Drivers of deforestation and Forest degradation in nghe an Province under the context of Climate action: case study in the Two communes of con cuong and Thanh chuong districts

directed and adapt to compete the working environment such as English proficiency, Japanese communication skills, having on time management, using basic computer skills p[r]

VIETNAM JAPAN UNIVERSITY

DO THI NHINH

DRIVERS OF DEFORESTATION AND FOREST DEGRADATION IN NGHE AN PROVINCE UNDER THE CONTEXT OF CLIMATE ACTION: CASE STUDY IN THE TWO COMMUNES OF CON CUONG AND

THANH CHUONG DISTRICTS

MASTER’S THESIS

VIETNAM JAPAN UNIVERSITY

DO THI NHINH

DRIVERS OF DEFORESTATION AND FOREST DEGRADATION IN NGHE AN PROVINCE UNDER THE CONTEXT OF CLIMATE ACTION: CASE STUDY IN THE TWO COMMUNES OF CON CUONG AND

THANH CHUONG DISTRICTS

MAJOR: CLIMATE CHANGE AND DEVELOPMENT

PLEDGE

I assure that this thesis is the result of my own research and has not been published anywhere The use of other research results and documents must comply with the regulations Citations and references to documents, books, research papers and websites must be in the list of references of the thesis

Author of the thesis

(Signature)

Đỗ Thị Nhinh

TABLE OF CONTENTS

PLEDGE i

LIST OF TABLES iv

LIST OF FIGURES v

LIST OF ABBREVIATIONS vi

ACKNOWLEDGMENT vii

CHAPTER 1 INTRODUCTION 1

1.1 Background and motivation 1

1.2 Research objectives and tasks 4

1.3 Research framework 5

1.4 Deforestation in the world 6

1.5 Deforestation in Vietnam 9

1.6 Impacts of deforestation on GHGs mitigation, global climate and impacts of climate change on the forests 10

1.6.1 Impacts of deforestation on GHGs mitigation 10

1.6.2 Impacts of deforestation on global climate 11

1.6.3 Impacts of climate change on the forests 14

1.7 Scope of the study 15

1.8 The research questions and hypothesis 17

1.9 Literature review 17

1.9.1 Technical component 18

1.9.2 Socio-economic component 19

1.9.3 Saola and its habitat distribution 24

CHAPTER 2 METHODOLOGY 26

2.1 Data collection 26

2.1.1 Remote sensing data 26

2.1.2 Socio-economic data 28

2.1.3 Bioclimate data 28

2.2 Methods to identify hotspots of deforestation 29

2.3 Methods to quantify the amount of CO2 emissions from deforestation 30

2.4 Methods to identify drivers of deforestation 30

2.5 Methods to project the habitat distribution of the Saola 30

CHAPTER 3 RESULTS AND DISCUSSIONS 32

3.1 Deforestation rates, hotspots and maps in Nghệ An 32

3.1.1 Deforestation at provincial scale 32

3.1.2 Deforestation at district scale 34

3.1.3 Deforestation at commune scale 37

3.2 GHGs emissions from forest cover loss in Nghệ An province 42

3.2.1 CO 2 emissions at provincial scale 42

3.2.2 CO 2 emissions at district scale 44

3.3 Identifications of deforestation drivers 49

3.3.1 Understanding about the study sites from ground 50

3.3.2 Dynamics of drivers of forest cover changes in Châu Khê and Thanh Sơn communes 55

3.3.3 Underlying forces of forest cover loss in Châu Khê and Thanh Sơn communes 61

3.3.4 Indigenous people and their awareness about the forests 62

3.4 Changes of Saola’s habitat distribution over years, from 2000 to 2080 65

3.4.1 Saola’s habitat in 2000 66

3.4.2 Saola’s habitat projected for 2020 67

3.4.3 Saola’s habitat projected for 2050 68

3.4.4 Saola’s habitat projected for 2080 69

3.5 The inter-linkages between socio-economics, deforestation, climate change, and Saola’s habitat distribution 71

3.5.1 Socio-economics, deforestation and climate change 73

3.5.2 Deforestation, climate change and changes of Saola’s habitat 73

3.5.3 Climate change, socio-economics and changes of Saola’s habitat 74

4.1 Proposals of solutions to deal with deforestation, forest degradation and Saola’s habitat 75

4.1.1 For the short-term perspective 75

4.1.2 For the medium-term perspective 77

4.1.3 For the long-term perspective 78

4.2 Recommendations for further research 79

CHAPTER 5 CONCLUSIONS 80

LIMITATIONS 83

REFERENCES 84

LIST OF PUBLICATIONS 90

APPENDIX 1: Matrix of Learning Outcome for the Master’s thesis 91

APPENDIX 2: Household survey questionnaire in English 94

APPENDIX 3: Household survey questionnaire in Vietnamese 95

LIST OF TABLES

Table 1.1 Classification of deforestation drivers 20

Table 1.2 Studies on deforestation in Vietnam 22

Table 2.1 Average income over years from 2009 to 2019, Thanh Sơn commune 28

Table 3.1 Communes of greatest primary forest area in Con Cuông, 2000-2017 37

Table 3.2 Fluctuations of forest cover in Thanh Sơn commune 39

Table 3.3 Percentage of primary forest in tree canopy cover, Nghệ An 41

2000-2017 41

Table 3.4 Tree cover depletion, AGB loss and CO2 emissions from AGB loss in Nghệ An province, 2000-2019 43

Table 3.5 Forest cover loss, AGB loss and CO2 emissions from AGB loss 45

Table 3.6 CO2 emission from forest cover loss in Thanh Chương, 2000-2019 47

Table 3.7 Average CO2 emissions from deforestation (unit: thousand tons) 48 Table 3.8 Income groups by village 57

