Forest Reference Level (FRL) of

Myanmar

Ministry of Natural Resources and Environmental Conservation

Myanmar
November 2018
 Table of Contents
List of Tables .................................................................................................................................. iii
List of Figures ................................................................................................................................. iii
List of Annex ................................................................................................................................... iii
Acronyms ........................................................................................................................................ v
Acknowledgements.............................................................................................................................. viii
Summary ................................................................................................................................................. 1
1. Introduction .................................................................................................................................... 4
1.1. Objectives of FRL ..................................................................................................................... 5
1.2. Summary of guidance for FRL development .......................................................................... 6
2. National Context ............................................................................................................................. 8
2.1. National circumstances ........................................................................................................... 8
2.2. Myanmar’s Nationally Determined Contributions-(NDC) submission and relation to the FRL . 10
2.3. Forests in Myanmar .............................................................................................................. 11
2.4. Forest Land Use and Land Cover by Forestry Legislation ..................................................... 14
2.5. REDD+ development in Myanmar......................................................................................... 15
3. Definitions ..................................................................................................................................... 16
3.1. Forest definition .................................................................................................................... 16
3.2. Deforestation definition........................................................................................................ 17
3.3. Enhancement of forest carbon stocks through reforestation/afforestation activities ........ 17
4. Scale .............................................................................................................................................. 17
5. Scope (activities, pools, gases) ...................................................................................................... 18
5.1. REDD+ activities .................................................................................................................... 18
5.2. Pools and gases ..................................................................................................................... 19
6. Reference period of FRL ................................................................................................................ 21
7. Data used in the FRL ..................................................................................................................... 21
7.1. Deforestation ........................................................................................................................ 21
7.1.1 Activity data ......................................................................................................................... 21
7.1.2 Emission factors ................................................................................................................... 35
7.2 Enhancement of forest carbon stocks ....................................................................................... 42
7.2.1 Activity data ....................................................................................................................... 43
7.2.2 Emission/ Removal factors ................................................................................................... 45
7.2.3 Methodology of calculation ................................................................................................. 45
7.2.4 Results on reference level from enhancement measures ................................................... 46
8. Myanmar Initial FRL ...................................................................................................................... 47

i
9. Uncertainty assessment of emission estimates ............................................................................ 48
10. Future improvement opportunities ................................................................................................ 50
10.1. For Activity Data ...................................................................................................................... 50
10.2. For Emission Factors ............................................................................................................. 52
10.3 Uncertainty assessment ............................................................................................................ 53
References: ........................................................................................................................................... 54
Annex .................................................................................................................................................... 57

ii
List of Tables
Table 2. 1: Forest Types and Vegetation Zones in Myanmar .................................................................. 11
Table 2. 2: Major Land Categories Reported in Statistical Year Book compiled by CSO......................... 15
Table 7. 1: Validation datasets used to assess 1,884 samples ................................................................ 29
Table 7. 2: Structure of accuracy assessment matrix .............................................................................. 32
Table 7. 3: Confusion matrix for 2005-2015 forest change map based on 1,884 stratified random
samples..................................................................................................................................................... 33
Table 7. 4: Stratified area estimates in hectare (ha) with confidence intervals (CI), weighted producer’s
accuracy and user’s accuracy under forest and non-forest cover classes for the years 2005 and 2015
.................................................................................................................................................................. 34
Table 7. 5: Stratified area estimates in hectare (ha) with confidence intervals (CI), weighted producer’s
accuracy and user’s accuracy for three land cover classes during 2005-2015. ...................................... 34
Table 7. 6: Results of the CO2 eq Mean Value tonnes per ha and weighted mean CO2 eq tonnes per ha
.................................................................................................................................................................. 40
Table 7. 7: Forest plantation areas as recorded by the Forest Department 2000- 2015 ....................... 43
Table 7. 8: Accumulated CO2e removal from forest plantations ............................................................ 46
Table 8. 1: Summary proposed FRL of Myanmar 2005-2015.................................................................. 48
Table 9. 1: Uncertainty Result for Activity Data in % ............................................................................... 50

List of Figures
Figure 7. 1:Steps followed to produce stratified random sample-based estimates of forest change
during 2005-2015 in Myanmar ................................................................................................................ 25
Figure 7. 2: Distribution of the 1,884 samples across Myanmar ............................................................ 27
Figure 7. 3: Decision tree for sample change analysis ............................................................................ 28
Figure 7. 4: Example of reference data (Google Earth Archive) .............................................................. 30
Figure 7. 5: Area estimates of forest cover change during 2005-2015, in million ha, from local land cover
maps and from stratified random sampling design ................................................................................ 35
Figure 7. 6: Location of forest inventory plots collected during 2005 to 2017 ...................................... 37
Figure 7. 7: Carbon stock enhancement from forest plantations 2000-2015 and reference level 2005-
2015 in tonnes of CO2e ........................................................................................................................... 47

List of Annex
Annex 1: Dominant Types of Forests in Myanmar ................................................................................. 57
Annex 2: Comparison of forest cover area and its percentage within and outside PFE (based on 2015
National Forest Cover Map) ..................................................................................................................... 58

iii
Annex 3: Datasets used to generate activity data for establishing a forest reference emission level in
Myanmar during 2005-2015 .................................................................................................................... 61
Annex 4: Harmonizing national land use categories with FRA and IPCC land use categories ............... 62
Annex 5: Wall-to-wall thematic raster maps of Myanmar under seven national land use/cover
categories at three years: 2005, 2010 and 2015 ..................................................................................... 63
Annex 6: Maps showing the positioning error in the wall-to-wall maps of Myanmar boundary .......... 64
Annex 7: A brief description of the simple and stratified random sampling estimators ....................... 65
Annex 8: Examples of sample assessment using validation datasets: ................................................... 68
Annex 9: Sample Plot Design for 40 Districts from Forest Management Inventory mentioning the
conversion factor to per ha ..................................................................................................................... 70
Annex 10: Plot Designs and Description ................................................................................................. 73
Annex 11: Summary of calculation steps for three carbon pools and National Emission Factor value
for Myanmar ............................................................................................................................................ 75
Annex 12: Description of the forest types and respective value of R (Ratio of BGB to AGB) and Litter
Range for each district ............................................................................................................................. 76
Annex 13: Net Biomass Increment and Removal Factor Calculation for Enhancement ........................ 80
Annex 14: Uncertainty Result for Emission Factor in % ......................................................................... 81

iv
Acronyms
AD Activity Data
AGB Above Ground Biomass
BGB Below Ground Biomass
BUR Biennial Update Report
CH4 Methane
CI Confidence Interval
CO Carbon Monoxide
CO2 eq Carbon Dioxide Equivalent
CSO Central Statistical Organization, Ministry of Planning and Finance (MOPF)
DBH Diameter at Breast Height
ECD Environmental Conservation Department, Ministry of Natural Resources and
Environmental Conservation (MONREC)
EF Emission Factor
EIA Environmental Impact Assessment
FAO Food and Agriculture Organization of the United Nations
FD Forest Department, Ministry of Natural Resources and Environmental
Conservation (MONREC)
FRA Forest Resource Assessment
FREL/ FRL Forest Reference Emission Level/ Forest Reference Level
GDP Gross Domestic Product
GFC Global Forest Change
GFOI Global Forest Observations Initiative
Gg Gigagram
GHGs Green House Gases
GIS Geographic Information System
GOFC-GOLD Global Observation of Forest and Land cover Dynamics
GPG Good Practice Guidance
Ha hectare
INDC Intended Nationally Determined Contribution
IPCC Intergovernmental Panel on Climate Change
IRS Indian Remote Sensing Satellites
LUKE Natural Resources Institute, Finland
LULUCF Land Use, Land Use Change and Forestry

v
MCCSAP Myanmar National Climate Change Strategy and Action Plan
MGD Methods Guidance Document
mm/year millimeter per year
MOALI Ministry of Agriculture, Livestock and Irrigation
MRV Measuring, Reporting and Verification
N2O Nitrous Oxide
NAPA National Adaptation Programme of Actions
NDC Nationally Determined Contribution
NFI/ NFMIS National Forest Inventory/ National Forest Monitoring and Information System
NFMS National Forest Monitoring System
NO2 Nitrogen Dioxide
MRRP Myanmar Reforestation and Rehabilitation Programme
PaMs Policy and Measures
PAS Protected Area System
PFE Permanent Forest Estate (RF + PPF)
PPF Protected Public Forest
REDD+ Reducing Emissions from Deforestation and Forest Degradation, and
Conservation, Sustainable Management of Forest and Enhancement of Forest
Carbon Stocks
REF Removal Factor
RF Reserved Forest
RS Remote Sensing
SBSTA Subsidiary Body for Scientific and Technological Advice
SE Standard Error of Sample Mean
SEPAL System for Earth Observation Data Access, Processing and Analysis for Land
Monitoring
SIS Safeguard Information System
SNC Second National Communication
STRS Stratified Random Sample
TA Technical Assessment
TWGs Technical Working Groups for REDD+ (Drivers and Strategy TWG, Stakeholder
Engagement and Safeguard TWG and MRV TWG)
UNFCCC United Nations Framework Convention on Climate Change

vi
UN-REDD United Nations Programme on Reducing Emissions from Deforestation and Forest
Degradation, and Conservation, Sustainable Management of Forest and
Enhancement of Forest Carbon Stock

vii
Acknowledgements
We, REDD+ MRV team, would like to thank Dr Nyi Nyi Kyaw, Director General, Forest
Department, Ministry of Natural Resources and Environmental Conservation for his
overall technical guidance and support for developing this report.

We also like to thank U Kyaw Kyaw Lwin, Deputy Director General, Forest Department
and Dr. Thaung Naing Oo, Director, Forest Research Institute, Ministry of Natural
Resources and Environmental Conservation for their valuable comments and
suggestions on submission.

We gratefully acknowledge the constant support of U Hla Maung Thein, Director

General of the Environmental Conservation Department, Ministry of Natural
Resources and Environmental Conservation as the National Focal Point for all
communications with the UNFCCC, without whom this work would not have been
possible.

Our sincere thanks also go to the FAO experts who are working with us in developing
and upgrading our National Forest Monitoring System, specifically: Mr. Franz Eugen
Arnold, Ms Thinn Thitsar Kyaw, Dr Abu Rushed Jamil Mahmood, Ms. Marieke Sandker,
Mr. Ben Vickers, Mr. Mathieu VanRijn and Ms Donna Lee for their important advice,
assistance, and reviews on the methodology and approaches developed and applied
for this submission.

During data collection as well the consultation meetings, many other organizations
and individuals provided valuable comments and suggestions to develop this first
Forest Reference Level for Myanmar. Words cannot fully express our thanks to those
who knowingly or unknowingly helped us during the preparation of this report.

Lastly, we are deeply grateful to the assessment team (AT), who critically reviewed
and revised several drafts of this report during 2018 with many valuable suggestions
and supportive comments: Ms. Fernanda Alcobe (Argentina) and Dr. Markus Didion
(Switzerland), Mr. Thiago De Araujo Mendes (Brazil), and Mr. Nalin Srivastava (UNFCCC

viii
secretariat) who shared their valuable time and knowledge with us for the successful
completion of this submission.

Contributions:
Forest Department, Ministry of Natural Resources and Environmental Conservation
(MONREC), Myanmar
 Dr. Nyi Nyi Kyaw, Director General (National REDD+ Focal Point)
 Dr. Thaung Naing Oo, Director, Forest Research Institute (Alternate National
REDD+ Focal Point and UN-REDD National Programme Director)
 U Kyaw Kyaw Lwin, Deputy Director General, Forest Department (Chairman of
the National REDD+ Task Force)
 U Tin Htun, Director, Planning and Statistics Division, Forest Department
(overall responsibility for National Forest Monitoring System in Myanmar)
 Dr. Myat Su Mon, Deputy Director (leader of the MRV technical working group
of the national REDD+ Programme in Myanmar)

Technical Team
 Dr. Myat Su Mon, Deputy Director (MRV Team Leader)
 U Saw Daniel, Assistant Director (Emission Factor Calculation Team)
 U Phone Htut, Retired Staff Officer (Activity Data Calculation Team)
 U Khine Zaw Wynn, Staff Officer (Activity Data Calculation Team)
 U Nay Lin Tun, Range Officer (Activity Data Calculation Team)
 MRV Team members of Forest Research Institute
 All staff of RS and GIS section, Computer section and Forest Inventory section
under Planning and Statistics Division, Forest Department

FAO Experts
 Mr. Franz-Eugen Arnold, Chief Technical Advisor for NFMS, NFI and FREL in
Myanmar

ix
 Mr. Ben Vickers – Regional Coordinator REDD+ Asia Pacific Region, FAO RAP
Bangkok
 Ms. Marieke Sandker, Forestry Officer, FAO Rome
 Dr. Abu Rushed Jamil Mahmood, International Consultant, Remote Sensing and
Land Cover Assessment Expert, FAO RAP, Bangkok
 Mr. Mathieu VanRijn, Forestry Officer, FAO RAP, Bangkok
 Ms. Thinn Thitsar Kyaw, Project Assistant, FAO’s UN-REDD Programme

x
xi
Summary
The Government of Myanmar is fully aware of the causes and potential impacts of
climate change. Myanmar actively participated in global climate change mitigation
efforts as a non-Annex 1 party. Nationally Determined Contribution (NDC) was
submitted in 2016. Under the NDC, forestry is a key sector and quantitative targets are
likely to be included. Myanmar’s Initial National Communication (INC) was submitted
to UNFCCC in 2012 and the Second National Communication (SNC) is now under
preparation. Currently, Myanmar views REDD+ initiatives as a contribution to the
green development of Myanmar as well as supporting the mitigation of, and
adaptation to, climate change. In addition, the Government of Myanmar stressed that
the national REDD+ Programme is critical to their mitigation and adaptation pledges
according to its country statement to COP 23.

Following the suggestion of Decision 12/CP.17, Myanmar prepared its FRL using a
stepwise approach. Myanmar submitted its initial FREL report on January 2018 and
current submission was revised one by following to the Assessment Team-AT of
UNFCCC. The FRL submission will be a benchmark for assessing its performance in
implementing REDD+ activities in contribution to climate change mitigation. The main
objective of the FRL submission is to support the climate change mitigation efforts
under the national context of Myanmar. Further objectives of the submission are;

- To assess and evaluate the performance of REDD+ policies and measures and
sustainable forest management practices
- To provide information on emission projections to stakeholders including policy
makers, government line departments, technicians and members of the public
on a clear, transparent and consistent basis.
- To facilitate access to potential funding sources for results-based payments and
to support efforts to reduce emissions from the forest and land use sector.

The development of the FRL was initiated by a group of experts; REEDD+ TWG on
Measuring, Reporting and Verification (MRV), representing a cross-section of

1
ministerial agencies and organizations. This submission is largely due to the effort and
commitment of the members of this TWG and reviewed by AT.

Myanmar FRL is national level as all the existing land and forest monitoring and
measurement capacities are at the national level. In addition, Initial National
Communication (INC) and Forest Resources Assessment (FRA) were also reported by
national level. Myanmar prioritized two of the five REDD+ activities as the focus of the
country’s first FRL submission; deforestation and enhancement of forest carbon stock
through afforestation/reforestation. The Government of Myanmar recognizes the
potential importance of plantation and forest restoration measures to climate change
mitigation efforts. January 2018 submission was focused mainly on deforestation due
to the existing data limitations. It is converted from the FREL to the FRL in the present
submission by the TWG’s efforts on historical forest enhancement and comments from
AT.

Three carbon pools; i.e. Above Ground Biomass (AGB), Below Ground Biomass (BGB)
and Litter were included in this FRL, using allometric equations derived from district
forest management inventory and according to IPCC Good Practice Guidance (2003).
Due to the limitation of existing national data sources on soil organic carbon on forest
soils are very localized, soil organic carbon was excluded in the submission. Deadwood
carbon pool was also excluded due to limited information/data at national level. In
absence of a reliable data sources for the estimation of non-CO2 gases, this FRL has
considered only CO2 gas. Strategic planning is currently being developed to facilitate
the inclusion of remaining carbon pools and non-CO2 emissions and so expected in
future submissions.

The reference period, from the year 2005 to 2015, was decided through a series of
consultation meetings, based on availability of the most reliable national scale existing
Activity Data (AD) and Emission Factors (EF), and to ensure the consistency with GHGs
Inventory reporting. It is expected that the database generated through this FRL

2
development process will benefit the SNC and the Biennial Update Report (BUR) to the
UNFCCC.

AD has been developed by estimating the extent of forest change measured as gross
area estimates of forest, non-forest and forest loss (deforestation) during 2005-2015,
excluding forest degradation, forest improvement and forest area gain. The amount
of forest loss has been estimated using a sample-based approach. Following the IPCC
(2003, 2006) guidelines and the GFOI (2016) methods guidance documents, the bias-
corrected gross forest loss is about 428,984 ha per year during the reference period
2005-2015.

This submission used 11,284 forest inventory plots data collected during 2005 to 2017
which were covered throughout the country. A combination of Tier 1 (for BGB and
Litter) and 2 (for AGB) approaches were used in estimating the carbon pools. A
weighted mean value of 125.43 tCO2 eq per ha is estimated for a national level EF in
Myanmar. Annual CO2 emission from gross forest loss during the historical reference
period 2005-2015 is estimated as 53,807,463 tonnes per year.

This submission followed the IPCC Good Practice Guidance and Uncertainty
Management of National Greenhouse Gas Inventories. The uncertainty only includes
sampling-based error through the propagation of errors (tier 1 approach) but no other
error sources (non-sampling errors). The % uncertainty of AD, i.e. Forest loss area is
8.97 % for this submission whereas the % uncertainty of the EFs is 13.03%, thus the
resulting combined overall error % of the emission estimate from deforestation is 15.06
for this submission.

For the carbon stock enhancement from forest plantation establishment, the average
annual removal for the ten-year reference period 2005 – 2015 has been calculated
with 3,351,332 tonnes of CO2e. Statistical uncertainty assessment for this figure is not
available since the data are not based on sampling but on aggregating records from
subnational reports on plantation establishment.

