IN DEGREE PROJECT ENVIRONMENTAL ENGINEERING,SECOND CYCLE, 30 CREDITS

, STOCKHOLM SWEDEN 2018

Impact of Priority and Protected Areas on Deforestation in Brazilian Legal Amazon

TITI SARI NURUL RACHMAWATI

KTH ROYAL INSTITUTE OF TECHNOLOGYSCHOOL OF ARCHITECTURE AND THE BUILT ENVIRONMENT

TRITA 18400

www.kth.se

Impact of Priority and Protected Areas on Deforestation in Brazilian Legal Amazon

TITI SARI NURUL RACHMAWATI

Supervisor FLÁVIO LUIZ MAZZARO DE FREITAS

Examiner ULLA MÖRTBERG

Degree Project in Environmental Engineering and Sustainable Infrastructure

KTH Royal Institute of Technology

School of Architecture and Built Environment

Department of Sustainable Development, Environmental Science and Engineering

SE-100 44 Stockholm, Sweden

2

TRITA-ABE-MBT 18400

i

Abstract

Legal Amazon, the Brazilian region where much of the global tropical forest is

located, has suffered from rapid deforestation for decades, undermining the

provision of ecosystem services and the conservation of biodiversity on local and

global scales. In order to prevent deforestation, the Brazilian government has

established priority and protected areas to ensure the preservation of high

biodiversity areas and ecosystem services. This study analyses whether the

establishment of priority and protected areas have an impact in preventing

deforestation, thus promoting biodiversity and ecosystem services. Furthermore,

this study also analyses the extent to which deforestation affects priority areas for

biodiversity conservation. Deforestation datasets from 2001 to 2014 of the Legal

Amazon was processed and analyzed. The total area and density of deforestation

were compared across three categories of land: (1) protected priority areas, (2)

unprotected priority areas, and (3) non-priority areas. Spatial methods of

geoprocessing and the statistical method one-way ANOVA were used to analyse the

deforestation trends. As a result, the deforestation density was found to be lowest

inside protected areas than in unprotected areas and non-priority areas. This implied

that land-use restrictions in protected areas had more impact compared to

unprotected areas and non-priority areas. Furthermore, deforestation has been

more intensive in regions of lower biodiversity importance. Despite this positive

evaluation, substantial tracts of forest had been converted within regions of high

biodiversity importance. Therefore, the regulation of priority and protected areas

must be evaluated and improved in the future.

Keywords

Deforestation, Legal Amazon, priority areas, protected areas, biodiversity, remote

sensing data

iii

Sammanfattning

Den brasilianska regionen Legal Amazon, där mycket av den globala andelen av

tropisk skog ligger, har drabbats av snabb avskogning i årtionden, vilket

underminerar tillhandahållandet av ekosystemtjänster och bevarandet av den

biologiska mångfalden både på lokal och global skala. För att förhindra avskogning

har den brasilianska regeringen etablerat prioriterade och skyddade områden för att

säkerställa bevarandet av områden med hög biologisk mångfald och

ekosystemtjänster. Denna studie analyserar huruvida etableringen av prioriterade

och skyddade områden påverkar förebyggandet av denna avskogning. Dessutom

analyserar denna undersökning också hur omfattningen av avskogningen påverkar

prioriterade områden för bevarande av biologisk mångfald. Avskogningsdata från

2001 till 2014 i Legal Amazon bearbetades och analyserades. Total areal och densitet

av avskogningen jämfördes mellan tre kategorier av mark: (1) skyddade prioriterade

områden, (2) oskyddade prioriterade områden och (3) icke prioriterade områden.

Rumsliga metoder för bearbetning och den statistiska metoden envägs-ANOVA

användes för att analysera avskogningstrenderna. Resultaten visar att

avskogningstätheten var lägst inom skyddade områden jämfört med oskyddade

områden och icke-prioriterade områden. Det innebär att

markanvändningsbegränsningar i skyddade områden har haft större inverkan

jämfört med oskyddade områden och icke prioriterade områden. Vidare har

avskogningen varit mer intensiv i områden med lägre värden för biologisk mångfald.

Trots denna positiva utvärdering hade ett betydande skogsområde omvandlats inom

områden med höga värden för biologisk mångfald. Därför måste regleringen av

prioriterade och skyddade områden utvärderas och förbättras i framtiden.

Nyckelord

Avskogning, Legal Amazon, prioriterade områden, skyddade områden, biologisk

mångfald, fjärranalysdata

iv

Acknowledgments

I would like to offer my sincere gratitude to my thesis supervisors, Flávio Luiz

Mazzaro De Freitas, for the opportunity to do this Master's thesis and providing me

with continuous supports and dedication throughout the project.

I would also like to thank Prof. Ulla Mörtberg as the examiner of this thesis,

for the supports and valuable feedbacks.

Finally, I would like to thank my parents, family, and friends for the endless

support and prayers.

Stockholm, July 2018

Titi Sari Nurul Rachmawati

v

Table of contents

Abstract ...................................................................................... i Keywords ................................................................................................. i

Sammanfattning ...................................................................... iii Nyckelord ............................................................................................... iii

Acknowledgments ................................................................... iv

Table of contents ...................................................................... v

List of figures .......................................................................... vii List of tables ............................................................................ ix

List of acronyms and abbreviations ....................................... xi 1 Introduction ....................................................................... 13

1.1 Background ................................................................................ 13 1.2 Aim and objectives .................................................................... 14

1.3 Relevance of the study .............................................................. 15 1.4 Scope and delimitations of the study ....................................... 15 1.5 Structure of the thesis ............................................................... 15

2 Background ....................................................................... 16

2.1 Drivers and actors of deforestation in Legal Amazon ............ 16

2.2 Set of policies and initiatives related to the protection of native vegetation ........................................................................ 18

2.2.1 Conservation framework in public lands .......................... 19

2.2.2 The conservation framework in private lands .................. 20

2.2.3 Deforestation monitoring system in Legal Amazon ......... 20 2.2.4 Supply chain interventions ............................................... 21

2.3 Priority areas establishment ..................................................... 21 2.4 Ecosystem services and biodiversity in Legal Amazon ......... 22

3 Methodology ...................................................................... 24

3.1 Description of the study area .................................................... 24 3.2 Overview of the research methodology ................................... 25

3.3 Literature review ........................................................................ 26 3.4 Data acquisition ......................................................................... 26 3.5 Dataset processing .................................................................... 27

3.5.1 Identifying deforestation in priority and protected areas .. 28 3.5.2 Identifying deforestation based on years ......................... 32

3.6 Analysis of variance (ANOVA) .................................................. 34

3.6.1 Defining independent and dependent variables .............. 34 3.6.2 Checking assumptions .................................................... 36 3.6.3 Stating null and alternative hypothesis ............................ 36 3.6.4 Calculating the test statistic ............................................. 36 3.6.5 Interpreting the results ..................................................... 37

3.7 Visualization with Tableau ........................................................ 38

4 Results ............................................................................... 39

vi

4.1 Deforestation trend based on area type ................................... 39

4.2 Deforestation trends and the biological importance of priority areas .............................................................................. 41

4.3 Deforestation differences of year 2006 and 2014 .................... 44

5 Discussion ......................................................................... 46

5.1 Deforestation trends analysis ................................................... 46 5.2 Limitations .................................................................................. 48 5.3 Suggestions for future work ..................................................... 48

6 Conclusions ....................................................................... 49

References .............................................................................. 51

vii

List of figures

Figure 1. Annual deforestation rate in the Brazilian Amazon (INPE, 2017) ........................................................................................ 17

Figure 2. Clear cut deforestation (Bergen, 2018) .......................................... 18 Figure 3. Sets of public policies and NGO initiatives ................................... 18 Figure 4. The relationship between priority and protected areas in Legal

Amazon .................................................................................... 19 Figure 5. Legal Amazon in Brazil ................................................................... 25 Figure 6. Overview of the research methodology .......................................... 25 Figure 7. Priority areas of Legal Amazon (MMA, 2006) ............................... 27 Figure 8. Data processing for identification priority and protected areas

in reference to deforestation ................................................... 28 Figure 9. Exporting deforestation data based on years ................................ 29 Figure 10. The deforestation layer in priority areas in 2006 with its

attribute table .......................................................................... 29 Figure 11. The process of creating polygons outside the priority areas

using the erase function ...........................................................30 Figure 12. Deforestation outside priority areas in 2006 with its attribute

table .......................................................................................... 31 Figure 13 Data processing for identifying deforestation ............................... 32 Figure 14. Deforestation in protected and unprotected areas ...................... 33 Figure 15. Overview of the ANOVA workflow ............................................... 34 Figure 16. Deforestation trend based on type of area ................................... 39 Figure 17. Deforestation trends based on areas of biological importance .... 42

ix

List of tables

Table 1. A slice of the final table of deforestation in priority areas ...............30 Table 2. A slice of the final table of deforestation outside priority areas ..... 31 Table 3. A Slice of the final table of deforestation based on type of area ..... 33 Table 4. Conducted one-way ANOVA ............................................................ 36 Table 5. Test statistic calculation of one-way ANOVA .................................. 37 Table 6. Deforestation based on type of area ............................................... 40 Table 7. Descriptive table of first analysis of one-way ANOVA ................... 40 Table 8. ANOVA table of the first analysis one-way ANOVA ....................... 41 Table 9. Multiple comparisons table of the first analysis with one-way

