Patterns and causes of deforestation in the Colombian Amazon

Patterns and causes of deforestation in the Colombian Amazon

Dolors Armenteras a,b,*, Guillermo Rudas c, Nelly Rodriguez a,Sonia Sua a, Milton Romero a

a Biological Resources Research Institute Alexander von Humboldt, GIS Unit, Carrera 7#35-20, Bogota, Colombia (South America)b Department of Geography, King’s College London Strand, London WC2R 2LS, UK

c Department of Economics, Javeriana University, Calle 40 N 6-23, Bogota, Colombia (South America)

Accepted 29 March 2005

Abstract

Ecosystem information on the Colombian Amazonia is poor in comparison with that on the Brazilian Amazon. We examined

patterns of ecosystem diversity, deforestation and fragmentation and provided an estimate on their possible causes through a

temporal and spatial analysis of biotic and abiotic data using remote sensing and geographical information systems in six pilot

areas covering a total of 4,200,000 ha. Ecological, demographic and socio-economic data were analysed to establish the local

conditions. We used a landscape ecology approach to calculate indicators of ecosystem diversity, cover and forest fragmentation

such as number of patches, mean patch size, mean shape index and mean nearest neighbour distance. Patterns of deforestation

did not run parallel to access roads; instead the typical pattern of unplanned colonization follows the only transportation network

existing in many areas in the Colombian Amazonia: rivers. In addition, we have used indicators of human influence such as

demographic pressure, quality of life and economic activity indicators. Results show that the extent and rate of change varies

between areas depending on population density. Annual deforestation rates were 3.73 and 0.97% in the high population density

growth areas of Alto Putumayo and Macarena respectively, and 0.31, 0.23, and 0.01% in the relatively unpopulated areas of

indigenous population. These changes are related to land use history as well as to environmental and historical socio-economic

factors such as oil extraction, deforestation, cattle ranching or illegal cropping. The current situation in the region suggests that

tropical deforestation rates in the Colombian Amazon are substantially higher than those found in previous studies in the rest of

the Amazon.

# 2005 Elsevier Ltd. All rights reserved.

Keywords: Fragmentation; Satellite imagery; Tropical deforestation; Land use change; Biodiversity; Indicators; Amazonia; Colombia

This article is also available online at:www.elsevier.com/locate/ecolind

Ecological Indicators 6 (2006) 353–368

* Corresponding author. Tel.: +57 1 6086900x238;

fax: +57 1 6086900.

E-mail addresses: [email protected],

[email protected] (D. Armenteras),

[email protected] (G. Rudas).

1470-160X/$ – see front matter # 2005 Elsevier Ltd. All rights reserved

doi:10.1016/j.ecolind.2005.03.014

1. Introduction

The destruction of tropical forests has received

worldwide attention due to the significance of forest

on global climate, carbon sequestration, water cycles,

biodiversity and the potential global effects on climate

.

D. Armenteras et al. / Ecological Indicators 6 (2006) 353–368354

change (Fearnside, 1995; Fearnside et al., 2001).

Globally, some estimates suggest that 9 million km2 of

tropical humid forests have been lost in less than 50

years and that current rates of extinctions are not only

high but accelerating (Pimm et al., 2001). Achard et al.

(2002) estimate an annual deforestation rate of 0.38%

of humid tropical forest in Latin America.

Deforestation has led to the fragmentation of

natural ecosystems throughout the world causing

further loss of original forests, reduction of the size of

forest fragments and increasing isolation. Most studies

on ecosystem cover and fragmentation are centred on

the quantification of those changes (Vogelmann, 1995;

Ranta et al., 1998; Sierra, 2000; Steininger et al.,

2001a, 2001b) and on the effects that these can have on

ecological processes (Klein, 1989; Carvalho and

Vasconcelos, 1999; Gascon et al., 1999; Davies and

Margules, 1998; Laurance et al., 1998, 2000; Nepstad

et al., 1999). The Amazon hosts over half of the

world’s remaining tropical forests and it is currently

subject to accelerating deforestation and changing

patterns of ecosystem loss (Laurance, 1998; Whit-

more, 1997; Lima and Gascon, 1999). Laurance et al.

(2002) suggest, among others, that Brazilian Amazo-

nia has the world’s highest absolute rates of forest

deforestation and fragmentation.

However, while Brazilian Amazonia deforestation

has been widely analysed (Fearnside, 1990; Fearnside

et al., 1990; Fearnside, 1995; Reis and Margulis, 1991;

Laurance, 1998; Parayil and Tong, 1998; Laurance

et al., 2002) and much emphasis has been placed on

this part of the world, information on other parts of the

Amazon, in particular the Colombian Amazonia, is

scarce or non-existent. While there are a number of

studies on social, demographic and economic deter-

minants of the Amazonian deforestation which

suggest that deforestation is primarily determined

by human population, accessibility, land use and land

tenure issues (Reis and Margulis, 1991; Fearnside,

1993; Wood and Skole, 1998; Laurance, 1998;

Fearnside, 2001; Nepstad et al., 2001; Portela and

Rademacher, 2001; Laurance et al., 2002), these

studies are largely confined to Brazil.

Colombia, having one of the most diverse regions

in flora and fauna in the world, has been identified as a

‘‘mega-diverse’’ country (IAvH, 1998). While tropical

ecosystem transformation is occurring all over the

tropics, the loss of biodiversity and landscape

transformation in Colombia remains largely unknown,

such that entire ecosystems could be under threat of

disappearance. Global conservation prioritisation

(Myers et al., 2000) proposed the northern Andes

and the Choco regions as two hotspots in Colombia,

while the Amazonia was classified merely as a ‘‘major

wilderness area’’. This ‘‘hotspot’’ approach is directed

towards decision-makers, but devalues and detracts

attention away from non-designated areas (Bates and

Demos, 2001).

There is evidence that ecosystems in half of the

Colombian Amazon are experiencing high rates of

deforestation. Ruiz (1989) estimated that 2.5 mil-

lion ha of forest were lost in the late 1980s in

Colombia. Sierra (2000) analysed the extent and rate

of deforestation and the level of forest fragmentation

in the Napo region of western Amazonia, which

included a small portion of the Colombian Amazon,

and concluded that deforestation was advancing faster

on the Colombian side of his study area (0.9%/year)

due to population growth from the foot of the Andes

towards the Amazon.

