ArticlePDF Available

Assessing the conservation status of mangroves in Rakhine, Myanmar

December 2023
Aquatic Conservation Marine and Freshwater Ecosystems 34(1)

DOI:10.1002/aqc.4058

License
CC BY-NC 4.0

Authors:

Calvin K. F. Lee

The University of Hong Kong

Emily Nicholson

University of Melbourne

Clare Duncan

Zoological Society of London

Hedley Grantham

Conservation International

Show all 7 authorsHide

Ecosystem degradation is a key challenge that human society faces, as ecosystems provide services that are tied to human well‐being. Particularly, mangrove ecosystems provide important services to communities but are suffering heavy degradation, loss and potential collapse due to anthropogenic activities. The IUCN Red List of Ecosystems is a transparent and consistent framework for assessing ecosystems' risk of collapse and is increasingly used to inform legislation and ecosystem management globally. Satellite data have become increasingly common in environmental monitoring due to their extensive spatial and temporal coverage. Here, recent advances in analyses using satellite‐derived data were implemented to reassess the conservation status of the ‘Rakhine mangrove forest on mud’, an important intertidal ecosystem in Myanmar, extending a previous national Red List assessment that assessed the ecosystem as Critically Endangered. By incorporating additional data sources and analyses, the extended assessment produced more robust results and reduced the uncertainty in the previous assessment. Overall, the ecosystem was assessed as Critically Endangered (range: Vulnerable to Critically Endangered) as a result of historical mangrove extent loss. Recent losses and biotic disruptions were also observed, which would have led to the ecosystem being assessed as Vulnerable. While the final outcome of the Red List assessment remained at Critically Endangered due to the historical state of the mangroves pre‐dating the temporal coverage from satellite data, the uncertainty of the ecosystem's status was reduced, and the reassessment highlighted the recent areal changes and mangrove degradation that has occurred. The importance of conducting reassessments when new data become available is discussed, and a template for future mangrove Red List assessments that use satellite data as their primary source of information to improve the robustness of their results is presented.

Content uploaded by Calvin K. F. Lee

Content may be subject to copyright.

RESEARCH ARTICLE

Assessing the conservation status of mangroves in Rakhine,

Myanmar

Calvin K. F. Lee

1,2

| Emily Nicholson

2,3

| Clare Duncan

| Hedley S. Grantham

David A. Keith

| Rob Tizard

| Nicholas J. Murray

School of Biological Sciences, University of

Hong Kong, Hong Kong, China

Centre for Integrative Ecology, School of Life

and Environmental Sciences, Deakin

University, Burwood, Victoria, Australia

School of Agriculture, Food and Ecosystem

Sciences (SAFES), Faculty of Science, The

University of Melbourne, Melbourne, Victoria,

Australia

Institute of Zoology, Zoological Society of

London, London, UK

Global Conservation Program, Wildlife

Conservation Society, New York, New York,

USA

Centre for Ecosystem Science, University of

NSW, Sydney, New South Wales, Australia

College of Science and Engineering, James

Cook University, Townsville, Queensland,

Australia

Correspondence

Calvin K. F. Lee, School of Biological Sciences,

University of Hong Kong, Hong Kong, China.

Email: leeckf@hku.hk

Funding information

Australian Research Council, Grant/Award

Numbers: LP170101143, FT190100234; HKU,

Grant/Award Number: 202011159154;

Australian Government Department of

Education

Abstract

1. Ecosystem degradation is a key challenge that human society faces, as

ecosystems provide services that are tied to human well-being. Particularly,

mangrove ecosystems provide important services to communities but are

suffering heavy degradation, loss and potential collapse due to anthropogenic

activities. The IUCN Red List of Ecosystems is a transparent and consistent

framework for assessing ecosystems' risk of collapse and is increasingly used to

inform legislation and ecosystem management globally.

2. Satellite data have become increasingly common in environmental monitoring due

to their extensive spatial and temporal coverage. Here, recent advances in

analyses using satellite-derived data were implemented to reassess the

conservation status of the ‘Rakhine mangrove forest on mud’, an important

intertidal ecosystem in Myanmar, extending a previous national Red List

assessment that assessed the ecosystem as Critically Endangered.

3. By incorporating additional data sources and analyses, the extended assessment

produced more robust results and reduced the uncertainty in the previous

assessment. Overall, the ecosystem was assessed as Critically Endangered (range:

Vulnerable to Critically Endangered) as a result of historical mangrove extent loss.

Recent losses and biotic disruptions were also observed, which would have led to

the ecosystem being assessed as Vulnerable.

4. While the final outcome of the Red List assessment remained at Critically

Endangered due to the historical state of the mangroves pre-dating the temporal

coverage from satellite data, the uncertainty of the ecosystem's status was

reduced, and the reassessment highlighted the recent areal changes and

mangrove degradation that has occurred.