Table 3.9 Timber income groups by village 58

Table 3.10 Principal Rotated Component Matrixa 60

Table 3.11 Land conflicts by village 62

Table 3.12 Ethnicity distribution by village 62

Table 3.13 Ethnicity groups and distribution of variables 63

Table 3.14 Suitability ranking and area of Saola’s habitat in 2000 66

Table 3.15 Suitability ranking and area of Saola’s habitat in 2020 67

Table 3.16 Suitability ranking and area of Saola’s habitat in 2050 68

Table 3.17 Suitability ranking and area of Saola’s habitat in 2080 69

Table 3.19 The interlink among deforestation, climate change and Saola’s habitat 72

LIST OF FIGURES

Figure 1.1 Research framework of the project 5

Figure 1.2 Deforestation worldwide 8

Figure 1.3 Radiative forcing induced by forest loss 13

Figure 1.4 Administrative map of Nghệ An province 15

Figure 1.5 Research flow of the Saola’s component 25

Figure 3.1 Map of primary forest cover change in Nghệ An, 2000-2017 33

Figure 3.2 Change rates of primary forest cover at district level, 2000-2017 34

Figure 3.3 Map of primary forest cover change at commune level, Con Cuông 2000-2017 35

Figure 3.4 Hotspots of primary forest cover loss in Thanh Chương, 2000-2017 36

Figure 3.5 Communes of highest forest loss in Thanh Chương, 2000-2017 38 Figure 3.6 Primary forests versus tree canopy cover in Nghệ An, 2000-2017 40

Figure 3.7 Tree cover loss, AGB loss and CO2 emissions from AGB loss in Nghệ An, 2000-2019 44

Figure 3.8 Tree cover loss, AGB loss and CO2 emissions from AGB loss in Con Cuông district during 2000-2019 46

Figure 3.9 Tree cover loss, AGB loss and CO2 emissions from ABG loss in Thanh Chương district, 2000-2019 48

Figure 3.10 GPS points of household investigations 51

Figure 3.11 Pictures of deforestation in Thanh Sơn commune, May 2020 54

Figure 3.12 Changes in primary forest cover, 2000-2017 56

Figure 3.13 Changes in tree canopy cover, 2000-2017 56

Figure 3.14 Suitability distribution map of Saola’s habitat in 2000 66

Figure 3.15 Suitability distribution map of Saola’s habitat for 2020 67

Figure 3.16 Suitability distribution map of Saola’s habitat in 2050 68

Figure 3.17 Suitability distribution map of Saola’s habitat in 2080 69

Figure 3.18 Suitability distribution of Saola’s habitat from 2020 to 2080 70

LIST OF ABBREVIATIONS

ER-PD Emissions Reduction Program Document

FAO Food and Agriculture Organization of United Nations FCPF Forest Carbon Partnership Facility

GLAD Global Land Analysis and Discovery

IPCC International Panel on Climate Change

IUCN International Union for Conservation of Nature

NDCs Nationally Determined Contributions

NTFPs Non-timber Forest Products

REDD Reducing Emissions from Deforestation and Forest

Degradation

UNFCCC United Nations Framework Convention on Climate Change

ACKNOWLEDGMENT

This study was implemented at Vietnam Japan University, under the Master’s Program in Climate Change and Development (MCCD) I am grateful to the university and the MCCD office for the research facilities and time for me to work on this project

Foremost, I would like to express the deepest gratitude to Assoc Prof Hà Quý Quỳnh, (from Vietnam Academy of Science and Technology) and Assoc., Prof Thorkil Casse (from Roskilde University in Denmark), my two supervisors for their enthusiastic guidance, encouragement and useful critiques to my master thesis I am thankful for their support and sharing of their projects’ data Without their great and patient instructions, this study would not have been feasible

My appreciation also extends to Prof Mai Trọng Nhuận for lots of his constructive comments to the earlier version of the thesis, so that I could revise the paper in a more logical and coherent structure Beyond, I am grateful to all the professors, lecturers (particularly Dr Kotera Akihiko and Dr Nguyễn Văn Quang) - and staff at the MCCD for their advice and kind assistance when I got into difficulties

I am indebted to the officials at Pù Mát national park in Nghệ An who were very supportive to my fieldwork in the province Further, I am thankful to all the informants in Con Cuông and Thanh Chương districts for their friendliness and time participating in the survey

Last but not the least, my special thanks should go to Prof Ito Tetsuji (together with Dr Yuki Ishikawa-Ishiwata) from Ibaraki University for various profound consultations during the entire project and for their timely motivation so that I could finalize this study Besides, I deeply appreciate the unconditional support

and great love of my family, my friends and those who kept me moving on The project would not have become possible without their encouragement The thesis is financially covered by the project “The emergence of the environmental state under authoritarianism (China, Vietnam and beyond)” funded by the Danish Ministry of Higher Education and Science.

CHAPTER 1 INTRODUCTION

1.1 Background and motivation

The planet is losing trees faster than ever Approximately thirteen million hectares (ha) of forests worldwide or estimated later, an area comparable to the size of Italy is being lost every year (FAO, 2010; Fred, 2018) In the context of climate change and sustainable development, forest management plays a crucial role via mitigating greenhouse gases (GHGs), mainly carbon, emissions that are fastening the warming of the globe in recent decades However, forests are also impacted by climatic changes and their contribution to mitigation strategies may be affected by stresses possibly inducing from it In terms of socio-economics, the global forests are also essential because they can provide goods, services, and other monetary values to various population groups Within the context of the worsening global warming, mitigation options have

to receive sought attention because of the substantial function of the forests in carbon cycle Thanks to the capacity of sequestering carbon in the trunks, roots and branches of the trees, the forests were either reported or forecasted to remove from 3,300 to 5,800 MtCO2/year between 1993-2003 (disregarding emissions from land-use change); 5,800 MtCO2/year in 1990s or averagely 5,380 MtCO2/year up until 2050 or potentially mitigating 11.679 MtCO2/ year

in 2010 (Denman et al., 2007; IPCC, 2007a)