3
Forest Reference Level (FRL) of Myanmar

1. Introduction

The Government of Myanmar is fully aware of the causes and potential impacts of
climate change. Myanmar actively participated in global climate change mitigation
efforts by ratification of the United Nations Framework Convention on Climate Change
(UNFCCC) in 1994 and the Kyoto Protocol in 2003 as a non-Annex 1 party. Currently,
Myanmar views REDD+ initiatives as a contribution to the green development of
Myanmar as well as supporting the mitigation of, and adaptation to, climate change.
Myanmar became a partner country of the UN-REDD Programme in December 2011
and has quickly taken steps to start implementing REDD+ Readiness activities.
Regarding Decision 1/CP.16, paragraph 71(b) of the 16th Conference of Parties to the
UNFCCC (COP 16) in Cancun, 2010, a Forest Reference Emission Level and/or Forest
Reference Level (FREL/FRL) is one of the four key elements to be developed to
participate in REDD+ and its submission is on a voluntary basis.

Myanmar submitted its Intended Nationally Determined Contribution (INDC) in 2015

as a contribution to global climate change mitigation and adaptation efforts. It was
confirmed as the country’s first Nationally Determined Contribution (NDC) after
ratification of the Paris agreement in 2016. Under the NDC, forestry is a key sector and
quantitative targets are likely to be included. Furthermore, in their statement to COP
23 in 2017, the Government of Myanmar stressed that the national REDD+ Programme
is critical to their mitigation and adaptation pledges. Myanmar’s Initial National
Communication (INC) was submitted to UNFCCC in 2012 and the Second National
Communication (SNC) is now under preparation.

Following the suggestion of Decision 12/CP.17, Myanmar prepared its FRL using a
stepwise approach. Myanmar prepared the FRL submission as a benchmark for
assessing its performance in implementing REDD+ activities in contribution to climate
change mitigation. The proposed FRL in this submission is entirely based on historical
data which Myanmar considers to be transparent. Nonetheless, the choice of using

4
average historical emissions as its benchmark was made after consideration of the
national circumstances and expected future development plans. This submission will
also be consistent with anthropogenic forest-related greenhouse gas emissions as
contained in the country’s GHGs inventories of the SNC. This submission covers all-
natural forests, covering approximately 52% of the total country land area in 2005. The
scope of the FRL submission covers two REDD+ activities; Deforestation and
Enhancement, three carbon pools (AGB, BGB and litter), and one gas as CO2 only, with
the final FRL calculation expressed in tonnes of carbon dioxide equivalent per year
(tCO2 eq). Since the submission covers emissions and removals of Greenhouse Gases
(GHGs), it is considered as a Forest Reference Level (FRL), rather than, a Forest
Reference Emission Level (FREL) and is referred to FRL throughout this document.
Myanmar intends to expand the scope of the FRL as more extensive and better-quality
data become available.

Moreover, the Government of Myanmar recognize the potential importance of

plantation and forest restoration measures to climate change mitigation efforts.
During the Technical Assessment (TA) process, there was improvement of data on
historical forest enhancement. Calculations of enhancement estimation and then
incorporation into the calculations was conducted as part of a revised submission in
order to convert the FREL to FRL.

1.1. Objectives of FRL

The main objective of the FRL submission is to support the climate change mitigation
efforts under the national context of Myanmar. The national REDD+ Programme is
critical to the mitigation and adaptation pledges according to the Myanmar Country
Statement submitted to COP 23. Further objectives of the submission are;

- To assess and evaluate the performance of REDD+ policies and measures and
sustainable forest management practices

5
 - To provide information on emission projections to stakeholders including policy
makers, government line departments, technicians and members of the public
on a clear, transparent and consistent basis.
- To facilitate access to potential funding sources for results-based payments and
to support efforts to reduce emissions from the forest and land use sector.

This is also one of the key technical reports to support the efforts of multiple
stakeholders to achieve the goal of REDD+. The country team interacted with AT of the
UNFCCC through technical assessment process and followed the preliminary questions
to improve the document and technical approaches. This submission will also provide
information to facilitate effective implementation of forest management plans in
order to contribute to climate change mitigation targets

1.2. Summary of guidance for FRL development

The following four major decisions at the level of the UNFCCC are related to the
development of FREL/FRLs and were considered during the process of developing
Myanmar´s FREL:

 Decision 4/CP.15 in Copenhagen recognizes that developing countries in

establishing FREL/FRLs should do so transparently considering historic data, and
adjust for national circumstances;
 Decision 1/CP.16, paragraph 71(b) of Cancun includes FREL/FRLs as one of the four
key elements to be developed for REDD+;
 Decision 12/CP.17 Durban provides guidance for modalities of development of
FREL/FRLs, as follows;
o Express FREL/FRLs in tonnes of carbon dioxide equivalent (tCO2 eq) per year,
in order to serve as benchmarks for assessing the country’s performance in
implementing REDD+ policies and measures
o Maintain consistency with anthropogenic forest-related greenhouse gas
emissions by sources and removals by sinks as contained in the country’s GHG
inventories

6
 o Follow a step-wise approach to national FREL/FRL development, enabling
Parties to improve FREL/FRLs by incorporating better data, improved
methodologies and, where appropriate, additional pools,
o Sub-national FREL/FRLs may be elaborated as an interim measure, while
transitioning to a national FREL/FRL, and
o Update FREL/FRLs periodically as appropriate, considering new knowledge,
new trends and any modification of scope and methodologies
 Decision 13/CP.19 Warsaw provides guidelines on procedures for the Technical
Assessment (TA) of submissions of FREL/FRLs, including:
o Each FREL/FRL submission shall be subjected to a technical assessment
o Submission is on a voluntary basis
o Technical assessment is possible also in the context of results-based payments
o A synthesis report on the TA process is prepared by the Secretariat, for
consideration by SBSTA after the first year of technical assessments
o Countries are invited to nominate experts to the roster for TA as well as to
support capacity-building efforts in relation to the development and
assessment of FREL/FRLs

The UNFCCC decisions considered at the country level in Myanmar can then be
summarized as follow:

 A transparent process in developing the FRL was applied

 The FRL is based on historical data
 The FRL is consistent with the upcoming GHG inventory under development for
the Second National Communication (coordinated by ECD),
 The FRL has been developed with a stepwise approach in mind as well as the
present measuring capacities in country (which are expected to improve over
the coming years). In this modified submission, Myanmar included two REDD+
Activities of Reducing Deforestation and Enhancement of forest carbon stock.
Myanmar will try to include other remaining activities by the chance of getting
the reliable data in the future.
7
 2. National Context

2.1. National circumstances

The Republic of the Union of Myanmar is a largely rural country, with 70% of its total
population of 51.5 million people living in rural areas. In Myanmar, there are around
135 different ethnic groups with 100 different languages. The rural population still
relies primarily on biomass for energy purposes, as only 30% of the total population
has access to electricity (Population Census, 2014). The rate of population growth is
0.8% per year with an increase of 2.5% in urban areas and a decrease of 0.1% in rural
areas. The agricultural sector is still a major contributor to the country’s economy,
accounting for roughly 30%1 of Gross Domestic Product (GDP).

Myanmar is in a process of economic and political reform with the overall goal of
becoming a modern, developed and democratic nation by 2030. The political reform
process is accompanied by a structural reform process of the economic sector with
strong and increasing foreign direct investment. However, significant challenges still
exist with wide socio-economic gaps and sub-national inequalities in poverty and other
human development indicators. In particular, geographically remote areas (e.g. Chin
and Rakhine states) are suffering from low levels of infrastructure and lack of basic
social services (especially health and education) and job opportunities.

Additionally, inequalities between women and men are particularly significant in the
country. Women, especially from forest-dependent communities, participate
unequally in socio-political and decision‐making processes, due in significant part to
an inferior status in a religious context and the institutionalization of the view within
society that gender inequality is not a problem. Policies and strategies are in the
process of being reviewed and updated for all sectors of the country in order to
support gender inclusiveness in the overall comprehensive national development plan
and the fulfilment of sector specific gender-related targets and goals.

1
http://www.csostat.gov.mm/

8
Current economic development is concentrated on the regions of the country which
are particularly exposed to climate hazards such as cyclones, heavy rain, flooding,
drought or erratic rainfall (e.g. the regions of Ayeyarwady, Bago, Mon, Rakhine and
the Central Dry Zone in general). Negative impacts on agriculture, fisheries, livestock
or forestry will be especially felt by the poor and smallholder farmers which constitute
the great majority of all farmers in the country. Those regions are also the ones with
the highest rates of deforestation over the last 10 - 15 years. The economic and social
circumstances in Myanmar as well as the cultural and ethnic diversity make climate
change mitigation (including REDD+) challenging and all the proposed policies and
measures need to be thoroughly screened for potential negative impacts on people’s
livelihoods.

In Myanmar, forestry sector has traditionally played a critical role in the development
and economic transformation of the country. Forest management is being focused on
sustainability of forest resources; such as sustainable production of goods and services
for local needs and export, and conservation of its ecosystem and environment. As
provision of forest resources, especially timber and fuelwood from natural forests
alone is insufficient to fulfill demands for forest products, establishment of forest
plantation by various objectives and scales was conducted since 1970s. Under
different reforestation and rehabilitation programs, FD has been establishing different
types of forest plantations such as commercial plantation, watershed plantation, local
supply plantation, industrial plantation within its limited manpower and budget. With
the objective of promoting private investment in plantation forestry, private
plantation program was launched in 2006. FD has been promoting community forestry
program since 1995. FD introduced very specific plan of 10-year Myanmar
Reforestation and Rehabilitation Program (MRRP) from 2017-2018 to 2026-2027 in
order to enhance economic and environmental conditions of the country through
national reforestation and rehabilitation program. Therefore, the establishment of
new forest plantations on heavily depleted forests (land cover is not fit with forest
cover definition) and the restoration of heavily depleted forests through reforestation,

9
enrichment planting and natural regeneration through silvicultural operations will
form an important part of the REDD+ strategy of Myanmar.

2.2. Myanmar’s Nationally Determined Contributions-(NDC) submission and

relation to the FRL

The 2016 NDC document of Myanmar formulates several actions relevant for climate
change mitigation. The main mitigation actions concerning forest and land use are as
follows:

 Fulfilment of the national Permanent Forest Estate (PFE) target by 2030 with
an increase of Reserve Forests (RF) and Protected Public Forests (PPF) to 30%
of the national land area and the Protected Area System (PAS) to 10% of the
national land area.
 Energy efficient cook stoves in order to reduce fuel wood for energy purposes,
especially for the Dry Zone of Myanmar. The target is to distribute 260,000 new
cook stoves between 2016 and 2030.

Based on the 2012 National Adaptation Programme of Actions (NAPA), adaptation

actions are described in the NDC for different levels of priority and sectors. Forest
preservation measures, together with resilience in the agricultural sector and early
warning systems, are among the first priorities. In addition, several strategies and
policies are in process of development, or already being implemented, in order to
support the achievement of targets such as those in the Myanmar National Climate
Change Strategy and Action Plan (MCCSAP); Green Economy Strategic Framework;
National Environmental Policy, Framework and Master Plan; Environmental
Conservation law; Environmental Impact Assessments (EIAs), and state of
Environment reports etc. According to the MCCSAP, the action area of “environment
and natural resources” includes REDD+ and LULUCF. For REDD+ specifically (and linked
to the overall MCCSAP goal) the following REDD+ goal is envisioned:

“The Land Use and Forestry Sector contributes to an overall low-carbon development
pathway of the Nation through reducing deforestation and forest degradation and

10
the related GHG emissions while enhancing the livelihood of forest dependent people
and communities as well as ensuring inclusive sustainable growth and development
of the country as a whole”.

For the policy area of Forest Management, the implementation of the National
Forestry Master Plan (2001-2030) is mentioned as well as the national Biodiversity
Strategy and Action Plan of 2015-2020.

The NDC does not lay out quantitative targets for emission reductions. Its mitigation
section focuses on forestry by maintaining Myanmar’s carbon-sink status. An update
of the NDC, using more concrete, quantifiable data, is currently in process.

2.3. Forests in Myanmar

Myanmar forests are diverse and varied in composition and structure and constitute a
valuable ecosystem due to their wide extent (between latitudes 9’ 55’’ – 28’ 15’’ N and
longitudes 92’ 10’’ – 101’ 10’’ E), varied topography and different climatic conditions.
The forests are distributed over three main climatically distinct regions (Tropical,
Subtropical and Temperate). The Forest Department of Myanmar recognized and
adopted the general description of eight dominant forest types, (Burmese Forester,
June 1956, Departmental Instructions for Forest Officers in Burma, Annex VIII, pages
214 – 217) as described in Table 2.1 and dominant types of forests in Myanmar map is
mentioned in Annex 1.

Table 2. 1: Forest Types and Vegetation Zones in Myanmar

Forest Corresponding Short description Typical Distinct species types
type forest types rainfall found
according according to NFI (mm/year)
to Davis, field instruction, *
1960 19852
Evergreen Evergreen forest, Occurs in Tanintharyi region and 2,500- Southern type:
hardwood giant; other lowland areas with high 4,000 Dipterocarpus, Hopea,
forests Evergreen forest, annual rainfall Shorea, Parashorea,
typical; In the Northern type
Bamboo forest Dipterocarpaceae are
(degraded rain replaced by Cedrela,
forests) Chukrasia, Dysoxylum

2
In use for district management plan inventories since the end of the last NFI

11
 Forest Corresponding Short description Typical Distinct species types
type forest types rainfall found
according according to NFI (mm/year)
to Davis, field instruction, *
1960 19852
Mangrove, Tidal forests in Irrawaddy delta >3,500 Avicennia, Bruguiera,
typical; region and other Coastal areas Rhizophora, Sonneratia,
Nipa, Heritiera,
Mangrove, high
(kanazo forest);
Swamp forests; On interior lowlands and some >3,500 Lagerstroemia, Amoora,
areas along the Coast covered Barringtonia, Xylia, few if
Evergreen by river and other non-saline 2,500- any Dipterocarpaceae
forests, riverine; water 4,000

Mixed Mixed deciduous In both, low land and upland 1,250- Tectona grandis, Xylia
evergreen forest, lower; formations characterized by 2,500 xylocarpa, Pterocarpus
and Mixed deciduous teak, occurred north and south macrocarpus, Gmelina
deciduous forest, upper of the central dry zone, lower arborea, Millettia
hardwood moist; parts of Shan hills, Arakan pendula
forests Yomas, Chin hills and most of
Mixed deciduous Bago Yoma
forest, upper
dry;
Dry Dipterocarp Edaphic forest type occurring 900-1,250 D. tuberculatus,
deciduous (indaing) forest, mostly on gravel and sandy obtusifolius, turbinatus,
hardwood high; soils, alatus,
forests Best sites north of central dry On drier sites D.
and Dipterocarp zone on alluvial soils in river tuberculatus with
scrubs (indaing) forest, valleys Pentacme, Emblica and
low other sp.
Hill forest, dry On dry slopes, ridges and >3,000 Xylia xylocarpa,
shoulders of mountains best Pterocarpus, Adina,
formations have teak too Shorea oblongifolia,
Tectona hamiltoniana,
Spondias, Terminalia,
Vitex,
Dry forest, than- Different dry to xerophytic <900 Than-Dahat forests
dahat; formations prevailing in the (Terminalia oliveri,
central dry zone and the foot Tectona hamiltoniana);
Dry forest, thorn; hills and low mountains of Shan Te scrub forests;
state (Diospyros burmanica,
Dry forest, Dalbergia, Acacia
aukchinsa – catechu, Limonia,
thinwin Zizyphus);
(Diospyros, Sha thorn and scrub
Milletia); forests (Acacia catechu,
A. leucocephala, T.
hamiltoniana, Zizyphus,
Limonia, Cassia)
Coastal Beach and dune Casuarina equisetifolia >3,500 Rhizophora apiculata,
Conifer forests formations in Coastal areas of Bruguiera gymnorhiza,
forests3 Rhakine and Tanintharyi, sea Heritiera fomes
face of Ayeyarwaddy Delta

3
Not in Davis´ classification

12
 Forest Corresponding Short description Typical Distinct species types
type forest types rainfall found
according according to NFI (mm/year)
to Davis, field instruction, *
1960 19852
Hardwood Hill forest, Similar to tropical type because >3,000 Temperate genera of
rain evergreen of layered structure but Quercus, Castanopsis,
forests distinctive because of mix Magnolia, Fraxinus, Celtis
between tropical and temperate mixed with tropical
genera. Occurs in areas with genera of Dipterocarpus,
abundant rainfall, fog, clouds Terminalia, Engelhardtia,
and moist soils on upper valleys Sterculia, Ficus
and lower mountains of Kachin
state, Naga hills, upper
Chindwin river valley.
Mountain Hill forest, Occur on slopes and tops of hills >3,000 Common genera are
forests evergreen and mountains where colder Quercus, Castanopsis,
and winter temperatures limit Magnolia, Acer, Alnus,
scrubs growth of typical tropical Prunus, Pyrus, Ulmus,
species, Salix, Podocarpus, but
also species of tropical
genera such as Bauhinia,
Engelhardtia,
Lagerstroemia,
Cinnamomon, Ficus,
Hill forest, dry On dry slopes and ridges, often >3,000 Quercus incana,
subject to clearing and fire, mix Rhododendron arborea,
of mainly deciduous with some Schima wallichii, Pinus
evergreen species, sometimes keysia, Kydia
scattered pine trees mixed with
oak and chestnut, when open
stands with savanna-like
vegetation and bracken fern,

Hill forest, pine Dominant species khasi pine >3,000 Quercus griffithi, Q.
(pinus keysia), often pure stands incana, Q. serrata), in
with open canopy, prevalent in some areas locally mixed
Shan state, Chin hills, and a few with Rhododendron
in Arakan Yomas, Mostly in maximum and Alnus
altitudes between 1350 m – nepalensis (Kachin)
2450 m, sometimes mixed with
a few hardwoods of low stature
Source: Davis, 1960, Kermode, 1964, Kress et.al. 2003, Departmental Instruction of FD

There is no recent study that provides an update of national-level distribution and

extent of these different forest types in Myanmar. According to available information,
the most abundant forest types are the tropical Mixed Evergreen and Deciduous
forest, both containing upland and lowland subtypes. These forest types are well
known for the occurrence of teak and other valuable timber species that are subject
to commercial timber logging. The subtropical forests, which are scattered over slopes

13
and peaks of hill and mountain ranges, e.g., in Chin and Shan states, are traditionally
under the influence of shifting cultivation carried out by local communities for their
livelihood. In recent years, logging has been extended to lowland Evergreen Hardwood
forests (e.g. Tanintharyi region, Southern Myanmar) sometimes followed by
conversion to oil palm and rubber plantation on accessible or degraded stands (Rao et
al., 2013). The tropical dry forest types in Myanmar, concentrated in the Central Dry
Zone of the country, are also affected by human activity (e.g. conversion to agriculture,
firewood collection) as well as forest fires. Although the latter can cause serious soil
degradation, they are part of the natural dynamics in some open dry or savannah-like
forests where species occur which can benefit from burning (Ratnam et al., 2011).
Forests on wetlands, especially mangroves along the coastlines and freshwater swamp
forests in river deltas (e.g. Ayeyarwady) are threatened by agriculture and aquaculture
(e.g., shrimp farming), unsustainable collection of firewood or coastal development
activities (Stibig et al., 2007).