ANOVA ..................................................................................... 41 Table 10. Deforestation trend based on areas of biological importance ....... 42 Table 11. Descriptive table of the second analysis with one-way ANOVA .... 43 Table 12. ANOVA table of the second analysis by one-way ANOVA ............ 43 Table 13. Multiple comparisons table of the second analysis with one-

way ANOVA ............................................................................. 44 Table 14. Descriptive table of the third analysis with one-way ANOVA ....... 44 Table 15. ANOVA table of the third analysis with one-way ANOVA ............ 45 Table 16. Multiple comparisons table of the third analysis with one-way

ANOVA ..................................................................................... 45

xi

List of acronyms and abbreviations

ANOVA Analysis of Variance

CBD Convention on Biological Diversity

CUFP Conservation Unit of Full Protection

CUSU Conservation Unit of Sustainable Use

DETER Real-Time System for Detection of Deforestation

GIS Geographic Information System

GWF Global Forest Watch

IBGE Brazilian Institute of Geography and Statistics

INPE Brazilian Spatial Research Institute

MMA Brazilian Ministry of Environment

NGO Non-Governmental Organization

Pg Petagrams

PPCDAm Action Plan for the Prevention and Control of Deforestation in

the Legal Amazon

PROBIO Project for the Conservation and Sustainable Use of Brazilian

Biological Diversity

PRODES Project for Monitoring the Brazilian Amazon Forest by Satellite

SNUC National System of Conservation Units

SPSS Statistical Package for the Social Sciences

MAPA Brazilian Agriculture Ministry

13

1 Introduction

1.1 Background

The Brazilian tropical forest located in the Legal Amazon plays a major role in the

world’s climate system, biodiversity, and ecosystem services by storing 52 Gt of

carbon in the above-ground biomass (Freitas et al., 2018) and hosting more than

40,000 different plants and more than 5,000 vertebrate species (Mittermeier et al.,

2002). The Brazilian Legal Amazon (Figure 5) also plays vital role in keeping the

water balance that gives services to human life on local and global scales (Davidson

et al., 2012).

Regardless of its richness of biodiversity, Legal Amazon has suffered from rapid

deforestation and land use change into farmland. From 2004 to 2017, the total

deforestation reached 142,714 km2 (INPE, 2017). Due to current deforestation and

forest degradation, many endemic species are affected and restricted to parts of their

habitat (Michalski and Peres, 2005). Specifically, trees and fauna mortality is

strongly affected by edge effects of forest fragmentation (Peres et al., 2010). Due to

national and global demand for food, timber and bioenergy (MAPA, 2016), more

areas of native forest may be converted leading to substantial impact for biodiversity

(Freitas et al., 2018).

To address this issue, the Brazilian government defined priority areas for

conservation, sustainable use and sharing of biodiversity benefits. This is one of the

efforts to fulfil the Convention on Biological Diversity (CBD) (UN 1992) that

represents a worldwide effort for the conservation and sustainable use of

biodiversity. The purpose of priority area establishment is to serve as a guideline of

policies that aims to find a balance between the demand for economic growth and

the conservation efforts.

The most recent nationwide database available on priority area was produced in

2006 by the Brazilian Ministry of Environment (MMA) when 900 priority areas were

established all around Brazil with 80 % located in the Amazon region (MMA, 2007).

Before the establishment of priority areas, the Brazilian government also established

protected areas through the National System of Conservation Units (SNUC,

Portuguese acronym). SNUC was stipulated in 2000 through Law 9,985, 19 July

2000. SNUC categorised protected areas as Indigenous Lands and Conservation

Units. These protected areas are included in the priority areas that was established

in 2006 (MMA, 2007) and protected by legal mechanisms that prevent the

conversion of native vegetation into agricultural land. On the other hand, the rest of

the priority areas named unprotected areas (for biodiversity conservation) are not

protected and available for legal deforestation (CBD, 2007).

14

A study by Soares-Filho et al. (2010) pointed out that protected areas in the Amazon

biome can avoid emissions as much as 47 ± 9 Pg of carbon. Furthermore, the

expansion of protected areas in 2006 have partially contributed to decreasing the

deforestation up to 75 % in the Amazon forest from 2004 to 2009 (INPE, 2009).

Several studies have quantified the effectiveness of protected areas in preventing

deforestation and protecting biodiversity (Soares-Filho et al., 2010, Nolte et al.,

2013; Brun et al., 2015; Miranda et al., 2016). These studies used varied methods

with different conclusions on protected areas’ efficiency from positive to no effect

(Soares-Filho et al., 2010) although most of the studies have shown positive impacts

of protected areas (Nolte et al., 2013; Brun et al., 2015; Miranda et al., 2016).

No studies could be found that analyse the extent to which deforestation is affecting

priority areas or if priority areas have contributed to reducing deforestation. The

impacts of priority areas and protected areas in preventing deforestation is still

uncertain even though it is highly essential to measure the effectiveness of

conservation strategies. The results of measurement can be used to justify further

investments and policies and to identify the need of adjustment in the existing

strategies (Mace et al., 2008). Hence, there is a need to address this issue and to

analyse the impacts of priority areas and protected areas in protecting biodiversity

from deforestation.

1.2 Aim and objectives

The aim of this thesis project was to analyse the impacts of priority and protected

areas for biodiversity conservation in preventing deforestation. Further, this study

investigates the efficiency of protected areas in preventing deforestation. An

investigation to which extent deforestation has affected priority areas for

biodiversity conservation was also conducted.

Thus, the research questions associated with the aim were:

1. Is the establishment of high biodiversity areas (priority areas) and land use

restrictions (protected areas) impactful in preventing deforestation?

2. Has deforestation affected areas of high biodiversity value?

The following objectives were established to meet the aim of this research:

1. To evaluate deforestation in relation to three categories of land tenure: (1)

protected priority areas, (2) unprotected priority areas, and (3) non-priority

areas.

2. To evaluate deforestation in relation to the biodiversity importance of the

native vegetation: (1) high importance, (2) very high importance, (3) extreme

high importance, and (4) no data.

15

3. To analyse the impacts of priority areas establishment in 2006 by analysing

the difference of deforestation in 2006 to the latest deforestation data in 2014.

1.3 Relevance of the study

The knowledge generated in this project may be used to support the management of

deforestation in priority and protected areas in the Legal Amazon. In the short term,

the gained knowledge will improve the understanding of the current situation, the

pattern of deforestation and how biodiversity is affected. In the long term, the

concepts and data-driven analysis presented here can serve as the basis for more

elaborated toolkits for evaluating the current public policies and producing

improved strategic plans, public policies and initiatives.

1.4 Scope and delimitations of the study

This study is limited to priority areas established in 2006 by the Brazilian Ministry

of Environment, and protected areas stipulated in 2000 through Law 9,985, 19 July

2000 in the Legal Amazon of Brazil. Furthermore, the deforestation datasets used

were provided by the INPE (Instituto Nacional de Pesquisas Espaciais) through

Project for Monitoring the Brazilian Amazon Forest by Satellite program (PRODES,

Portuguese acronym) where deforestation is available yearly from 2001 to 2014.

1.5 Structure of the thesis

This master’s thesis comprises six chapters. Chapter 1 presents the background, aim,

and objectives of the study. Chapter 2 describes the theoretical background and

related works to biodiversity and historical deforestation of the Legal Amazon and

discusses the major policies and initiatives related to land use governance in the

study region. Chapter 3 provides the study area and detailed description of the

research methodology employed to conduct the spatial and statistical analysis.

Chapter 4 presents a summary of the major results, while Chapter 5 discusses the

results in relation to existing literature and suggests possible future work which

could be built upon this project. Finally, Chapter 6 gives the key conclusion.

16

2 Background

This chapter provides the history of deforestation in Legal Amazon, which is the area

of this study. Furthermore, this chapter also explains the set of policies and

initiatives that has been established by government and private sector to address

deforestation in the Legal Amazon. Lastly, this chapter provides an overview on the

biodiversity in the study area and the establishment of priority and protected areas

as the means to protection such biodiversity.

2.1 Drivers and actors of deforestation in Legal Amazon

Deforestation in the Amazon region is closely related to the development of road

infrastructure in the region (Fearnside, 2017). An important event which indirectly

has triggered the deforestation process in the Amazon region was the construction

of Trans Amazon Highway (BR-230) in 1972 (Fearnside, 2017). The highway crossed

over the states of Paraiba, Ceara, Piau, Maranhao, Tocantins, Para, and Amazonas.

Due to the economic crisis, it was only paved from the beginning to Maraba.

However, it was enough to give access to public and private companies to penetrate

the Legal Amazon (Fearnside, 2017). Besides the highway construction, other major

drivers are agricultural intensification, cattle ranching and soybean expansion,

timber business, policy failures and land tenure establishment (Jean et al., 2011).

The complex natures and interconnectedness between these drivers resulted in

widely varied deforestation rates over the decades.

Until 1970, deforestation in Legal Amazon covered approximately 98,000 km2

(Fearnside, 2009). Afterwards, deforestation increased until 1987 because of fiscal

incentives by the government to the large ranches before it decreased due to the

economic recession until 1991 (Fearnside, 2017). However, it increased after 1991 as

the economy recovered and reached its peak in 1995 (29,100 km2) as shown in Figure

1. Deforestation decreased drastically in 1996 before it increased gradually until it

reached the peak again in 2004 (27,772 km2) because of the strengthening economy

and rising commodity prices (Fearnside, 2017). At that span of time, technological

advances favoured large scale expansion of soy and soy prices spiked in 2004

(Nepstad et al., 2014).

