmer: legal type of farmer; Tenure: tenure regime.
M. Vallejos and others

farmers. This subclass had the lowest dominance of
physical person as legal type of farmer (Farmer).
This means that other legal types of farmer are
dominant (for example, societies, cooperatives,
nonprofit institutions, national public entities).
Subclass 3b (Low intensity livestock dominance) was
located in census tracts in the southern part of the
Dry Chaco, corresponding with the Arid Chaco.
This subclass had the lowest productivity (PPm)in
the whole region, and the highest proportion of
cattle breeding activity (Calf) of Class 3. Subclass 3c
(Incipient agriculture) was located in census tracts in
the northern part of the Dry Chaco, where agri-
culture is expanding. This subclass had the highest
productivity (PPm) and agricultural production
(Crops) within Class 3, but also has the highest le-
vels of structural poverty in the whole region (Pov).
DISCUSSION
Mapping SEFT allowed us to characterize and
understand the heterogeneity of the social–eco-
logical systems in the Chaco. We identified three
classes and eight subclasses, in a nested level of
Figure 5. Social–ecological functional types in the Argentine Chaco at census tract resolution (corresponding to the year
2002). Eight subclasses of SEFT are shown in the map: 1a, 1b, 1c, 2a, 2b, 3a, 3b and 3c. The polar diagrams at the right of
the figure illustrate the contribution of the 17 variables for each subclass of SEFT. All the variables are scaled from 0 to 1.
Variables: 1. Pop (population density); 2. Pov (poverty); 3. Lab (permanent labor); 4. School (number of schools); 5. Nf
(native forest area); 6. PPm (mean annual primary productivity); 7. PPs (annual seasonality of primary productivity); 8.
Irrig (irrigated area); 9. Tract (number of tractors per area); 10. Cattle (stocking rate); 11. Defor (natural cover
transformation); 12. Crops (annual crops area); 13. Forage (forage crops area); 14. Calf (% cow pregnancy); 15. Roads
(roads density); 16. Farmer (legal type of farmer); 17. Tenure (land tenure regime).
Social-Ecological Functional Types

detail at the census tract resolution. These classes
represent relatively homogeneous units in terms of
its social–ecological functioning. Our results
showed that ecological variables defined the first
level of heterogeneity (classes), while human and
interaction variables defined a second level of
heterogeneity (subclasses). This supports not only
the importance of considering the human dimen-
sion when zoning the territory but also suggest a
hierarchy of the controls that determine the spatial
distribution of SEFT.
Class 1 is associated with Agricultural functioning
systems, where there is a high pressure on the
environment, high seasonality of the productivity
and intermediate levels of well-being. Class 2 is
associated with Extensive cattle functioning systems,
where there is a high livestock production, a lower
pressure on the environment in relation to the first
class, high mean annual productivity and low sea-
sonality of the productivity. Class 3 is associated
with Forest functioning systems, where there is a high
native forest area, high structural poverty, pre-
dominance of de-capitalized producers and low
mean annual productivity. Class 2 and 3 are clearly
associated with biogeographical areas (Humid and
Dry Chaco, respectively), while Class 1 is inter-
spersed mainly over Class 3, reflecting the advance
of agriculture mainly in the Dry Chaco. In this re-
gion, agriculture is expanding over areas with less
suitability, and land clearing dynamics are associ-
ated with the proximity to already cleared areas,
defining a frontier-advancement pattern which
suggests the prevalence of a contagion process
(Volante and others 2016). In the last two decades,
agriculture has drastically expanded into the Chaco
ecoregion, due to favorable political and economic
factors in Argentina, increasing soybean prices and
new genetically modified varieties, between other
aspects (Pengue 2005;Ca
´ceres 2015; Piquer-Ro-
drı´guez and others 2018). Although it was not
possible to capture the current state of social–eco-
logical systems in this study, we can assume that
the expanding agri-business (that is, Subclass 1c)
has spread out in many census tracts of the study
area in the present.