2.4. Forest Land Use and Land Cover by Forestry Legislation

Myanmar has a decentralized statistical system where the Central Statistical

Organization (CSO) and line ministries have responsibilities for data collection in their
respective domains. As a decentralized statistical system, statistical production
activities are scattered across the ministries and agencies (Mon. M.S, 2017). According
to the Forest Law, all forest areas and forest tree cover are subject to declaration as
Permanent Forest Estate (PFE) and administered by the Forest Department (FD) under
MONREC. The FD is responsible for protection and conservation of biodiversity and
sustainable management of the country’s forest resources through establishment of
PFE. PFEs include all forested areas on Land at the Disposal of the Government and
constitutes Reserved Forest (RF), Protected Public Forest (PPF) and the Protected
Areas System (PAS)4. RF and PPF are given a similar legal status under the Forest Law.
The status of land as PFE indicates administrative responsibility by FD but does not

4
See annex 2, 2 (a) and 2 (b)

14
directly imply any information regarding tree cover. Table 2.2 shows major land
categories defined by different Ministries and Departments which are concerning with
management on land resources of Myanmar;

Table 2. 2: Major Land Categories Reported in Statistical Year Book compiled by CSO

No Major Land Categories 2013-2014 2015-2016

1 Forest Land (RF, PPF, PAS) 27.64% 27.42%
2 Agriculture Land (6 sub categories) 18.08% 18.41%
Vacant Land/Virgin Land/ Cultivable 29.75% 29.54%
3
Waste Land (2 sub categories)
4 Other Land (13 sub categories) 24.53% 24.62%

2.5. REDD+ development in Myanmar

Myanmar is a signatory to the UNFCCC, having ratified the convention in November

1994 and signed the Kyoto protocol in 2003. The government of Myanmar together
with many key stakeholders is aware of the causes and potential impacts of climate
change and is striving to reduce its GHG emissions and contribute to climate change
mitigation. Myanmar became a partner of the UN-REDD Programme in December
2011. During 2012 and 2013, a REDD+ Readiness Roadmap was developed through a
national multi-stakeholder consultation process5. The process of implementing the
Roadmap is currently underway, in order to establish the national REDD+ design
elements according to the Warsaw framework for REDD+ (COP 19). A stakeholder
engagement process has been established and a safeguards roadmap has been
developed with the objectives of clarifying safeguards in the national context and
setting the ground for a future Safeguard Information System (SIS). REDD+ strategy
development began in 2016 with a comprehensive driver analysis and a first draft of a
REDD+ strategy document was developed by the end of the first quarter of 2017

5
The proceedings of the workshops are available on the country page of Myanmar on the UN-REDD workspace
(www.unredd.net)

15
through series of discussions with relevant Technical Working Groups6 and wider
stakeholder consultations. Sub-national consultations on the REDD+ strategy began in
the last quarter of 2017 and are expected to be completed in 2018. Action plans for a
National Forest Monitoring System (NFMS) and FREL/FRL were developed during 2015
and implementation is ongoing.

The development of the FRL was initiated by a group of experts representing a cross-
section of ministerial agencies and organizations. The REDD+ TWG on Measuring
Reporting and Verification (MRV) provided technical guidance and direction on the
implementation of both the NFMS and FREL/FRL action plans. The MRV TWG provides
a forum to access national technical capacities and institutional arrangements within
Myanmar, for both NFMS and FREL/FRLs development for REDD+. This submission is
largely due to the effort and commitment of the members of this TWG.

3. Definitions

Definitions used for the FRL are also consistent with those used in relation to SNC and
NDC preparation, including as follows;

3.1. Forest definition

The definition of ‘forest’ in Myanmar follows that used for the FAO Forest Resource
Assessment (FRA): “Land spanning more than 0.5 hectares with trees higher than 5
meters and a canopy cover of more than 10 percent or trees able to reach these
thresholds in situ”. In addition to land cover, land use is also considered in identifying
areas that fall under this forest definition. It covers temporarily de-stocked land for
which the long-term use remains forest. Therefore, it does not include land that is
predominantly under agricultural or urban land use. This definition is also applied by
FD, MONREC for satellite image classification in the national forest resource
assessment. In the context of this submission, ‘forest’ refers to all areas under forest

6
There are three Technical Working Groups for REDD+ in Myanmar, which are (1) Stakeholder Engagement
and Safeguard TWG, (2) Drivers and Strategy TWG and (3) Monitoring, Reporting and Verification TWG. Detail
TORs are available through http://www.myanmar-redd.org/.

16
cover which meet the above criteria, both under PFE and outside PFE as mentioned in
section 2.4.

3.2. Deforestation definition

Deforestation is defined as the conversion of forest land use to non–forest land use
(i.e., 100% loss of all three carbon pools7 included in EF calculations of this submission).
This FRL submission considers only the complete conversion of forest land use to other
land use during the period 2005-2015, not including land that was temporarily de-
stocked (and subsequently restocked) during the reference period.

3.3. Enhancement of forest carbon stocks through reforestation/afforestation

activities

Enhancement of forest carbon stocks through reforestation/afforestation is defined

as the establishment of new carbon sequestration capacities resulting from the
change of non-forest land use to forest land use. Although enhancement of forest
carbon stocks also occurs in forest land remaining as forest land, this submission
excludes this latter aspect of enhancement due to data limitation. However, it will be
included once the measuring capacities through an improved NFMS are established,
most likely after completion of the first measurement cycle of the upcoming National
Forest Inventory after 2022-23.

4. Scale

In accordance with the draft REDD+ strategy, Myanmar FRL is national level. A national
scale FRL is appropriate as all the existing land and forest monitoring and
measurement capacities are at the national level and there is currently limited MRV
capacity at the sub-national level. In addition, the impact of the REDD+ strategy
implementation is expected to be measured at national level.

7
Only three carbon pools of AGB, BGB and Litter are only considered in this submission and the remaining two
carbon pools will be likely to consider in the future.

17
5. Scope (activities, pools, gases)

5.1. REDD+ activities

Myanmar prioritized two of the five REDD+ activities as the focus of the country’s first
FRL submission: (1) deforestation and (2) enhancement of forest carbon stock through
afforestation/reforestation, due to the following reasons;

- Deforestation is estimated to be the main contribution to the total emissions

from the land use, land use change and forestry (LULUCF) sector in Myanmar;
- Measurement and monitoring in the impact and effectiveness of the existing
Myanmar Reforestation and Rehabilitation Programme (MRRP) is critical.
- For the remaining REDD+ activities (forest degradation, sustainable forest
management, and conservation of forest carbon stocks), not enough reliable
data are available, since measurement capacities have not yet been fully
established.
- Consistency with the Draft REDD+ strategy and Draft NDC which formulates
quantitative targets for reducing deforestation and enhancement of forest
carbon stocks through the targets of the 30-year Forestry Master Plan (2001-
2030) and the MRRP (2017-2026).

In common with many other developing countries, Myanmar experiences both

deforestation and forest degradation from various anthropogenic drivers. The
measurement of deforestation, however, is possible using data and methodologies
currently available at the national level, whereas considerable further work is required
before the change of carbon stocks in forest remaining as forest can be measured with
confidence. Moreover, it is estimated that the impact of deforestation, in recent
years, has had more of an impact on forest carbon stocks than forest degradation. The
definition of forest degradation, and methodologies to measure and monitor it, are
currently under discussion and will be included in future submissions. The definition
of conservation of forest carbon stocks and sustainable management of forests, in the
context of REDD+, will also be further considered, but at present it is assumed that the

18
impact of these two REDD+ activities will be captured by measurement of
deforestation and forest degradation.

The MRRP was initiated in 2017 and is intended to continue to 2026-2027. The MRRP
has been accorded priority in national forestry sector policy in recognition of the fact
that the provision of forest products and services from natural forests is insufficient to
meet demand. Under the MRRP, establishment of new plantations in degraded forest
areas and restoration of natural forests by silvicultural practices are being conducted
through systematic planning, implementation and monitoring. Enhancement of forest
carbon stocks was therefore prioritized as the second activity in the FRL submission
for Myanmar, initially with a focus on afforestation/ reforestation.

5.2. Pools and gases

Three carbon pools of Above Ground Biomass (AGB), Below Ground Biomass (BGB)
and Litter are included in this FRL, using the country district forest management
inventory data. The allometric equations and the default values according to IPCC
Guidances (IPCC 2003 and 2006) are used to calculate the country emission factor. This
submission omitted Dead Wood (DW) and Soil Organic Carbon (SOC) carbon pools due
to limited information/ data at national level. Plans are currently being developed to
facilitate the inclusion of the remaining two carbon pools in forthcoming submissions.
Because of the following reasons, Myanmar decided to exclude the SOC in this FRL
submission:
 The existing national data sources on SOC on forest soils are very localized and
only conducted for a few selected areas prior to forest plantation
establishment.
 The data collection protocols for SOC are not standardized and are inconsistent
over the years, depending on the project and on the involved technical
personnel.

19
  The use of Tier 1 default emission factors for SOC stocks as provided by IPCC
guidelines (IPCC-GPG, 2003)8 appear as substantially high in C-stock and do not
seem to represent average conditions of forest soils in Myanmar. The
application of default reference values of SOC stocks under native vegetation
of between 31 and 66 tonnes of C per ha for different tropical ecoregions (dry,
moist, wet) and different soil types (high and low activity clay, sandy soils)
would result at average in higher below ground SOC stocks than the average C-
stocks in above ground living biomass, which is very unlikely to be the case. The
only conditions where below ground SOC stock is likely higher than the above
ground C-stock is in mangroves and peatlands which both represent minor parts
of the forest vegetation in Myanmar, when compared with total forest cover.

Therefore, the inclusion of SOC, will be postponed until Myanmar is in the condition
to present more reliable data on SOC in forest and non-forest soils. Presently the
Forest Department, through the Forest Research Institute, FRI, is working on
developing standardized methodologies for soil data collection in upland mineral soils
and in Mangroves, which will be applied in the upcoming National Forest Inventory.
Results from the soil data analysis will likely not be available before the years 2022/
2023.

The default values for dead organic matter stocks, particularly DW, are not provided
since these are highly variable and site-specific, depending on forest type and age,
disturbance history and management regime. In addition, data on coarse woody debris
decomposition rates are scarce and thus, IPCC explains, it was deemed that globally
applicable default factors and associated uncertainty estimates cannot be developed
(IPCC 2006, Volume 4 Chapter 2.2.1).

Myanmar submitted only CO2 gas in this initial FRL although there are also non- CO2
emissions from LULUCF. Myanmar’s INC report included non- CO2 gases from biomass
burned due to land clearing and forest fire. These non-CO2 gases included CH4, N2O

8
Table 3.2.4 page 3.43 GPG, 2003

20
and NO2 and the total combined emissions were 637 Gg9 while that of CO2 was 102,264
Gg. Based on the data in the INC, therefore, this submission considers the contribution
of non-CO2 gases to be insignificant.

6. Reference period of FRL

Through a series of consultation meetings, the proposed reference period was

identified as the period from the years 2005 to 2015, due to the availability of the most
reliable national existing Activity Data (AD) and Emission Factors (EF) for this period.
Consistency with GHGs Inventory reporting was also considered for this submission.
Currently the SNC is being prepared based on 2010 data sets by using IPCC GPG 2003
and IPCC GL 2006 in order to ensure consistency with the development of AD and EF
for the FRL.

Although the reference period does not overlap with the INC, which used data from
the year 2000, The INC used EF based on IPCC global default factors and AD based on
projected data from FRA reports. This FRL used AD estimation based on an un-biased
sampling approach and EF calculations from plot data of available inventories of
district management plans. The EFs are therefore considered as nationally specific
data and more accurate compared to the previous GHGs inventory. Data generated in
this FRL development process will also benefit the current process of developing a
Second National Communication (SNC) and the Biennial Update Report (BUR) to the
UNFCCC.

7. Data used in the FRL

7.1. Deforestation

7.1.1 Activity data

According to the revised IPCC Guidelines for National Greenhouse Gas Inventories,
Activity Data (AD) are defined as data on the magnitude of human activity resulting in
emissions or removals taking place during a given period of time. The emissions

9
1 Gg = 1000 ton

21
include human activities resulting from deforestation and from forest degradation
while the removals include forest gain or enhancement of canopy cover. In this report
the AD has been developed by estimating the extent of forest change measured as
area estimates of forest, non-forest and forest loss during 2005-2015, excluding forest
degradation, forest improvement and forest area gain. The amount of deforestation
(forest loss) has been estimated using a sample-based approach. The data sets used
to generate AD are listed in Annex 3.

7.1.1.1 Rationale of sample-based approach:

Wall-to-wall maps for the years 2005, 2010 and 2015 have been prepared by the
Remote Sensing and GIS (RS/GIS) unit of Myanmar Forest Department (FD) (see Annex
3). These wall-to-wall maps are pixel-based and were produced through supervised
maximum-likelihood classifiers using imagery from Landsat (30 m), for the years 2005
and 2015, and using imagery from IRS (23.5 m) for the year 2010. Those maps were
classified according to FRA categories (Annex 4).

The post-classification change detection method is recognized as a suitable option to

estimate changes within and across different land cover types (IPCC, 2006), using the
available temporal maps. In this method, the AD for each mapping year could be
derived using estimates from each map on various land-cover classes, as sum of areas
of map units assigned to map classes are characterized as pixel counting.

However, while preparing the AD for FRL, several problems with these wall-to-wall
maps (Annex 5) have been detected. For instance, they were produced by different
people in the RS & GIS unit of FD without defining standard operating procedures that
could be followed to maintain quality control or could be reproduced in the context of
a long-term NFMS. Moreover, the mapping datasets (satellite imagery) used to
produce those land cover maps were not from the same reference year. For example,
the year 2005 map was produced using Landsat imagery collected during 2004-2006.

Following a detailed evaluation, errors in the wall-to-wall maps (see Annex 6)

appeared to be substantial because the maps have shown inconsistencies in geo-

22
locations both within individual maps and across the three temporal maps (2005,
2010, 2015). The causes of these inconsistencies are uncertain and might be attributed
to, for example, lack of consistent application of mapping between time periods,
inconsistencies in classification procedures, and inconsistencies in map qualities (IPCC,
2006).

According to IPCC General Guidelines (GL) (2006), AD should be neither over- nor
under-estimates (without bias or quantification of bias) and uncertainty should be
reduced as much as practically possible. The existing wall-to-wall maps, generally,
make no provision for accommodating the effects of map classification errors (Foddy,
2010). Moreover, the map accuracies (error matrix) can inform only about thematic
error issues but they do not produce the information necessary for calculating
sampling errors and the associated confidence intervals (Olofsson et al., 2013).
Therefore, the pixel-counting-based wall-to-wall approach provides no quantification
of sampling errors and no assurance that estimates are unbiased or that uncertainties
are reduced (Stehman, 2005; GFOI, 2016). The bias resulting from applying pixel
counting to obtain the area of a land cover class is labeled as “measurement bias”
rather than as “estimator bias” because a pixel count represents a complete census of
the region and therefore is not a sample-based estimator. Gallego's (2004) review
provides an excellent summary of many of the area estimation options, including a
critique of pixel counting and an overview of estimators combining ground and remote
sensing information, as well as a review of methods for small area estimation
applicable when interest lies in small geographic regions that receive few sample units.

We therefore used a sample-based approach as an independent and consistent

method to derive estimates on areas of stable forest, stable non-forest and forest loss
between the years 2005 and 2015. Among various types of probability-based sampling
design, stratified random sample (STRS) design has been used. The STRS offers the
option to increase the sample size in change class and forest loss proportional to the
total area and reduce the standard errors of the class-specific accuracy estimates for
comparatively rare classes such as deforestation. In addition, STRS is one of the easier

23
survey designs to implement having unbiased variance estimators (Olofsson et al.,
2014).

In order to develop an appropriate sample survey design, a notion of the existing

variability distribution of our parameters of interest is necessary. For this purpose, a
combination of multiple maps within the same period where used to identify a
spatially-explicit forest change map for the period of interest. The forest change areas
often occupy a small proportion of the landscape and assumptions of the STRS design
largely depend on accurate delineation of the change strata. Given the limitations of
the wall-to-wall maps, as described above, Global Forest Change (GFC) maps10 (Hansen
et al., 2013) were used to generate the strata map for the years 2005-2015. The freely
available GFC maps, year 2000 tree canopy cover map and annualized loss map, have
been produced following a transparent and consistent methodology and, therefore,
offer a complete application of the independent sample-based approach. The GFC
maps were adjusted to the forest definition adopted by Myanmar; a minimum of 5 m
canopy height and a minimum tree canopy cover of 10% within a 0.5 ha area. Figure
7.1 illustrates the steps followed to generate stratified random sample-based
estimates of forest change between 2005 and 2015.