17

Figure 1. Annual deforestation rate in the Brazilian Amazon (INPE, 2017)

Eventually, in 2004 the Brazilian government launched the Action Plan for the

Prevention and Control of Deforestation in the Legal Amazon (PPCDAm). Several

studies pointed out that deforestation from 2004 to 2012 not only decreased because

of PPCDAm but that it was also triggered by the reduction of commodity prices,

especially soybeans and beef, and an initiative by the Brazilian Central Bank (BACEN

3545/2008) (Fearnside, 2017; Assuncao, 2015). The Bank blocked loans that gave

significant impact to large agriculture and ranching companies (Fearnside, 2017).

The Brazilian Institute for the Environment and Renewable Natural Resources also

initiated a policy to blacklist municipalities with high deforestation rates who

suffered restrictions on agricultural credits (Fearnside, 2017; Nepstad, 2014).

However, since 2009 the deforestation have stabilized at 5,000-7,000 km2 (INPE,

2017) as shown in Figure 2, and often occurs as small patches in remote parts of

Legal Amazon by smallholders (Assunção et al., 2017). Looking back at the

deforestation from 1970 until 2014, deforestation and forest degradation are

concentrated in the eastern and southern parts of Legal Amazon, mostly in the Mato

Grosso, Tocantins, and Maranhao states (Morton et al., 2006). Based on actors, the

deforestation between 2004 and 2011 was dominated by large holders and only small

parts by smallholders (Godar et al., 2014).

Nevertheless, the current deforestation data that has been collected has a limitation.

Both the current deforestation monitoring system named PRODES and Real-Time

System for Detection of Deforestation (DETER, Portuguese acronym) which is

21

,05

0

17

,77

0

13

,73

0

11

,03

0

13

,78

6

14

,89

6

14

,89

6

29

,05

9

18

,16

1

13

,22

7 1

7,3

83

17

,25

9

18

,22

6

18

,16

5

21

,39

4 25

,24

7

27

,42

3

18

,84

6

14

,10

9

11

,53

2

12

,91

1

7,4

64

7,0

00

6,4

18

4,5

71

5,8

91

5,0

12

6,2

07

7,8

93

6,6

24

-

5,000

10,000

15,000

20,000

25,000

30,000

35,000

19

88

19

89

19

90

19

91

19

92

19

93

19

94

19

95

19

96

19

97

19

98

19

99

20

00

20

01

20

02

20

03

20

04

20

05

20

06

20

07

20

08

20

09

20

10

20

11

20

12

20

13

20

14

20

15

20

16

20

17

De

fore

stat

ion

(km

2/y

ear

)

Year

18

explained in Subchapter 2.2 can only detect clear-cutting deforestation (Figure 2)

and not forest degradation or tree loss by selective logging. Furthermore, PRODES

can only identify deforestation of at least 6.25 hectares while DETER can only

identify deforestation of at least 25 hectares (Assunção et al., 2017).

Figure 2. Clear cut deforestation (Bergen, 2018)

2.2 Set of policies and initiatives related to the protection of native vegetation

The Brazilian government has stipulated several legal policies to regulate a

framework for conservation in public and private lands. Besides that, the Brazilian

government also established priority areas that are public areas with high richness

of biodiversity and created for environmental conservation. On Legal Amazon scale,

remote-sensing based monitoring of deforestation was also established through the

PRODES and DETER programmes. On the other hand, the Non-Governmental

Organization (NGO) sector also set initiatives concerning the supply chain of soy and

cattle that put pressure on the deforestation actors. The set of policies and initiatives

as explained above that will be discussed in this study are shown in Figure 3.

Figure 3. Sets of public policies and NGO initiatives

Public Policies

Conservation Framework

In Public Land

In Private Land

Deforestation Monitoring System

PRODES DETER

Priority Areas

NGO Initiatives

Supply Chain Intervention

Soy Moratorium

Cattle Agreement

19

2.2.1 Conservation framework in public lands

Public lands in Brazil are categorised as Conservation Units, Indigenous Lands and

Military Lands. In 2000, the Brazilian government stipulated the National System

of Conservation Units (SNUC) through Law no 9,985 (MMA, 2017). The SNUC law

gives clear guidance in the procedure of creation, implementation, and management

of protected areas. Based on the government level, the areas were categorised as

federal, state, municipal, and district protected areas that were managed in an

integrated manner (MMA, 2017). The protected areas under the Law covered 17.2 %

of the continental area and 1.5 % of the marine area of Brazilian territory (MMA,

2017).

Besides that, protected areas were also categorised based on conservation type called

(1) conservation unit of full protection (CUFP) and (2) conservation unit of

sustainable use (CUSU) (MMA, 2015). CUFP serves the purpose of parks and

reserves while CUSU has a function of natural resources extraction in a sustainable

manner (MMA, 2017). A study by Nepstad et al., (2006) pointed out that the location

and category of protected areas affect the effectiveness in preventing deforestation.

Protected areas located near deforestation frontiers have a higher chance of

deforestation (Nepstad et al., 2006).

In 2004, the Ministry of the Environment stipulated Decree No. 5,092 that stated

the establishment of 900 priority areas. Furthermore, there were initiatives to

include the existing protected areas into priority areas. Finally, in 2006, 490

protected areas were included in priority areas in the Amazon biome. The relation

between both areas is displayed in Figure 4.

Figure 4. The relationship between priority and protected areas in Legal

Amazon

Concerning Indigenous Lands, the Constitution was stipulated in 1988. The

Constitution gives the right to indigenous people that they have usufruct rights to

the land including the soil, the rivers and the lakes existing therein but not to possess

20

mineral resources inside the lands (Ramos, 1988). Because of that, Indigenous Lands

have the lowest risk of deforestation (Nepstad et al., 2006). A study by Ricketts et al.

(2010) showed that the probability of deforestation was 7 to 11 times lower within

Indigenous Lands than protected areas in public and private lands.

2.2.2 The conservation framework in private lands

To ensure the preservation of native vegetation in private property, the Brazilian

government stipulated the Brazilian Forest Act (Law no 4,771) in 1965 which acted

as the central framework to limit agricultural expansion over private lands. However,

this law was highly ineffective in stopping deforestation due to its strict rules and

weak enforcement, resulting in very low compliance (Fearnside, 2017). In 2012, the

Act was replaced by the New Forest Act (Law no. 12,651), which brought a number

of changes to facilitate compliance and balance production and conservation

(Sparovek et al., 2012; Freitas, 2017; Fearnside, 2017).

The main feature of the New Act is the Rural Environmental Registry (CAR,

Portuguese acronym) which regulates the protection and registration mechanisms

of permanent preservation areas and legal reserves on private territory (Freitas et

al., 2018). Studies pointed out that the New Forest Act has both drawbacks and

positive aspects compared to the previous Act (Sparovek et al., 2012; Soares-Filho et

al., 2014). It reduces restriction for deforestation on steep hillsides and near rivers

and approves illegal clearing done by 2008. Nonetheless, it introduced more efficient

command and control regulations to reduce illegal deforestation (Soares-Filho et al.,

2014).

2.2.3 Deforestation monitoring system in Legal Amazon

The Brazilian government launched a diverse set of policy interventions named the

Action Plan for the Prevention and Control in the Legal Amazon (PPCDAm) in 2004

specifically for the Legal Amazon (MMA, 2013). PPCDAm runs in inter-ministerial

format between federal, state, and municipal governments as well as environmental

organisations and civil society (Assunção et al., 2017). PPCDAm has three main

policies: land tenure regulation (MMA, 2013), land use monitoring (Arima et al.,

2014), and sustainable agricultural promotion (Dalla-Nora et al., 2014).

The projects under the second policy of PPCDAm are PRODES and DETER

(Assunção et al., 2017). PRODES is a deforestation monitoring programme covering

the Legal Amazon. It uses satellite images of LANDSAT with a spatial resolution of

20 to 30 m. PRODES releases deforestation data annually, and it has been run since

21

1988. PRODES has a limitation as it only identifies clear-cut deforestation with

cleared areas of at least 6.25 hectares (Assunção et al., 2017).

Meanwhile, DETER is a near-real-time monitoring system of deforestation based on

satellite images. DETER’s sources images are from the MODIS sensor on the Terra

satellite. Although DETER releases monthly data which is more often than PRODES,

it has a limitation as it can only detect deforestation greater than 25 hectares due to

that the spatial resolution of the satellite images is 250 m (Assunção et al., 2017).

Both the PRODES and DETER programmes have contributed to the planning and

development of public policies in the Legal Amazon.

2.2.4 Supply chain interventions

In addition to governmental policies, NGOs also put pressure on deforestation

actors, such as private companies and commodity exporters. In 2006, the NGO

Greenpeace succeeded in pressuring major soy exporters to sign a “soy moratorium”

that stated that exporters cannot buy soy commodities produced from illegal

deforestation that happened after 2006 (Fearnside, 2017). Afterwards, the

moratorium was renewed in 2016 and has had an effect on reducing deforestation

(Adario, 2016; Gibbs et al., 2015). In 2009, cattle agreements named “term of

adjustment of conduct” and “zero deforestation agreement” were signed, leading to

a significant effect in the reduction of deforestation (Fearnside, 2017).

2.3 Priority areas establishment

A report by MMA (2007) stated that the priority areas were first established between

1998 and 2000 under Project for the Conservation and Sustainable Use of Brazilian

Biological Diversity (PROBIO). The result was stipulated through Law no. 5,092 that

state a 900 priority areas establishment. The purpose of the priority areas

establishment was to serve as a guideline of policies especially in conserving

biodiversity but also maintaining economic growth. The results of these priority

areas were updated in 2006 by MMA. The scope of the project covered the entire

area of Brazil. The result of this process was an increase of priority areas (MMA,

2007).