Detailed and updated studies are still lacking in

Colombia. There are no regional geographic databases

of current information on the dynamics and patterns of

land cover change and the levels and patterns of

fragmentation and ecosystem integrity for this part of

the world (Sierra, 2000). Hence the aim of this study is

to provide an estimate of both deforestation and

ecosystem transformation and probable causes of

change using an analysis which incorporates both

biotic and anthropogenic data for a portion of the

Colombian Amazon. An approach that incorporates

both factors is critical for understanding of the

consequences of human activities on the natural

environment in the Colombian Amazon. As a first

attempt towards monitoring and analysing the state of

natural ecosystems in the Amazonia, we have chosen

to use remote sensing (RS) and a geographic

information system (GIS). In addition, we have

evaluated possible economic and socio-demographic

determinants of deforestation.

The patterns of ecosystem diversity, the spatial

patterns of deforestation and the resulting forest

fragmentation patterns that occur in Colombian

rainforests are totally different to those documented

in Brazil (Batistella et al., 2000) or even in Ecuador

(Sierra, 2000). In Colombia, pasture-led deforestation

D. Armenteras et al. / Ecological Indicators 6 (2006) 353–368 355

for ranching activities occurs, but in the absence of a

clear state policy, spontaneous colonization has also

occurred since the early 1970s and both follow rivers

as the only existing communication network in some

areas.

The more recent – and more significant – threats to

the eastern slopes of the Andes and thus the adjacent

Amazon lowland are the cropping of illicit cultivars.

Coca cultivation in the Andes, in particular in Peru,

Bolivia and Colombia, has been expanding over the

last 20 years, resulting in the destruction of an area of

2.4 million ha of tropical forest (United States

Department of State, 1999). This illegal cropping is

located primarily in remote tropical forest areas and in

mountainous terrain outside governmental control.

Cultivation of illegal crops, therefore, extends beyond

the traditional frontier forests, becoming a serious

threat to the most isolated pristine areas where

terrestrial transport access does not exist. Until today

most parts of the Colombian Amazon have been

passively protected due to their relative inaccessibility.

Understanding the human dimension in the

deforestation of the Colombian Amazonia will be

an important contribution to the knowledge of the

Amazon. The analysis we present here merges

datasets from satellite based estimates of land cover

change and ecosystem fragmentation with demo-

graphic and socio-economic indicators and has the

potential to contribute to global environmental

modelling efforts currently underway (International

Geosphere-Biosphere Programme (IGBP), Interna-

tional Human Dimension Programme on Global

Environmental Change (IHDP), etc).

2. Methods

2.1. Study area

Colombia extends between 12826046 N and

4813030 S, and 66850054 E and 79802033 W. It is

the fourth largest country in South America, after

Brazil, Argentina and Peru, and covers an area of

approximately 1,142,000 km2. Colombia is a geogra-

phically diverse country. The western part is mostly

mountainous but major parts of the country are plains

located below 500 m. Colombia embraces 7% of the

Amazonian basin (Domınguez, 1987).

Due to variation in geology and geomorphology, the

region yields environments with varying drainage

systems and soil qualities. This has led to very

significant differences in ecosystem composition and

structure that supports a high degree of biological

diversity. The region can be divided into five broad

vegetation categories (Kalliola et al., 1993; Domın-

guez, 1987; Prance, 1985; Huber, 1981; Sierra, 1999):

(1) L
owland forests <600 m (Kalliola et al., 1993),
which can be either riparian (Varzea), periodically

flooded forest (Bosques Temporalmente Inund-

ables, moist (Igapo) or permanently flooded forest

(Bosques Inundables);

(2) U
pland forests, differentiated into riparian upland
forest complexes (Campinarana, Bosques de

Tierra Firme, Bosques de Colinas) and montane

upland forest complexes (Piedemonte, Sierra);

(3) I
solated summits, occurring in the western Ama-
zon, with Tepui vegetation and montane savannas

with high biological endemism (Tepuis, Pantepui);

(4) L
arge scale dry and humid savannas also found in
the western Amazon (for example the Llanos of

Colombia and Venezuela);

(5) V
arious types of aquatic and swamp vegetation
complexes along the major rivers such as the

Amazon.

This study focused on six pilot areas (Fig. 1)

covering a total area of 4,200,000 ha (9% of the

colombian Amazonia), with sizes ranging from

approximately 626,786 ha for the smallest pilot area

(Mitu), to 802,047 ha in the Alto Putumayo region.

These areas are slightly different in environmental and

vegetation conditions although all six areas belong to

the northwestern Amazonia. The Alto Putumayo area

has a strong Andean influence and belong to the

Andean-Amazonian region, the Macarena has both

influence of the Serrania de la Macarena and of the

Amazonian lowlands. Pure and Chorrera are typical

Amazonian lowland regions and both Mitu and Inirida

have influence of the Guyana Shield and the

transitional area between the savannahs of the

Orinoquia and the Amazonian forests. These pilot

areas were selected in agreement with different

national, regional and local stakeholders, according

to current local knowledge, institutional interest and

due to the total lack of information in other areas. All


Fig. 1. Location of the study pilot areas (1–6) and 10 protected areas (national park, PNN and national nature reserve, RNN) in the Colombian

Amazonia.

of the study areas were delimited without considering

any kind of political boundary. We also analysed the

information available in the ten national protected

areas in the Amazon. These areas represent over 65%

of all current protected areas in Colombia (Fig. 1).

2.2. Data collection

2.2.1. Ecosystem mapping

Remote sensing data from satellite imagery for the

period 1985–2001 (Landsat MSS, TM, ETM) were

used to generate ecosystem maps for all the pilot areas.

In the case of La Chorrera, cloud free satellite

information was only available for the year 1985.

Major ecosystem types were determined by a

combination of supervised classification and manual

interpretation of satellite images supplemented with

secondary information on climate and geomorphol-

ogy, vectorisation and finally ground-truthing. Areas

transformed by human activities were defined using

the spectral characteristics of deforested sites. This

classification included agricultural areas. Standard

methods of accuracy assessment, based on contin-

gency tables, were used. Ground truth data were taken

at 250 points at each of the five different transects, one

for each pilot area. In one of the sites (Macarena) field

work was cancelled due to increased social unrest, so

accuracy value could not be calculated. Extensive field

work was carried out between June and October 2001

in order to verify classes. Misclassified polygons were

identified and corrected manually in a GIS. Overall

accuracy after correction with ground truth data was

93% in the 1980s and 95% in the 2000s. We identify

accuracy of 1980s map based on natural ecosystems


that remain untouched in the 2002. We used both

ERDAS Imagine (ERDAS Inc., 2000) remote sensing

processing software and the GIS software Arcview

(ESRI, 2000) to integrate the data using standard GIS

features.

As a result of this interpretation, maps of ecosystem

cover for two different time periods for five of the six

pilot areas were produced at the 1:250,000 scale.