The degree of anthropization and mean annual
productivity were important variables to explain
the first two axes in the ordination (32% of the
total variance). Biophysical and social variables
showed a low-to-moderate level of covariation,
evidencing their high complementarity for the
definition of the SEFT classes. In fact, the seven of
axes needed to explain 75% of the spatial vari-
ability in the principal component analysis reveals
the complex and multidimensional nature of so-
cial–ecological systems. Although in this case the
reduction in dimensionality was not possible, we
support the use of multivariate analyses in order to
understand the structure of the data and, hence,
the importance of each variable for the definition of
SEFT.
When studying the correlation of variables at the
census tract resolution, we observed that census
tracts located in disadvantaged environments (that
is, with low productivity) were also those that
showed less pressure on the environment (low
population density, technification level and natural
cover transformation). Moreover, we observed that
in these areas, where agri-business agriculture is
not feasible and the natural cover has not been
replaced, the levels of well-being and development
are low (high poverty, low employment and edu-
cational infrastructure). This reflects a clear con-
nection between land quality and poverty,
suggesting a long-term ‘accumulation by dispos-
session’ process (Harvey 2004;Ca
´ceres 2015),
where smallholders are on poor land because they
have been expelled by large farmers. This process of
marginalization of the less capitalized stakeholders
in lower-quality lands entails a potential risk of
poverty traps (Aghion and Durlauf 2005), as mar-
ginal producers make use of the surrounding forest
intensifying the degradation of natural resources,
which in turn feeds back the level of poverty
(Duraiappah 1998). This situation is difficult to
overcome without an appropriate redistribution of
wealth, and the implementation of incentives for
marginalized lands. The results presented here
could be used to identify the areas prone to poverty
traps to apply specific interventions oriented to-
ward promoting rural development and halting
environmental degradation.
The identification of SEFT using administrative
units as entities is useful for land-use planning
because it represents the space where interests and
problems relevant to local stakeholders connect
with decision makers (Martı´n-Lo
´pez and others
2017). Studying social and ecological aspects to-
gether at lower resolutions (for example, land-
holding level) also matters because of its direct link
with the unit in which stakeholders make deci-
sions. This would require the collection of primary
data from interviews and field surveys, and the use
of high spatial resolution satellite information. The
choices over scale, extent and resolution critically
affect the type of patterns that are observed, be-
cause patterns that appear at one level of resolution
or extent may be lost at lower or higher scales
(Gibson and others 2000). So, the observed pat-
terns cannot be extrapolated to other scales
M. Vallejos and others

(Peterson and others 1998). Complex systems must
be analyzed and managed by performing a simul-
taneous analysis at various scales and approaches
(Viglizzo and others 2005; Cumming and others
2006). At a regional scale, coarse resolutions are
enough to answer questions about the general
patterns of interaction between social and natural
systems, but to study smaller geographic areas,
more detailed resolutions will be required (Gibson
and others 2000).
Despite the complexity of social–ecological sys-
tems, it is possible to assess them in a degree of
simplicity necessary for understanding, but also
with the required complexity to develop policies for
sustainability (Holling 2001). Nevertheless, con-
ceptualizing the systems functioning is a matter of
both scientific knowledge and values. The defini-
tion of the system limits, the selection of variables
and the classification methods are always depen-
dent on the observer and his/her specific interests
or problems to be tackled. Depending on the
objectives and resolution of the study other criteria
for the selection of variables may be used. Once the
classification has been conducted and distinct
modes of functioning were identified, the question
of which of those modes of functioning is prefer-
able (or desirable) is still a matter of evaluation by
human observers (Theobald and others 2005).
Each mode of functioning can be considered as
‘proper functioning’ for its kind of system and state,
or not. Each productive system has ‘winners’ and
‘losers’ among the parts involved and impacted,
and also in terms of the services that human may
derive from the system (Jax 2010). Although the
definition of SEFT is not absolutely neutral (since
the choice of the unit to be classified, the variables
for its classification and also the number of classes
are made by humans), the advantage of using SEFT
for understanding the social–ecological hetero-
geneity of the territory lies in the transparency of
the process. Therefore, it could be a useful tool to
guide participatory processes in land-use planning
and to define sustainable policies in more trans-
parent and integrative ways (for example, seeking
for consensus with regard to the expansion of
particular activities in a SEFT class), as a function of
the overall performance or functioning of the sys-
tem.