The process of establishing priority areas in 2006 was conducted in three stages:

technical meetings, data processing, and regional seminars (MMA, 2007). In the

technical meetings, three broad categories were established to determine the

biological importance of priority areas. The classes are (1) biodiversity, whether there

are endemic species, endangered species, etc., (2) sustainable use, whether there are

22

species, economic, or cultural importance, etc., and (3) processes, whether the area

possesses environmental services, biogeochemical cycles, etc. (MMA, 2007). As a

result, the biological importance was categorised as extremely high, very high, high,

and no data. While the priority areas were produced during data processing, in

regional seminars, the urgency of action to priority areas was established. The

urgency of action was categorised as extremely high, very high, and high (MMA,

2007).

The process itself involved the participation of various stakeholders with different

interests (government officials, academic experts, environmental organisations,

traditional communities and the private sector) (Castro and Albernaz, 2016).

Furthermore, the result was strongly based on the experience of the stakeholders and

available information of biodiversity in each biome.

Priority areas themselves were defined as areas with high quality of native

ecosystems and biodiversity richness and were created for environmental

conservation. Priority areas were divided into two types named protected and

unprotected areas. Protected areas as defined in the SNUC (Law no. 9,985, 2000)

belong to priority areas and are protected by legal mechanisms that prevent the

conversion of native vegetation into agricultural land. On the other hand,

unprotected areas are areas also with high biodiversity value but are not protected

and available for legal deforestation (CBD, 2007).

Currently, the second update of priority areas is ongoing. The update for the Cerrado,

Pantanal, and Caatinga biomes was done in 2016 while the update for the Amazon,

Pampa, Atlantic forest, and Coastal and Marine Zones is expected to be finished by

the end of 2018 (MMA, 2017).

2.4 Ecosystem services and biodiversity in Legal Amazon

Legal Amazon hosts the largest area of tropical rainforest in the world. It stores

carbon as much as 54 Gt and maintains habitat for 5,526 vertebrate species

(Mittermeier et al., 2002). Legal Amazon consists of three biomes: the Amazon and

parts of the Cerrado and Pantanal biomes. A study by Vieira et al. (2008) pointed

out that the Amazon biome consists of at least 50,000 plant species of which 30,000-

35,000 endemic. Meanwhile, the Cerrado biome accommodates about 90,000 insect

species, 150 mammal species, and 550 bird species, which are also endemic (Ratter

et al., 1997). Lastly, the Pantanal biome contains several aquatic species and stores a

large quantity of carbon (Liengaard et al., 2012). The diverse plant and vertebrate

species in these three biomes are highly threatened by deforestation.

23

Considering the capacity above of Legal Amazon, deforestation in Legal Amazon will

not just affect this area but also the sustainability of the world. Deforestation means

loss of ecosystem services including its roles in storing carbon, maintaining

biodiversity and recycling water (Fearnside, 2017). Due to the current deforestation

and forest degradation, many endemic species are affected and restricted to parts of

their habitat (Michalski and Peres, 2005). Also, especially tree and fauna mortality

is strongly affected by edge effects of forest fragmentation (Peres et al., 2010).

The main drivers of deforestation in Legal Amazon are human activities, mainly in

the form of land use and land cover changes from rainforest to agricultural land and

road and urban infrastructure, which are likely to continue to happen in the future

(Goldewijk, 2001). These activities resulted in direct effect on ecosystem dynamics,

among others nitrogen dynamic (Melillo et al., 2001), water balance, biodiversity

loss (Foley et al., 2005), as well as forest degradation, fragmentation and destruction

(Freitas et al., 2018). A study by Barlow et al., (2016) revealed that biodiversity is

decreasing at alarming rates.

The policies and initiatives to protect biodiversity in the Legal Amazon should at

least maintain core ecosystem services, among others nutrient cycling, water

balance, and hydrological cycles. They should as well involve environmental

organisations, the private sector and the civil society in the conservation efforts

(Ometto et al., 2011).

24

3 Methodology

3.1 Description of the study area

The study area was the Legal Amazon region, an administrative division of Brazil

which consists of seven states (Acre, Amapa, Amazonas, Para, Rondonia, Roraima,

and Tocantins) and parts of two states (Mato Grosso and Maranhao). Legal Amazon

is established as a legal region to serve the purpose of regional planning and public

policy. It covers 560 million hectares which are 59 % of the total area of Brazil.

Around 24 million inhabitants are living in Legal Amazon. Legal Amazon comprises

three biomes including the Amazon and parts of the Cerrado and the Pantanal

biomes.

85 % of the Amazon biome is covered with native vegetation which is dominated by

rainforest (IBGE, 2004). On the other hand, the Cerrado biome is dominated by

savanna and seasonal tropical forest which holds many endemic species, while the

Pantanal biome consists mostly of tropical wetland with parts covered by semi-

deciduous forest (IBGE, 2004; Ratter et al., 1997).

Legal Amazon is famous and known for its significant importance in the provision of

ecosystem services, such as storing carbon, maintaining biodiversity, and recycling

water (Fearnside, 2017). However, due to national and global demands for food,

timber and bioenergy, agricultural expansion is expected to happen which gives

pressure to the pristine forest of Legal Amazon (MAPA, 2016). As a result, the forest

has undergone rapid deforestation and changed into agricultural and farmland.

Respectively, in Legal Amazon, priority and protected areas was established with the

main purpose of protecting the biodiversity.

25

Figure 5. Legal Amazon in Brazil

3.2 Overview of the research methodology

The following workflow diagram as shown in Figure 6 illustrates the overview of the

methodology. After conducting a literature review, the datasets were processed using

GIS software (ESRI, 2011) to produce the data needed to be further processed

statistically, using SPSS software (IBM Corp., 2016) and visualised using the Tableau

software (Tableau Software, 2017). The final results of the analysis are figures, tables,

and maps for demonstrating the effect of deforestation on priority areas in Legal

Amazon spatially and statistically.

Figure 6. Overview of the research methodology

26

3.3 Literature review

A literature review was conducted covering relevant scientific literature, policy

documents, governmental reports and websites as well as environmental

organisation reports. An extensive review was carried out to understand the history

of deforestation in Legal Amazon that is addressed by sets of diverse policies by

governments and initiatives by environmental organisations. In addition, a review of

governmental official reports was done to understand the establishment of priority

areas in Legal Amazon. Lastly, a review of scientific papers was done to comprehend

the method to address the efficiency of priority and protected areas in preventing

deforestation and protecting biodiversity.

3.4 Data acquisition

Four main datasets were used in the spatial and statistical analysis produced in this

study, including a) a dataset of priority areas for conservation in the Legal Amazon

conducted by the Brazilian Ministry of Environment (MMA) in 2006; b) a dataset of

deforestation of Legal Amazon from 2001 to 2014 that is yearly conducted by INPE

(Instituto Nacional de Pesquisas Espaciais) through the PRODES program; c) an

area of Legal Amazon by INPE and d) a dataset of municipalities in Legal Amazon by

the Brazilian Institute of Geography and Statistics (IBGE) in 2018. All datasets were

set in a projected coordinate system named South America Albers Equal Area Conic

so that the total area of the polygons could be calculated.

The priority areas dataset has 1,125 priority areas and consist of 616 protected areas

and 509 unprotected areas. The priority areas are displayed as polygons which are

features of closed shape. Each polygon has data of its biological importance and

location in the biome. The dataset was derived from the MMA website. On the other

hand, the municipalities’ datasets contained 1,046 municipalities’ polygons and was

derived from the IBGE website. Meanwhile, the area of the Legal Amazon dataset is

also displayed as polygon features showing border lines of Legal Amazon and also

states border lines inside Legal Amazon.

27

Figure 7. Priority areas of Legal Amazon (MMA, 2006)

Lastly, the deforestation dataset were derived from the Global Forest Watch (GFW)

website instead of the INPE website because GFW provides pre-processed

deforestation data in the format needed for conducting the analysis of this study.

Nevertheless, the dataset from GWF is derived from the dataset from INPE,

presenting the same information. There are 1,075,127 polygons in the deforestation

dataset which contains data of the year the deforestation occurred from 2001 to

2014, the location in the biome, and the total area. Respectively, deforestation

defined here is only clear-cutting deforestation and not forest degradation and tree

loss by selective logging (Assunção et al., 2017).

3.5 Dataset processing

After data acquisition, the datasets were processed using GIS software. The process

was divided into two parts including identifying deforestation in reference to priority

and protected areas and identifying deforestation in reference to the years that the

deforestation occurred.

28

3.5.1 Identifying deforestation in priority and protected areas

In identifying priority and protected areas with reference to deforestation, the

datasets of deforestation, priority areas, and municipalities were added into GIS

software. The overall dataset processing of this part is shown in Figure 8.

Figure 8. Data processing for identification priority and protected areas in

reference to deforestation

The deforestation dataset which consists of deforestation from 2001 to 2014 was

exported based on year. Exporting data works by allowing the user of GIS software

to optionally choose specific data, change the format of the data, and produce new

datasets. In this study, data was exported by choosing data based on year (Figure 9).

As a result, 14 layers of deforestation from year 2001 to 2014 were produced.

However, since priority areas used in this study was established in 2006, only

deforestation datasets from 2006 to 2014 were used.