Ecosystems were identified and classified into biomes

adapting Walter’s classification (1985; Walter and

Breckle, 1986) as follows (a biome is defined here as

an assembly of ecosystems with similar structural and

functional characteristics):

� A
mazonia and Orinoquia Tropical Forests;
� A
mazonian Orobiomes;
� A
ndean Orobiomes;
� A
mazonian Helobiomes;
� A
mazonian Litobiomes;
� P
einobiomes;
� T
ransformed ecosystems.
Maps of deforestation were also produced for each

of the pilot areas. In order to facilitate reporting,

ecosystem classes were aggregated into (a) biomes and

(b) three major ecosystem types (natural, transformed

and water). Landscapes dominated by land uses

associated with agriculture, pasture or urban sites were

assigned the category of transformed ecosystems.

2.2.2. Census estimates and other socio-economic

indicators

Demographic and economic structure indicators

were derived from the population and agricultural

census at the municipal level, the only source of this

information for the Amazon in Colombia (CGR, 1951;

DANE, 1964, 1973, 1985 and1993). The population

data is split into ‘rural’ and ‘total’ for each of the

municipal areas in the pilot areas. The information on

population quality of life was obtained directly from

the Colombian Departamento Nacional de Planea-

cion. This provided us with a synthetic index that

includes information on education, family size,

household building quality material, water availabil-

ity, garbage collection, household density and income.

We used four types of indices to relate levels of these

indicators to levels of deforestation with the municipal

area as the spatial analysis unit:

� a
n index of quality of life, with values between 0
and 100 that represent the minimum and maximum

possible level of population quality of life

respectively.

� d
emographic indices expressed as absolute popula-
tion (number of inhabitants), population density

(inhabitants/km2) as well as annual population

growth rate (%/year).

� a
n economic activity index, or the percentage of
land area devoted to ranching and farming.

� a
violence index, the annual percentage of deaths
that were violent deaths.

Table 1 presents the results of the demographic and

socio-economic indicators generated for the pilot a-

reas and protected areas analysed in this study.

2.3. Data analysis

Our goal was to offer region specific information to

support decision-making in the Colombian Amazon

with maximum cost effectiveness under budget

restrictions. Furthermore, the data had to be as up

to date as possible and presented in an easily

interpretable way. This paper focuses on quantifying

both ecosystem changes and the changes in the spatial

patterns of ecosystems that have taken place over time

in the Amazon. It also points out how they might be

related to the changes in the demographic and

economic structure of this area. We reported

quantitative data of land cover change over the last

20 years in this part of the world. The measures were

generated as part of the Indicators Project (Armenteras

et al., 2002; Rudas et al., 2002) at the Biological

Resources Research Institute Alexander von Hum-

boldt of Colombia. Biotic indicators such as ecosys-

tem extent (ha), change rates (%) and fragmentation

indices were based on major biomes types map that we

derived from the remote sensing and ground-truthing

studies described earlier. In addition, we also analysed

ecosystem information for the ten national protected

areas obtained from a general ecosystem map of

Colombia (Etter, 1998).

With the exception of fragmentation indicators,

which were calculated using the software Fragstats

(McGarigal and Marks, 1995), the other measures

were calculated using standard GIS functions in

Arcview and ERDAS Imagine and statistical analy-

tical tools such as SPSS v.10.

D.

Arm

entera

set

al./E

colo

gica

lIn

dica

tors

6(2

00

6)

35

3–

36

83

58

Table 1

Socio-economic, demographic and biological indicators: pilot and protected areas in the Amazon

Pilot (PL) and

protected areas (PT)a

Quality of

life index

Population (1993)

(number of

inhabitants)

Population density

(inhabitants/km2)b

Annual population growth (%) Pasture area

(% total area)

Violent

deaths (%)

Natural ecosystems

remaining

(% total area) 2000s

Rural Total Rural Total Rural Total 1973–1985 1985–1993 1973–1993

Alto Putumayo (PL) 42.9 62.4 213549 388062 9.98 18.11 3.7 4.0 3.8 56.6 27.9 27.9

Macarena (PL) 33.4 50.5 96690 127306 2.45 3.16 8.2 6.0 7.4 26.1 38.0 68.5

Inirida-Mataven (PL) n.a. n.a. 13942 18367 0.76 1.00 2.6 6.8 4.3 0.0 0.0 94.3

La Chorrera (PL) n.a. n.a. 20146 22963 0.30 0.32 3.6 7.9 5.3 0.0 0.0 98.3c

Mitu (PL) n.a. n.a. 9768 13896 0.47 0.67 7.7 �4.5 2.8 0.0 0.0 91.2

Pure (PL) n.a. n.a. 3028 5227 0.13 0.22 8.3 4.7 6.9 0.0 0.0 99.2

Amacayacu (PT) n.a. 71.7 14539 35083 1.90 5.10 8.8 5.3 7.4 0.4 0.0 99.7

La Paya (PT) 78.6 64.7 32778 54740 2.42 4.03 12.2 �6.6 4.7 1.5 15.1 86.6

Nukak (PT) n.a. 61.8 28628 33424 1.18 1.36 0.0 0.0 0.0 13.4 65.3 96.8

Sumapaz (PT)d 44.3 61.0 78597 132308 2.76 4.23 3.8 �2.0 1.5 16.3 24.9 99.1

Chiribiquete (PT) n.a. 57.5 8280 10078 0.15 0.18 6.6 8.4 7.3 1.4 31.2 100.0

La Macarena (PT) 36.0 42.4 91162 116369 2.85 3.63 6.6 2.8 5.1 24.8 26.8 79.9

Los Picachos (PT) 35.2 42.1 89963 115084 2.19 2.77 8.3 1.8 5.7 14.1 62.2 97.0

Tinigua (PT) 34.5 40.1 61275 76874 2.45 3.08 7.6 1.0 5.0 9.2 42.4 82.4

Puinawal (PT) n.a. n.a. 22422 26847 0.33 0.39 �1.7 12.7 4.1 0.0 0.0 100.0

Cahuinari (PT) n.a. n.a. 3322 4489 0.06 0.08 19.7 10.0 15.8 0.0 0.0 100.0

n.a.: not available.a Total indicators for municipal areas with territory in selected pilot and protected areas.b Weighted mean by municipal territory participation in total pilot or protected area.c 1980s data (no information available for the 2000s).d Without Bogota’s indicators.