The huge evidenced depletion and degradation of the planet’s forests and its consequences has drawn great attention to forests restoration and afforestation

in the Paris Agreement 2015, the United Nations (UN) 2030 Agenda for Sustainable Development (Goal 13a, Goal 15.2), and the 2014 New York Declaration on Forests at the UN Summit as the forests absorb approximately one-third of the CO2 emitted from fossil fuels combustion on a yearly basis (IUCN, 2016) A restoration of 350 million hectares of the world’s deforested

and degraded land by 2030, according to the 2011 Bonn Challenge, could sequester up to 1.7 giga tons of carbon dioxide equivalent every year to meet the objectives of hampering the increase of the global temperature to well below 2 degrees Celsius (0C) of the Paris Agreement (Bonn Challenge, 2011) Though determinants of forest depletion and losses have been listed, they have not been explored thoroughly despite the fact that the rate of deforestation and forest degradation is still on increase, particularly in developing countries including Vietnam according to various inventory sources (FAO, 2010) Using satellite image analysis at the global scale, Hansen et al (2013) concluded that Vietnam classifies as a top nation for gross tree cover loss in the beginning of the 21st century (p.851)

Due to limited studies on deforestation in Vietnam at national scale, I used QGIS 3.10 to analyze the satellite images from Global Land Analysis and Discovery (GLAD) about the national forest loss from 2000 to 2017 Results show that the rate of primary forest depletion in Vietnam is (-0.78), equivalent

to a loss of 1,105,000 hectares of the intact forests over 18 years In terms of tree canopy cover, the loss rate is smaller (-0.07) and Vietnam lost about 32,474,966 hectares of the total vegetation cover during 2000-2017

Later, Khuc et al., 2018 singled out Nghệ An as a province exhibiting the most deforestation between 2001-2010 among all Vietnamese provinces (p.133) Therefore, it draws the concern about Vietnamese people’s urgent actions on climate change mitigation for sustainable development

* Significance of the study

Deforestation has happened at an alarming rate and it contributes up to 17% to global warming (IPCC, 2015), inducing disasters hitting more frequently and intensively though unevenly at the global scale Therefore, this research with a case study in Nghệ An aims to find out the driving forces of forest cover loss

in order to come up with recommendations to stop deforestation, which could hamper the increase of the global temperatures to well below 20C as stated in the Paris Agreement This nature-based mitigation method, at the same time could benefit countless people living close to the forest edges with sources of food and income from NTFPs and also a good way to conserve wildlife

* Terms definition and explanation:

Before going specifically to the study, the project should come up with a brief explanation of certain terms used in the report

In general, forests are defined as areas which are covered with no less than 10%

of the tree cover of at least 5 meters in height It consists of both natural forests and anthropogenic plantations

In the tropical and sub-tropical regions, the forests account for between 10-40% canopy cover are classified as open canopy forest, and between 40-100% tree cover is defined as closed canopy forest (Matthews et al., 2000) Forests are endangered by deforestation and forest degradation Deforestation often results from forest degradation (Asner et al., 2005) When deforestation describes the total clearance of canopy cover to lower than 10% (Mayaux et al., 2006), forest degradation is identified by a remarkable decline in either tree density or in the shift of canopy cover from closed forests to open or fragmented forests (Karjalainen et al., 2003; Achard et al., 2004) In other words, forest degradation can gradually lead to deforestation, which may result in shifting the forest land

to land use change, depending on purpose, agricultural cultivation, plantations, industrial facilities or urbanization in which the forests can never regrow as it used to

In the specific case of Vietnam, the “forest”, according to the Forestry Law of Vietnam (16/2017/QH14) is defined as “an ecosystem including forest flora and fauna, fungi, microorganisms, forestland and other environmental factors

in which the main component is one or some species of trees, bamboo or arecaceae whose height is determined according to the flora of the soil or rocky mountain, submerged land, sandy land or other typical flora; with inter-regional area of at least 0.3 ha; canopy of at least 0.1”

Last but not least, afforestation could also help to preserve the natural habitat

of numerous flora and fauna species living in the forest environment, especially the critically endangered animals like Saola

1.2 Research objectives and tasks

The study is designed to meet the following objectives:

 Evidence-based findings about the rates of deforestation and the hotspots of forest chopped down in the last 20 years in a particular location of Vietnam This can be attained by using Geographical Information System (GIS) to analyze yearly satellite images of forest cover;

 Identification of drivers, both direct and underlying, of deforestation and forest degradation in the research site This can be achieved by field visit to the visible changes indicated from the satellite images analysis mentioned above People interviewing/ household survey and data collection would be carried out for analysis;

 Estimated changes in natural habitat of Saola, a critically endangered animal species, in Nghệ An province in 2020, 2050 and 2080 in the context of recent and future climate change by applying GIS and bioclimate modeling Main tasks of the exercise include:

 Calculate the deforestation rates in Nghệ An at provincial, district and commune levels, and then work out the hotspots of deforestation in the province;

 Collect socio-economic data in the two communes, the first commune

is with the least forest cover changes and the second one with greatest tree canopy loss (so-called a hotspot) of forest depletion to compare and analyze the causes of forest depletion in the research sites;

 Calculate the changes of Saola’s habitat areas in the years of 2020, 2050 and 2080;

 Recommend solutions to tackle deforestation and forest degradation in the two communes and to conserve the Saola species