A tree cover map of 2000 and gross forest cover loss data from 2001 to 2015 were
used to produce forest change (loss) strata map from 2005 to 2015. A forest gain map
was not used in stratification because the GFC gain map has not been separated by
the years of gain, which was required to identify the amount of forest gain
(enhancement) between 2005 and 2015. Through the Stratified Area Estimator –
Design tool within FAO’s System for Earth Observation Data Access, Processing and
Analysis for Land Monitoring (SEPAL11), a total of 1,884 stratified random samples
were generated using the GFC-based strata map of 2005-2015.

10
https://earthenginepartners.appspot.com/science-2013-global-forest
11
https://sepal.io/

24
The validation process followed recognized design considerations in which three
distinctive and integral phases are identified: sampling design, response design, and
analysis and estimation (Stehman and Czaplewski, 1998).

Figure 7. 1:Steps followed to produce stratified random sample-based estimates of forest

change during 2005-2015 in Myanmar

Step 4: Straified Estimator

Step 3: Sample Assessment
Analysis
Sample assessment: Prepare master sample assessment file
Using Collect Earth Tool in Open Foris. By combining all files from different
Interpreters' check every sample for a Forest interpreters following quality control
or NonForest cover, using available high-
quality validation datasets in Google Earth,
Bing Map and Google Earth Engine; accessible Stratified Estimator - Analysis: Using SEPAL
through Collect Earth. platform (http://sepal.io)

Quality Control
Sample-based area estimates of three land
Following a standard operating procedure, all cover classes: Forest, NonForest and Loss were
interpreters were allocated 70 samples generated from a stratified randon design
randomly selected which followed by sampling estimators and from a theoretical
consistance check. simple random design sampling estimators.
All assessed samples were randomly rechecked
for quality control through open discussuion
for assurance of quality check.

Step 1: Forest Change Map 2005-

Step 2: Generate Samples
2015
Forest/NonForest Map 2005
Generate stratified samples
Forest/NonForest map 2000 (≥10% canopy
The stratified forest change map 2005-2015
cover threshold from the tree cover 2000
Sample size formula of Cochran (1977)
map) + Gross forest cover loss event (loss
Sample generation: Sepal platform
year) 2001-2005 (Hansen et al., 2013)
Forest/NonForest Map 2015 (http://sepal.io)
Forest/NonForest map 2005 + Gross forest Sample survey design
cover loss event 2006-2015 (Hansen et al., using survey designed in Open Foris Collect
2013)
Stratification: Forest change map 2005-
2015
Stratified Forest Change Map 2005-2015:
Three strata (Forest, NonForest, Loss) were
generated from Forest/NonForest 2005 Map
and Forest/NonForest 2015 Map

25
7.1.1.2 Sample design
The sampling design refers to the methods used to select the locations at which the
reference data are obtained, in this case, the methods through which the 1,884
samples were derived from the GFC-based strata map of 2005-2015 using SEPAL’s
Stratified Area Estimator – Design tool. By default, this tool allocates a minimum of 50
samples in the smallest stratum, following the Cochran (1977) formula (see Equation
1 below) (Olofsson et al., 2014), which in this case is the forest loss stratum. However,
a total of 300 samples, out of 1,884 samples, were generated for the forest loss
stratum with an aim to reduce standard error for the change user’s accuracy estimate.
The sample size is much higher than a sample size of 50-100 suggested by Olofsson et
al. (2014) for change stratum using the variance estimator for user’s accuracy. Figure
7.2 shows the distribution of the samples across Myanmar.
Equation 1

2
(∑ 𝑊𝑖 𝑆𝑖 )2 ∑ 𝑊𝑖 𝑆𝑖
𝑛= ≈ ( )
2 1 𝑆(𝑂̂)
[𝑆(𝑂̂)] + ( ) ∑ 𝑊𝑖 𝑆𝑖2
𝑁

Where: n= the number of required sample units

N= total number of all possible sample units in the area of interest
𝑆(𝑂̂)= the standard error of the estimates over all accuracy that we would like to achieve
𝑊𝑖 = mapped proportion of area of class i,

𝑆𝑖 = standard deviation of stratum i, 𝑆𝑖 = √𝑈𝑖 (1 − 𝑈𝑖 )

26
 Figure 7. 2: Distribution of the 1,884 samples across Myanmar

(a) Stratified forest change map 2005-15 (b) Stratified random sample

Figure 7.3 illustrates a change decision tree where the 2005 land cover is forest. There
will be equivalent decision trees for other scenarios e.g. intact forest to degraded
forest, and forest to non-forest land cover types. These assessments allow change in
major land cover categories to be reported and areas under different stratum
estimated.

27
 Figure 7.3: Decision tree for sample change analysis

Has the land cover changed between

2005 and 2015?

[NO] - Stable condition in

[YES] - Forest in 2005 but both years.
NonForest in 2015 Was land cover forest in
2015?

[YES]- Loss of forest [NO] - Stable

[YES] - Stable forest
(deforestation) nonforest

7.1.1.3 Response design

The desired goal of this validation was to derive a statistically robust and quantitative
assessment of the uncertainties associated with the forest area change estimates.
Several factors potentially impact on the quality of forest mapping (GOFC-GOLD,
2016), namely:

 The spatial, spectral and temporal resolution of the imagery

 The radiometric and geometric pre-processing of the imagery
 The automated and manual procedures used to interpret the forest map
category
 Thematic standards (i.e. minimum mapping unit and land use definitions)
 The availability of field reference data for evaluation of the results.

Approaches were used to minimize these sources of error following IPCC and GOFC-
GOLD good practice guidelines, as appropriate. However, the quality of reference data
and the sample selection for accuracy assessment of the change area were slightly
compromised by the restricted availability of high-spatial resolution archived imagery
in Google Earth and Bing Map across Myanmar.

28
Through a collect survey design form using Open Foris Collect12, the two Land cover
types (forest and non-forest) were assessed within each sample through an expert
image interpretation of medium (15m pan-sharpened Landsat) to very high (<1m)
spatial resolution satellite data. The map and reference datasets used in the change
assessment are listed in Table 7.1. The reference datasets have sufficient temporal
representation consistent with the change period: 2005-2015. The collect survey
design form has been set for each reference label to allow an interpreter-specified
confidence level of high, medium or low. Figure 7.4 shows an example of reference
data available within Google Earth-based Collect Earth System13 used to interpret land
use/cover and monitor changes with time. The figure also illustrates sample no. 1517
with temporal resolution of Google Earth imagery, used for sample assessment during
2005-2015.

Table 7. 1: Validation datasets used to assess 1,884 samples

Type Data types Spatial Source

resolution
Stratified Landsat scenes captured in 2005 30-m USGS Earth
and 2015 Explorer
High-resolution RGB imagery from 10-m to 30-cm Google Earth Pro/
various satellite sensors, such as Bing Maps
SPOT, GeoEye-1, WorldView-
1/2/3, Digital Globe, IKONOS, etc.
Validation
Landsat time series archive: 2005- 30-m (15-m Google Earth
2015 pan-sharpened) Engine
Landsat/Sentinel
MSI 2 Archive

12
http://www.openforis.org/tools/collect.html
13
http://www.mdpi.com/2072-4292/8/10/807/html

29
 Figure 7. 4: Example of reference data (Google Earth Archive)

2005 2015

The Collect Earth system works through a combination of freely available platforms
includes Google Earth, Google Earth Engine and Bing Maps. The temporal consistency
of the system has been ensured through the Google Earth Engine (GEE) script,
prepared for the assessment of 1884 samples, to produce reference data using the
best available cloud-free pixels during the end of the year. Therefore, it produces
annual composite for each sample in GEE and the composite is bias towards the end
of each year within the reference period. The composite was designed to produce a
false colour composite (FCC) imagery through a combination infra-red, short-wave
infra-red and red spectral reflectance in the red, green and blue channels, respectively.
The FCC imagery from infra-red, short-wave infra-red and red reflectance bands in the
red, green and blue channels produce the best visualization to discriminate between
forest and non-forest. The high-spatial resolution imagery in Google Earth and Bing
Maps also helped to cross-check the samples with uncertainty in discrimination. Figure
7.4 illustrates the process of consistency check between the two years (year 2005 and
year 2015) through the provision of freely available high-spatial resolution imagery
from Google Earth and Bing Maps. It is important to note that the sample assessment
protocol also has the provision to record the type of reference data used (e.g. Digital
Globe or Landsat) to take decisions on a sample by the interpreter.

30
The forest cover assessment team in the RS-GIS unit of FD took special care on quality
control and on quality assurance. To ensure consistency among different interpreters,
Myanmar has taken the advantages of in-house capacities and followed two
approaches, includes: consistency check among interpreters before sample
assessment and quality check after the sample assessment. Figure 7.1 Sample
Assessment (Step 3) explained the methods followed before and after the sample
assessment. It is important to note that the forest cover assessment team has received
several national trainings from FAO on sample-based assessment through the
Myanmar UN-REDD national programme (Sep 2015; Nov-Dec 2016 and August-Sep
2017) and a few members in the team have also attended various international
training programmes, all of which were specifically designed to interpret and validate
change areas (such as changes of Forest to Non-Forest or changes of Non-Forest to
Forest) and/or across the different IPCC land cover/use categories from imagery using
Collect Earth. All interpreters, moreover, have sound knowledge on Myanmar forest
conditions and so have provided additional advantages to separate forest from non-
forest.

7.1.1.4 Analysis design

Table 7.2 shows the generic structure of error matrix that has been used to derive
sample-based area estimates. Grey coloured cells represent map areas that have been
validated as correct. Orange coloured cells, however, are either false positives or false
negatives. For example, cell 𝑝12 is false negative and cell 𝑝21 is false positive.
Interpretation of these data assumes that the reference data are error free and that
the sampling is unbiased and of enough size. Nevertheless, the confusion matrix
provides a simple and convenient method to illustrate the nature of any disagreement
between the stratified map and the reference data.

The accuracy of a class is expressed in two ways: user's and producer's accuracies. The
producer's accuracy provides a measure of accuracy of the classification scheme. The
producer’s accuracy is also known as the error of omission because areas that have
been incorrectly classified are “omitted” from the correct class. This accuracy indicates
31
how well the sample points falling on a given land cover type are classified, i.e., it is
the probability of how well the reference data fitted the map.

Table 7. 2: Structure of confusion matrix

Reference data using Collect Earth

System User’s accuracy
Total
Non- (𝑈𝑖 )
Forest Loss
Forest
𝑝11
Forest 𝑝11 𝑝12 𝑝13 𝑝1.
𝑝1.
Stratified 𝑝22
Loss 𝑝21 𝑝22 𝑝23 𝑝2.
map 𝑝2.
Non- 𝑝33
𝑝31 𝑝32 𝑝33 𝑝3.
Forest 𝑝3.
Total 𝑝.1 𝑝.2 𝑝.3 1
Producer’s accuracy 𝑝11 𝑝22 𝑝33 𝑂𝑣𝑒𝑟𝑎𝑙𝑙 𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦
(𝑃𝑖 ) 𝑝.1 𝑝.2 𝑝.3 = 𝑝11 + 𝑝22 + 𝑝33

7.1.1.5 Estimation and uncertainty

GFOI 2016 methods guidance document (MGD) version 2 was used to derive sample-
based area estimates and uncertainty of the sample-based area estimates using
reference data label of the 1,884 samples. The reference data of each sample derived
from sample assessment is defined as the best available determination of the ground
condition at a specified location. The reference data label is assumed to be correct,
but it is well-known that reference inference error could often occurs and could impact
greatly on evaluations of land cover and land cover change by remote sensing (Foody,
2010). The way uncertainty is addressed depends on the inference framework
employed. In this report, a design-based inference has been used (Särndal et al., 1992)
in which the uncertainty associated with the estimator is defined as the variability of
the estimates over the set of all possible samples that could have been obtained for
the chosen sampling design and population sampled. Other sources of uncertainty
could often be present. For example, error in the reference class label, geo-location
error, and mis-matched classification legends may generate additional uncertainty.
But only the variability attributable to sampling in the uncertainty analysis was
considered. Stratified estimators were used to derive sample-based estimates and a
32
brief description of the stratified random sampling estimators has been described in
Annex 7.

7.1.1.6 Results
The error matrix of the 1,884 assessed samples is summarized in table 7.3. The
reference datasets were used to generate sample-based estimates along with the
associated confidence intervals for these sample-based area estimates. The user’s
accuracy, or commission error, represents an over-estimation of forest cover
compared with forest loss and non-forest cover classes. For example, 188 samples out
of 831 were detected as forest cover when they were not. Specifically, 49 of these
samples were actually forest loss in 2015. Three examples of sample assessment using
different validation datasets have been illustrated in Annex 8, including forest 2005 to
forest 2015; non-forest 2005 to non-forest 2015; and forest 2005 to non-forest 2015.

Table 7. 3: Confusion matrix for 2005-2015 forest change map based on 1,884 stratified
random samples

Reference data
User's
Forest Loss Non-forest Total
accuracy
Forest 0.32 0.03 0.10 0.44 0.71
GFC Map
Loss 0.04 0.07 0.05 0.16 0.44
(Modified)
Non-forest 0.16 0.02 0.22 0.40 0.55
Total 0.52 0.11 0.37 1.00
Producer's accuracy 0.61 0.62 0.60 0.61

Table 7.4 provides stratified area estimates of forest and non-forest for the years 2005
and 2015, respectively. Forest cover estimates reported to FRA (FA0, 2015) were 33.32
million ha in the year 2005 and 29.04 million ha in the year 2015, much lower than
sample-based forest cover estimates. Table 7.5 provides more detailed estimates on
various parameters from the sample-based assessment, considering the weighted
producer accuracy (proportional to the area per class) for forest, non-forest and loss
strata. This corresponds to the interpretation of the results from the perspective of
the reference data. In general, it indicates that the forest loss stratum has

33
 comparatively lower producer’s accuracy in spatial detection of information.
Therefore, the sample-based estimates showed a much wider (25%) confidence
interval in the forest loss stratum compared to stable forest and stable non-forest
strata. Apart from the usual subjective differences between estimators, it is assumed
that the uncertainty might also be associated with misinterpretation of samples, lack
of high spatial resolution imagery for some samples, misinterpretation of some forest
types with non-tree vegetation cover, and seasonal variations.

Table 7. 4: Stratified area estimates in hectare (ha) with confidence intervals (CI), weighted
producer’s accuracy and user’s accuracy under forest and non-forest cover classes for the
years 2005 and 2015
GFC Map
Accuracy (modified) Sample-based estimates
Land cover
area (ha)
classes
Weighted MOE
Producer’s User’s Area (ha) SE (ha) CI (ha)
Producer’s (%)
Year 2005
Stable forest 0.71 0.77 0.77 44705401 44107555 999206 1958444 4.4
Non-forest 0.60 0.53 0.55 22952351 23550197 999206 1958444 8.3
Year 2015
Stable forest 0.61 0.57 0.71 42894890 39841930 1082252 2121216 5.3
Non-forest 0.74 0.77 0.63 24762862 27815821 1082252 2121216 7.6

Table 7. 5: Stratified area estimates in hectare (ha) with confidence intervals (CI), weighted
producer’s accuracy and user’s accuracy for three land cover classes during 2005-2015.

Accuracy GFC Map Sample-based estimates

Land cover classes Produce Weighted (modified) MOE
User’s area (ha) Area (ha) SE (ha) CI (ha)
r’s Producer’s (%)
Stable forest 0.61 0.76 0.71 42894890 40446950 789576 1547568 4
Loss 0.62 0.18 0.44 1766799 4289839 392682 769656 18
Stable non-forest 0.60 0.55 0.55 22996062 22920963 751052 1472062 6

Along with the forest and non-forest classes, the sample-based area estimate and
associated confidence interval of the forest loss class have been improved through
four approaches: increasing sample number to 300; using local knowledge and
information on areas of forest loss and cross-checking existing national maps that can
be used as proxy for the investigation of forest loss, using high spatial-resolution
imagery available in Google Earth for those samples, and following a quality control

34
procedure. The stratified area estimates with confidence intervals outlined in table 7.5
were calculated following the IPCC (2003, 2006) guidelines and the GFOI (2016)
methods guidance documents. It is important to underline that the map estimates are
bias-corrected (‘adjusted’) considering the national forest definition, which includes
land classification as well as tree cover.

In conclusion, the area of forest loss is about 428,984 ha per year over the period 2005-
2015. In comparison, the forest loss between 2005 and 2015 reported to FRA 2015
(based on locally-produced wall-to-wall maps) was 428,000 ha per year which is about
793 ha per year lower than the sample-based estimates and has been found
comparable between the two independent estimates. Figure 7.5 shows a comparison
of estimates from local maps and estimates after bias-correction for forest, non-forest
and forest loss classes. Given the current limitations of local maps, as explained above,
it was decided to use the stratified area estimated of forest loss as activity data.

Figure 7. 5: Area estimates of forest cover change during 2005-2015, in million ha, from local
land cover maps and from stratified random sampling design

50

40 GFC map (modified) estimates

Area (million ha)

Sample-based estimates
30

20

10

0
Forest Gross loss NonForest
Strata

7.1.2 Emission factors

The first national scale ground survey-based forest inventory with a probability
sampling approach was initiated in 1980-81 under a UNDP/FAO project (first phase)
which eventually was extended until 1992 (second phase). The target precision for the
key parameter (timber volume) was an error margin of 20% (at 95% confidence level)

35
at the 50,000-ha level. However, the NFI work was never fully finished for several
reasons.

Since then, the FD has carried out district level inventories for periodic management
planning purposes including the definition of annual allowable cut (AAC) planning and
the development of stand and stock tables at the forest management unit (FMU) level.
The inventory design is based on the former NFI design with a systematic distribution
of plots within two basic strata: closed forests (>= 40% tree cover) and open forests
(>=10% - <40% tree cover). For the plot design, over the years four different types
were in use: (1) the 1.05 ha L shape original NFI plot, (2) a nested rectangular 1 ha plot
with two sub-plots in the upper right edge of the main plot area, (3) a circular 50 m
radius plot and (4) a rectangular 1-acre size plot.