The rate and extent of natural ecosystem loss and

fragmentation was calculated in five of the pilot areas,

because for one (La Chorrera) cloud free satellite

information was not available. Matrix information was

generated to obtain total area of each ecosystem type

and estimates of patterns and average rates of

deforestation throughout the study period (1985–

2001) for each pilot area. In order to calculate the

average annual deforestation rate, we assumed that

this rate is not constant and the following formula was

employed:

Loss rate ¼ ½LnðAt1Þ � LnðAt0Þ� � 100

t1 � t0

;

where A equals the ecosystem area (ha), t1 final year

and t0 initial.

Forest fragmentation was analysed for the period

1985–2001 using the following landscape metrics:

fragment number, size (mean patch size, patch size

standard deviation), shape (mean shape index, mean

perimeter/area ratio and mean patch fractal dimension)

andedge (totaledge,edgedensityandmeanpatchedge).

The results for each index were grouped using 0.5

standard deviation limits into three classes (high,

average, and low). We constructed maps showing the

degree of transformation and the fragmentation of

natural ecosystems for each pilot area using these three

categoriesandcombiningtheresults foreachindexintoa

single fragmentation level that could be illustrated in a

map.Fragstatssoftwarewasalsousedtostudylandscape

diversity within the five sites, and three different indices

of landscape diversity were used in order to compare the

pilot areas: the number of ecosystem types, the Shannon

diversity and the evenness index.

The smallest spatial unit for which economic and

demographic data is available is the municipio, which

has a purely administrative boundary. The census data

were aggregated into a single record for each pilot area

and protected area by weighting the indices by the

percentage of the total area of interest that is covered

by the municipio. This aggregated record data

therefore refers to the characteristics of all the

municipios within these pilot and protected areas.

In order to analyse the impact of human pressures

on natural ecosystems and find the possible determi-

nants of forest ecosystem loss we undertook simple

ordinary least squares (OLS) regression analysis.

Demographic and socio-economic data were treated as

independent variables. The percentage of ecosystem

loss (NED, natural ecosystem degradation, includes

only changes from natural to anthropogenic) was used

as the dependent variable. We analysed the correlation

between variables and undertook complementary

regression analysis to clarify the levels of statistical

confidence in the relationships between some of the

analysed variables. Further, we analysed the most

significant determinant of deforestation and projected

the ecosystem changes over 50 years from the present,

assuming that the same tendencies will prevail.

3. Results and discussion

The most representative biomes of the six pilot areas,

are tropical humid forests (61.81%) and helobiomes of

the Amazonia (11.57%). The pilot areas with the

highest percentage of natural ecosystems over the 15-

year period of analysis (1985–2000) are Pure (99.23%),

Inırida-Mataven (94.37%) and Mitu (91.23%)

(Table 2). The area of greatest transformation is the

zone of Alto-Putumayo near the Andes, with only 28%

of natural ecosystems left in 2001, followed by

Macarena with 68.57% (Table 2). These areas also

had the highest average annual rate of natural

ecosystem loss: the highest rate corresponds to the

Putumayo (3.73%), followed by Macarena (0.97%) and

followed by Mitu (0.31%), Inırida-Mataven (0.23%)

and finally Pure with 0.01% annual loss rate (Table 2,

Fig. 2).

In general, the relative degree of fragmentation of

each site follows the same order as the above-mentioned

deforestation rates, with the highly fragmented natural

ecosystems in Putumayo and Macarena pilot areas and

less fragmentation in the other three. The pattern of

fragmentation follows the colonization and develop-

ment associated with the rivers, the only transportation

network in Amazonia (Fig. 3). This pattern of

fragmentation and deforestation is clearly very different

to the ‘‘fishbone’’ patterns in areas of the Brazilian

Amazonia and some parts of Ecuador (Sierra, 2000),

where the construction of roads is one of the main

drivers of deforestation (e.g. in Rondonia, Batistella

et al., 2000), a determinant that is not apparent yet in the

Colombian Amazonia. These differences may reflect

different periods in the evolution of fragmentation that

in Colombia may be in an earlier stage than some areas


Table 2

Extent, natural ecosystem (NE) cover and landscape level metrics for natural ecosystems in six pilot areas of the Colombian Amazonia

Pilot areas Study

area (ha)

Natural ecosystems

remaining (ha)

(% of study area)

NEa annual

loss rate (%)

Number of

NEa

Shannon’s

diversity index

Shannon’s

evenness index

1980s 2000s 1980s 2000s 1980s 2000s 1980s 2000s

La Macarena 713,386 560,658 (78.6%) 489179 (68.5%) 0.97% 41 41 3.32 3.28 0.89 0.88

Mataven-Inirida 640,887 628,026 (97.9%) 604826 (94.3%) 0.23% 41 41 3.01 3 0.81 0.81

Mitu 626,786 595,480 (95.0%) 571384 (91.2%) 0.31% 39 39 2.96 2.97 0.81 0.81

Pure 705,056 700,770 (99.4%) 699638 (99.2%) 0.01% 40 40 2.80 2.80 0.78 0.76

Putumayo 802,477 336, 912 (42.2%) 224799 (27.9%) 3.74% 45 45 3.34 3.28 0.88 0.77

Chorrerab 712,116 700,057 (98.3%) – – 30 – 2.66 – 0.77 –a NE, natural ecosystems.b No information available for the 2000s.

Fig. 2. Deforestation patterns and natural ecosystem loss in five pilot areas of the Colombian Amazonia for a period of 15–20 years (between the

1980s and 2000s).


Fig. 3. Degree of natural ecosystem fragmentation in the Macarena Pilot Area for the year 2000.

in Brazil. Further, there are several small, scattered and

isolated deforestation patches usually related to coca

growing in the Amazon, present in some of the areas but

to a major degree present in Alto Putumayo and

Macarena areas.

Table 2 also summarizes the indices of landscape

diversity for all pilot areas: number of ecosystem types,

Shannon’s diversity (SD) and Shannons evenness (SE)

index. The areas all have similar ecosystem richness,

but ecosystem diversity and evenness are high in the

more threatened Putumayo and Macarena areas with an

SD ranging between 3.28 and 3.34 for the two areas, and

an SE of between 0.88 and 0.89 for the Macarena (the

highest), and between 0.77 and 0.88 for the Putumayo

area. On the contrary, the less degraded areas have

lower SD and SE: (a) Mataven-Inirida has an SD of 3

and an SE of 0.81, (b) Mitu has an SD of 2.97 and SE of

0.81 and (c) Pure, the most preserved pilot area has an

SD of 2.80 and SE of 0.76.