Impacts of deforestation & CC

on habitat distribution of Saola

Carbon emissions from deforestation in Nghệ An

Carbon emissions from deforestation in Nghệ An

by district

Climatic changes

Recommendations for CC mitigation

& wildlife conservation

1.4 Deforestation in the world

Internationally, the Global Forest Watch (GFW) and the Global Forest Resources Assessment (FRA) are the two main sources of tree loss data and they are increasingly contradictory with each other One, the GFW paints a bleak picture of the forest cover which, from satellite imaginary analysis conducted by the World Resources Institute, declines by over 29 million hectares in 2017 (almost 1.5 times bigger than in 2015) This analysis was then supported by ground truthing with strong evidence in Southeast Asia where natural forests are continuously being converted to palm oil plantations

The other key data source for forest depletion, the FRA is compiled from government inventories input by the UN FAO, gives out a less gloomy situation It estimates the annual net tree loss, taking the forest regrowth into account, at slightly over 3 million hectares (equivalent to one tenth of the forest loss calculated by Global Forest Watch) Then, it concludes that deforestation has declined by over 50 percent in the last decade

On one hand, the satellite-based GFW is more rigorous to work on the Landsat photos It aims to work out the quantity of the forest area that has vanished since the previous year without working out the reasons why or how it happens The FRA, on the other hand, is largely dependent on the data of registered land use rather than actual tree cover For example, it includes also even the treeless areas due to logging, provided that the governments still classify it as productive forest which is claim to regrow and be logged again in the future The problem is that the two organizations, after 20 years of agreement, now convey with very different and irreconcilable pictures about the vision of the planet’s forests, one is highly negative and the other is rather positive

* World forest cover change in the 21 st century by Hansen et al 2013

By working on Landsat satellite images resolution of 30m, Hansen et al 2013 developed a global forest map covering both loss and gain between 2000 and

2012 A total of 230 million hectares of forests were lost and 80 million hectares

of the forests were re-generated The distribution of loss and gain, nonetheless, differs from region to region The tropical forests (32%) experienced loss and gain and also the ratio of loss to gain (3.6 for >50% of forest cover, 210,000 hectares/year) at the highest level of the four climate zones (tropical, sub-tropical, temperate and boreal), which indicated a dynamics in deforestation drivers Sub-tropical forests, with rare long-lived intact forests, underwent the greatest proportional losses of tree cover and lowest ratio of loss to gain (1.2 for >50%) The temperate climatic zone, nonetheless, has a low rate of deforestation and relatively insignificant ratio of loss to gain (1.6 for >50% of forest land), mainly damaged by typhoons and fires, making the re-generating process slower and therefore the ratio of loss to gain is rather high (2.1 over

>50% of forest area) Global deforestation maps can be seen in the following page

Figure 1.2 Deforestation worldwide (Source: Hansen et al., 2013)

Examination of deforestation rates at an annual basis shows that the tropical forests

in Southeast Asia are the most seriously endangered While the forest depletion in Latin America and Africa suffered from the deforestation rate of about 0.4 to 0.5%, the net tree loss in Southeast Asia is increasing tremendously in the past two decades with a rise from 0.83% between 1990-2000 to 0.98% during 2000 – 2005 per year (FAO, 2006, Achard et al., 2002; Mayaux et al., 2005) Subsequently, nowadays, only 33.4% of the land in South and Southeast Asia are covered by the forests (FAO, 2006) Main driving forces of deforestation are anthropogenic reasons such as, but not limited to, slash-and-burn cultivation; conversion from natural forests to plantations of rubber, acacia, pepper, palm; resettlement policies, logging (whether at vast scale or selectively) legally or illegally (Christanty, 1986; Sunderlin WD et al 1996; FWI/GFW, 2002; Schroeder-Wildberg et al., 2003; Goldammer, 2007)

1.5 Deforestation in Vietnam

For years, World Bank (WB) has shown special attention to Climate Change and acknowledged the urgent need for reducing carbon emissions via Reducing Emissions from Deforestation and Forest Degradation (REDD+) Acknowledging the urgent need for reducing carbon emissions via REDD+, WB via the Forest Carbon Partnership Facility (FCPF) has invested in REDD+ programs in 47 countries across the globe including Vietnam

Apart from participation to WB’s REDD+, Vietnam is also one of the early countries participating in UN-REDD project and afterwards WB REDD+ The nation was claimed to become ready for the REDD+ programs credit to the net increase of the forest cover nationwide (Patrick Meyfroidt et al 2009, Tatarski al 2016) According to the WB version of REDD+, the country (here Vietnam) enters phase II by an application to the Carbon Fund In 2016, Vietnam submitted an

Emissions Reduction Program Document (ER-PD) for the WB REDD+ system (Republic of Vietnam, 2016) for a possible second phase with a funding request

of $50 million from the Carbon Fund A technical assessment report from WB raised several concerns including matching appropriate interventions to the drivers

of deforestation remains unresolved, and the transformation of forest into plantation continues apace (FCPF, 2016) before the country could qualify for additional financial aid for REDD+ After revisions, WB decided that Vietnam could get the funding in 2018 (FCPF, 2018) despite the technical comment saying that the ER-PD fails to explicitly clarify the applied methodology to rank the various drivers in the six provinces1 “with additional reports on drivers, with data collected from national and provincial government reports and with outcomes of consultations conducted in the last two years at all levels in the six provinces" (ER-

PD, p.22, 23)

1.6 Impacts of deforestation on GHGs mitigation, global climate and impacts

of climate change on the forests

1.6.1 Impacts of deforestation on GHGs mitigation

Tropical Forests and Climate Change (mitigation: 8%, solution: 23%)