29

Figure 9. Exporting deforestation data based on years

Hereafter, the layer of deforestation of each year and priority areas dataset were set

as input features for the geometric intersection function. The intersection function

works by producing overlapping features or portions of features in both layers as the

output feature class. As a result, 9 layers of deforestation that occurred only in

priority areas were produced. An example of deforestation that occurred in priority

areas in 2006 along with its attribute table is displayed in Figure 10. The

deforestation is displayed as polygons. Then the area of these polygons were

calculated in km2 using the calculate geometry function.

Figure 10. The deforestation layer in priority areas in 2006 with its attribute

table

30

Then, the tables of each layer were extracted into a database file format which is

needed so that all the tables can be joined in the next step. A database file is one of

the file types in GIS which only contains tables. Next, the tables were joined into a

table of priority areas dataset using the join function. Joining data works by

appending the fields of one table to those of another through an attribute or field

common to both tables. In this case, the common field was ID of priority areas and

the tables of deforestation from 2006 to 2014 were joined into a table of priority

areas. As a result, the map and table of deforestation from 2006 to 2014 in each

polygon of the priority areas were produced. Table 1 shows a slice of 20 records from

the final dataset.

Table 1. A slice of the final table of deforestation in priority areas

A similar procedure was done to produce a map and table of deforestation in non-

priority areas. In this case, the municipalities’ boundaries was used. The

municipalities’ dataset was processed together with the priority areas dataset with

the erase function in GIS to produce polygons in non-priority areas as shown in

Figure 11. Erase works by overlaying the polygons of the erase coverage with the

features of the input coverage. The portions of the input coverage features outside

the erase polygons outer boundaries were copied to the output coverage. As a result,

a layer of polygons outside priority areas was produced.

Figure 11. The process of creating polygons outside the priority areas using

the erase function

31

Hereafter, the layer of deforestation for each year from 2006 to 2014 and the layer

of polygons outside the priority areas dataset were set as input features for the

geometric intersection function. As a result, 9 layers of deforestation occurred

outside the priority areas were produced. An example of deforestation occurred

outside the priority areas in 2006 along with its attribute table is displayed in Figure

12. The deforestation is displayed as polygons. Then the area of these polygons were

calculated in km2 using the calculate geometry function.

Figure 12. Deforestation outside priority areas in 2006 with its attribute table

Then, the tables of each layer were extracted in a database file format. Next, the

tables were joined into a table of polygons outside the priority areas dataset using

the join function. In this case, the common field to join the data was ID of

municipalities. Then, the tables of deforestation from 2006 to 2014 were joined into

a table of polygons outside priority areas. As a result, the map and table (Table 2) of

deforestation from 2006 to 2014 in each polygon outside priority areas were

produced. Table 2 shows a slice of 20 records from the final dataset.

Table 2. A slice of the final table of deforestation outside priority areas

32

Finally, the dataset of deforestation inside and outside priority areas were combined

and became the final dataset containing 616 protected areas, 509 unprotected areas,

and 504 non-priority areas. Information about the biological importance, total area

in km2, and total deforestation of each year from 2006 to 2014 were attached to each

polygon. Then, the final dataset were converted to an excel file (Appendix 1) and

processed further in SPSS software which will be explained in Subsection 3.6.

3.5.2 Identifying deforestation based on years

In identifying deforestation with reference to the year the deforestation occurred, the

dataset of deforestation and priority areas were added into the GIS software. The

overall dataset processing of this part is shown in Figure 13.

Figure 13 Data processing for identifying deforestation

33

First, the priority areas was exported into 2 types of areas, among others protected

areas and unprotected areas. The priority areas were also processed with the Legal

Amazon area with the erase function to produce a layer of non-priority areas. Then,

the deforestation dataset was intersected with each layer of protected areas,

unprotected areas, and non-priority areas. The results were merged to produce a

deforestation dataset with additional data of area type to each deforestation polygon.

Lastly, the area was calculated in km2.

Figure 14. Deforestation in protected and unprotected areas

The final dataset (Table 3) contained 1,171,986 deforestation polygons with

information of year of deforestation, total area in km2, biome, area type, and

biological importance. It was attached to each polygon and converted into a csv file

and processed further in the Tableau software, as explained in Subsection 3.7.

Table 3. A Slice of the final table of deforestation based on type of area

34

3.6 Analysis of variance (ANOVA)

An analysis of variance (ANOVA) was performed with the SPSS software to analyse

the data resulting from the data processing in the GIS software. ANOVA was used to

test whether there are statistically significant differences between the means of two

or more independent variables in the dependent variable. ANOVA works by

examining the variance between the data samples to variances within particular

samples. In this study, one-way ANOVA was applied to one independent variable

and one dependent variable. The workflow diagram as shown in Figure 15 illustrates

the steps of conducting one-way ANOVA.

Figure 15. Overview of the ANOVA workflow

3.6.1 Defining independent and dependent variables

In this study, three one-way ANOVA analyses were conducted. First, one-way

ANOVA was conducted to see whether there are statistically significant differences

between deforestation in protected, unprotected, and non-priority areas. The dataset

from the data processing based on type of area was used, which contained 1,629 areas

(616 protected areas, 509 unprotected areas, and 504 non-priority areas).

Information about the biological importance, total area in km2, and total

deforestation of each year from 2006 to 2014 were attached to each areas.

In order for the deforestation of each areas to be compared in one-way ANOVA,

deforestation density (𝐷𝑒𝑓𝑑𝑒𝑛) was established. Deforestation density is defined as

the density of total deforestation from 2006 to 2014 to the total area of the polygon

based on type of area. The equation is stated as:

𝐷𝑒𝑓𝑑𝑒𝑛 = 𝑇𝑜𝑡𝑎𝑙 𝑑𝑒𝑓𝑜𝑟𝑒𝑠𝑡𝑎𝑡𝑖𝑜𝑛2006−2014 (𝑘𝑚2)

𝐴𝑟𝑒𝑎 𝑜𝑓 𝑝𝑜𝑙𝑦𝑔𝑜𝑛 (𝑘𝑚2) × 1000

35

The deforestation density was multiplied by 1000 so that the value is not too small

and can be analysed in SPSS. As a result, there are 1,629 deforestation density values

calculated for each type of area. The deforestation density was set as a dependent

variable while the type of area was the independent variable.

Regarding the second analysis, one-way ANOVA was conducted to see whether there

are statistically significant differences between deforestation in areas based on

importance (high, very high, extremely high, and no data). In this second analysis,

deforestation density was also used as the dependent variable while biological

importance was the independent variable. The dataset from data processing based

on type of area was used, which contained of 1,125 areas (249 areas of high

importance, 270 areas of very high importance, 590 areas of extremely high

importance, and 16 areas of no data) with information about the biological

importance, total area in km2, and total deforestation of each year from 2006 to 2014

attached to each area.

One-way ANOVA was conducted to analyse the impact of priority areas

establishment in 2006 in preventing deforestation. This aim was answered by

analysing the difference of deforestation in 2006 as the year of priority areas

establishment to the latest deforestation data in 2014. In this third analysis,

deforestation density in 2006 and 2014 were established. Then, deforestation

density in 2014 was reduced with deforestation density in 2006. The equation is

stated as:

𝐷𝑒𝑓𝑑𝑒𝑛2014− 𝐷𝑒𝑓𝑑𝑒𝑛2006

= (𝑇𝑜𝑡𝑎𝑙 𝑑𝑒𝑓𝑜𝑟𝑒𝑠𝑡𝑎𝑡𝑖𝑜𝑛2014 (𝑘𝑚2)

𝐴𝑟𝑒𝑎 𝑜𝑓 𝑝𝑜𝑙𝑦𝑔𝑜𝑛 (𝑘𝑚2) × 1000) − (

𝑇𝑜𝑡𝑎𝑙 𝑑𝑒𝑓𝑜𝑟𝑒𝑠𝑡𝑎𝑡𝑖𝑜𝑛2006 (𝑘𝑚2)

𝐴𝑟𝑒𝑎 𝑜𝑓 𝑝𝑜𝑙𝑦𝑔𝑜𝑛 (𝑘𝑚2) × 1000)

Finally, one-way ANOVA was run with the difference of deforestation density in year

2014 with year 2006 as the dependent variable and type of area as the independent

variable. The dataset used in this analysis was the same as the dataset from the first

analysis. The summary of the one-way ANOVA analysis is shown in Table 4.

36

Table 4. Conducted one-way ANOVA

Analysis Independent variable Dependent variable

Deforestation trend based on

type of area

Type of area (protected,

unprotected, and non-priority

area)

Deforestation density

Deforestation trend based on

biological importance of the

area

Importance of area (high, very

high, extreme high, and no

data)

Deforestation density

Deforestation trend based on

difference between

deforestation density in year

2014 and 2006

Type of area (protected,

unprotected, and non-priority

area)

Difference between

deforestation density of 2014

and deforestation density of

2006

3.6.2 Checking assumptions

Three assumptions of the data samples must be fulfilled before one-way ANOVA can

be conducted: (1) the samples are normally distributed, (2) the samples have

homogeneity of variance, and (3) the samples are independent of one another and

randomly selected. The final dataset was checked and fulfilled the assumptions.