Landscape diversity results are especially surpris-

ing considering that although ecosystem richness is

very similar between areas, ecosystem diversity and

evenness are higher in the more threatened areas such

as the Putumayo and Macarena regions. In fact, this

result is logical because these two areas are the closest

to the Andes and, being topographically much more

heterogeneous than the lowlands of the Amazon areas,

they contain a higher number of different environ-

ments and ecosystems.

In the Putumayo region there has been a major

decrease in the percentage of natural ecosystems

(from 42 to 28%), coincident with the annual change

rate of 3.73% discussed above. Fragmentation of

natural forest has increased dramatically. Current rural

population density is 9.98 inhabitants/km2 and has

increased from 1.98 (1951) to 2.97 (1964) to 4.63

(1973) to 7.24 (1985). There has also been a

substantial decrease in the percentage of natural


Fig. 4. Evolution of rural population density around pilot areas over

the last five population censuses in Colombia. Source: Rudas et al.

(2002) and Armenteras et al. (2002).

ecosystems (from 79 to 68%) similar to the Macarena

region, although the annual change rate is 0.86% which

is lower than Macarena. Natural ecosystems are

becoming increasingly fragmented. Rural population

density in the Macarena region is lower than in the

Putumayo pilot area at 2.45 inhabitants/km2, a popula-

tions increased from 0.56 in 1973 to 1.51 in 1985.

The results also show that areas that are more

degraded coincide with areas of high population

density pressure (Fig. 4) and low quality of life.

Information analysis results demonstrate a statistically

significant relationship between demographic pres-

sures and the percentage of natural cover lost (Fig. 2).

In fact, as shown in Eqs. (1)–(4) of Table 3, there is a

statistically significant correlation of both absolute

population and population density with natural

ecosystem degradation (NED). Furthermore the result

is significant for total population (TP), rural popula-

tion (RP) as well as for total and rural population

density (TPD and RPD). It is important to note that for

models with demographic variables the intercept was

not significant, and so these models were re-estimated

ignoring the intercept. This is based on the assumption

that with zero population, ecosystem degradation

would also be zero.

Contrary to the significant results of the above-

mentioned demographic variables, annual population

growth rates (APG) are not significantly related to

natural ecosystem degradation NED (Eq. (6), Table 4).

This is because high growth rates are present in both

areas with current high population density (e.g. Alto

Putumayo and Macarena), and in areas with low

population and naturally high growth rates (Vaupes,

La Chorrera and Inırida-Mataven). In the former case,

the results indicate that not only do the areas have high

populations but also that the growth processes are still

highly dynamic. In the latter case this result reflects

the fact that although important changes in population

are taking place, population density remains low.

The significant effect of economic activity on

natural ecosystems is also reflected in these results.

Eq. (5) (Table 4) shows that the pasture area (PA, main

legal land use activity) has a significant positive

relationship with degraded ecosystems. However, it is

not appropriate, to undertake a multivariate analysis

using this variable and population variables since

these variables are highly correlated (Eqs. (7)–(10),

Table 4). Another significant result is that neither

quality of life nor violence levels are statistically

related with ecosystem degradation (Eqs. (11) and

(12), Table 4).

Overall, population processes and the main

economic activity have a very significant impact on

natural ecosystems degradation in these areas. For

instance, an increment of one inhabitant per square

kilometer would generate a loss of natural ecosystems

of more than 7% (Eq. (4b), Table 3). As a result, a

deforestation simulation model was developed in

order to project future tendencies of natural ecosystem

degradation in the area. Since rural population density

(RPD) was the most significant determinant of natural

ecosystem degradation (NED) with a r2 = 0.86

(Eq. (4a), Table 3), we used this factor to simulate

ecosystem loss. The values were area weighted.

Roughly speaking, NED equals 6.61� RPD. RPD was

linearly projected over the next 10, 20, 30, 40 and 50

years using the minimum (0.33%), maximum

(17.25%) and average (6.8%) RPD of the last five

demographic censuses in the study area. We then

constructed three different future scenarios assuming

that in the 1990s the ecosystem degradation was zero.

The results of applying this model can be seen in

Fig. 5. According to the model, the 16 sites of the

Colombia Amazonia studied will suffer significant

natural ecosystem cover loss (Fig. 5). In fact, only

under a scenario of low rural population density will

Amazonian forests be relatively protected. Both

average and high rural population density projections


Table 3

Statistical results of the regression analysis of population and economic indicators as predictors of natural ecosystem degradation

Equation Est Var Predictor Coefficient t P > jtj n F P > F R2R2

adjusted

(1a)

y ¼ 1:75TP

NED 16 59.35 0.000 0.81 0.80

Intercept �1.76 �0.65 0.526

TP 1.75 7.70 0.000

(1b)

y ¼ 1:66TP

NED 16 90.62 0.000 0.86 0.85

TP 1.66 9.52 0.000

(2a)

y ¼ 2:97RP

NED 16 48.15 0.000 0.77 0.76

Intercept �3.42 �1.10 0.290

RP 2.97 6.94 0.000

(2b)

y ¼ 2:65RP

NED 16 69.87 0.000 0.82 0.81

RP 2.65 8.36 0.000

(3a)

y ¼ 3:76TPD

NED 16 48.49 0.000 0.78 0.76

Intercept �0.17 �0.06 0.953

TPD 3.76 6.96 0.000

(3b)

y ¼ 3:74TPD

NED 16 77.69 0.000 0.84 0.83

TPD 3.74 8.81 0.000

(4a)

y ¼ 7:08RPD

NED 16 6.30 0.000 0.83 0.82

Intercept �2.25 �0.88 0.393

RPD 7.08 8.32 0.000

(4b)

y ¼ 6:61RPD

NED 16 102.08 0.000 0.87 0.86

RPD 6.61 10.10 0.000

(5)

y ¼ 1:08PA

NED 16 55.75 0.000 0.80 0.79

Intercept 0.11 0.04 0.968

PA 1.08 7.47 0.000

NED = natural ecosystem degradation (%) in 1990s. TPD = total population density, inhabitants/km2 (1993). TP = total population, �10.000

inhabitants (1993). RPD = rural population density, inhabitants/km2 (1993). RP = rural population, �10.000 inhabitants (1993). PA = pasture

area (%). NED = natural ecosystem degradation which is the percentage of natural land cover converted to anthropogenic land cover; TP = total

population; TPD = total population density; RP = rural population and PA = percentage of the area under pasture. Sample size = 16 (6 study

plots and 10 natural protected areas).