Effective management of the world’s carbon cycle in the coming decades would enable the prevention of the excessive concentration of carbon dioxide in the atmosphere The accumulated amount of carbon emitted from fossil fuels by far surpasses the emissions from both deforestation and other land use changes at global scale (Le Quéré et al 2016) This sometimes can lead policymakers or people to the misunderstanding that the role of the forests in mitigating GHGs emission or climate change is just minor because, according to Seymour and

Busch 2016, the net balance between emissions from tropical forest loss and carbon sequestration stays at 8 percent of the aggregated emissions Consequently,

it may lead various communities worldwide to underestimate the opportunities for climate change mitigation via the tropical forests because by halting forest depletion, emissions from deforestation would go down while the sequestration capacity of the trees would be enhanced The net contribution of tropical forests

to climate change mitigation is actually the subtraction of emissions to sinks Therefore, reducing deforestation can bring two-fold benefits of 1) hampering the carbon stored in the trees to be emitted into the atmosphere due to woodcutting, and 2) the increase in the size of canopy cover by replanting and conserving the forests can remove more carbon emitted from other sources

According to Griscom et al (2017), the land use sector can globally remove 11.3 gigaton (Gt) CO2 per year 2030, equivalent to 37% of the total amount of carbon

to reduce by 2030 to hamper warming to below 2°C Of this total, almost thirds (approximately 7.1 Gt CO2) can be stored via conserving, restoring and improving the management of tropical forests with attention to mangroves and petlands This means that though they illustrate only 8 percent of net emissions, tropical forests (combined with wetlands) can contribute around 23% of the total mitigation needed by 2030 to keep the increase of the global temperatures to well below 20C This forest-based mitigation strategies is a cost-effective measure to achieve NDCs commitment to UNFCCC via the Paris Agreement; it has not received proper attention and efforts (less than 3% of climate mitigation budget, Buchner et al 2015 and Climate Focus, 2017) from governments

two-1.6.2 Impacts of deforestation on global climate

Deforestation not only weakens the carbon sequestration capacity of the carbon sinks, it can also lead to dryer and hotter climate conditions at various scales

According to the report compiled by World Resources Institute, recent studies have proved the strong inter-link between forest loss and climatic abnormalities caused by the changes of the leaf surface albedo for the radiative balance, the movement of air, water, and heat through the process of evaporation and transpiration in the leaves and roots of the trees This, at some stage, can lead to unprecedented precipitation patterns causing floods or drought and subsequently soil erosion and loss of income for many people at local and regional scales Explanations can be presented briefly as in the figures below

Figure 1.3 Radiative forcing induced by forest loss (Source: Processed Infographics adapted from Bonan (2008), figure 2 by

Michael W et al (2018) - World Resources Institute)

1.6.3 Impacts of climate change on the forests

Forests impact climate, and climate also influences the forests in return In this section, I would like to follow the discussion about the forest-climate interactions, with a focus on the tropical forest from the Yale School of the Environment (2020)

At the global scale, the rise of carbon could potentially benefit forest development

by providing a “fertilization” pool to “feed” the trees Climate change, however, could also lead to higher temperatures and expand water stress which decreases the forest growth Studies in the Amazon project that rising temperatures can reduce 10-20% of precipitation, causing water shortage for the trees to grow In this perspective, reduced forest cover leads to more carbon emissions, less water circulation, further drying, and even forest fires Tropical rain forests, however, are claimed to be more resilient to the changing climate, drought and may continue

to store carbon for longer

Climatic changes also impacts various endangered species, especially rare animals living dependently on tropical forests Scientists have documented the possible extinction of tens of species such as harlequin frogs, flying foxes, etc due to a synergy of the diseases and climatic degradation

Lots of programs have paid attention to the problems of tropical forests and climate change via mitigation and adaption The REDD+ mechanism, for instance, offers payment for ecosystem services to halt deforestation and hamper carbon emissions Adaptation programs like agroforestry and community forestry which can benefit forest communities with plenty of alternatives to improve economic development, food supply, and sustainable forest management

1.7 Scope of the study

Based on an exploratory analysis of drivers of deforestation and forest degradation

in Vietnam at national level, Khuc et al (2018, p.133) came up with a conclusion that Sơn La and Nghệ An were the two provinces with the highest levels of deforestation and forest degradation during 2001-2010

In order to have a closer look into the changes in the forest area and its quality over twenty recent years, I decided to take one the two mentioned provinces, Nghệ

An to conduct the study, applying both GIS and socio-economic methods

Figure 1.4 Administrative map of Nghệ An province

* Characteristics of the study sites:

Nghệ An province (19°20′N 104°50′E), the biggest province of Vietnam, situated

in the North Central coast region of Vietnam With the population of 3,327,791people, the province is divided into 1 capital city (Vinh), 3 district-level towns

(Cửa Lò, Thái Hòa, Hoàng Mai) and 17 districts2 These 17 districts are further subdivided into 17 commune-level towns, 431 communes, and 32 wards Nghệ

An covers one session of the Trường Sơn mountain range so 83% of its area is mountainous

The study selected Châu Khê commune (in Con Cuông district) and Thanh Sơn (in Thanh Chương) to conduct a case study on drivers of deforestation and forest degradation Below I would like to provide some brief information about these two communes only (Reasons for this selection will be discussed in the later components)

Châu Khê is one of the seven communes in Con Cuông district, located in the

buffer zone of Pù Mát national park It covers an area of 44,058 ha and is one of the two communes with greatest forest cover The commune consists of 10 villages with a population of 5,578 people (ethnic minorities account for 73%) 51% of the population live in poverty (Socio-economic report of Pù Mát national park, 2011)

Thanh Sơn was established as a commune in Thanh Chương district in 2009

(Decree 07/NĐ-CP, 2009) by mobilizing 6,739 ha of natural land from Hạnh Lâm commune and 648 ha of natural land from Thanh Mỹ commune Thanh Sơn is the resettlement commune for people coming from Tương Dương district (3,650 people from Kim Đa commune and 1,598 people from Hữu Dương – the two communes were already in dissolution due to the construction of Bản Vẽ hydropower plant in Tương Dương district, Nghệ An) Thanh Sơn is classified into 16 villages, which were then re-divided into 7 villages in 2019 and 72% of

Lộc, Nghĩa Đàn, Quế Phong, Quỳ Châu, Quỳ Hợp, Quỳnh Lưu, Tân Kỳ, Thanh Chương, Tương Dương and Yên Thành

the population belongs to Thái group (Socio-economic report of Thanh Sơn commune, 2009-2019)

1.8 The research questions and hypothesis

To get understanding about the forests in Nghệ An, I would like to answer the two following research questions:

1 What are the possible change rates of forest cover during 2000 - 2019 in Nghệ An at the provincial, district and commune levels?