Land-use planning implies a public policy that
must reconcile the process of economic develop-
ment and the conservation of natural capital
through the regulation of land use, with the ulti-
mate goal of increasing the human well-being and
the equity in the distribution of cost and benefits
associated with land use. The identification and
mapping of SEFT constitutes a social–ecological-
based zoning process, where territorial units have a
similar behavior or functioning in terms of its so-
cial–ecological vulnerability, adaptability and resi-
lience (Chapin and others 2009). The framework
developed here might help decision makers to
understand the social–ecological context to estab-
lish restrictions, incentives or to promote an effec-
tive spatial arrangement of activities. For example,
the area of the triple border of the Santiago del
Estero, Chaco and Santa Fe provinces shares the
same biophysical characteristics. However, provin-
cial boundaries define different SEFT due to dif-
ferences in the human aspects or the interaction
between the biophysical and ecological compo-
nents (Figure 5). Whereas in eastern Santiago del
Estero SEFTs associated with an expanding agri-
business were dominant, in western Chaco domi-
nant SEFT were associated with more traditional
middle-scale agriculture. In northern Santa Fe, in
contrast, dominant SEFT were associated with
subsistence activities or cattle breeding. All these
SEFT present contrasting socio-ecological arrange-
ment and, consequently, different potential roles or
functions in the region (for example, providing
commodities for exports, meat or areas for conser-
vation). The presences of distinct SEFT would also
indicate differences in stakeholders and actual or
potential conflicts. Such discrimination clearly
indicates the need of different intervention policies.
In short, the use of SEFT facilitates a more inclusive
understanding of the territory in a comprehensive
and transparent framework and could be useful to
identify areas within which to develop similar
management policies in accordance with the
objectives of the land-use planning. Improving ac-
cess to scientific information could help decision
makers anticipate potential consequences of rural
land-use change and in doing so, avoid unintended
ecological and social effects. Finally, the developed
framework can be applied to understand the
heterogeneity, complexity and trends of social–
ecological systems in other regions of the world.
ACKNOWLEDGEMENTS
This research was supported by Consejo Nacional
de Investigaciones Cientı´ficas y Te
´cnicas (Ar-
gentina), Universidad de Buenos Aires (Argentina).
This work was carried out with the aid of a grant
from the Inter-American Institute for Global
Change Research (IAI) CRN III 3095, which is
supported by the US National Science Foundation
(Grant GEO-1128040), and an associated comple-
mentary project financed by CONICET. The project
Social-Ecological Functional Types

CGL2014-61610-EXP, which is supported by Di-
reccio
´n General de Investigacio
´n Cientı´fica y Te
´c-
nica, Ministerio de Economı´a y Competitividad
(Spain) also provided fundings for the research. We
would also like to thank anonymous referees and
the editors, who have made valuable comments to
a previous version of the manuscript, and to Dr.
Ignacio Gasparri who also made useful comments.
REFERENCES
Aghion P, Durlauf SN. 2005. Handbook of Economic Growth.
Amsterdam: North Holland.
Aguiar S, Texeira M, Paruelo JM, Roma
´n ME. 2016. Conflictos
por la tenencia de la tierra en la provincia de Santiago del
Estero.Su relacio
´n con los cambios en el uso de la tierra. In:
Transformaciones agrarias argentinas durante las u´ ltimas
de
´cadas: una visio
´n desde Santiago del Estero y Buenos Aires.
FAUBA. Buenos Aires. pp 199–225.