Table 4

Control analysis statistics of regressions between population and economic indicators and natural ecosystems degradation


adjusted

(6)

y ¼ �0:72APG

PA 1.09 9.41 0.000

NED 16 0.25 0.623 0.02 �0.05

Intercept 15.11 1.65 0.120

APG �0.72 �0.50 0.623

(7a)

y ¼ 1:54TP

PA 16 151.84 0.000 0.92 0.91

Intercept �1.13 �0.77 0.968

TP 1.54 12.32 0.000

(7b)

y ¼ 1:48TP

PA 16 236.05 0.000 0.94 0.94

TP 1.48 15.36 0.000


Table 4 (Continued )


adjusted

(8a)

y ¼ 2:70RP

PA 16 210.97 0.000 0.94 0.93

Intercept �3.04 �2.25 0.041

RP 2.70 14.52 0.000

(8b)

y ¼ 2:41RP

PA 16 245.58 0.000 0.94 0.94

RP 2.41 15.60 0.000

(9a)

y ¼ 3:02TPD

PA 16 39.15 0.000 0.74 0.72

Intercept 1.10 0.44 0.668

TPD 3.02 6.26 0.000

(9b)

y ¼ 3:15TPD

PA 16 67.70 0.000 0.82 0.81

TPD 3.15 8.23 0.000

(10a)

y ¼ 5:87RPD

PA 16 72.35 0.000 0.84 0.83

Intercept �0.90 �0.44 0.670

RPD 5.87 8.51 0.000

(10b)

y ¼ 5:68RPD

PA 16 119.38 0.000 0.89 0.88

RPD 5.68 10.93 0.000

(11)

y ¼ 0:18VD

NED 16 0.69 0.420 0.05 �0.02

Intercept 7.48 1.15 0.269

VD 0.18 0.83 0.420

(12)

y ¼ �0:66QLI

NED 10 0.01 0.933 0.001 �0.12

Intercept 19.65 0.48 0.641

QLI �0.06 �0.09 0.933

(13)

y ¼ �0:98QLI

VD 10 3.32 0.106 0.29 0.21

Intercept 87.79 2.89 0.641

QLI �0.98 �1.82 0.106

(14a)

y ¼ 0:97TPD

VD 16 0.51 0.489 0.03 �0.03

Intercept 17.93 2.53 0.024

TPD 0.97 0.71 0.489

(14b)

y ¼ 1:09TPD

VD 16 5.29 0.036 0.26 0.21

TPD 1.09 9.41 0.000

(15a)

y ¼ 2:79RPD

VD 16 1.33 0.268 0.09 0.02

Intercept 15.57 2.15 0.050

RPD 2.79 1.15 0.268

(15b)

y ¼ 6:07RPD

VD 16 8.49 0.011 0.36 0.32

RPD 6.07 2.91 0.011

NED = natural ecosystem degradation in 1990s (%). APG = anual population growth, 1973–993 (%). TP = total population,�10,000 inhabitants

(1993). PA = pasture area (%). RP = rural population, �10,000 inhabitants (1993). VD = violent death (%). TPD = total population density,

inhabitants/km2 (1993). QLI = quality of life index (0 < QLI < 100; best = 100). RPD = rural population density, inhabitants/km2 (1993).

NED = natural ecosystem degradation which is the percentage of natural land cover converted to anthropogenic land cover; TP = total

population; TPD = total population density; RP = rural population, PA = percentage of the area under pasture; APG = annual population growth;

QLI = quality of life. Sample size = 16 (6 study plots and 10 natural protected areas).


Fig. 5. Three possible scenarios of natural ecosystem degradation in the study area under three different projections of rural population density

(area weighted).

are very pessimistic and suggest that in 50 years

between 85 and 100% of natural ecosystems will be

lost.

Furthermore, the model shown in Eq. (5) (Table 3)

confirms the hypothesis that pasture and cattle

ranching are the main cause of ecosystems degrada-

tion in the Colombian Amazon. Both variables were

measured by independent procedures, and the model

estimates that an increase of one percent of pasture

areas would produce a decrease of one percent of

natural ecosystem cover. However, statistical informa-

tion on the area of illicit crops per municipality is not

available and this must be taken into account. It was

impossible to incorporate these data into the analysis

to further explain deforestation rates in the study area.

4. Conclusions

Ecosystem loss rates reported in this project are

much higher than recent estimates which suggest an

annual rate of change of forest cover of between

�0.38% (Achard et al., 2002) and �0.4% in tropical

South America (FAO, 2001). This suggests that

greater attention at a national and international level

should be directed towards the Colombian Amazon.

Furthermore, not only the extent and rate of

deforestation is worrying but also the degree of

fragmentation.

Clearly, the Macarena and Alto Putumayo areas

have undergone dramatic changes since the mid-1980s

mainly due to deforestation associated with, cattle

farming and illegal cropping. Both predominant land

uses are located near the Andes where more than half

of the population lives. In the 1950s oil extraction

brought population immigration to the Alto Putumayo

region. Since the 1970s, partly due to a lack of

government policies, illegal activities such as coca

growing have been taking place and violent para-

military and guerrilla groups have spread. The current

patterns on the landscape are clearly due to forest

extraction associated with cattle farming combined

with illegal cropping. In the 1970s and 1980s there

was an unsuccessful attempt at establishing slash and

burnt agriculture in the Macarena region. This land

slowly transformed into grazing encouraged partly by

a land reform. Recently, these areas have also been

transformed by illegal cropping due to, in part, by

establishment of a peace talk territory with no

government presence that has lasted a few years.

Analysts have applied different approaches to study

change and the causes of change in natural ecosys-

tems. Over the 15-year period covered by this study,

there has been an increase in the magnitude of natural

ecosystem loss. This loss is a function of the change in

pressure on the natural resources in the region. The

population processes and the main economic activity

related to population have very significant impacts on

natural ecosystem degradation in these areas.

We expect that the GIS methods developed during

this project will prove to be a very valuable tool for

policy and decision makers in the Amazon. We expect

that they will assist in both identifying further data and

solving interpretation problems, as well as developing

new indicators according to the area local conditions

and data availability. This study covered less than 10%


of Amazonia. Clearly, many important sites could and

need to be studied further.

Ecosystem loss rates reported in this project are

much higher than recent estimates which suggest an

annual rate of change of forest cover of between

�0.38% (Achard et al., 2002) and �0.4% in tropical

South America (FAO, 2001). This suggests that

greater attention at a national and international level

should be directed towards the Colombian Amazon.

Furthermore, not only the extent and rate of

deforestation is worrying but also the degree of

fragmentation. This is critically important due to its

effect on forest degradation and ecosystem function-

ality.

Hopefully, the results of this work will provide

some much-needed answers in our efforts to under-

stand the spatial pattern and probable causes of

deforestation. At the same time we also hope that

natural resource managers and planners use the

information presented in this paper as to undertake

biodiversity policies in the Colombian Amazonia.