For the emission factor calculation in this FRL, the best available data are therefore the
management plan inventory data, roughly for the same reference period as the activity
data. District level inventories were carried out in 40 districts out of 68 districts during
2005 and 2017. For the remaining 28 districts, no full inventory data are available yet.
However, the management plan inventory data covered 11 out of 15 states and
regions of Myanmar. The inventories also represented all tropical and sub-tropical
forest types. There were included most of the areas where forest cover change has
occurred during the reference period.

When the first cycle of upcoming NFI is finalized and the measurement results are
available, Myanmar will replace the EF calculated on existing data by more accurate
and unbiased values covering the full national scale. Notwithstanding, this will
probably not be possible before the year 2022/23.

This submission drew upon the data generated from 11,284 inventory plots of district
level forest inventory that were collected during 2005 to 2017 (Figure 7.6). Some of
the sample plots are also located outside PFE. The sample size, sample design, tree
measurement and the prorated per ha value for four different sample plot designs for

36
40 districts are described in the Annex 9. Detailed description of different sample plot
designs is mentioned in Annex 10.

Figure 7. 6: Location of forest inventory plots collected during 2005 to 2017

7.1.2.1 Methodology for Emission Factor Calculation

As described above, Myanmar decided to report three carbon pools, i.e. AGB, BGB and
Litter in FRL. A combination of Tier 1 and 2 approaches were used in estimating these

37
carbon pools using the following information from the management plan inventory
datasets of 40 districts:

- ID of Tree/Stand and its Location (District ID, Latitude and Longitude)

- Tree Type/Categories (by two categories, i.e. stand DBH of 20 cm and above,
and that of 10 cm to 19 cm)
- Forest Types (by 19 Types that were mentioned in Section 2.3)
- Tree Species Code (based on Forest Inventory Manual of FD)
- Diameter at Breast Height (DBH) in cm

Tree height was not collected in previous district management inventories. DBH is only
recorded as key parameter and timber volume estimation was done based on existing
national bole volume equations. Therefore, the equation with DBH only as entrance
variable can be used for choosing the allometric equation to estimate AGB.

Different allometric equations for AGB value were tested as well as the national bole
volume equations with corresponding expansion factors. Finally, the equation14 of
AGB= EXP ((-2.289 + 2.649 * LN (DBH) – 0.021 * (LN(DBH))2) were chosen since no
major differences seemed to exist between this equation and more elaborate
calculations based on national bole volume equations or other pantropical equations
(e.g., Chave et al, 2005, Chan et al, 2013 and Table 4.A.1 GPG-LULUCF, IPCC 2003
according to each forest types in Myanmar).

The Below Ground Biomass (BGB) value is estimated by multiplying the mean AGB and
the ratio of BGB to AGB (R) for each forest types. Myanmar uses the forest types
classification according to Field Instruction 1985 as mention in Section 2.3. Although
the forest types in the instruction are different to the global forest types, they have
similar characteristics so the forest types in Myanmar are simplified according to
global categories in order to select the appropriate ratio of BGB to AGB.

14
Table 4.A.1: Allometric Equations for Estimating Above Ground Biomass of Tropical and Temperate
Hardwood and Pine Species from IPCC Good Practice Guidance for LULUCF

38
The sum the AGB and BGB values (tonnes per ha) were converted into carbon tonnes
per ha by the multiplication with the default value of carbon fraction of dry matter;
i.e. 0.47. The default values (Carbon tonnes per ha) for litter, according to respective
forest types, are used and then later the total emission in Carbon tonnes per ha value
for each district is got by summing of the Carbon tonnes per ha values of all three
carbon pools. The default factor of 3.67 is used to convert the Carbon tonnes per ha
value to carbon dioxide tonnes per ha value. The brief step by step calculation for three
carbon pools are mentioned in the Annex 11. The respective allometric equations
applied in AGB calculation, the correspondent values of R and Litter in accordance with
Myanmar forest types are mentioned in the Annex 12.

Weighting is necessary because areas of districts are different as well as forest cover
areas are also different. Therefore, there were not the same number of forest
inventory samples. Weighting is carried out to get the EF values of different districts
as a proportional weight rather than just calculating an arithmetic mean of the sum of
the district values. Weighting is calculated as per the following steps:

𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 𝑖𝑛 ℎ𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = 𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑙𝑜𝑡𝑠 𝑥 𝑝𝑙𝑜𝑡 𝑠𝑖𝑧𝑒 … (1)
𝑇𝑜𝑡𝑎𝑙 𝐶𝑂2 𝑇𝑜𝑛 𝑃𝑒𝑟 𝐻𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 … … … … … … … … … … … … … … … … … … … (2)
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2)
𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝐶𝑂2 𝑚𝑒𝑎𝑛 𝑣𝑎𝑙𝑢𝑒 𝑝𝑒𝑟 ℎ𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = … … … … . (3)
𝑡𝑜𝑡𝑎𝑙 𝑛𝑜: 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑙𝑜𝑡𝑠
𝑂𝑣𝑒𝑟𝑎𝑙𝑙 𝑤𝑒𝑖𝑔ℎ𝑡𝑖𝑛𝑔 𝑣𝑎𝑙𝑢𝑒 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (3) 𝑥 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (1) … … … … … . . … . (4)

𝒕𝒐𝒕𝒂𝒍 𝒐𝒗𝒆𝒓𝒂𝒍𝒍 𝒘𝒆𝒊𝒈𝒉𝒕𝒊𝒏𝒈 𝒗𝒂𝒍𝒖𝒆

𝑾𝒆𝒊𝒈𝒉𝒕𝒆𝒅 𝑴𝒆𝒂𝒏 (𝑵𝒂𝒕𝒊𝒐𝒏𝒂𝒍 𝑬𝒎𝒊𝒔𝒔𝒊𝒐𝒏 𝑭𝒂𝒄𝒕𝒐𝒓) =
𝒕𝒐𝒕𝒂𝒍 𝒔𝒂𝒎𝒑𝒍𝒆 𝒔𝒊𝒛𝒆

Only one national level EF will be used since national data sources for stratification by
forest type are currently unavailable. In future FRL submissions, some stratification
maybe available to develop more disaggregated data and improve the resulting
emission and removal estimates.

39
7.1.2.2 Results
After using the steps mentioned in the methodology, the following table gives the values of tCO2 eq per ha representing for three
carbon pools, for the respective forest districts, including the number of sample plots used for the calculation. Myanmar will use
weighted mean values of tCO2 eq per ha for a national level EF based on 40 districts, i.e. 125.43 tCO2 eq per ha (Table 7.6).
Table 7. 6: Results of the CO2 eq Mean Value tonnes per ha and weighted mean CO2 eq tonnes per ha

Total No Weighted CO2

Sample Size in Total CO2 Ton Per Ha Overall
No. District Name of sample Plot Size Mean Value per
Ha (for all sample plots) weighting
plots: ha
1 Bago 567 1 567 32,353.79 57.06 32,353.79
2 Dawei 109 0.4047 44.11 58,412.59 535.90 23,639.58
3 Falam 159 0.4047 64.35 39,151.93 246.24 15,844.79
4 Hinthada 68 0.7854 53.41 8,756.80 128.78 6,877.59
5 Bhamo 427 1.05 448.35 84,784.33 198.56 89,023.55
6 Myitkyina 200 1.05 210 49,934.00 249.67 52,430.70
7 Katha 661 1 661 65,077.32 98.45 65,077.32
8 Shwebo 174 1 174 16,122.59 92.66 16,122.59
9 Monywa 190 1 190 20,930.61 110.16 20,930.61
10 Magwe 8 1 8 511.32 63.92 511.32
11 Gangaw 311 1 311 38,631.23 124.22 38,631.23
12 Minbu 467 1 467 51,309.52 109.87 51,309.52
13 Pakkoku 33 1 33 2,661.06 80.64 2,661.06
14 Thayet 274 1 274 29,057.04 106.05 29,057.04
15 Myaungmya 10 0.7854 7.85 224.74 22.47 176.51
16 Dakinathiri 105 1.05 110.25 13,792.70 131.36 14,482.34
17 Ottarathiri 67 1.05 70.35 7,948.02 118.63 8,345.42
Taungoo: Nay Pyi Taw 10 1.05 10.5 1,001.73 100.17 1,051.82

40
 Taunggyi: Nay Pyi Taw 20 1.05 21 2,959.88 147.99 3,107.87
18 Pyarpon 47 0.4047 19.02 704.34 14.99 285.05
19 Pyay 430 1 430 35,286.86 82.06 35,286.86
20 Sittwe 6 1.05 6.3 5,063.63 843.94 5,316.81
21 Taungoo 962 0.7854 755.55 77,606.43 80.67 60,952.09
22 Thandwe 147 0.7854 115.45 12,852.01 87.43 10,093.97
23 Tharyarwaddy 446 0.7854 350.28 31,618.45 70.89 24,833.13
24 Kalay 869 1 869 103,861.64 119.52 103,861.64
25 Khamti 951 1 951 80,905.13 85.07 80,905.13
26 Mawlaik 971 1 971 141,341.44 145.56 141,341.44
27 Tamu 45 1 45 3,295.45 73.23 3,295.45
28 Linkhay 234 1.05 245.7 22,868.07 97.73 24,011.47
29 Taunggyi North 225 1.05 236.25 44,405.31 197.36 46,625.58
30 Taunggyi South 334 1 334 50,596.34 151.49 50,596.34
31 Kyaukme 519 1.05 544.95 163,846.43 315.70 172,038.75
32 Pathein 76 0.7854 59.69 4,854.90 63.88 3,813.04
33 Kyaukphyu 219 0.4047 88.63 20,536.31 93.77 8,311.04
34 Maungdaw 52 1.05 54.60 5,159.97 99.23 5,417.97
35 MyaukOo 77 1.05 80.85 7,672.97 99.65 8,056.62
36 KyaukSe 163 1 163 20,389.03 125.09 20,389.03
37 Mandalay 19 1 19 1,362.88 71.73 1,362.88
38 Meiktila 90 1 90 5,882.73 65.36 5,882.73
39 Pyin Oo Lwin 465 1 465 52,438.86 112.77 52,438.86
40 Yamethin 77 1 77 4,955.61 64.36 4,955.61
11,284 10696.46 1,421,126 5,884.26 1,341,706.12
Arithmetic mean 147.11 125.43
weighted mean
by sample size

41
7.2. Enhancement of forest carbon stocks

Following the discussions and recommendations of the first technical assessment (TA)
during 19-23 March 2018, Myanmar has developed a reference level for enhancement
from the establishment of forest plantations. During the technical exchange with the
AT, Myanmar questioned the necessity of using historical average removals as
benchmark for performance on enhancement as a result of REDD+ implementation.
The reason to question this is that for calculating the amount of removals achieved
since REDD+ implementation, removals from the past bear no influence on the amount
of new removals as a result of REDD+. As such, Myanmar was wondering whether it
would not be the best and most straightforward to assume a zero baseline, accounting
only new removals from REDD+ implementation and no removals from continued
growth in existing plantations. In response to Myanmar’s proposal, the AT noted that
decision 12/CP.17 (annex) requires Parties to present accurate information on their
FRLs. The AT further indicated that a zero baseline would not be accurate because
forest enhancement efforts have been ongoing since at least 2004. Secondly, the AT
suggested that a zero baseline would not take into account the effect of forest age and
areas available for forest plantation in the reference period and in the future. The AT
argued that the growth rate of young forests is higher than that of old forests and,
hence, achieving removals with younger forests is easier than with older forests.
Ignoring this would likely result in an overestimation of enhancements. The AT also
commented that a similar limitation applies to the area available for forest
enhancement, since if a significant extent of area suitable for plantation is already
used for reforestation or afforestation in the reference period, it would be more
difficult to increase the area available for forest enhancement, thus impacting the
carbon removals in the future. Following the suggestions by the AT, Myanmar
therefore calculated historical removals during the reference period.

Forest plantations in Myanmar are established on land that was normally heavily
degraded or on grass, savannah and bush/scrub land, often accompanied by bamboo
with only occasional occurrence of small trees, i.e. areas that fall outside of the

42
 definition of forests. During the establishment of plantations, site preparation takes
place that includes the removal of existing above ground biomass on the area usually
with the use of fire.

As mention in Section 2.1, plantations are established for different purposes (e.g.
commercial timber production, watershed protection, industrial and fuelwood use,
village forests and for Mangrove restoration etc). The species used are normally native
tree species which also occur in natural forests, such as teak (Tectona grandis),
ironwood/pyinkado (Xylia xylocarpa), rosewood (Dalbergia sp.), padauk (Pterocarpus
macrocarpus), mangrove species (Rhizophora sp., Avicennia sp, Bruguiera sp.), among
others. To a minor extent, fast growing exotic tree species such as Eucalypts or Acacias
are also planted, mostly for fuelwood and for industrial purposes.

7.2.1 Activity data

There are four major types of state-owned forest plantation established by the FD
under the government budgets (here after government plantation) from 1980s and
plantation database are managed by Natural Forest and Plantation Division of FD.
Starting from 2006, government policies encouraged private sector investment in
establishing commercial forest plantations. Following table presents forest plantation
areas of FD established and that of private sector;

Table 7. 7: Forest plantation areas as recorded by the Forest Department 2000- 2015

Total Cumulative Cumulative

total private total area
Year Government plantation 2005 - 15
Plantation Area ha
Plantation Area area

2000-2001 30,718 30,718 30,718

2001-2002 30,756 30,756 61,474

2002-2003 31,396 31,396 92,870

2003-2004 30,441 30,441 123,310

2004-2005 31,974 31,974 155,285

2005-2006 33,201 33,201 188,486 188,486

43
2006-2007 28,328 113 28,441 216,927 216,927

2007-2008 25,670 2,765 26,666 243,593 243,593

2008-2009 26,504 5,087 29,373 272,966 272,966

2009-2010 25,349 9,933 32,271 305,237 305,237

2010-2011 17,729 11,189 26,189 331,426 331,426

2011-2012 14,253 13,110 24,846 356,273 356,273

2012-2013 9,248 11,032 17,264 373,536 373,536

2013-2014 11,281 12,283 18,354 391,890 391,890

2014-2015 8,806 10,895 14,133 406,023 406,023

2015-2016 7,805 10,371 12,354 418,377

total 363,458 86,779 418,377

Average 308,636
05-15

For the enhancement estimation, data available in FD were checked against

consistency with other recordings and reporting (FRA, statistical year book), the results
of which are presented in table 7.7. Unlike the INC, which considered all accumulated
plantation areas from 1963 up to 2000, a cut has been made and a zero baseline
effectively been established for the year 2000 for the present reference level of carbon
stock enhancement in this submission. The reasons for this are: (1) the difficulties to
assume the level of reliability of area data for the years before 1990s and (2) an
unknown number of older plantations have likely been disappeared and transformed
in other land uses according to expert judgements from FD. Therefore, it can be
assumed to be on the safe side with carbon stock calculations from enhancement if
only the more recent plantations are considered and the likelihood of overestimating
the carbon removal from forest plantations for the reference period 2005-2015
substantially be reduced.

44
7.2.2 Emission/ Removal factors

For the emission/ removal factors (EF/ REF) from enhancement by forest plantations
two aspects need to be considered:

 Biomass consumption during site preparation, and

 Net biomass increments per year and per ha on the plantation sites

For the biomass consumption from fire during site preparation, the mean default value
for tropical savannah grassland of 10 tonnes of dry matter biomass per ha is used
(table 3A1.13, page 3.181, IPCC, 2003 since this figure appears as more realistic than
the rather high value of 42.2 tonnes per ha used for similar calculations in the INC.

For the net biomass increment from plantations, the lower default value from table
4.10, IPCC Guideline, 2006 for tropical dry forests and tropical moist deciduous forests
with 7 tonnes biomass dry matter per ha has been used. This value is lower than the
one used in the INC where approximately 10 tonnes of dry biomass increment per ha
were applied, thus eventual loss, mortality and removal during the rotation period is
cautiously accounted for.

Although the calculations start from the year 1999-2000, only the accumulated carbon
stock between 2005 and 2015 is included in the reference level. For the details of the
calculations, please refer to the Annex 13.

7.2.3 Methodology of calculation

The actual C-removal in this reference level has been calculated with the equation15
as follows:

15
Based on equation 2.9 (Annex 2: Summary of Equations page A2.5 of 2006 IPCC guidelines for National
Greenhouse Gas Inventories) but modified and adapted.

45
 Equation 2: for CO2 removal calculation from historic enhancement efforts in Myanmar

t2
∑ (At x Bmic - St x BmCo) x CvfC x CvfCO2
t1
aRt1 – t2 =
t2 - t1
Where
aRt1 – t2 Average net removal of CO2 from forest plantations for the reference period
t1 First year of the reference period
t2 Last year of the reference period
At Accumulated area planted in year t in ha
St Area of site preparation in year t in ha (for plantation in year t+1)
Bmic Net biomass increments per ha and year (table 4.10, IPCC 2006; table 3A.6, IPCC 2003)
Bmco Biomass consumption during site preparation per ha and year (table 3A1.13, IPCC
2003)
CvfC Conversion factor biomass dry matter in C (0.47)
CvfCO2 Conversion factor C in CO2 (44/12)

7.2.4 Results on reference level from enhancement measures

For calculating the reference level from enhancement measures, the historical average
method is applied for the same reference period of deforestation. The accumulated
carbon stock from plantations in this period amounts to 33,513,321 tonnes of CO2e
with an annual mean over the reference period of 10 years of 3,351,332 tonnes of
CO2e as showed in the following table;

Table 7. 8: Accumulated CO2e removal from forest plantations

Year Accumulated CO2e removal in tonnes

1999-2000 -529,857
2000-2001 -159,613
2001-2002 200,717
2002-2003 596,270
2003-2004 937,362
2004-2005 1,302,264
2005-2006 1,785,250
2006-2007 2,159,285
2007-2008 2,434,561
2008-2009 2,739,229
2009-2010 3,233,790
2010-2011 3,573,172

46
 2011-2012 4,003,959
2012-2013 4,193,607
2013-2014 4,488,027
2014-2015 4,902,442
Sum 2005 - 2015 33,513,321
Average Annual CO2e removal 3,351,332

Figure 7. 7: Carbon stock enhancement from forest plantations 2000-2015 and reference
level 2005-2015 in tonnes of CO2e

8. Myanmar Initial FRL

Myanmar proposes an initial FRL by historical average of emissions during the

reference period from 2005 to 2015. The stratified area estimate of annual
deforestation is estimated with 428,984 ha per year during 2005-2015. Weighted
mean values of tCO2 eq per ha result in a national EF value of 125.43 tCO2 eq per ha.
Therefore, annual CO2 emission from deforestation during the historical reference
period 2005-2015 is estimated as 53,807,463 tonnes of CO2 eq per year with an
associated uncertainty of +/- 15.06%.