2 What are the possible drivers of deforestation and forest degradation? And contribution of each force into the whole picture?

3 Can the bioclimate, forest cover and socio-economics impact the habitat of

a mammal species called Saola?

The project is based on the hypothesis: Nghệ An is a transition province in terms

of shifting from deforestation to afforestation in the last 20 years, 2000-2019

1.9 Literature review

The forests-climate interactions are not just one-way The IPCC report and other studies describe how the health and functioning of trees, animals and various forest ecosystems are impacted in the context of increasing extreme weather events (heat waves, droughts, floods) in terms of frequency, severity and time duration Forests also become more vulnerable to new pests and diseases which are both increasing and expanding in the warmer climatic conditions However, the most noticeable effect of climate degradation on the forests could be the increasingly high risk of forest fires due to low humidity (dryer climate) and drought, especially combined with deforestation (slash-and-burn agricultural cultivation) and forest degradation when left-overs of timber fragments get dry and become extremely flammable Forest fires are a remarkable source of global emissions, as long-lasting fires have

been reported in carbon-rich tropical forests in Indonesia, Brazil or Australia recently

1.9.1 Technical component

Through a thorough review article, Anthea et al (2017) summarized various advanced technologies/approaches that have been applied quite efficiently in the attempts to monitor forest losses, including multi-resolution optical, synthetic aperture radar (SAR) and LiDAR data In fact, there is no single method that can help to track forest degradation, mainly due to the specific nature of the degradation type or process and the timeframe over which it is observed Two main approaches: 1) where detection is indicated by a change in canopy cover or proxies, and second, the quantification of loss (or gain) in above ground biomass (AGB) are assessed to be useful for application into this exercise of monitoring deforestation and forest degradation

Remote sensing can detect the degradation that has a visible impact on the forest canopy At present, using available satellite and forest inventory data can help to estimate the forest areas, carbon stocks and changes over years Inventory data, however, is often missing And one possible issue is that pictures should be taken

in a spontaneous cloud-free seasonal period of time Apart from that, high resolution images are needed but it may be costly to obtain

For under canopy deforestation such as understory grazing or fuel wood extraction, LIDAR seems to be an efficient tool for scanning the quality of the forest for calculations or estimations In case this sounds to be too technical and difficult to get proper facilities, then ground truthing or conducting production/consumption surveys could be organized for further clarification

1.9.2 Socio-economic component

Numerous peer-reviewed articles have come out, providing different views and opinions about drivers of deforestation and forest degradation By looking at two review articles (one by Jonah et al, 2017 over 121 articles and the other by Janice

et al, 2015 over 19 papers) which both apply meta-analysis from all the available resources into their studies, it is obvious that there are two different approaches when looking into drivers of forest losses

In the first approach, Jonah et al (2017), after analyzing more than one hundred relevant articles published between 1996 – 2013 on worldwide determinants of deforestation and forest degradation, summarized the drivers into five groups that encourage land users to deforest They include biophysical characteristics (slope, elevation, wetness, soil suitability, distance to water), market demand for commodities (agriculture, timber), built infrastructure (proximity to roads, proximity to urban areas), demographic and socio-economic characteristics (community demographics, population, poverty and income, rural income support, payments for ecosystem services), and ownership and management rights (protected areas, law enforcement, community forest management and land tenure security) The other approach, according to Janice et al, 2015, illustrates direct and indirect drivers of deforestation through benefits by classifying one main driver into sub-drivers from publications on REDD+ projects elsewhere in the world as follows:

Table 1.1 Classification of deforestation drivers

1 Deforestation – agriculture conversion

Government support for commercial agriculture Indirect

2 Forest degradation – logging

3 Forest degradation – other direct drivers

4 Deforestation and forest degradation – other indirect drivers

5 Deforestation and degradation indirect driver – poverty

(Source: Janice et al, 2015)

* Studies on deforestation in Vietnam

In the study from Lào Cai, Jardin et al., 2013 entered data on ethnicity, rice production, and households with electricity, village altitude and deforestation data

in an attempt to refute the hypothesis that poverty is the origin of deforestation Nguyen, 2019 used accessibility to roads, income and population density and poverty rate to explain the deforestation in Hòa Bình province Likewise, Nguyen,

2019 could not find any significance between deforestation rates and poverty Chi

et al., 2013 focused on the probability of land transformation and concluded that minority groups are most likely to convert scrub and forest land (in here broader than forest) to agriculture fields than villages dominated by Kinh people From a slightly different angle, Cochard et al 2017 in another country wide study concluded that extensive forest areas today (2013) are confined to higher elevation areas Provinces with increasing agriculture productivity experienced higher forest losses Support for the view of competition between natural forest and plantations the authors could not confirm; a fact that is a bit surprising Finally, the study argued that illegal logging for high value timber continues to be a serious problem

in protected areas Ankersen et al., 2015 compares dense forest with open forest land in a study from Con Cuông district in Nghệ An and observes most deforestation was concentrated around the settlement area Results are shown in Table 1.2