Remote sensing offers rapid and accurate sampling

and is perhaps the only affordable means of looking at

processes over large spatial areas. GIS analysis and

mapping are a good way of informing and warning

managers. It is imperative to keep in mind that

indicators have to be quantitative so that results can be

compared over time and space. It is important to

continue further interpretation, modeling and predic-

tions of the dynamics of our natural resources through

analysis involving both remotely sensed and social

data. Although this research is just a small contribu-

tion to the still scarce knowledge regarding the

Colombian Amazonia, we hope that this article will

capture the attention of both public and researchers

towards this part of the world.

Acknowledgements

This work was the result of the project Diseno e

Implementacion del Sistema de Indicadores de Segui-

miento de la Polıtica de Biodiversidad en la Amazonia

Colombiana (Instituto Humboldt – Ministerio del

Medio Ambiente, Credito BID 774 OC/CO). Our

thanks to Fernando Gast, General Director of Instituto

Alexander von Humboldt. We want to specially thank

Juan Carlos Betancourth and Carol Franco and also

Pedro Botero and Jaime Forero for their work in this

project. We also want to thank the following participant

institutions: CDA, Corpoamazonia, Cormacarena,

Instituto Sinchi, Unidad de Parques and the Ministerio

del Medio Ambiente. Finally we want to thank Sophia

Burke and Scott Newey for their corrections and

comments on the manuscript.

References

Achard, F., Eva, H.D., Stibig, H., Mayaux, P., Gallego, J., Richards,

T., Malingreau, J.P., 2002. Determination of deforestation rates

of the world’s humid tropical forests. Science 297, 999–1002.

Armenteras, D., Rodriguez, N., Rudas, G., Sua, S., 2002. An

indicator system with GIS to explain and prevent biodiversity

losses in the Colombian Amazonia. In: Proceedings of the 22nd

ESRI Users Conference, San Diego, CA, July.

Bates, J.M., Demos, T.C., 2001. Do we need to devalue Amazonia

and other large tropical forests? Divers. Distrib. 7, 249–255.

Batistella, M., Brondizio, E.S., Moran, E.F., 2000. Comparative

analysis of landscape fragmentation in Rondonia, Brazilian

Amazon. Int. Arch. Photogr. Rem. Sens. Spatial Inform. Sci.

33, 148–155.

Carvalho, K.S., Vasconcelos, H.L., 1999. Forest fragmentation in

central Amazonia and its effects on litter-dwelling ants –

analysis by ant communities. Biol. Conserv. 91 (2), 151–157.

Contralorıa General de la Republica (CGR), 1951. XII Censo de

Poblacion y I de Vivienda.

Davies, K.F., Margules, C., 1998. Effect of habitat fragmentation on

carabid beetles: experimental evidence. J. Anim. Ecol. 67, 460–

471.

Departamento Administrativo Nacional de Estadıstica (1964, 1973,

1985 & 1993). Censos de poblacion de 1964, 1973. 1985 y 1993.

Colombia.

Domınguez, C., 1987. Colombia y la Panamazonia. In: Colombia

Amazonica. Universidad Nacional de Colombia y Fondo FEN de

Colombia, Bogota, Colombia.

Etter, A., 1998. Mapa general de ecosistemas de Colombia

(1:1.500.000). In: IAvH (1998) Informe nacional sobre el estado

de la biodiversidad 1997. Instituto de Investigacion de Recursos

Biologicos Alexander von Humboldt, PNUMA and Ministerio

de Medio Ambiente, 3 vol. Bogota, Colombia.

FAO, 2001. Global Forest Resource Assessment. FAO, Rome.

Fearnside, P.M., 1990. The rate and extent of deforestation in

Brazilian Amazonia. Environ. Conserv. 17 (3), 213–226.

Fearnside, P.M., 1993. Deforestation in Brazilian Amazonia: the

effect of population and land tenure. Ambio 22 (8), 537–545.

Fearnside, P.M., 1995. Potential impacts of climatic change on

natural forests and forestry in Brazilian Amazonia. Forest Ecol.

Manage. 78, 51–70.

Fearnside, P.M., Lima de Alencastro Graca, P.M., Alves Rodrıguez,

F.J., 2001. Burning of Amazonian rainforests: burning efficiency

and charcoal formation in forest cleared for cattle pasture near

Manaus, Brazil. Forest Ecol. Manage. 146, 115–128.


Fearnside, P.M., 2001. Land-tenure issues as factors in environ-

mental destruction in Brazilian Amazonia: the case of Southern

Para. World Dev. 29 (8), 1361–1372.

Fearnside, P.M., Tardin, A.T., Meira Filho, L.G., 1990. Deforesta-

tion Rate in Brazilian Amazonia. Instituto de Pesquisas Espa-

ciais (INPE), Sao Jose dos Campos, Sao Paulo, 8 pp.

Gascon, C., Lovejoy, T.E., Bierregaard Jr., R.O., Malcolm, J.R.,

Stouffer, P.C., Vasconcelos, H.L., Laurance, W.F., Zimmerman,

B., Tocher, M., Borges, S., 1999. Matrix habitat and species

richness in tropical forest remnants. Biol. Conserv. 91 (2), 223–

229.

Huber, O., 1981. Significance of savanna vegetation in the Amazon

territory of Venezuela. In: Prance, G.T. (Ed.), Biological Diver-

sification in the Tropics. Columbia University Press, New York.

Instituto, de Investigacion de Recursos Biologicos Alexander von

Humboldt (IavH), 1998. Informe nacional sobre el estado de la

biodiversidad 1997. Instituto de Investigacion de Recursos

Biologicos Alexander von Humboldt, PNUMA and Ministerio

de Medio Ambiente, 3 vol. Bogota, Colombia.

Kalliola, R., Maarit, P., Danjoy, W., 1993. Amazonia Peruana.

Vegetacion humeda tropical en el llano subandino. Proyecto

Amazonia Universidad de Turku, Findland.

Klein, B.C., 1989. Effect of forest fragmentation on dung and

carrion beetle communities in central Amazonia. Ecology 70

(6), 1715–1725.

Laurance, W.F., Delamonica, P., Laurance, S.G., Vasconcelos, H.L.,

Lovejoy, T.E., 2000. Conservation: rainforest fragmentation

kills big trees. Nature (404), 836.

Laurance, W.F., Laurance, S.G., Delamonica, P., 1998. Tropical

forest fragmentation and greenhouse gas emissions. Forest Ecol.

Manage. 110 (1), 173–180.