3.6.3 Stating null and alternative hypothesis

After checking the assumptions, the null and alternative hypothesis were made. The

null hypothesis (𝐻0) states that the population means are equal as stated:

𝐻0 ∶ 𝜇1 = 𝜇2 = 𝜇3

The alternative hypothesis (𝐻𝑎) is the opposite of the null hypothesis which states at

least there is one mean of the population which is not statistically equal as stated:

𝐻𝑎 ∶ 𝐴𝑡 𝑙𝑒𝑎𝑠𝑡 𝑜𝑛𝑒 𝑚𝑒𝑎𝑛 𝑖𝑠 𝑛𝑜𝑡 𝑠𝑡𝑎𝑡𝑖𝑠𝑡𝑖𝑐𝑎𝑙𝑙𝑦 𝑒𝑞𝑢𝑎𝑙

3.6.4 Calculating the test statistic

The next step was calculating the test statistic concerning the variance between the

data samples to variances within particular samples. The summary of the test

statistic calculation is shown in Table 5.

37

Table 5. Test statistic calculation of one-way ANOVA

Source Sum of Squares Degrees of Freedom

Variance Estimate (Mean Square)

F Ratio

Between 𝑆𝑆𝐵 𝐾 − 1 𝑀𝑆𝐵 = 𝑆𝑆𝐵

𝐾−1 𝑀𝑆𝐵

𝑀𝑆𝑊

Within 𝑆𝑆𝑊 𝑁 − 𝐾 𝑀𝑆𝑊 =

𝑆𝑆𝑊

𝑁 − 𝐾

Total 𝑆𝑆𝑇 = 𝑆𝑆𝐵 + 𝑆𝑆𝑊 𝑁 − 1

Where: SSB = Sum of square between the samples

SSW = Sum of square within the samples

SST = Sum of square in total

K = Number of columns in the data table

N = Total number of observations

MSB = Mean square between the samples

MSW = Mean square within the samples

Then, the obtained F ratio was compared with Fcv. Fcv has two degrees of freedom

(df1 and df2), where df1 is the numerator degrees of freedom equal to K-1 and df2 is

the denominator degrees of freedom equal to N-K. The value of Fcv can be found in

corresponding to α = 5% by using the F tables. The null hypothesis is rejected if: F

(observed value) > FCV (critical value) which means that there is a statistical

significant difference between the means of the groups. As ANOVA is an omnibus

test statistic which mean it cannot tell which specific groups are statistically

significant different from each other. Therefore, a post hoc test named the Tukey

post hoc test was used in this study to determine which of the groups differed from

each other.

3.6.5 Interpreting the results

The one-way ANOVA process resulted in three types of tables: descriptive tables,

ANOVA tables, and multiple comparisons tables. The results are presented in the

Results chapter.

38

3.7 Visualization with Tableau

The Tableau software was used to produce interactive visualizations and tables of the

data resulting from the data processing in GIS. The results are presented in the

Results chapter.

39

4 Results

This chapter presents the results of visualization using the Tableau software and the

one-way ANOVA results using the SPSS software.

4.1 Deforestation trend based on area type

Figure 16 illustrates the deforestation in protected priority areas, unprotected

priority areas and non-priority areas from year 2001 to 2014. Overall, the

deforestation increased from 2001 and reached the most peaked deforestation in

2004. Afterwards, it decreased drastically from 2004 to 2006 although it increased

a little in 2008 before deforestation decreased and stabilized from 2009 to 2014.

Figure 16. Deforestation trend based on type of area

Half of the deforestation from 2001 to 2014 occurred in non-priority area as

illustrated in Table 6, where 94,260 km2 of forest had been deforested. Without land-

use restrictions, deforestation that reached to 43% still took place in unprotected

priority areas. On the other hand, protected areas with land-use restrictions are the

area with the lowest occurred deforestation which is only 7 % of the total

deforestation.

40

Table 6. Deforestation based on type of area

Year Deforestation (km2/year)

Protected priority areas

Unprotected priority areas

Non-priority areas

2001 859 7,484 11,176

2002 1,456 9,577 13,622

2003 1,695 10,863 13,567

2004 2,043 11,753 13,086

2005 1,399 10,422 11,882

2006 730 4,975 5,171

2007 985 5,116 5,362

2008 997 6,423 5,934

2009 716 2,801 3,076

2010 619 3,161 2,551

2011 560 2,551 2,598

2012 462 2,071 1,913

2013 621 2,387 2,386

2014 421 2,072 1,936

SUM 13,563 81,656 94,260

% 7% 43% 50%

Furthermore, the ANOVA analysis shows that there is a significant difference of

deforestation density based on type of area. This one-way ANOVA produced three

tables including descriptive, ANOVA, and multiple comparisons table. Table 7

displays that protected priority areas, unprotected priority areas, and non-priority

areas had different mean values of deforestation density. Non-priority areas had the

highest deforestation density (94) which is very high compare to the deforestation

density of unprotected priority areas (16,2) and even more to protected priority areas

(7).

Table 7. Descriptive table of first analysis of one-way ANOVA

The ANOVA analysis displayed in Table 8 shows that there is a statistically

significant difference between types of area determined by one-way ANOVA

(F(2,1626) = 42.019, p<0.001) shown by the lower significance value (p<0.001)

which was below 0.05.

41

Table 8. ANOVA table of the first analysis one-way ANOVA

Furthermore, a Tukey post hoc test revealed which areas differed from each other.

The result as shown in Table 9 shows statistically significant differences in

deforestation density between protected priority areas and non-priority areas

(p<0.05) as well as between unprotected priority areas and non-priority areas

(p<0.05). However, there are no significant differences in deforestation density

between protected priority areas and unprotected priority areas (p>0.05) as shown

in Table 9.

Table 9. Multiple comparisons table of the first analysis with one-way ANOVA

4.2 Deforestation trends and the biological importance of priority areas

The diagram in Figure 17 illustrates the deforestation trend in relation to the

biological importance of priority areas. Overall, deforestation that occurred in areas

of very high and extremely high importance is more or less similar while

deforestation occurred less in areas of high importance and almost none in areas

with no data of importance.

42

Figure 17. Deforestation trends based on areas of biological importance

Most of deforestation occurred in areas with very high importance (43 %) and

extremely high importance (38 %). On the other hand, deforestation occurred only

by 17 % in areas with high importance and 2 % in areas with no data of importance

as shown in Table 10.

Table 10. Deforestation trend based on areas of biological importance

Year Deforestation (km2/year)

High Very High Extremely High No Data

2001 1,895 3,398 2,983 66

2002 2,436 4,638 3,905 54

2003 1,960 5,547 4,934 117

2004 1,957 6,470 5,181 187

2005 2,019 5,112 4,503 187

2006 1,018 2,438 2,132 116

2007 1,183 2,508 2,302 108

2008 1,255 2,986 3,077 102

2009 555 1,559 1,313 90

2010 622 1,303 1,775 80

2011 406 1,352 1,265 87

2012 450 945 1,011 127

2013 397 1,251 1,205 155

2014 362 1,086 921 124

SUM 16,515 40,593 36,507 1,600

% 17% 43% 38% 2%

43

Furthermore, the second analysis with one-way ANOVA shows that there is a

significant difference of deforestation density based on biological importance of the

area. This one-way ANOVA produced three tables including descriptive, ANOVA,

and multiple comparisons tables. Table 11 displays that areas of high, very high,

extremely high, and no data of importance had different mean values of

deforestation density. Deforestation density of areas with high (15.57) and very

importance (15.19) was more or less similar. The deforestation density of areas with

extremely high (7.68) and no data (7.32) of importance had lower value than the

areas mentioned before.

Table 11. Descriptive table of the second analysis with one-way ANOVA

The ANOVA analysis displayed in Table 12 shows that there is a statistically

significant difference between areas of different biological importance determined

by one-way ANOVA (F(3,1121)=5.979, p<0.001) shown by a lower significance value

(p<0.001) which is below 0.05.

Table 12. ANOVA table of the second analysis by one-way ANOVA

Furthermore, the Tukey post hoc test revealed which areas differed from each other.

The result as shown in Table 13 shows statistically significant differences in

deforestation density between areas of high and extremely high importance

(p=0.003) and areas of very high and extremely high importance (p=0.004).

44

Table 13. Multiple comparisons table of the second analysis with one-way

ANOVA

4.3 Deforestation differences of year 2006 and 2014

The third analysis with one-way ANOVA showed that there is a significant difference

of deforestation density between deforestation that occurred in 2006 and in 2014.

This one-way ANOVA produced three tables including descriptive, ANOVA, and

multiple comparisons tables. Table 14 displays that the differences of deforestation

density was highest in non-priority areas (-3.88) while it was very low in protected

areas which is only -0.37.

Table 14. Descriptive table of the third analysis with one-way ANOVA

The ANOVA analysis displayed in Table 15 shows that there is a statistically

significant difference between type of area determined by one-way ANOVA

(F(2,1629)=4.203, p=0.015) shown by lower significance value (p<0.015) which is

below 0.05.

45

Table 15. ANOVA table of the third analysis with one-way ANOVA

Furthermore, a Tukey post hoc test revealed which areas differed from each other.

The result as shown in Table 16 shows statistically significant differences in

deforestation density between protected and non-priority areas (p=0.013).

Table 16. Multiple comparisons table of the third analysis with one-way

ANOVA

46

5 Discussion

This study focused on analysing the impact of priority and protected areas by the

deforestation trend in Legal Amazon. Furthermore, this study also analysed the

impact of deforestation in areas based on biological importance. Finally, analysis

regarding the implication of priority areas establishment was also conducted.