Aguiar S, Mastrangelo ME, Garcı´a-Collazo MA, Camba-Sans G,
C.E. M, Ciuffoli L, Schmidt M, Vallejos M, Langbehn L, Ca
´-
ceres D, Merlinsky G, J.M. P, Seghezzo L, Staiano L, Texeira
M, Volante J, Vero
´n S. 2018. ¿Cua
´l es la situacio
´n de la Ley de
Bosques en la Regio
´n Chaquen
˜a a diez an
˜os de su sancio
´n?
Revisando su pasado para discutir su futuro. Ecologı´a Austral,
en prensa.
Alcaraz-Segura D, Paruelo JM, Cabello J. 2006. Identification of
current ecosystem functional types in the Iberian Peninsula.
Global Ecology and Biogeography 15:200–12.
Alessa L, Kliskey A, Brown G. 2008. Social-ecological hotspots
mapping: A spatial approach for identifying coupled social-
ecological space. Landscape and Urban Planning 85:27–39.
Arneth A, Brown C, Rounsevell MDA. 2014. Global models of
human decision-making for land-based mitigation and adap-
tation assessment. Nature Climate Change 4:550–7.
Baldi G, Houspanossian J, Murray F, Rosales AA, Rueda CV,
Jobba
´gy EG. 2015. Cultivating the dry forests of South
America: Diversity of land users and imprints on ecosystem
functioning. Journal of Arid Environments 123:47–59.
Berkes F, Folke C. 1998. Linking social and ecological systems:
management practices and social mechanisms for building
resilience. New York: Cambridge University Press.
Berkes F, Colding J, Folke C. 2003. Navigating Social-Ecological
Systems. Cambridge: Cambridge University Press.
Bivand R, Keitt T, Rowlingson B. 2015. Rgdal: Bindings for the
Geospatial Data Abstraction Library. R package version 1.0-6.
http://CRAN.R-project.org/package=rgdal. Accessed 1 July
2019.
Bonan GB, Levis S, Kergoat L, Olson KW. 2002. Landscapes as
patches of plant functional types: An integrating concept for
climate and ecosystem models. Global Biogeochemical Cycles
2 16:1–18.
Ca
´ceres DM. 2015. Accumulation by Dispossession and Socio-
Environmental Conflicts Caused by the Expansion of
Agribusiness in Argentina. Journal of Agrarian Change
15:116–47.
Chapin FS, Folke C, Kofinas GP. 2009. A Framework for
Understanding Change. In: Principles of Ecosystem Steward-
ship. New York, NY: Springer New York. pp 3–28.
Cumming GS, Cumming DHM, Redman CL. 2006. Scale Mis-
matches in Social-Ecological Systems: Causes, Consequences,
and Solutions. Ecology and Society 11:1–14.
Duraiappah AK. 1998. Poverty and environmental degradation:
A review and analysis of the nexus. World Development
26:2169–79.
Ellis EC, Ramankutty N. 2008. Putting people in the map:
anthropogenic biomes of the world. Frontiers in Ecology and
the Environment 6:439–47.
FAO. 1996. Agro-ecological zoning guidelines. Rome: Food and
Agriculture Organization of the United Nation. Soil Resources,
Management and Conservation Service.
Garcı´a Collazo MA, Panizza A, Paruelo JM. 2013. Ordenamiento
territorial de bosques nativos: Resultados de la zonifcacio
´n
realizada por provincias del norte Argentino. Ecologia Austral
23:97–107.
Giarracca N, Teubal M, Palmisano T. 2008. Paro agrario: cro
´nica
de un conflicto alargado. Realidad Econo
´mica 237:33–54.
Gibson CC, Ostrom E, Ahn TK. 2000. The concept of scale and
the human dimensions of global change: a survey. Ecological
Economics 32:217–39.
Grau HR, Aide TM, Gasparri NI. 2005. Globalization and Soy-
bean Expansion into Semiarid Ecosystems of Argentina.
AMBIO: A Journal of the Human Environment 34:265–6.