Laurance, W.F., 1998. A crisis in the making: responses of Ama-

zonian forests to land use and climate change. TREE 13 (10),

412–451.

Laurance, W.F., Albernaz, A.K.M., Schroth, G., Fearnside, P.M.,

Bergen, S., Venticinque, E.M., Da Costa, C., 2002. Predictors of

deforestation in the Brazilian Amazon. J. Biogeogr. 29 (5), 737–

748.

Lima, M.G., Gascon, C., 1999. The conservation value of linear

forest remnants in central Amazonia. Biol. Conserv. 91, 241–

247.

Mcgarigal, K., Marks, B.J., 1995. Fragstats: spatial pattern analysis

program for quantifying landscape structure. Gen. Tech. Re

PNW-GTR-351. U.S. Departament of Agriculture, Forest Ser-

vice, Pacific Northwest Research Station, Portland, OR, USA.

Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca,

G.A.B., Kent, J., 2000. Biodiversity hotspots for conservation

priorities. Nature 403, 853–858.

Nepstad, D.C., Veroassimo, V., Alencar, A., Nobre, C., Lima, E.,

Lefebvre, P., Schlesinger, P., Potterk, C., Moutinho, P., Mendoza,

E., Cochrane, M., Brooks, V., 1999. Large-scale impoverishment

of Amazonian forests by logging and fire. Nature 398, 505–

508.

Nepstad, D.C., Carvalho, G., Barros, A.C., Alencar, A., Capobianco,

J.P., Bishop, J., Moutinho, P., Lefebvre, P., Silva Jr., U.L., Prins,

E., 2001. Road paving, fire regime feedbacks, and the future of

Amazon forests. Forest Ecol. Manage. 154, 395–407.

Parayil, G., Tong, F., 1998. Pasture-led to logging-led deforestation

in the Brazilian Amazon: the dynamics of socio-environmental

change. Glob. Environ. Change 8 (1), 63–79.

Pimm, S.L., Ayres, M., Balmford, A., Branch, G., Brandon, K.,

Brooks, T., Bustamante, R., Costanza, R., Cowling, R., Curran,

L.M., Dobson, A., Farber, S., da Fonseca, G.A.B., Gascon, C.,

Kitching, R., McNeely, J., Lovejoy, T., Mittermeier, R.A.,

Myers, N., Patz, J.A., Raffle, B., Rapport, D., Raven, P., Roberts,

C., Rodrıguez, J.P., Rylands, A.B., Tucker, C., Safina, C.,

Samper, C., Stiassny, M.L.J., Supriatna, J., Wall, D.H., Wilcove,

D., 2001. ENVIRONMENT: Can We Defy Nature’s End?

Science 293, 2207–2208.

Portela, R., Rademacher, I., 2001. A dynamic model of patterns of

deforestation and their effect on the ability of the Brazilian

Amazonia to provide ecosystem services. Ecol. Model. 143,

115–146.

Prance, G.T., 1985. A review of the phytogeographic evidences for

pleistocene climate changes in the neotropics. Ann. Missouri

Botanical Garden 69 .

Ranta, P., Blom, T., Niemela, J., Joensuu, E., Siitonen, M., 1998. The

fragmented Atlantic rain forest of Brazil: size, shape and dis-

tribution of forest fragments. Biodiv. Conserv. 7, 385–403.

Reis, E.J., Margulis, S., 1991. Perspectivas Economicas de desflor-

estamiento de Amazonia. Textos para Discussao No. 215.

Insituto de Pesquisa Economica Aplicada (IPEA), Brasilia, 47

pp.

Rudas, G.D., Armenteras, S.M., Sua, Rodrıguez, N., 2002. Indi-

cadores de Seguimiento de la Polıtica de Biodiversidad en la

Amazonia Colombiana 2001. Informe Final de Resultados.

Proyecto Diseno e Implementacion del Sistema de Indicadores

de Seguimiento de la Polıtica de Biodiversidad en la Amazonia

Colombiana. Instituto Humboldt, CDA, Corpoamazonia,

Cormacarena, Instituto Sinchi, Unidad de Parques, Ministerio

del Medio Ambiente (Credito BID 774 OC/CO), Bogota,

Colombia.

Ruiz, J., 1989. Desarrollo regional en Colombia: ordenamiento

ambiental y desarrollo regional. Cuadernos de Industria y Econ-

omia Rural 22.

Sierra, R. (Ed.), 1999. Propuesta preliminar de un sistema de

clasificacion de vegetacion para el Ecuador continental. Project

INEFAN/GEF-BIRF and EcoCiencia, Ecuador.

Sierra, R., 2000. Dynamics and patterns of deforestation in the

western Amazon: the Napo deforestation front, 1986–1996.

Appl. Geogr. 20, 1–16.

Steininger, M.K., Tucker, C.J., Ersts, P., Killeen, T.J., Villegas, Z.,

Hecht, S.B., 2001a. Clearance and fragmentation of tropical

deciduous forest in the Tierras Bajas, Santa Cruz, Bolivia.

Conserv. Biol. 15 (4), 856–866.

Steininger, M.K., Tucker, C.J., Townshend, J.R.G., Killeen, T.J.,

Desch, A., Bell, V., Ersts, P., 2001b. Tropical deforestation in the

Bolivian Amazon. Environ. Conserv. 28 (2), 127–134.

U.S. Department of State, 1999. 1999 International Narcotics Con-

trol Strategy Report. USA.

Vogelmann, J.E., 1995. Assessment of forest fragmentation in

Southern New England using remote sensing and geographic

information systems technology. Conserv. Biol. 9 (2), 439–

449.


Walter, H., 1985. Vegetation of the Earth and Ecological Systems of

the Geo-biosphere, 3rd ed. Springer-Verlag, New York, 318 pp.

Walter, H., Breckle, S.W., 1986. Ecological Systems of the Geobio-

sphere. Tropical and Subtropical Zonobiomes, vol. 2. Springer-

Verlag, Berlin/Heidelburg.

Whitmore, T.C., 1997. Tropical forest disturbance, disappearance,

and species loss. In: Laurance, W.F., Bierregaard, Jr., R.O.

(Eds.), Tropical Forest Remnans: Ecology. Management and

Conservation of Fragmented Communitites, pp. 3–12.

Wood, C.H., Skole, D., 1998. Linking satellite, census and

survey data to study deforestation in the brazilian amazon. In:

Liverman, D. (Ed.), People and Pixels: Linking Remotes

Sensing and Social Sciences. National Research Council,

USA, pp. 70–93.

Patterns and causes of deforestation in the Colombian Amazon

Documents