5.1 Deforestation trends analysis

Priority and protected areas are areas of high biodiversity value, which are

established to conserve biodiversity and maintain economic growth in a sustainable

way (CBD, 2007). While protected areas under SNUC Law no. 9,985 are included in

public lands, priority areas under Decree no. 5,092 in 2006 are to a large extent are

included in undesignated lands (Damasceno et al., 2017; Damasceno, 2016). This

undesignated land has unclear land-use rights with high rates of illegal deforestation

although most of the areas contained rich biodiversity (Damasceno et al., 2017;

Damasceno, 2016). Therefore, in a way to guide the creation of protected areas,

priority areas were established. Protecting a legalizing is crucial considering that

clear ownership of land enable more intensive monitoring by the government and

indigenous people (Damasceno et al., 2017; Nepstad et al., 2006; Soares-Filho et al.,

2010).

Differences in regulation and ownership of land between protected priority areas,

unprotected priority areas, and non-priority areas resulted in different deforestation

trends in each area. The results demonstrates that there is not much difference

between total deforestation in unprotected areas (77,041 km2, 42 %) and non-

priority areas (94,260 km2, 51 %). In contrast, total deforestation in protected areas

was much smaller (13,563 km2, 7%) than unprotected and non-priority areas. This

indicates that protected areas have impact while priority areas that are unprotected

do not bring much impact in preventing deforestation.

However, based on deforestation density, priority areas do have an impact. The

mean of deforestation density in unprotected priority areas was 16.2 which is 5.8

times lower than in non-priority area that had a value of 94. Further, deforestation

density in protected areas was 7 which is 2 times lower than deforestation density in

unprotected areas. It can be concluded that priority areas have an impact in

preventing deforestation. Even so, protected areas have more significant impact in

slowing down deforestation. These results converge with previous research by

Nepstad (2014) and Soares-Filho et al. (2010). Nepstad (2014) argued that protected

areas expansion reduced the area of undesignated lands while Soares-Filho et al.

(2010) suggested that expansion of protected areas led to 37 % of reduction in

deforestation. The most likely reason for the significant impact of protected areas

compared to priority areas is that protected areas are monitored more intensively by

the government and indigenous people. This makes it more difficult for deforesters.

47

In contrast, the priority areas that are mostly undesignated lands give more

opportunities for deforesters, who promote deforestation in this areas. Thus, highly

biodiverse areas under priority areas for biodiversity conservation may need to be

converted to protected areas to facilitate surveillance and ensure its preservation of

nature in these regions.

The spatial analysis showed that deforestation was concentrated in non-priority

areas that surround the priority and protected areas. This finding is consistent with

studies by Armsworth et al. (2006) and Wittemyer et al. (2008) who stated that

deforestation still occurred and only moved to adjacent non-priority areas. Even

though protected areas are restricted from land-use, there are still small patches of

deforestation. Deforestation that occurred in restricted land-use of protected areas

shows the necessity to increase the monitoring in these vast areas and also increase

the integration of governmental actors (the army, police, and agencies) as well as

indigenous people (References).

Based on total deforestation, 43 % deforestation occurred in areas of very high

importance and 38 % in areas of extremely high importance. On the other hand,

deforestation occurred less in areas of high importance (17 %). Thus, deforestation

affected areas of high biodiversity value regardless the biological importance status

of the areas. Furthermore, based on deforestation density, deforestation density in

areas of extremely high importance (7.68) was 2 times lower than in areas of high

(15.57) and very high importance (15.19). It can be concluded that deforestation

affected more in areas with less biodiversity value.

Finally, the impact of Decree No 5,092 in 2006 was analysed by calculating the

difference of deforestation density between deforestation that occurred in 2006 and

2014. It is assumed that if priority areas are effective, the difference of deforestation

density in priority areas will be higher than in non-priority areas. However, the

results demonstrated that the difference of deforestation density was highest in non-

priority areas (-3.88) while it was very low in protected areas which is only -0.37.

The most likely reason for these results are that priority areas are not necessarily the

only factor that affected the drop of deforestation. Nepstad (2014) stated that the

drop of deforestation occurred because of interconnectedness of supply chain

intervention, enforcement of law, economic conditions as well as the expansion of

protected areas.

It is worth highlighting that deforestation is still behind the Brazilian government

target, as Brazil has set a target in the Conference of Parties of the United Nations

Framework Convention on Climate Change in 2009 to decrease the national GHG

emissions by 2020 (Jean et al., 2011). This target can be accomplished by reducing

deforestation in Legal Amazon by 80 % over a decade (Jean et al., 2011). Therefore,

the regulation of priority and protected areas must be evaluated and improved in the

future.

48

5.2 Limitations

This study had several limitations. Firstly, this study only considered deforestation

as defined by the PRODES system even though PRODES can only detect

deforestation as large as 6.25 hectares. This underestimates the loss through forest

degradation, selective logging, and decline of conservation capacity (Nepstad et al.,

1999). Secondly, the simple inside-outside comparison conducted in this study may

have overestimated the effect of priority and protected areas (Andam et al., 2008).

This study do not consider that location of priority and protected areas influences

the chance of deforestation. Remote areas are and areas surrounded by other areas

are less likely to be deforested even without priority and protected areas

establishment (Pressey et al., 1993; DeFries et al., 2005). Furthermore, this kind of

establishments can displace the deforestation to non-priority areas as Joppa et al.

(2008) stated as ‘neighbourhood leakage’.

5.3 Suggestions for future work

Changes in deforestation trends in Legal Amazon can be attributed to various drivers

such as highway construction, agricultural intensification, cattle ranching and

soybean expansion, land tenure establishment, and policy failures (Jean et al., 2011).

The complex nature and interconnectedness between these drivers has resulted in

widely varied of deforestation rates over the decades. However, this study only

considered priority and protected areas establishment as a driver that affect

deforesters in choosing the type of area to be deforested while this might not be the

case in actuality. Nevertheless, this study is able to draw deforestation trends in

accordance to priority and protected areas from 2001 to 2014. However, this study

can be further developed by considering other drivers which will become a more

representative input as so produce more actual results. Otherwise, this study can also

analyse deforestation in priority and protected areas while at the same time

considering the size, the location, the purpose, and the responsible party of the areas

which should produce a better assessment of the impacts of priority and protected

areas.

49

6 Conclusions

This thesis provided series of spatial and statistical analysis to analyse the impact of

priority and protected areas on deforestation trends in the Brazilian Legal Amazon.

From the outcomes of this thesis, it is possible to conclude that deforestation had

been statistically more intensive in non-priority areas when compared with priority

areas. Furthermore, land-use restrictions in protected areas had more impact

compared to unprotected areas.

In addition, the statistical analysis of deforestation according to biological

importance of the priority areas showed that deforestation had been more intensive

in areas classified as having lower biological importance. Yet, substantial tracts of

forest had been converted within areas of high biological importance. Areas of

extremely high importance for biodiversity had 2 times lower deforestation density

than areas of high and very high importance.

Nonetheless, the statistical analysis regarding implication of Decree no. 5,092 in

2006 showed that the difference in deforestation density that occurred in 2006 and

2014 was highest in non-priority areas while it was very low in protected areas. It can

be concluded that priority areas not necessarily affected the deforestation drop but

might this may have been caused by other policies. This implies that the regulation

of priority and protected areas can be improved in the future. Therefore, the creation

of protected areas constitute a key towards the preservation of biodiversity.

51

References

Adario, P., 2016. The soy moratorium, 10 years on: How one commitment is stopping

Amazon destruction? Amsterdam, the Netherlands: Greenpeace International.

Andam, K.S., Ferraro, P.J., Pfaff, A., Sanchez-Azofeifa, G.A. & Robalino, J., 2008.

Measuring the effectiveness of protected areas networks in reducing

deforestation. Proceedings of the National Academy of Sciences USA, 105,

16089–16094.

Arima EY, Barreto P, Araújo E, Soares-Filho B., 2014. Public policies can reduce

tropical deforestation: Lessons and challenges from Brazil. Land Use Policy

41:465–473.

Armsworth, P.R., Daily, G.C., Kareiva, P. & Sanchirico, J.N., 2006. Land market

feedbacks can undermine biodiversity conservation. Proceedings of the

National Academy of Sciences USA, 103, 5403–5408.

Assunção J, Gandour C, Rocha R, 2015. Deforestation slowdown in the Brazilian

Amazon: prices or policies? Environment and Development Economics, 20,

697-722.

Assunção, J., Gandour, C., Pessoa, P., & Rocha, R., 2017. Property-level assessment

of change in forest clearing patterns: The need for tailoring policy in the

Amazon. Land Use Policy, 66, 18-27. doi:10.1016/j.landusepol.2017.04.022

Barlow, J., Lennox, G.D., Ferreira, J., Berenguer, E., Lees, A.C., Nally, R.M.,

Thomson, J.R., Ferraz, S.F.d.B., Louzada, J., Oliveira, V.H.F., Parry, L., Ribeiro

de Castro Solar, R., Vieira, I.C.G., Aragão, L.E.O.C., Begotti, R.A., Braga, R.F.,

Cardoso, T.M., Jr, R.C.d.O., Souza Jr, C.M., Moura, N.G., Nunes, S.S., Siqueira,

J.V., Pardini, R., Silveira, J.M., Vaz-de-Mello, F.Z., Veiga, R.C.S., Venturieri, A.,

Gardner, T.A., 2016. Anthropogenic disturbance in tropical forests can double

biodiversity loss from deforestation. Nature 535, 144-147.

Bergen, M., 2018. Stopping deforestation in the Amazon by 2020? Human Nature –

Conservation International Blog. Available at:

https://blog.conservation.org/2016/08/stopping-deforestation-in-the-

amazon-by-2020-its-possible/ [Accessed 29 Jun. 2018].

Brasil, 2012. Brazilian Native Vegetation Protection Law "LEI Nº 12.651, DE 25 DE

MAIO DE 2012", Law Nº 12.651, 25-May-2012, Brasilia, Brazil.