Guerry AD, Polasky S, Lubchenco J, Chaplin-Kramer R, Daily
GC, Griffin R, Ruckelshaus M, Bateman IJ, Duraiappah A,
Elmqvist T, Feldman MW, Folke C, Hoekstra J, Kareiva PM,
Keeler BL, Li S, Mckenzie E, Ouyang Z, Reyers B, Ricketts TH,
Rockstro
¨m J, Tallis H, Vira B. 2015. Natural capital and
ecosystem services informing decisions: From promise to
practice. PNAS 112:7348–55.
Hair J., Black WC, Babin BJ, Anderson RE. 2010. Multivariate
Data Analysis. A Global Perspective. 7th Edition, Pearson
Education, Upper Saddle River.
Hamann M, Biggs R, Reyers B. 2015. Mapping social-ecological
systems: Identifying ‘green-loop’ and ‘red-loop’ dynamics
based on characteristic bundles of ecosystem service use.
Global Environmental Change 34:218–26.
Hansen MC, Potapov PV, Moore R, Hancher M, Turubanova SA,
Tyukavina A, Thau D, Stehman SV, Goetz SJ, Loveland TR,
Kommareddy A, Egorov A, Chini L, Justice CO, Townshend
JRG. 2013. High-resolution global maps of 21st-century forest
cover change. Science (New York, NY) 342:850–3.
Harvey D. 2004. The New Imperialism: Accumulation by Dis-
possession. Socialist Register 40:63–87.
Herrero-Ja
´uregui C, Arnaiz-Schmitz C, Reyes MF, Telesnicki M,
Agramonte I, Easdale MH, Schmitz MF, Montes C, Aguiar M,
Go
´mez-Sal A. 2018. What Do We Talk about When We Talk
about Social-Ecological Systems? A Literature Review. Sus-
tainability 10:2950.
Hijmans RJ. 2015. Raster: Geographic Data Analysis and
Modeling. R package version 2.4-18. http://CRAN.R-project.
org/package=raster. Accessed 1 July 2019.
Holling CS. 2001. Understanding the Complexity of Economic,
Ecological, and Social Systems. Ecosystems 4:390–405.
INDEC. 2001. Censo Nacional de Poblacio
´n, Hogares y Viviendas
2001.
INDEC. 2002. Censo Nacional Agropecuario 2002.
Instituto Geogra
´fico Nacional. 2015. Red vial. http://www.ign.g
ob.ar/NuestrasActividades/InformacionGeoespacial/CapasSIG.
Accessed 1 July 2019.
M. Vallejos and others

Jax K. 2010. Ecosystem Functioning. Cambridge: Cambridge
University Press.
LART 2012. Laboratorio de Ana
´lisis Regional y Teledeteccio
´n.
‘Monitoreo de desmonte’ Project. http://monitoreodesmon
te.com.ar/. Accessed 1 July 2019.
Laterra P, Nahuelhual L, Vallejos M, Berrouet L, Arroyo Pe
´rez E,
Enrico L, Jime
´nez-Sierra C, Mejı´a K, Meli P, Rinco
´n-Ruı´z A,
Salas D, S
ˇpiric
´J, Villegas JC, Villegas-Palacio C. 2019. Linking
inequalities and ecosystem services in Latin America.
Ecosystem Services 36:100875.
Leake A. 2008. Los pueblos indigenas cazadores-recolectores del
Chaco Salten
˜o: poblacio
´n, economı´a y tierras. Salta, Argenti-
na: Fundacio
´n Asociana: Instituto Nacional de Asuntos
Indı´genas, Universidad Nacional de Salta.
Legendre P, Legendre L. 1998. Numerical ecology. Amsterdam:
Elsevier.
Liu J, Dietz T, Carpenter SR, Alberti M, Folke C, Moran E, Pell
AN, Deadman P, Kratz T, Lubchenco J, Ostrom E, Ouyang Z,
Provencher W, Redman CL, Schneider SH, Taylor WW. 2007.
Complexity of coupled human and natural systems. Science
(New York, NY) 317:1513–16.
Maechler M, Rousseeuw P, Struyf A, Hubert M, Hornik K. 2015.
cluster: Cluster Analysis Basics and Extensions. R package
version 2.0.3. http://CRAN.R-project.org/package=cluster..