From sample-based estimation, it was not possible to develop a forest gain map for
the period 2005 to 2015 and therefore development of a reference level for
enhancement of forest carbon stocks based on existing forest areas is impossible.

47
 Efforts are ongoing to improve the data available and this will be added to future
submission in due course.

For carbon removal from historic enhancement efforts based on forest plantation
establishment and for the same historic reference period the proposed amount is:
3,351,332 tonnes of CO2e. This level would be used to measure additional carbon
removal as a result of the 10-year MRRP that is also part of the national REDD+ strategy
of Myanmar.

Table 8. 1: Summary on proposed FRL of Myanmar with reference period of 2005-2015

Annual area Average EF/REF in Total Upper limit Lower limit

of accumulated tonnes CO2e/ tonnes tonnes tonnes
deforestation afforestation ha CO2e CO2e CO2e
area in ha

Deforestation 428,984 125.43 53,807,463 63,243,891 44,371,035

Enhancements 308,636 10.86 3,351,332

9. Uncertainty assessment of emission estimates

The uncertainty assessment for activity data and emission factor (deforestation) was
conducted at tier 1 level according to the IPCC Good Practice Guidance and
Uncertainty Management in National Greenhouse Gas Inventories (IPCC, 2000, 2006).
The calculations include only uncertainty derived from sampling and do not include
other error sources (e.g., errors from allometric equation, human errors or others).

Uncertainty estimation for activity data and removal factors from enhancement
cannot be conducted since the data are not derived from sampling, but from annual
records of the FD of planted areas from subnational offices, which then are aggregated
at national level.

In calculation of the % uncertainty of activity data (deforestation estimate), the

following equation was applied,

48
Equation 316
1
(95% 𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 𝑊𝑖𝑑𝑡ℎ)
% 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 = 2 𝑥 100
𝜇

Where µ=mean of the distribution

As described in the equation, the 95% Confidence Interval for all 40 districts are
calculated with the equations of:

Equation 4
𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
95% 𝑈𝑝𝑝𝑒𝑟 𝐿𝑖𝑚𝑖𝑡 𝑜𝑓 𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 = 𝑀𝑒𝑎𝑛 𝐶𝑂2 𝑇𝑜𝑛 𝑃𝑒𝑟 𝐻𝑎 + 1.96 ∗ ( )
√𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 𝐼𝑛 𝐻𝑎

𝑆𝑡𝑎𝑛𝑑𝑎𝑟𝑑 𝐷𝑒𝑣𝑖𝑎𝑡𝑖𝑜𝑛
95% 𝐿𝑜𝑤𝑒𝑟 𝐿𝑖𝑚𝑖𝑡 𝑜𝑓 𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 = 𝑀𝑒𝑎𝑛 𝐶𝑂2 𝑇𝑜𝑛 𝑃𝑒𝑟 𝐻𝑎 − 1.96 ∗ ( )
√𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 𝐼𝑛 𝐻𝑎

95% 𝐶𝑜𝑛𝑓𝑖𝑑𝑒𝑛𝑐𝑒 𝐼𝑛𝑡𝑒𝑟𝑣𝑎𝑙 𝑊𝑖𝑑𝑡ℎ = 95% 𝑈𝐶𝐼 − 95% 𝐿𝐶𝐼

The uncertainty of the overall estimates for AD and EF was calculated by error
propagation with the following equation,
Equation 5

√(𝑈1 ∗ 𝑥1 ) 2 + (𝑈2 ∗ 𝑥2 ) 2 + ⋯ + (𝑈𝑛 ∗ 𝑥𝑛 ) 2

𝑈 𝑡𝑜𝑡𝑎𝑙 17 =
𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛

𝑈 𝑡𝑜𝑡𝑎𝑙 = the percentage uncertainty in the sum of the quantities

𝑥𝑛 𝑎𝑛𝑑 𝑈𝑛 are the uncertain quantities and the percentage uncertainties associated with
them, respectively.

Table 9.1 shows the % uncertainty of AD, forest loss area, i.e. 8.97% for this submission
whereas the % uncertainty of Emission factor from 40 district forest-inventories is
13.03% (detail calculation in Annex 14).

16
Box 5.2.1: Chapter 5 of IPCC Good Practice Guidance for LULUCF
17
Table 6.1: Tier 1 Uncertainty Calculation and Reporting under IPCC Good Practice Guidance and Uncertainty
Management in National Greenhouse Gas Inventory

49
 Table 9. 1: Uncertainty Result for Activity Data in %

Area in ha (estimate) Standard Error (ha) Uncertainty %

Forest 40,446,950 789,576 1.92
Forest loss (deforested area) 4,289,839 392,682 8.97
Non-forest 22,920,963 751,052 3.21
Total area 67,657,752

Combined uncertainty was finally estimated by using the uncertainty of AD, i.e.,
uncertainty % of forest loss (deforestation) and the uncertainty of EF estimates as
follows:

Equation 6

𝐶𝑜𝑚𝑏𝑖𝑛𝑒 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 (%) = √𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 % 𝑜𝑓 𝐴𝐷2 + 𝑈𝑛𝑐𝑒𝑟𝑡𝑎𝑖𝑛𝑡𝑦 𝑜𝑓 % 𝐸𝐹 2

Regarding the above calculation, uncertainty of AD, deforestation area estimation and
EF were 8.97 % and 12.10% respectively and therefore, combined uncertainty % of
15.06 % is estimated for this submission.

10. Future improvement opportunities

10.1. For Activity Data

In absence of a reliable baseline map of forest gain during the reference period 2005-
2015 as a gain stratum, much uncertainty exists in identification of forest gain
(enhancement) classes while using remote sensing technologies. Only based on
remote sensing, there are difficulties in distinguishing between afforestation and
growing cycles of the forest plantations, and also difficulties in identifying the
ecological pattern of forest regrowth following deforestation. Indeed, seasonality
(leaf phenology) and soil moisture variations may have played a role in the other
misclassifications such as dry forest types or teak plantation. Therefore, forest gain
has not been considered for sample-based estimates and has eventually been
identified as one of the areas of future improvement for the FRL.

50
There is a need to develop a standard operating procedure to detect land cover change
under the six IPCC land cover classes through a combination of remote sensing-based
and ground-based information to provide a robust estimate of carbon emissions and
removals.

The existing land cover maps do not allow precise estimates of forest cover change,
either loss or gain. Improvement of area estimated through generation of national
maps is a priority. National maps will be developed with a standard mapping
procedure and so is expected to reduce uncertainty in next reporting.

Further developments may include analysis of forest degradation with a particular

attention to the definition in the national context (e.g. the types of plantations which
can be classified as forest). The coming NFI will include other woodland and areas with
trees outside forests. Therefore, data sets concerning with values for areas which have
lost forest cover in the past, will be developed in the future.

Specific activities which are either planned or ongoing for improvement of AD include
the following:

- The local technical team is trying to standardize the image interpretation

methodology in order to reduce the effect of interpreter and human bias in
image interpretation.
- In order to set-up the long-term assessment, FAO Open Foris: Collect Earth
System has been introduced and sample plots have been set up throughout the
country.
- Since the available datasets are almost from satellite images, perennial crops
and home gardens cannot be clearly distinguished from forest cover. The future
plan is to collect the attributes of recorded land areas for perennial crops/home
garden and create a spatial database through which they can be reliably
identified against satellite imagery.

51
 - Detailed land use maps are not currently available. Ongoing land use
assessment using RapidEye imagery which is carried out under the guidelines
of National Land Use Policy 2016, will assist in the development of these maps.
- AD calculations will be refined based on administration boundaries, i.e. states
and on regions.
- FD is also now recording spatial data of plantations. Plantation areas will also
be included in the future NFI. The NFI results of which may make it easier to
subdivide plantation areas in age classes.
- Stepwise approach should be applied to improve the current FRL over time
through incorporating more REDD+ activities, better quality data, higher tier
level in terms of methodologies and inclusion of additional pools.
- Although this submission is at national level, strengthening of land and forest
monitoring and measurement capacities under various projects, like the
National Forest Inventory/ National Forest Monitoring and Information System
(NFI/ NFMIS) and OneMap Myanmar, future FRL submissions may be divided
into sub-national levels based on the available improved datasets.
- Spatially explicit area estimates for plantation data is also an area of future
improvement. Plantations will be included in the upcoming National Forest
Inventory as a distinct stratum for which data and results be produced.
The REDD+ MRV team will set-up better coordination between FD and ECD to
avoid inconsistencies of data sources for the AD between the FRL and the GHG
inventory included in Myanmar’s INC. FD is planning to achieve consistency in
data and methods between the FRL and the national GHG inventory to be
included in the future NC/BUR.

10.2. For Emission Factors

FD conducted district forest inventories every year in available districts and many
forest parameters are available. On the other hand, there is no database management
system or standardization of parameter coding system. The forthcoming NFI/NFMIS

52
project will focus on the national forest monitoring and information system and will
provide more qualified data and information to inform future FRL submissions.

The NFI/NFMIS project will also improve accuracy of geo-location of the sample plots
and integrated application of remote sensing data/ satellite and forest inventory data
for effective estimation of forest resources.

Although Myanmar has partially collected soil data, the confidence of using those data
is still limited. But in the future, the national data on soil organic carbon will be possible
to be collected with the support of the Finland Forest Research Institute (LUKE) and
the NFI/NFMIS project. The future NFI will include SOC measurements in uplands and
in also Mangroves. Special attention will be given to SOC in Mangroves as soil organic
carbon pools in soil sediments of Mangroves are very important and currently data are
not available.

Removal factor calculation for enhancement measures can be improved once results
from further cycles of the NFI are available18. In the meantime, default values for
biomass increment and biomass consumption from site preparation need to be
applied.

10.3 Uncertainty assessment

For future uncertainty assessment, the possibility of moving to tier 2 assessment will
be evaluated applying an analysis based on the Monte Carlo approach once the NFI
data are available.

18
From the second cycle onwards, when first results from permanent sample plots in plantations allow for
stock difference measurements.

53
References:
Chan, N., Takeda, S., Suzuki, R. and Yamamoto, S. (2013): Establishment of allometric
models and estimation of biomass recovery of swidden cultivation fallows in
mixed deciduous forests of the Bago Mountains, Myanmar, Forest Ecology and
Management 304 (2013) 427-436
Chave, J., Andalo, C., Brown, S., Cairns, M. A., Chambers, J. Q., Eamus, D., Folster, H.,
Fromard, F., Higuchi, N., Kira, T., Lescure, J.-P., Nelson, B. W., Ogawa, H., Puig,
H., Riera, B. and Yamakura, T. (2005). Tree allometry and improved estimation
of carbon stocks and balance in tropical forests, validity of DBH range 5-156 cm,
Oecologia 145: 87-99
Cochran, W.G. (1967). Sampling Techniques. John Wiley and Sons Ltd. 413 pp.
Cochran, W.G. (1977). Sampling Techniques, Third Edition, New York: John Wiley &
Sons
Davis, J. H. (1960). The Forests of Burma, University of Mandalay and University of
Florida, http://www.burmalibrary.org/docs13/The-Forest-of-Burma-ocr2.pdf
FAO (2005, 2010, 2015). Global Forest Resource Assessment – Country reports
Myanmar. FAO, Rome. www.fao.org/forest-resouces-assessment/en/
FD and FAO, (2016). Data Book with the Results of the Project, “Strengthening
Myanmar’s National Forest Monitoring System- Land Use Assessment and
Capacity Building” (TCP/MYA/3501), December 2016
Foody, Giles M. (2010). Assessing the accuracy of land cover change with imperfect
ground reference data, Remote Sensing of Environment, 114, 2271-2285 (2010)
Gallego, F.J. (2004). Remote sensing and land cover area estimation. International
Journal of Remote Sensing, 25(15), 3019–3047
GFOI (2016). Methods and Guidance Document v2, Chapter 5.1.5 Estimating
uncertainty of area and change in area
GOFC-GOLD, (2016). A sourcebook of methods and procedures for monitoring and
reporting anthropogenic greenhouse gas emissions and removals associated
with deforestation, gains and losses of carbon stocks in forests remaining

54
 forests, and forestation. GOFC-GOLD Report version COP22-1, (GOFC-GOLD
Land Cover Project Office, Wageningen University, The Netherlands).
Hansen, M.C., P.V. Potapov, R. Moore, M. Hancher, S. A. Turubanova, A. Tyukavina,
D. Thau, S. V. Stehman, S. J. Goetz, T. R. Loveland, A. Kommareddy, A. Egorov,
L. Chini, C. O. Justice, and J. R. G. Townshend. (2013) “High resolution global
maps of 21st century forest cover change.” Science 342 (15 November): 850-53;
http://earthengine-partners.appspot. com/science-2013-global-forest
Hilden, M., Maekinen, K., Jantunen, J., Jokinen, M., Lilja, R., Than, M. M., Rantala, S.
and Aung, T. (2016). Needs assessment for the effective implementation of the
Environmental Conservation Law in Myanmar. UNDP, Yangon. 124 pp.
http://www.mm.undp.org/content/dam/myanmar/docs/Publications/EnvEngy
/UNDP_MM_Needs_Assessment_Environmental_Conservation_Law_web.pdf
IPCC (2000). Good practice guidance and uncertainty management in national
greenhouse gas inventories. http://www.ipcc-nggip.iges.or.jp/public/
gp/english/index.html
IPCC (2003). Good practice guidance for land use, land use change and forestry.
http://www.ipcc-nggip.iges.or.jp/public/gpglulucf/gpglulucf_contents.html
IPCC (2006). Guidelines for national Greenhouse Gas Inventories. Volume 4 :
Agriculture, Forestry and Other Land Use. http://www.ipcc-nggip.iges.or.jp/
public/2006gl/ vol4.html
Kress, W. J., DeFilipps, R. A., Ellen, F. and Kyi, Y. Y. (2003): A Checklist of the Trees,
Shrubs, Herbs, and Climbers of Myanmar
MOECAF (2012). Myanmar's first national communication under the United Nations
Framework Convention on Climate. MMR/COM/1 B. Nay Pyi Taw, 268pp. (url:
http://unfccc.int/resource/docs/natc/mmrnc1.pdf, date of access: 12
December 2017)
MOECAF (2013). Myanmar REDD+ Readiness Roadmap. Nay Pyi Taw. 148 pp.
MOECAF (2015a). Development of a National Forest Monitoring System for Myanmar.
Forest Department – UN-REDD Programme, Nay Pyi Taw, 84 pp.

55
MOECAF (2015b). Forest Reference (Emissions) Level action plan for Myanmar. Forest
Department – UN-REDD Programme, Nay Pyi Taw, 39 pp.
Mon, Myat Su, (2017), Evidence-based Land Use Planning Process: Piloting in Bago
Region, Myanmar, Paper presented at “2017 World Bank Conference on Land
and Poverty”, The World Bank - Washington DC, March 20-24, 2017
MONREC (2017). Country Statement on Climate Change COP 23, http:// unfccc.int/
files/meetings/bonn_nov_2017/statements/application/pdf/myanmar_cop23c
mp13cma1-2_hls.pdf
MONREC-ECD (2016). Myanmar Climate Change Strategy and Action Plan, MCCSAP,
Final Draft 2016 – 2030. Nay Pyi Taw, 180 pp.
MONREC-FD (2016). National Reforestation and Rehabilitation Programme of
Myanmar, MRRP, MONREC, Nay Pyi Taw.
MONREC-FD (2011, 2016). Myanmar’s Biodiversity Strategy and Action Plan.
MOECAF/ MONREC, Nay Pyi Taw. https://www.cbd.int/nbsap/ about/latest/
default. shtml #mm
Olofsson, P., Foody, G. M., Stehman, S. V. and Woodcock, E. C. (2013). Making
better use of accuracy data in land change studies: Estimating accuracy and
area and quantifying uncertainty using stratified estimation, Remote Sensing of
Environment 129, 122-131
Olofsson, P., Foody, G.M., Herold, M., Stehman, S. V., Woodcock, C. E. and Wulder,
M. A. (2014). Good practices for estimating area and assessing accuracy of land
change, Remote Sensing of Environment 148, 42-57:
Sarndal, C. E., Swensson, B. and Wretman, J. (1992). Model Assisted Survey Sampling.
Springer, NY.
Stehman, S. (2005). Comparing estimators of gross change derived from complete
coverage mapping versus statistical sampling of remotely sensed data. Remote
Sensing of Environment, 96, 466–474.
UN-REDD Programme. Myanmar REDD+ Strategy Draft (2018). http://www.
myanmar-redd.org/

56
Annex
Annex 1: Dominant Types of Forests in Myanmar

(a) Sources from Departmental Instructions

(b) A Checklist of the Trees, Shrubs, Herbs, and Climbers of Myanmar:
contributions from the United States National Herbarium: Volume 45:1-
590
(a) (b) Distribution of Veg

Distribution of Vegetation Types in Myanmar.