Castro, R. and Albernaz, A., 2016. Consistency and use of information about threats

in the participatory process for identification of priority conservation areas in

the Brazilian Amazon. Journal for Nature Conservation, 30, pp.44-51.

52

CBD, 2007. Review of implementation of the programme of work on protected areas

for the period 2004–2007 UNEP/CBD/WGPA/2/2.

Dalla-Nora E.L., de Aguiar A.P.D., Lapola D.M. and Woltjer G., 2014. Why have land

use change models for the Amazon failed to capture the amount of

deforestation over the last decade? Land Use Policy 39:403–411.

Damasceno, R., 2016. Insecure land rights in Brazil: Consequences for rural areas

and challenges for improvement. In: Núcleo de Avaliacao de Politicas

Climaticas. PUC-Rio, Climate Policy Initiative.

Damasceno, R., Chiavar, J., & Lopes, C. L., 2017. Evolution of land rights in Brazil.

In: Nucleo de Avaliacao de Politicas Climaticas. PUC-Rio, Climate Policy

Initiative.

Davidson EA, et al., 2012. The Amazon basin in transition. Nature 481(7381):321–

328.

DeFries, R., Hansen, A., Newton, A. and Hansen, M., 2005. Increasing Isolation of

Protected Areas in Tropical Forests Over the Past Twenty Years. Ecological

Applications, 15(1), pp.19-26.

ESRI, 2011. ArcGIS desktop, Release 10 (Tenth edition). Redlands, CA:

Environmental Systems Research Institute.

Fearnside, P., 2017. Deforestation of the Brazilian Amazon. Oxford Research

Encyclopedia of Environmental Science.

Fearnside, P.M., 2009. Carbon benefits from Amazonian forest reserves: Leakage

accounting and the value of time. Mitigation and Adaptation Strategies for

Global Change 14(6):557-567. [doi: 10.1007/s11027-009-9174-9].

Foley JA, Defries R, Asner GP et al., 2005. Global consequences of land use. Science

309, 570–574.

Freitas, F. L. M., 2017. Brazilian land use policies and the development of ecosystem

services. Stockholm: KTH Royal Institute of Technology.

Freitas, F. L. M., Englund, O., Sparovek, G., Berndes, G., Guidotti, V., Pinto, L. F. G.,

& Mörtberg, U., 2018. Who owns the Brazilian carbon? Global change biology.

doi:10.1111/gcb.14011

Gibbs, H.K., L. Rausch, J. Munger, I. Schelly, D.C. Morton, P. Noojipady, B. Soares-

Filho, P. Barreto, L. Micol, and N.F. Walker., 2015. Brazil’s soy moratorium.

Science 347:377-378. [doi: 10.1126/science.aaa0181].

53

Godar, J., Gardner, T.A., Tizado, E. J., & Pacheco, P., 2014. Actor-specific

contributions to the deforestation slowdown in the Brazilian Amazon. Proc

Natl Acad Sci U S A, 111(43), 15591-15596. doi:10.1073/pnas.1322825111

Goldewijk, K.K., 2001. Estimating global land use change over the past 300 years:

The HYDE Database. Global Biogeochemical Cycles 15, 417-433.

IBM Corp., 2016. IBM SPSS Statistics for Windows, Version 24.0. Armonk, NY: IBM

Corp.

Instituto Nacional de Pesquisas Espaciais, 2009. PRODES: Assessment of

deforestation in Brazilian Amazonia. Available at

http://www.obt.inpe.br/prodes/.

INPE (2017) Projeto PRODES: Monitoramento da floresta Amazônica Brasileira por

satélite. Available at www.obt.inpe.br/prodes.

Jean P.O., Ana P.D.A., and Luiz A.M., 2011. Amazon deforestation in Brazil: effects,

drivers and challenges, Carbon Management, 2:5, 575-585, DOI:

10.4155/cmt.11.48

Joppa, L.N., Loarie, S.R. & Pimm, S.L., 2008. On the protection of ‘protected areas’.

Proceedings of the National Academy of Sciences USA, 105, 6673–6678.

Liengaard, L., Nielsen, L.P., Revsbech, N.P., Priemé, A., Elberling, B., Enrich-Prast,

A., Kühl, M., 2012. Extreme Emission of N(2)O from Tropical Wetland Soil

(Pantanal, South America). Frontiers in Microbiology 3, 433.

MAPA, 2016. Brasil Projeções do agronegócio 2015/2016 a 2025/2026, 7ª ed.

Ministério da Agricultura, Pecuária e Abastecimento, Brasília-DF.

Melillo, J., Steudler, P., Feigl, B., Neill, C., Garcia, D., Piccolo, M., Cerri, C. and Tian,

H., 2001. Nitrous oxide emissions from forests and pastures of various ages in

the Brazilian Amazon. Journal of Geophysical Research: Atmospheres,

106(D24), pp.34179-34188.

Michalski, F., and C.A. Peres., 2005. Anthropogenic determinants of primate and

carnivore local extinctions in a fragmented forest landscape of southern

Amazonia. Biological Conservation 124: 383–396. [doi:

10.1016/j.biocon.2005.01.045].

Mittermeier, R.A., Mittermeier, C.G., Gil, P.R., Pilgrim, J., Fonseca, G.A.B., Brooks,

T. & Konstant, W.R., 2002. Wilderness: Earth’s Last WildPlaces. CEMEXS.A.,

Mexico City.

54

MMA (Ministério do Meio Ambiente), 2007. Áreas Prioritárias para Conservac¸

ão,Uso Sustentável e Repartic¸ ão de Benefícios da Biodiversidade Brasileira.

pp. 29–38.Secretaria de Biodiversidade e Florestas.

MMA, 2015. Cadastro Nacional de Unidades de Conservação. MMA, Brasília, DF,

Brazil.

MMA, 2013. Action plan for prevention and control of the Legal Amazon

deforestation (PPCDAm): 3rd phase (2012-2015) towards sustainable use and

forest conservation (MMA).

MMA, 2017. National Biodiversity Strategy and Action Plan. MMA.

Morton D.C., DeFries R.S., Shimabukuro Y.E. et al., 2006. Cropland expansion

changes deforestation dynamics in the southern Brazilian Amazon. Proc. Natl

Acad. Sci. USA 103(39), 14637–14641.

Nepstad, D. C., A. Verissimo, A. Alencar, C. Nobre, E. Lima, P. Lefebvre, P.

Schlesinger, C. Potter, P. Moutinho, E. Mendoza, M. Cochrane, and V. Brooks.

1999. Large-scale impoverishment of Amazonian forests by logging and fire.

Nature 398:505–508.

Nepstad, D.C., S. Schwartzman, B. Bamberger, M. Santilli, D. Ray, P. Schlesinger, R.

Lefebvre, A. Alencar, E. Prinz, G. Fiske, and A. Rolla., 2006. Inhibition of

Amazon deforestation and fire by parks and indigenous lands. Conservation

Biology 20:65-73. [doi: 10.1111/j.1523-1739.2006.00351.x].

Nepstad, D., McGrath, D., Stickler, C., Alencar, A., Azevedo, A., Swette, B., and Hess,

L., 2014. Slowing Amazon deforestation through public policy and

interventions in beef and soy supply chains.(Report)(Author abstract). Science,

344(6188), 1118. doi:10.1126/science.1248525

Peres C.A., Gardner T.A., Barlow J. et al., 2010. Biodiversity conservation in

humanmodified Amazonian forest landscapes. Biol. Conserv. 143 (10), 2314–

2323.

Pressey, R.L., Humphries, C.J. & Margules, C.R., 1993. Beyond opportunism: key

principles for systematic reserve selection. Trends in Ecology and Evolution, 8,

124–128.

Ratter, J.A., Ribeiro, J.F., Bridgewater, S., 1997, The Brazilian Cerrado Vegetation

and Threats to its Biodiversity. Annals of Botany 80, 223-230

Ramos, A.R., 1998. Indigenism: Ethnic Politics in Brazil. University of Wisconsin

Press, Madison, Wisconsin.

55

Ricketts T.H., Soares-Filho B., da Fonseca G.A.B. et al., 2010. Indigenous lands,

protected areas, and slowing climate change. PLoS Biol. 8(3), e1000331.

Rylands, A. and Brandon, K., 2005. Brazilian Protected Areas. Conservation Biology,

19(3), pp.612-618.

Soares-Filho, B., Moutinho, P., Nepstad, D., Anderson, A., Rodrigues, H., Garcia, R.,

and Maretti, C., 2010. Role of Brazilian Amazon protected areas in climate

change mitigation. Proceedings of the National Academy of Sciences of the

United States, 107(24), 10821. doi:10.1073/pnas.0913048107

Sparovek G., Berndes G., Barretto A. G. D. O. P., Klug I. L. F., 2012. The revision of

the Brazilian Forest Act: increased deforestation or a historic step towards

balancing agricultural development and nature conservation? Environmental

Science & Policy, 16, 65-72.

Tableau Software. 2017. Tableau Software version 10.4.

Vieira ICG, Toledo PM, Silva JMC, Higuchi H., 2008. Deforestation and threats to

the biodiversity of Amazonia. Braz. J. Biol. 68(Suppl. 4), 949–956.

Wittemyer, G., Elsen, P., Bean, W.T., Coleman, A., Burton, O. & Brashares, J.S.,

2008. Accelerated human population growth at protected area edges. Science,

321, 123–126.

TRITA-ABE-MBT 18400

www.kth.se