Accessed 1 July 2019.
Martı´n-Lo
´pez B, Palomo I, Garcı´a-Llorente M, Iniesta-Arandia I,
Castro AJ, Garcı´a Del Amo D, Go
´mez-Baggethun E, Montes C.
2017. Delineating boundaries of social-ecological systems for
landscape planning: A comprehensive spatial approach. Land
Use Policy 66:90–104.
Mastrangelo ME, Laterra P. 2015. From biophysical to social-
ecological trade-offs: integrating biodiversity conservation and
agricultural production in the Argentine Dry Chaco. Ecology
and Society 20:art20.
Millennium Ecosystem Assessment. 2005. Ecosystems and hu-
man well-being: synthesis. In: Island Press; Washington D.C.,
USA. Oxford University Press, New York, USA.
Ministerio de Educacio
´n, Argentina 2012. (Data unpublished).
Mu
¨ller OV, Berbery EH, Alcaraz-Segura D, Ek MB, Mu
¨ller OV,
Berbery EH, Alcaraz-Segura D, Ek MB. 2014. Regional Model
Simulations of the 2008 Drought in Southern South America
Using a Consistent Set of Land Surface Properties. Journal of
Climate 27:6754–78.
Murray F, Baldi G, von Bernard T, Viglizzo EF, Jobba
´gy EG.
2016. Productive performance of alternative land covers along
aridity gradients: Ecological, agronomic and economic per-
spectives. Agricultural Systems 149:20–9.
OECD. 2015. How’s Life? 2015: Measuring Well-being. Paris:
OECD Publishing.
Oliver TH, Heard MS, Isaac NJB, Roy DB, Procter D, Eigenbrod F,
Freckleton R, Hector A, Orme CDL, Petchey OL, Proenc¸a V,
Raffaelli D, Suttle KB, Mace GM, Martı´n-Lo
´pez B, Woodcock
BA, Bullock JM. 2015. Biodiversity and Resilience of Ecosys-
tem Functions. Trends in Ecology & Evolution 30:673–84.
Olson DM, Dinerstein E, Wikramanayake ED, Burgess ND,
Powell GVN, Underwood EC, Damico JA, Itoua I, Strand HE,
Morrison JC, Loucks CJ, Allnutt TF, Ricketts TH, Kura Y, La-
moreux JF, Wettengel WW, Hedao P, Kassem KR. 2001.
Terrestrial Ecoregions of the World: A New Map of Life on
Earth. BioScience 51:933.
Ostrom E. 2009. A General Framework for Analyzing Sustain-
ability of Social-Ecological Systems. Science 325:419–22.
Paolasso P, Bolsi A, Gasparri I, Longhi F. 2012. La pobreza en el
nordeste argentino: cambios y persistencias durante la primera
de
´cada del siglo XXI. In: Estudios y contribuciones en home-
naje a la doctora Norma Cristina Meichtry. Resistencia: Ed.
Con Texto. p 111–136.
Paruelo JM, Jobba
´gy EG, Sala OE. 2001. Current Distribution of
Ecosystem Functional Types in Temperate South America.
Ecosystems 4:683–98.
Paruelo JM, Vero
´n SR, Volante JN, Seghezzo L, Vallejos M,
Aguiar S, Amdan L, Baldassini P, Ciuffolif L, Huykman N,
Davanzo B, Gonza
´lez E, Landesmann J, Picardi D. 2011.
Conceptual and Methodological Elements for Cumulative
Environmental Effects Assessment (CEEA) in Subtropical
Forests. The Case of Eastern Salta. Argentina. Ecologia Austral
21:163–78.
Paruelo JM, Texeira M, Staiano L, Mastra
´ngelo M, Amdan L,
Gallego F. 2016. An integrative index of Ecosystem Services
provision based on remotely sensed data. Ecological Indicators
71:145–54.
Pebesma EJ, Bivand RS. 2005. Classes and methods for spatial
data. R package version 3.1–1. http://CRAN.R-project.org/pac
kage=sp.