5. The importance of conducting reassessments when new data become available is

discussed, and a template for future mangrove Red List assessments that use

satellite data as their primary source of information to improve the robustness of

their results is presented.

Received: 17 May 2023 Revised: 16 November 2023 Accepted: 29 November 2023

DOI: 10.1002/aqc.4058

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial License, which permits use, distribution and reproduction in any

medium, provided the original work is properly cited and is not used for commercial purposes.

Aquatic Conserv: Mar Freshw Ecosyst. 2023;1–17. wileyonlinelibrary.com/journal/aqc 1

KEYWORDS

ecosystem risk assessment; ecosystem risk of collapse; mangrove ecosystem; Rakhine,

Myanmar; Red List of Ecosystems; satellite time series

1|INTRODUCTION

Ecosystems around the world continue to be threatened by

anthropogenic activity (Intergovernmental Science-Policy Platform on

Biodiversity and Ecosystem Services [IPBES], 2019). These threatening

processes have led to reductions in biodiversity and decreased capacity

for the delivery of ecosystem services, with important impacts on

human well-being (Cardinale et al., 2012). To support effective

conservation decision-making, there is a need for transparent and

consistent methods for assessing the status of ecosystems based on

sound ecological knowledge. The IUCN (International Union for the

Conservation of Nature) Red List of Ecosystems was developed to

assess and identify ecosystems at risk of losing biodiversity and

ecological functions (Keith et al., 2013). The Red List was designed to

be applicable to any terrestrial, marine or freshwater ecosystems and

since its inception has been applied to >4000 ecosystems across >100

countries (Bland et al., 2019;http://iucnrle.org).

The Red List of Ecosystems is a standardized framework that

enables estimates of relative risk of ecosystem collapse; collapse is

the endpoint of ecosystem decline when defining biotic or abiotic

features are lost and the characteristic native biota are no longer

sustained (Keith et al., 2013). The Red List enables consistent

national- and international-level ecosystem assessments that can be

used to inform legislation and ecosystem management (Bland

et al., 2019; Keith et al., 2013). The Red List of Ecosystems

framework comprises five criteria (each with additional sub-criteria)

that reflect symptoms of ecosystem change. Criterion A uses

measures of change in ecosystem area over time, where ecosystems

with greater areal losses are at a higher risk of collapse. Criterion B

identifies ecosystems at risk of collapse from spatially explicit,

stochastic threats using specific metrics and thresholds of ecosystem

size, where smaller ecosystems are at a higher risk of collapse.

Criterion C estimates the risk associated with environmental

degradation related to key physical and abiotic processes, and

Criterion D estimates the risk associated with degradation to key

biota and/or ecological interactions or processes, where the loss of

either of these leads to a transformation of the identity of the

ecosystem. Finally, Criterion E allows the use of simulation models to

directly estimate an ecosystem's probability of collapse within a fixed

time frame (Keith et al., 2013). Full details of the criteria can be found

in the Red List of Ecosystems guidelines (Bland et al., 2017; Keith

et al., 2015; Rodríguez et al., 2015).

Myanmar is one of the most forested countries in Southeast Asia

(Leimgruber et al., 2005), supporting a large number of endemic

species with important economic and cultural significance to the

country (Aung, 2007; Murray et al., 2020). Despite the importance of

Myanmar's ecosystems, they are facing increasing anthropogenic

threats as the country continues to develop and its population grows

(Veettil et al., 2018; Webb et al., 2014). A national Red List

assessment of all terrestrial ecosystems in Myanmar was completed in

2020 to support conservation efforts (Murray et al., 2020). These

assessments provide information for all 64 terrestrial and coastal

ecosystems in Myanmar, highlighting the ecosystems most at risk,

along with data-deficient ecosystems that will require additional

research attention to inform appropriate conservation actions.

Mangrove ecosystems occur globally along tropical and warm

temperate coastlines and play critical economic and ecological roles

for human communities and the surrounding ecosystems. They

provide a wide range of ecosystem services, acting as sources of food

and fuel for local communities; nursery sites for ecologically,

subsistence and commercially important faunal species; and coastal

protection from storm events and are carbon-rich ecosystems aiding

in climate regulation (Goldberg et al., 2020; Lee et al., 2014;

Richards & Friess, 2016; Veettil et al., 2018). Myanmar is one of the

most mangrove-rich countries in the world (Estoque et al., 2018), and

mangrove ecosystems are particularly important along the coast of

the country, as the majority of the human communities here rely on

mangroves in their daily activities (Storey, 2015). Despite their

importance, Myanmar is a hotspot for mangrove loss (De Alban

et al., 2020; Goldberg et al., 2020) and is one of six nations in

Southeast Asia that together contribute to nearly 80% of total global

anthropogenic mangrove loss over the past two decades (Goldberg

et al., 2020).