Table 1.2 Studies on deforestation in Vietnam

Ankersen

et al.,

2015

Landsat images and Participatory Land Use Mapping

Dense forest is concentrated around Pù Mát national park

Focus on REDD+ reference period and no discussion of

deforestation drivers

Chi et al.,

2013

Landsat images and secondary social data

Deforestation is correlated to:

Slope (-) Ethnicity/ minorities (+)

Ethnicity of households was the major finding

Cochard

et al.,

2017

Secondary social data from 63 provinces

Primary forest has increased in remoted areas

Useful observations but

no focus on deforestation drivers

Jardin et

al., 2013

Landsat images and secondary social data

Poorest communities have reach a certain degree of sustainability

Narrow focus on the poverty/ethnicity variable

Khuc et

al., 2018

Forest maps and secondary social data

Deforestation is correlated to:

Poverty (+), Agriculture (+), Migration (+)

Forest maps instead of satellite images and strange definition of forest governance Khuc et

al., 2020

Forest inventory maps and SPOT images, household data

Plantations can provide new opportunities for poor households

Natural forest cover increase in this article (different from the

2018 article) No discussion is included Nguyen,

2019

Landsat images and secondary social data

Deforestation is correlated to:

Slope (-) Labour in industry (+) Income from forest (+) Poverty (-)

Sustainability is on the increase in the province (Hòa Bình)

Signaling increasing sustainability in forest management seems odd, unless expansion

of plantations covers for the sustainability indicator

What do we learn from the literature review? First, use of time series data in a test

of deforestation factors is only possible if independent variables include secondary official data Primary data collected at the commune, district or province level reveal more nuances, but they are snap photos at a given period in time Linking deforestation rates to primary data requires other methods (see below) Second, disagreement among observers are notable as to whether Vietnam currently experiences an increase or decrease in forest cover nationwide or by province One explanation, though probably not the only one, is a reference to a broad concept

of forest cover, including both primary forest and plantations If plantations replace primary forest, the total forest cover might well increase over time making Vietnam look like a transition country A transition country defines a nation changing management of forests from a deforestation to an afforestation phase Third, poverty either influences the deforestation rate with a correlation coefficient that is positive or negative significant In some papers, the official poverty status

of households weighs in as the chosen indicator Other scholars focus on the relative poverty or income level among households in a village, commune or district Higher-level or aggregated analyses are likely to demonstrate a positive correlation to deforestation Local level analyses seem to reach the opposite conclusion A micro-macro paradox is clearly distinguishable in the case of the deforestation-poverty relationship

The research will not try to explain why conclusions from papers on deforestation drivers seldom end up stressing unanimity Both on the national, provincial and commune level, this research will use the same satellite photo images and in defining poverty I will focus on the village level

1.9.3 Saola and its habitat distribution

In May 1992, a new species of bovid (called the Saola Pseudoryx nghetinhensis) was found in Vũ Quang nature reserve in Hà Tĩnh province of Vietnam (Dung et al., 1993, 1994) Since then, lots of international attention has been paid to this discovery and surveys were conducted in Vũ Quang and adjacent areas including Nghệ An in Vietnam and also in the neighbouring Laos, to investigate further the distribution, behavior and circumstance of this species in 1992 and 1993 The survey in Nghệ An province was carried out by a joint team from two organizations (the former Ministry of Forestry and World Wildlife Fund (WWF), coming to a conclusion that more than 20 specimens of Saola remained (Dung et al., 1993, 1994) However, if they still exist today remains questionable Later, from the study in Laos implemented by the Wildlife Conservation Society of New York and the Department of Forestry of the Lao People’s Democratic Republic (Schaller and Rabinowitz, 1995), 3 more were confirmed to exist These surveys eventually concluded the existence of 23 Saola in the areas adjacent to the Vu Quang nature reserve in Vietnam, with extension to the south and west both

In Vietnam, studies show that the historical distribution of Saola stretched from Quế Phong district, Nghệ An to Tây Giang district, Quảng Nam province Nonetheless, it has shifted the current distribution from Chu river in Quế Phong district, Nghệ An province to A Vương commune, Tây Giang district, Quảng Nam

An estimation indicates that there are some 120-160 Saola individuals exist in Vietnam (in 50 communes of 20 districts and 6 provinces3 in the central of Vietnam) Particularly, they are distributed in Pù Hương nature reserve, Pù Mát

national park, southwestern Quảng Bình – northern Hương Hóa area and Thừa Thiên Huế - northern Quảng Nam area

Two places which has special importance for Saola conservation in Vietnam are: 1) Thừa Thiên Huế - northern Quảng Nam area harboring about 40-50 Saola individuals and 2) southwestern Quảng Bình - north Hương Hóa area harboring about 20-30 Saola individuals Later, they are learnt to live in Pù Mát national park (Nghệ An), Vũ Quang nature reserve, northern end of the Trường Sơn mountain range, south of the Cả river (Dawson, 1995)

Structure of the Saola’s component can be presented in the following figure

Figure 1.5 Research flow of the Saola’s component

Validate the result

Species location recorded map Change to coordination format;

(in CSV)

Run Maxent sorfware

Data processing,(clip study area, change data format choise the date, time )

2050, 2080)

Species distribution map

GIS database: Environment factors, report…

Results; Maps Data, chart

Export data, analysis data

CHAPTER 2 METHODOLOGY

2.1 Data collection

2.1.1 Remote sensing data

* Forest time-series satellite images:

Time-series satellite images containing data about primary forests and forest cover over 20 years were downloaded from GLAD, session “Annual continuous fields

of woody vegetation structure in the lower Mekong region” (https://glad.geog.umd.edu/dataset/mekong) Pictures were processed and characterized into primary forest and tree canopy cover during 2000-2017 by the University of Maryland In total, 36 pictures (18 each forest type) were downloaded Accuracy 81% All data layers collected are in the geographic coordinates using WGS84 reference system with the pixel size of 0.00025 x 0.00025 degree The yearly data are available in 8-bit GeoTiff format that is proper for visualization and analysis in most of GIS software

While the primary forests are described as long-lived natural forests without no hints of disturbance or regeneration since 1985 and their spectral signature indicates a mature forest, the tree canopy cover characterize trees higher than 5 meters processed annually by airborne lidar Resolution of 30mx30m applies for both two types

The objective of this exercise jointly conducted by GLAD lab (from the University of Maryland) and SEVIR-Mekong is to provide consisted time-series

of forest cover structure data suitable for land cover change monitoring at the national and regional scales, especially the yearly dynamics of woody vegetable structure and pristine forest extent Furthermore, presented data can be useful for

assessing the national forest management policy effects, international forest resources reporting, and greenhouse gases emission estimation

* GHGs emissions data

Recently updated data was collected from Global Forest Watch (https://www.globalforestwatch.org/), a US-based forest monitoring organization with input forest cover satellite images also from the University of Maryland from which I also collected the satellite images for analysis and work out deforestation rates as well as hotspots in the technical component

Global Forest Watch (GFW) is an online platform that provides data and tools for monitoring forests and carbon emissions from deforestation By harnessing cutting-edge technology, GFW allows anyone to access near real-time information about where and how forests are changing around the world and how much carbon

it has released back into the atmosphere

2.1.2 Socio-economic data

Primary data collected from a household survey covering 118 participants in 4 villages (Khe Bu, Khe Nà, Bình Yên and Thanh Tiến) belonging to two communes (Châu Khê and Thanh Sơn) in Con Cuông and Thanh Chương districts The questionnaire consists of five main contents about household’s income and assets, forest land and use, forest access regulations, people’s awareness about forest changes and biodiversity changes over time The questionnaire ends with a question about the preferable priority of local people between environmental conservation and economic development In-depth interviews were also conducted with village leaders, forest rangers (from both Pù Mát national park and protective forest management) and commune officials responsible for forest land allocation and management in the two districts

The secondary data about income and socio-economic growth was then collected from local authorities though the accuracy is questionable For example, income figures in Thanh Sơn show a gradual increase trend since its establishment (2009)

to May 2020 (See table 2.1 below) Another thing for consideration is that the figures are on average for the whole commune4, which unfortunately cannot illustrate the differences among households’ income or how it should be translated

in understanding the drivers of deforestation

Table 2.1 Average income over years from 2009 to 2019, Thanh Sơn commune

Year Average annual

income

Unit (‘Dong/ per head)

4 Commune is the smallest official administrative unit in the Vietnamese administrative system

BIO0 = Annual Mean Temperature

BIO1 = Mean Diurnal Range [Mean of monthly (max temp - min temp)] BIO2 = Isothermality (BIO2/BIO7) (*100)

BIO3 = Temperature Seasonality (standard deviation *100)

BIO4 = Max Temperature of Warmest Month

BIO5 = Min Temperature of Coldest Month

BIO6 = Temperature Annual Range (BIO5-BIO6)

BIO7 = Mean Temperature of Wettest Quarter

BIO8 = Mean Temperature of Driest Quarter

BIO9 = Mean Temperature of Warmest Quarter

BIO10 = Mean Temperature of Coldest Quarter

BIO11 = Annual Precipitation

BIO12 = Precipitation of Wettest Month

BIO13 = Precipitation of Driest Month

BIO14 = Precipitation Seasonality (Coefficient of Variation)

BIO15 = Precipitation of Wettest Quarter

BIO16 = Precipitation of Driest Quarter

BIO17 = Precipitation of Warmest Quarter

BIO18 = Precipitation of Coldest Quarter

2.2 Methods to identify hotspots of deforestation

Technical component: In order to get understanding about forest loss rates in Nghệ

An over the last two decades, QGIS version 3.10 was used to calculate the amount

of annual primary forest depletion in all the districts of the province, from 2000 to

2017 with the satellite pictures input from GLAD The imagery analysis showed that Thanh Chương is the district where the deforestation rate is the highest of all districts, while Con Cuông has the greatest cover of pristine forests Afterwards,

GIS once again helped to calculate deforestation rates in all the communes of these two districts so that I could select the two communes: Thanh Sơn in Thanh Chương exhibiting the most deforestation and Châu Khê in Con Cuông has the biggest intact forest area of all the communes During field visits to the two communes, ground truthing was also carried out to check the state of the forests, deforestation area and the land use change

2.3 Methods to quantify the amount of CO 2 emissions from deforestation

Statistics is applied in this component with the dataset collected from GFW to present the amount of CO2 emissions from the whole province of Nghệ An and then the two districts of Con Cuông and Thanh Chương

2.4 Methods to identify drivers of deforestation

Social-economic component: RCT (Randomized Controlled Trial) was deployed for comparison between Thanh Sơn and Châu Khê Data was collected from 118 households by questionnaire interviewing (questionnaires are almost identical for the two communes with 58 participants in Châu Khê and 60 in Thanh Sơn) In-depth interviews with forest and local authorities were also carried out Major challenge is to measure in compatible terms, the significance and magnitude of the underlying forces of deforestation

2.5 Methods to project the habitat distribution of the Saola

Maxent software can be used to project the niches and distribution of Saola species

in Nghệ An over decades by applying a machine-learning technique called maximum entropy modeling From a set of environmental (e.g., climate) grids and georeferenced occurrence localities, the model expresses a probability distribution where each grid cell has a predicted suitability of conditions for the species