Pengue W. 2005. Transgenic crops en Argentina: the ecological
and social debt. Bulletin of Science, Technology and Society
25:322.
Peterson G, Allen CR, Holling CS. 1998. Ecological Resilience,
Biodiversity, and Scale. Ecosystems 1:6–18.
Piquer-Rodrı´guez M, Butsic V, Ga
¨rtner P, Macchi L, Baumann
M, Gavier Pizarro G, Volante JN, Gasparri IN, Kuemmerle T.
2018. Drivers of agricultural land-use change in the Argentine
Pampas and Chaco regions. Applied Geography 91:111–22.
QGIS Development Team. 2016. QGIS Geograhic Information
System. Open Source Gepatial Foundation Project.
R Core Team. 2015. R: A language and environment for statis-
tical computing. R Foundation for Statistical Computing,
Vienna, Austria. http://www.R-project.org/.
Raudsepp-Hearne C, Peterson GD, Bennett EM. 2010. Ecosys-
tem service bundles for analyzing tradeoffs in diverse land-
scapes. Proceedings of the National Academy of Sciences
107:5242–7.
Redaf. 2011. Conflictos sobre Tenencia de Tierra y Ambientales
en la Regio
´n del Chaco argentino. 2
o
Informe Datos relevados
hasta Agosto de 2010 Red Agroforestal Chaco Argentina
Observatorio de Tierras, Recursos Naturales y Medioambiente.
Resilience Alliance. 2007. Assessing and managing resilience in
social-ecological systems: Volume 2 supplementary notes to
the practitioners workbook.
Senado y Ca
´mara de Diputados de la Nacio
´n Argentina. 1968.
Ley 17.622 ‘Marco legal de las estadı´sticas oficiales’, Boletı´n
Oficial de la Repu´ blica Argentina.
Senado y Ca
´mara de Diputados de la Nacio
´n Argentina. 2007.
Ley N°26.331 ‘Presupuestos Mı´nimos de Proteccio
´n Ambi-
ental de los Bosques Nativos’, Boletı´n Oficial de la Repu´ blica
Argentina.
Theobald DM, Spies T, Kline J, Maxwell B, Hobbs NT, Dale VH.
2005. Ecological support for rural land-use planning. Ecolog-
ical Applications 15:1906–14.
Va
´clavı´k T, Lautenbach S, Kuemmerle T, Seppelt R. 2013.
Mapping global land system archetypes. Global Environmen-
tal Change 23:1637–47.
Vallejos M, Volante JN, Mosciaro MJ, Vale LM, Bustamante ML,
Paruelo JM. 2015. Transformation dynamics of the natural
Social-Ecological Functional Types

cover in the Dry Chaco ecoregion: A plot level geo-database
from 1976 to 2012. Journal of Arid Environments 123:3–11.
Viglizzo EF, Pordomingo AJ, Buschiazzo D, Castro MG. 2005. A
methodological approach to assess cross-scale relations and
interactions in agricultural ecosystems of Argentina. Ecosys-
tems 8:546–58.
Volante JN, Alcaraz-Segura D, Mosciaro MJ. 2012. Ecosystem
functional changes associated with land clearing in NW Ar-
gentina. Agriculture, Ecosystems & Environment 154:12–22.
Volante JN, Mosciaro MJ, Gavier-Pizarro GI, Paruelo JM. 2016.
Agricultural expansion in the Semiarid Chaco: Poorly selec-
tive contagious advance. Land Use Policy 55:154–65.
Walker BH. 1997. Functional types in non-equilibrium ecosys-
tems. In: Smith TM, Shugart HH, Woodward FI, editors. Plant
Functional Types: their Relevance to Ecosystem Properties
and Global Change. Cambridge. p 369.
Ward JH. 1963. Hierarchical Grouping to Optimize an Objective
Function. Journal of the American Statistical Association
58:236–44.
Wickham H. 2009. ggplot2: elegant graphics for data analysis.
New York: Springer. p 2009.
M. Vallejos and others