Source: A Checklist of the Trees, Shrubs, Herbs, and Climbers of

Myanmar. Contributions from the United States National
Herbarium. Volume 45: 1-590

57
Annex 2: Comparison of forest cover area and its percentage within and outside PFE (based
on 2015 National Forest Cover Map)

Inside PFE Forest Outside PFE Forest

Land Cover All Areas (ha)
Areas (ha) Areas (ha)
Closed Forest 12160128.74 5229098.53 6931030.21

Open Forest 16914653.10 6570101.26 10344551.84

Mangrove 486935.54 148968.45 337967.09

Other wooded land 19528356.46 5695377.45 13832979.00

Other lands 16737085.37 1579981.09 15157104.29

Water Body 1694333.67 209363.00 1484970.67

Snow 127075.42 101890.75 25184.68

Total Area 68248983.33 19534780.52 48714202.81

% of country area 100.00 28.62 71.38

Total Forest Area 29561717.38 11948168.24 17613549.15

% of forest cover
44.13 16.69 27.44
based on country area

58
Annex 2 (a):

59
Annex 2 (b)

60
Annex 3: Datasets used to generate activity data for establishing a forest reference emission
level in Myanmar during 2005-2015

no Data type Format Note

1 Wall-to-wall maps:
Land cover map from Forest
Department
Land cover map 2005 (30 m-Landsat- Raster The maps were generated using
based) supervised maximum likelihood classifier
Land cover map 2015 (23.5 m-IRS- Raster into seven land cover classes which
based) include: Close forest, Open Forest, Other
Land cover map 2015 (30-m-Landsat- Raster wooded land, Mangrove, water, snow,
based) and others.
2 Sample-based estimates
2.1 Global Forest Change (Stratifier)
Tree canopy cover for year 2000 Raster Tree cover in the year 2000, defined as
(treecover2000) canopy closer for all vegetation taller
than 5m in height. Encoded as a
percentage per output grid cell, in the
range of 0-100.
Year of forest cover loss event 2001- Raster Forest loss during the period 2000-2015
20015 (loss year) defined as a stand-replacement
disturbance, or a change from a forest to
non-forest state. Encoded as either 0 (no
loss) or else a value in the range of 1-15,
representing loss detected primarily in
the year 2001-2015, respectively.
2.2 1,884 stratified random samples Vector The samples were generated following
and csv stratified random sampling design using
stratified forest change map of 2005-
2015, derived from Global Forest Change
Maps. The samples were checked
against high-quality reference data.

61
Annex 4: Harmonizing national land use categories with FRA and IPCC land use categories

Land use code National land IPCC Forest Resources Assessment

use/cover categories
1 Closed forest Forest land Forest (includes primary forest, other naturally
8 Mangrove regenerated forest, and planted forest)
2 Open forest
3 Other wooded land Other wooded land
11 Grassland Grassland
4 Cropland Cropland Other land with tree cover
5 Other lands Other lands Other land
9 Snow
6 Settlements Settlement
7 Wetland Wetland
10 Water Inland water bodies

62
Annex 5: Wall-to-wall thematic raster maps of Myanmar under seven national land
use/cover categories at three years: 2005, 2010 and 2015

63
Annex 6: Maps showing the positioning error in the wall-to-wall maps of Myanmar
boundary

64
 Annex 7: A brief description of the simple and stratified random sampling estimators

Stratified estimators

Stratified estimators of the mean (𝜇̂ 𝑆𝑇𝑅 ) and the variance of the estimate of the
mean (𝑉𝑎̂𝑟 (𝜇̂ 𝑆𝑇𝑅 )) are provided by Cochran (1977) as,
𝐻

𝜇̂ 𝑆𝑇𝑅 = ∑ 𝑤ℎ 𝜇̂ ℎ -----------(𝐸𝑞. 3)
ℎ=1

and
2
𝑎̂ℎ
𝑉𝑎̂𝑟 (𝜇̂ 𝑆𝑇𝑅 ) = ∑𝐻 2
ℎ=1 𝑤ℎ -----------(𝐸𝑞. 2)
𝑛ℎ

Where
𝑛ℎ
1
𝜇̂ ℎ = ∑ 𝑦ℎ𝑖 ------------(𝐸𝑞. 4)
𝑛ℎ
𝑖=1

And
𝑛ℎ
1
𝜎̂ℎ2 = ∑(𝑦ℎ𝑖 − 𝜇̂ ℎ )2 -------------(𝐸𝑞. 5)
𝑛ℎ − 1
𝑖=1

ℎ = 1, … … , 𝐻denotes strata;

𝑦ℎ𝑖 is the 𝑖 𝑡ℎ sample observation in the ℎ𝑡ℎ stratum;

𝑤ℎ is the weight for the ℎ𝑡ℎ stratum;

𝑛ℎ is the number of plots assigned to the ℎ𝑡ℎ stratum; and

𝜇̂ ℎ and𝜎̂ℎ2 are the sample estimates of the within-strata means and variance,
respectively.

Using the notation of Eq. 1 (see section 7.1.2), and adding the subscript j to indicate
reference class j,

65
 𝑛ℎ
1
𝜇̂ ℎ𝑗 = ∑ 𝑦ℎ𝑗𝑖 ------------ (𝐸𝑞. 6)
𝑛ℎ
𝑖=1

But because

1 𝑖𝑓 ℎ = 𝑗
𝑦ℎ𝑗𝑖 = { ------------- (𝐸𝑞. 7)
0 𝑖𝑓 ℎ ≠ 𝑗

Equation (1) can be expressed as,

𝑛ℎ𝑗
𝜇̂ ℎ𝑗 = -------------- (𝐸𝑞. 8)
𝑛ℎ

So that from Eq. 3

𝐻 𝐻 𝐻
𝑛ℎ𝑗
𝜇̂ 𝑗 = ∑ 𝑤ℎ . 𝜇̂ ℎ𝑗 = ∑ 𝑤ℎ . = ∑ 𝑃̂ℎ𝑗 ----------------(𝐸𝑞. 9)
𝑛ℎ
ℎ=1 ℎ=1 ℎ=1

The area for reference class j is estimated as the product of 𝜇̂ 𝑗 and the total
area(𝐴𝑡𝑜𝑡 ).For example the estimated deforestation 𝐴̂1 =
𝑃𝑟𝑜𝑑𝑢𝑐𝑒𝑟 ′ 𝑠 𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦 𝑜𝑓 𝑑𝑒𝑓𝑜𝑟𝑒𝑠𝑡𝑎𝑡𝑖𝑜𝑛 𝑐𝑙𝑎𝑠𝑠 ∗ 𝐴𝑡𝑜𝑡

Confidence Interval

Using the notation of Eq. 5 and again adding the subscript to denote reference class
j,
𝑛ℎ
2
1 2
𝜎̂ℎ𝑗 = ∑(𝑦ℎ𝑖𝑗 − 𝜇̂ ℎ𝑗 ) -------------(𝐸𝑞. 10)
𝑛ℎ𝑗 − 1
𝑖=1

Noting from Eq. 7, the Eq. 10 can be expressed as,

𝑛ℎ
1
𝜎̂ℎ2 = ∑ 𝜇̂ ℎ𝑗 . (1 − 𝜇̂ ℎ𝑗 ) -------------(𝐸𝑞. 11)
𝑛ℎ − 1
𝑖=1

So that from Eq. 2

66
 𝐻 𝐻
𝑎̂ℎ2 𝜇̂ ℎ𝑗 . (1 − 𝜇̂ ℎ𝑗 )
𝑉𝑎̂𝑟 (𝜇̂ 𝑗 ) = ∑ 𝑤ℎ2 = ∑ 𝑤ℎ2 .
𝑛ℎ 𝑛ℎ𝑗 − 1
ℎ=1 ℎ=1
𝐻 2
𝑤ℎ . 𝑝̂ℎ𝑗 − 𝑝̂ℎ𝑗
= ∑ -----------(𝐸𝑞. 12)
𝑛ℎ𝑗 − 1
ℎ=1

And standard error,

𝑆𝐸(𝜇̂ 𝑗 ) = √𝑉𝑎̂𝑟 (𝜇̂ 𝑗 )------------(𝐸𝑞. 13)

From Eq. 13 so that the standard error of the estimated area of forest loss is

𝑆𝐸(𝐴̂1 ) = 𝑆𝐸 (𝜇̂ 1 ) ∗ 𝐴𝑡𝑜𝑡

At 95% confidence interval of the estimates area of forest loss is ±1.96 ∗ 𝑆𝐸(𝐴̂1 )

Strata (j) 𝜇̂ 𝑗 𝑆𝐸(𝜇̂ 𝑗 ) 𝜇̂ 𝑗 (ha) 95% Confidence interval

(proportion) (proportion) Lower (ha) Upper (ha)

Stable Forest

Stable Non-Forest

Loss

67
 Annex 8: Examples of sample assessment using validation datasets:

(i) Forest 2005 > Forest 2015; (ii) Non-Forest 2005 > Non-Forest 2015; (iii) Forest 2005 > Non-
Forest 2015 (sample # 408). 30-m spatial resolution Landsat False Color Composite
(NIR_SWIR1_R) and high spatial resolution Google Earth natural color (RGB) imagery were
used as validation datasets.

(i) Forest 2005 > Forest 2015

2005 Landsat 5 TM 2005 Google Earth Imagery

2015 Landsat 7 ETM+ 2015 Google Earth Imagery

(ii) Non-Forest 2005 > Non-Forest 2015

2005 Landsat 5 TM 2005 Google Earth RGB Imagery

68
2015 Landsat 7 ETM+ 2015 Google Earth Imagery

(ii) Forest 2005 > Non-Forest 2015 Time-series imagery from Google Earth

2005 2014

2013 2017

69
 Annex 9: Sample Plot Design for 40 Districts from Forest Management Inventory mentioning the conversion factor to per ha

Sr. State/ Sub-

No. Region District Year Sample Design Interval Plots Shape Size Measurement
3000 yard/ L-shape:
1 Kachin State Myitkyina 2004-2007 Systematic 2743.2 m 7 IL 7 RU 1.05 ha DBH >= 20 cm
Bhamo IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
Sagaing
2 Region
Upper Tamu, Mawlaik 100m x
Chindwin and Kalay 2014 Systematic 2000 m 3 Square 100m DBH >= 20 cm
Khamti 2015 Square 50m x 50m DBH bt 5 cm to 19 cm
Lower Shwebo and < 5 cm, Regen &
Chindwin Monywa 2015 Square 25m x 25m Bamboo
Katha 2015
3 Bago Region Tharyarwaddy 2011 Systematic 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
100m x
Bago 2012 Systematic 2000 m 3 Square 100m DBH >= 20 cm
Square 50m x 50m DBH bt 5 cm to 19 cm
< 5 cm, Regen &
Square 25m x 25m Bamboo
Taungoo 2010 Systematic 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
100m x
Pyay 2017 Systematic 2000 m 3 Square 100m DBH >= 20 cm
50m x 50m DBH bt 5 cm to 19 cm
< 5 cm, Regen &
25m x 25m Bamboo
100m x
Magwe Minbu 2013 Systematic 2000 m 3 Square 100m DBH >= 20 cm
4 Region Thayet 2013 Square 50m x 50m DBH bt 5 cm to 19 cm
< 5 cm, Regen &
Gangaw 2013 Square 25m x 25m Bamboo
70
 Magwe 2013
Pakkoku 2013
3000 yard/ L-shape:
5 Nay Pyi Taw Ottarathiri 2008 Systematic 2743.2 m 7 IL 7 RU 1.05 ha DBH >= 20 cm
Dakinathiri 2008 IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
Taungoo 2008 IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
Taunggyi
3000 yard/ L-shape:
6 Shan State Kyaukme 2007 Systematic 2743.2 m 7 IL 7 RU 1.05 ha DBH >= 20 cm
Linkhay 2007 Systematic IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
Taunggyi North
(Yasauk and
Ywangan) 2007 IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
Taunggyi South
(Kalaw,
Naungshwe,
Phekon, Pinlaung 100m x
and Taunggyi) 2011 Systematic 2000 m 3 Square 100m DBH >= 20 cm
Square 50m x 50m DBH bt 5 cm to 19 cm
< 5 cm, Regen &
Square 25m x 25m Bamboo
7 Rakhine State 3000 yard/ L-shape:
Sittwe 2005 Systematic 2743.2 m 7 IL 7 RU 1.05 ha DBH >= 20 cm
IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
Maungdaw 2005 Systematic 3000 yard/ 7 L-shape:
2743.2 m IL 7 RU 1.05 ha DBH >= 20 cm
IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
Thandwe 2015 One Shot 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
Kyaukphyu 2015 One Shot 2000 m 1 Square 1 ac plot all DBH classes (>=5cm)

71
 3000 yard/ L-shape:
MyaukOo 2005 Systematic 2743.2 m 7 IL 7 RU 1.05 ha DBH >= 20 cm
IL RU 1, 4, 7 15m radius DBH bt 10 cm to 19 cm
IL RU 1, 4, 7 10m radius DBH bt 5 cm to 9 cm
8 Tanintharyi Dawei 2015 One Shot 2000 m 1 Square 1 ac plot all DBH classes (>=5cm)
9 Chin Falam 2015 One Shot 2000 m 1 Square 1 ac plot all DBH classes (>=5cm)
10 Ayeyarwaddy Myaungmya 2015 One Shot 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
Pyarpon 2015 One Shot 2000 m 1 Square 1 ac plot all DBH classes (>=5cm)
Pathein 2015 One Shot 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
Hinthada 2015 One Shot 2000 m 1 Circular 50m radius all DBH classes (>=5cm)
100m x
11 Mandalay KyaukSe 2015-2016 Systematic 2000 m 3 Square 100m DBH >= 20 cm
Mandalay 50m x 50m DBH bt 5 cm to 19 cm
< 5 cm, Regen &
Meiktila 25m x 25m Bamboo
Pyin Oo Lwin
Yamethin

Summary on Plot Design

Plot Design Plot Area Plot Dimensions

Systematic (Square)  1 ha  100mx100m
 0.25 ha  50mx50m
One Shot (Circular)  0.7854 ha  50m radius
One Shot (Square)
 0.4047 ha
 1 acre (0.4047 ha)
Systematic (L-shaped)  1.05 ha  30 m x 50 m for each of the
seven sub-plots)
 0.0707 ha  50 m radius for each of
three sub-plots

72
Annex 10: Plot Designs and Description

The sample plots are of circular, square or rectangular shape. The strip is a special
shape of rectangular which is particularly used in forested areas that are not easily
accessible.

A sample tree is considered to fall inside a plot of given boundaries, if the center of
the bole at the base of the tree falls inside the plot. Consequently, each sample plot
contains edge trees with a growing space which is partly located outside the plot
boundaries. One of the important non-sampling errors in forest inventories is the
incorrectly omitting or including such edge trees.

Circular sample plots are often preferred to other plot shapes because they have the
smallest perimeter for a given area. Circular plots, therefore, tends to produce less
borderline trees than other plot shapes for the same plot size. A further advantage of
circular plots is that they are less time consuming to establish than square or
rectangular plots. In stands without undergrowth, the plot boundaries can be
conveniently located with the aid of optical devices.

In many tropical forests, however, it is more convenient to lay out square or

rectangular plots or strip-shaped sampling units. Due to the adverse environmental
conditions and the need for a larger plot size, utilizing circular sample units creates
difficulties in tropical forest. Square or rectangular shape can be used in inventories of
forest plantations where the trees were planted in rows. In this case, boundaries of
sample units should be established in the middle of two rows.
50 m

C 25 m

50m B

100 m
A

100 m
Circle shape Design Square shape Design

73
All DBH classes are collected in 50 m radius Circular plots whereas, the trees with DBH
200 mm and above are collected within square A area (100m x 100m: 1 Ha), trees
within 50 mm and 199 mm collected within square B area (50m x 50m: 0.25 Ha) and
the regeneration and bamboo are collected in square C area (25m x 25m: 0.0625Ha).

L-shaped design

The Sampling unit is composed by a strip of 15m wide to the left and right of a center
line which runs 175 meters in East-West and North-South Direction. It is distributed
systematically in a grid of 3 km x 3 km in the forest area with a sampling intensity 0.11
percent. The strip has an Inverted L-shape and is divided into seven units of size 30m
x 50m equal to 1.05 ha as shown in figure.

In the three special sample plots (15 meters radius circular plots numbered 1, 4 and 7
as in the figure, trees having diameters of 10 cm and above are enumerated. In the
special circular plots, the enumerated trees are also labelled with aluminum tags and
their position is recorded. (Source: Brief on National Forest Inventory, NFI, Forest
Resources Development Service, Rome, June 2007)

74
 Annex 11: Summary of calculation steps for three carbon pools and National Emission Factor value for Myanmar

District Three Carbon Pools (AGB, BGB and Litter)

Management
A B C D E F G
Inventory Data
DBH in cm Above Ground Below Ground AGB and BGB Litter Emission Emission CO2 for sample
(National Data) Biomass Biomass (Tonnes Per (C Tonnes Per (C Tonnes Per C Tonnes Per CO2 Tonnes plots
(Kg Per Ha) Ha) Ha) Ha) Ha Per Ha
>= 20 cm EXP ((-2.289+2.649 x Mean AGB Tonnes Per (AGB+BGB) x 0.47 Table 2.2: Tier 1 C+D E x 3.67 G x sample
10 cm to 19 cm LN (DBH) - 0.021 x (LN Ha x R (Ratio of below default values for number of each
(DBH)) ^2)19 ground biomass to Litter: 2006 IPCC forest types of
Tree data base, above ground Guidelines each district
summarize Per Plot biomass)
to get Per Ha (in each forest type)

The above step by step approach is used to get the total CO2 Tonnes Per Ha for each district.
For the National Emission Factor value, the following steps are taken:
𝑆𝑎𝑚𝑝𝑙𝑒 𝑆𝑖𝑧𝑒 𝑖𝑛 ℎ𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = 𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑙𝑜𝑡𝑠 𝑥 𝑝𝑙𝑜𝑡 𝑠𝑖𝑧𝑒 … … … … … … … … … … … … . . (1)
𝑇𝑜𝑡𝑎𝑙 𝐶𝑂2 𝑇𝑜𝑛 𝑃𝑒𝑟 𝐻𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … … (2)
𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (2)
𝑊𝑒𝑖𝑔ℎ𝑡𝑒𝑑 𝐶𝑂2 𝑚𝑒𝑎𝑛 𝑣𝑎𝑙𝑢𝑒 𝑝𝑒𝑟 ℎ𝑎 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = … … … … … … … … … … . (3)
𝑡𝑜𝑡𝑎𝑙 𝑛𝑢𝑚𝑏𝑒𝑟 𝑜𝑓 𝑠𝑎𝑚𝑝𝑙𝑒 𝑝𝑙𝑜𝑡𝑠
𝑂𝑣𝑒𝑟𝑎𝑙𝑙 𝑤𝑒𝑖𝑔ℎ𝑡𝑖𝑛𝑔 𝑣𝑎𝑙𝑢𝑒 𝑓𝑜𝑟 𝑒𝑎𝑐ℎ 𝑑𝑖𝑠𝑡𝑟𝑖𝑐𝑡𝑠 = 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (3) 𝑥 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (1) … … … … … … … … … … … … … … … (4)

𝒕𝒐𝒕𝒂𝒍 𝒐𝒗𝒆𝒓𝒂𝒍𝒍 𝒘𝒆𝒊𝒈𝒉𝒕𝒊𝒏𝒈 𝒗𝒂𝒍𝒖𝒆

𝑾𝒆𝒊𝒈𝒉𝒕𝒆𝒅 𝑴𝒆𝒂𝒏 (𝑵𝒂𝒕𝒊𝒐𝒏𝒂𝒍 𝑬𝒎𝒊𝒔𝒔𝒊𝒐𝒏 𝑭𝒂𝒄𝒕𝒐𝒓) =
𝒕𝒐𝒕𝒂𝒍 𝒔𝒂𝒎𝒑𝒍𝒆 𝒔𝒊𝒛𝒆

19
Table 4. A. 1; Allometric Equations for Estimating Above Ground Biomass of Tropical and Temperate Hardwood and Pine Species from IPCC Good Practice Guidance for LULUCF

75
 Annex 12: Description of the forest types and respective value of R (Ratio of BGB to AGB) and Litter Range for each district

Forest Type according to NFI Litter Range

Forest Type for
No. Field Instruction 1985 of Districts R Value Range (Tonnes C per
BGB Calculation
Myanmar ha)

Mangrove, typical Kyaukphyu

1 Mangrove, high (kanazo Tropical Rain Forest 0.37 2.1
Pyarpon
forest)

Bago, Dawei, Bhamo, Katha,

Shwebo, Minbu, Thayet,
Swamp forest Myaungmya, Kalay, Mawlaik,
Taunggyi South, Kyaukme, Pathein,
3 Maungdaw, MyaukOo, Pyin Oo Lwin Tropical Rain Forest 0.37 2.1

Bago, Myaungmya, Thandwe, Kalay,

Evergreen forest, riverine Khamti, Mawlaik, Tamu, Taunggyi
South
Bhamo, Myitkyina, Katha, Thayet,
Myaungmya, Pyay, Thandwe,
Evergreen forest, typical
Khamti, Kyaukme, Pathein,
Kyaukphyu
Bago, Bhamo, Myitkyina, Katha,
4 Evergreen forest, giant Minbu, Thayet, Myaungmya, Kalay, Tropical Rain Forest 0.37 2.1
Khamti, Mawlaik, Tamu, Pathein
Bago, Katha, Minbu, Thayet, Pyay,
Bamboo forest (degraded rain Tharyarwaddy, Kalay, Khamti,
forests) Mawlaik, Tamu, Pathein, Pyin Oo
Lwin

76
 Bago, Hinthada, Bhamo, Katha,
Shwebo, Monywa, Magwe, Gangaw,
Minbu, Thayet, Dakinathiri,
Ottarathiri, Pyay, Taungoo,
Mixed deciduous forest, lower Tharyarwaddy, Kalay, Khamti,
Mawlaik, Tamu, Linkhay, Taunggyi
North, Taunggyi South, Kyaukme,
Maungdaw, MyaukOo, KyaukSe, AGB< 125
Pyin Oo Lwin, Yamethin Tonnes/ Ha =
0.20 (0.09-
Tropical Moist
5 Bago, Hinthada, Bhamo, Myitkyina, 0.25) AGB> 125 2.1
Deciduous Forest
Katha, Shwebo, Monywa, Gangaw, Tonnes/Ha =
Minbu, Pakkoku, Thayet, Dakinathiri, 0.24 (0.22-
Ottarathiri, Taungoo: Nay Pyi Taw, 0.33)
Taunggyi: Nay Pyi Taw, Pyay, Sittwe,
Mixed deciduous forest, upper
Taungoo, Thandwe, Tharyarwaddy,
moist
Kalay, Khamti, Mawlaik, Tamu,
Linkhay, Taunggyi North, Taunggyi
South, Kyaukme, Kyaukphyu,
Maungdaw, MyaukOo, KyaukSe,
Meiktila, Pyin Oo Lwin, Yamethin

77
 Bago, Hinthada, Bhamo, Myitkyina,
Katha, Shwebo, Monywa, Magwe,
Gangaw, Minbu, Pakkoku, Thayet, AGB< 125
Dakinathiri, Ottarathiri, Taungoo: Tonnes/ Ha =
Nay Pyi Taw, Pyay, Taungoo, 0.20 (0.09-
Mixed deciduous forest, upper Tropical Moist
6 Tharyarwaddy, Kalay, Khamti, 0.25) AGB> 125 2.1
dry Deciduous Forest
Mawlaik, Tamu, Linkhay, Taunggyi Tonnes/Ha =
North, Taunggyi South, Kyaukme, 0.24 (0.22-
Pathein, MyaukOo, KyaukSe, 0.33)
Mandalay, Meiktila, Pyin Oo Lwin,
Yamethin

Bhamo, Myitkyina, Katha, Shwebo,

Monywa, Gangaw, Minbu, Pakkoku,
Thayet, Dakinathiri, Pyay, Taungoo,
Dipterocarp (indaing) forest,
Tharyarwaddy, Kalay, Khamti,
high
Mawlaik, Tamu, Linkhay, Taunggyi
AGB< 20
North, Taunggyi South, Kyaukme,
Tonnes/ Ha =
KyaukSe, Mandalay, Pyin Oo Lwin
0.56 (0.28-
7 Bhamo, Myitkyina, Katha, Shwebo, Tropical Dry Forest 0.68) AGB> 20 2.1
Monywa, Gangaw, Minbu, Pakkoku, Tonnes/Ha =
Thayet, Dakinathiri, Ottarathiri, 0.28 (0.27-
Pyay, Taungoo, Tharyarwaddy, 0.28)
Dipterocarp (indaing) forest,
Kalay, Khamti, Mawlaik, Tamu,
low
Linkhay, Taunggyi North, Taunggyi
South, Kyaukme, KyaukSe,
Mandalay, Meiktila, Pyin Oo Lwin,
Yamethin
Shwebo, Gangaw, Minbu, Pakkoku, AGB< 20
8 Dry forest, than-dahat Taungoo, Taunggyi North, Taunggyi Tropical Dry Forest Tonnes/ Ha = 2.1
South, KyaukSe, Pyin Oo Lwin 0.56 (0.28-

78
 Bago, Gangaw, Minbu, KyaukSe, Pyin 0.68) AGB> 20
Dry forest, thorn Tonnes/Ha =
Oo Lwin
0.28 (0.27-
Shwebo, KyaukSe, Mandalay, Pyin
Dry forest, aukchinsa-thinwin 0.28)
Oo Lwin
Falam, Bhamo, Katha, Minbu,
Thayet, Ottarathiri, Pyay, Kalay,
Tropical Mountain 0.27 (0.27-
9 Hill forest, evergreen Khamti, Mawlaik, Taunggyi North, 2.8
Systems 0.28)
Taunggyi South, KyaukSe, Meiktila,
Pyin Oo Lwin, Yamethin
AGB< 20
Bago, Katha, Shwebo, Gangaw,
Tonnes/ Ha =
Minbu, Thayet, Pyay, Taungoo,
0.56 (0.28-
Kalay, Khamti, Mawlaik, Linkhay,
10 Hill forest, dry Tropical Dry Forest 0.68) AGB> 20 2.1
Taunggyi North, Taunggyi South,
Tonnes/Ha =
KyaukSe, Meiktila, Pyin Oo Lwin,
0.28 (0.27-
Yamethin
0.28)
AGB< 50
Tonnes/ Ha =
0.40 (0.21-
1.06) AGB 50-
Monywa, Gangaw, Minbu, Kalay, Temperate: 150 Tonnes/ Ha
11 Hill forest, pine 4.1
Taunggyi North, Taunggyi South Conifers = 0.29 (0.24-
0.50) AGB >150
Tonnes/ Ha =
0.20 (0.12-
0.49)

79
 Annex 13: Net Biomass Increment and Removal Factor Calculation for Enhancement

Notes:
Planting/ site Plantation Total biomass Total C in Total CO2e in
preparation area total net biomass d.m accumulated d.m tonnes tonnes tonnes
Year Ha 1999-2000 2000-2001 2001-2002 2002-2003 2003-2004 2004-2005 2005-2006 2006-2007 2007-2008 2008-2009 2009-2010 2010-2011 2011-2012 2012-2013 2013-2014 2014-2015 2005 - 2015
2000-01 30,718 - 307,181 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 215,027 2,150,268 1,010,626 3,708,998
2001-02 30,756 - 307,562 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 215,293 2,152,931 1,011,878 3,713,591
2002-03 31,396 - 313,956 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 219,769 2,197,689 1,032,914 3,790,794
2003-04 30,441 - 304,405 213,084 213,084 213,084 213,084 213,084 213,084 213,084 213,084 213,084 213,084 213,084 213,084 2,130,835 1,001,493 3,675,478
2004-05 31,974 - 319,743 223,820 223,820 223,820 223,820 223,820 223,820 223,820 223,820 223,820 223,820 223,820 2,238,199 1,051,953 3,860,669
2005-06 33,201 - 332,013 232,409 232,409 232,409 232,409 232,409 232,409 232,409 232,409 232,409 232,409 2,324,089 1,092,322 4,008,821
2006-07 28,441 - 284,414 199,089 199,089 199,089 199,089 199,089 199,089 199,089 199,089 199,089 1,507,392 708,474 2,600,100
2007-08 26,666 - 266,659 186,661 186,661 186,661 186,661 186,661 186,661 186,661 186,661 1,226,630 576,516 2,115,814
2008-09 29,373 - 293,730 205,611 205,611 205,611 205,611 205,611 205,611 205,611 1,145,548 538,407 1,975,955
2009-10 32,271 - 322,712 225,898 225,898 225,898 225,898 225,898 225,898 1,032,679 485,359 1,781,267
2010-11 26,189 - 261,892 183,324 183,324 183,324 183,324 183,324 654,729 307,723 1,129,342
2011-12 24,846 - 248,461 173,923 173,923 173,923 173,923 447,231 210,198 771,428
2012-13 17,264 - 172,638 120,846.69 120,847 120,847 189,902 89,254 327,562
2013-14 18,354 - 183,538 128,476.48 128,476 73,415 34,505 126,634
2014-15 14,133 - 141,326 98,928.32 - 42,398 - 19,927 - 73,132
406,023 - 307,181 - 92,535 116,364 345,684 543,430 754,980 1,034,988 1,251,832 1,411,422 1,588,051 1,874,770 2,071,524 2,321,270 2,431,217 2,601,905 2,842,160 19,429,138 9,131,695 33,513,321
19,429,138 9,131,695 3,351,332
annual net increment rate (Y) in biomass dry matter tonnes per ha = 7 (according to Table 4.10, IPCC 2006; Table 3A.6, IPCC 2003)

biomass consumption from site preparation biomass dry matter tonnes per ha = 10 (according to Table 3A1.13, IPCC 2003)

80
 Annex 14: Uncertainty Result for Emission Factor in %

Mean Confidence Level

Sample CO2 eq (95%) Width/
Plots Plot Standard %
No Districts Size in Tonnes Confidence U x Xn (U x Xn) ^2
(n) Size Deviation Uncertainty
Ha Per Upper Lower Interval
Ha20
1 Bago 567 1 567 41.02 38.30 44.18 37.87 6.31 7.69 315.29 99,408.64
2 Dawei 109 0.4047 44.1123 385.55 308.31 476.53 294.57 181.97 23.60 9,098.27 82,778,547.65
3 Falam 159 0.4047 64.3473 185.81 181.23 230.09 141.53 88.56 23.83 4,428.18 19,608,738.88
4 Hinthada 68 0.7854 53.4072 100.90 65.50 118.47 83.34 35.13 17.41 1,756.61 3,085,682.03
5 Myitkyina 200 1.05 210 96.99 82.50 108.15 85.83 22.32 11.50 1,115.79 1,244,993.41
6 Bhamo 427 1.05 448.35 152.36 96.34 161.28 143.44 17.84 5.85 891.79 795,294.70
7 Katha 661 1 661 73.49 44.24 76.86 70.11 6.75 4.59 337.25 113,738.51
8 Shwebo 174 1 174 68.36 58.30 77.02 59.69 17.32 12.67 866.24 750,368.64
9 Monywa 190 1 190 84.60 34.98 89.57 79.63 9.95 5.88 497.36 247,366.59
10 Gangaw 311 1 311 96.15 71.96 104.15 88.15 15.99 8.32 799.74 639,591.49
11 Magwe 8 1 8 46.85 32.13 69.11 24.59 44.52 47.52 2,226.23 4,956,078.42
12 Minbu 467 1 467 81.07 68.20 87.26 74.89 12.37 7.63 618.57 382,629.12
13 Pakkoku 33 1 33 57.99 36.75 70.52 45.45 25.08 21.62 1,253.81 1,572,030.22
14 Thayet 274 1 274 78.85 51.35 84.93 72.77 12.16 7.71 607.99 369,647.99
15 Myaungmya 10 0.7854 7.854 2.04 15.14 12.62 (8.55) 21.17 519.29 1,058.51 1,120,448.33
16 Dakinathiri 105 1.05 110.25 102.84 73.17 116.50 89.18 27.32 13.28 1,365.85 1,865,549.50
17 Ottarathiri 67 1.05 70.35 92.11 69.50 108.35 75.87 32.48 17.63 1,624.17 2,637,941.38
Taungoo: Nay Pyi Taw 10 1.05 10.5 77.07 29.17 94.71 59.42 35.28 22.89 1,764.19 3,112,350.50
Taungyi: Nay Pyi Taw 20 1.05 21 116.92 37.29 132.86 100.97 31.90 13.64 1,594.75 2,543,236.62
18 Pyarpon 47 0.4047 19.0209 5.32 6.96 8.45 2.20 6.25 58.75 312.69 97,774.36
19 Pyay 430 1 430 61.51 64.47 67.60 55.42 12.19 9.91 609.33 371,281.91
20 Sittwe 6 1.05 6.3 674.39 655.33 1,186.13 162.65 1,023.47 75.88 51,173.74 2,618,751,828.74
21 Taungoo 962 0.7854 755.5548 60.76 52.35 64.49 57.03 7.47 6.14 373.31 139,360.36
22 Thandwe 147 0.7854 115.4538 62.85 49.81 71.93 53.76 18.17 14.46 908.65 825,637.30

20
Mean CO2 eq ton per ha is only for Above Ground Biomass Pool as national data are used only for AGB calculation and other remaining four pools are IPCC default values
81
23 Tharyarwaddy 446 0.7854 350.2884 52.55 71.39 60.03 45.07 14.95 14.23 747.60 558,911.31
24 Kalay 869 1 869 91.46 78.06 96.65 86.27 10.38 5.68 519.04 269,399.13
25 Khamti 951 1 951 61.90 55.83 65.45 58.35 7.10 5.73 354.87 125,931.78
26 Mawlaik 971 1 971 110.77 68.99 115.11 106.43 8.68 3.92 433.95 188,316.28
27 Tamu 45 1 45 53.07 24.20 60.14 46.00 14.14 13.32 707.02 499,875.86
28 Linkhay 234 1.05 245.7 73.29 33.62 77.49 69.08 8.41 5.74 420.33 176,674.23
29 Taunggyi_North 225 1.05 236.25 152.97 70.08 161.91 144.03 17.87 5.84 893.68 798,655.52
30 Taunggyi South 334 1 334 115.70 280.59 145.79 85.61 60.18 26.01 3,009.21 9,055,334.17
31 Kyaukme 519 1.05 544.95 245.45 164.62 259.27 231.63 27.64 5.63 1,382.14 1,910,309.77
32 Pathein 76 0.7854 59.6904 41.03 29.89 48.53 33.53 15.00 18.28 749.87 562,305.64
33 Kyaukphyu 219 0.4047 88.6293 70.48 46.13 80.09 60.88 19.21 13.63 960.34 922,246.74
34 Maungdaw 52 1.05 54.6 76.25 34.34 85.36 67.14 18.22 11.95 910.94 829,804.39
35 MyaukOo 77 1.05 80.85 76.44 41.96 85.59 67.29 18.29 11.97 914.71 836,687.16
36 KyaukSe 163 1 163 96.80 88.94 110.46 83.15 27.31 14.10 1,365.36 1,864,220.60
37 Mandalay 19 1 19 52.06 62.24 80.05 24.08 55.97 53.75 2,798.63 7,832,350.74
38 Meiktila 90 1 90 47.41 51.40 58.03 36.79 21.24 22.40 1,061.91 1,127,662.07
39 Pyin Oo Lwin 465 1 465 85.98 97.45 94.83 77.12 17.72 10.30 885.76 784,565.34
40 Yamethin 77 1 77 44.26 39.22 53.02 35.50 17.52 19.80 876.06 767,480.94
11,284 4,353.63 2,777,220,256.96

√(𝑈1 ∗ 𝑥1 ) 2 + (𝑈2 ∗ 𝑥2 ) 2 + ⋯ + (𝑈𝑛 ∗ 𝑥𝑛 ) 2 52,699.34

√(𝑈1 ∗ 𝑥1 ) 2 + (𝑈2 ∗ 𝑥2 ) 2 + ⋯ + (𝑈𝑛 ∗ 𝑥𝑛 ) 2

𝑈 𝑡𝑜𝑡𝑎𝑙 21 = Uncertainty value for overall emission factor is accepted as: 12.10
𝑥1 + 𝑥2 + ⋯ + 𝑥𝑛

21
Table 6.1: Tier 1 Uncertainty Calculation and Reporting under IPCC Good Practice Guidance and Uncertainty Management in National Greenhouse Gas Inventory
82