Studies on the mangrove ecosystems in Myanmar are often at a

national level (De Alban et al., 2020;Estoqueetal.,2018) or focused on

the Ayeyarwady delta (Webb et al., 2014; Win et al., 2020), and the

mangroves on the west coast of Myanmar along the Bay of Bengal are

relatively less studied. Neighbouring mangrove ecosystems in the

Sundarbans and Bangladesh to the north and along the Ayeyarwady

delta to the south have experienced well-documented mangrove losses

in the past few decades (De Alban et al., 2020; Sievers et al., 2020).

Here, the focus is on the ‘Rakhine mangrove forest on mud’,oneof

four mangrove ecosystems in Myanmar (Murray et al., 2020). The

national Red List assessment for this ecosystem assessed it as Critically

Endangered (Vulnerable–Critically Endangered) (Murray et al., 2020),

making it one of the most at-risk ecosystems assessed in Myanmar.

However, it also contained considerable uncertainty: For example, the

estimated change in distribution over 50 years was based on only three

Landsat images from 1988, 2000 and 2015 (Storey, 2015), while the

historical mangrove change since 1750 was based on questionnaires

that were from only five villages in the northern parts of the state

(Storey, 2015). The assessment of mangrove ecosystem degradation

was also based on a global study of mangroves that lack regional

context.

2LEE ET AL.

10990755, 0, Downloaded from https://onlinelibrary.wiley.com/doi/10.1002/aqc.4058, Wiley Online Library on [29/12/2023]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License

Knowledge and data on mangrove status are rapidly improving

due to improved satellite analysis methodology, larger satellite image

archives, and a better understanding about mangrove threats and

degradation dynamics. This assessment aims to use these advances to

provide the first-recorded reassessment of any ecosystem in Asia

under the Red List of Ecosystems criteria, revealing finer scale

patterns of degradation and areal change while demonstrating how

ecosystem status assessments can integrate newly acquired

knowledge.

The Red List guidelines were used to integrate diverse sources of

independent data and evidence, including various temporal scales

of mangrove distribution maps from freely available sources; results

from previous assessments of the ecosystem and newly developed

methods, including a dense time-series satellite remote sensing data

to estimate ecosystem area trends (Lee et al., 2021); and a mangrove

conceptual model to train a mangrove degradation model specific for

this ecosystem (Lee et al., 2021) into a single outcome assessing the

risk of ecosystem collapse. Within the context of Red List

assessments, reassessments are vital as ecosystems continue to

change and/or additional data or novel analytical methods become

available, potentially changing the outcome of the assessment. In this

study, the same ecosystem description as the previous assessment

was used to ensure identical units of assessment and that any

differences in the outcome are due to the added information or new

methods. By conducting a reassessment, we aimed to reduce the

uncertainty associated with the initial assessment and further improve

our understanding of the status of the ecosystem, the threats it faces,

the primary biotic and abiotic factors driving the risk of ecosystem

collapse and the conservation actions that will be required to mitigate

and reduce this risk.

2|MATERIAL AND METHODS

The Red List of Ecosystems criteria according to the IUCN guidelines

(Bland et al., 2017) was applied to assess the risk of collapse of the

principal ‘Rakhine mangrove forest on mud’ecosystem along

Myanmar's north-west coast (Figure 1), hereafter referred to as

‘Rakhine mangroves’. Existing data for the region that were

potentially suitable for assessing each of the five criteria

were reviewed, including data collated by Murray et al. (2020). These

data were supplemented with additional analyses of satellite data

using recently developed methods for mapping time-series ecosystem

change and degradation; the breakdown of the workflow is shown in

Figure 2. A detailed description of the assessment, including methods

and findings, can be found in the Supporting Information.

2.1 |Ecosystem description

Rakhine mangrove ecosystem is defined by the extent of mangrove-

dominated vegetation along Myanmar's coastline within the Bay of

Bengal, including all mangroves on a muddy substrate within Rakhine

state and Bassein, Ayeyarwady (Murray et al., 2020). The ecosystem

is classified under the IUCN Global Ecosystem Typology as functional

group MFT1.2 intertidal forests and shrublands of the brackish

intertidal biome (Keith et al., 2020; MFT1.2, https://global-

ecosystems.org/explore/groups/MFT1.2), and under the IUCN

Habitats Classification Scheme (Version 3.1) as habitat type 12.7

(Mangrove Submerged Roots) (IUCN, 2012). It occurs within the Bay

of Bengal marine ecoregion (Eco ID 321) (Spalding et al., 2007).

Rakhine mangroves differ from neighbouring mangrove

ecosystems by occurring across four geomorphic settings (deltaic,

open coast, lagoonal and estuarine), in contrast to the solely deltaic

mangroves of the Sundarbans and Ayeyarwady (Worthington

et al., 2020). The ecosystem consists of at least 28 true mangrove

species (Table S1), including the Critically Endangered Bruguiera

hainesii and Sonneratia griffithii (IUCN, 2020; Myint & Stanley, 2011).

The faunal diversity of the ecosystem is also high, including at least

62 species of fin fish; five species of crustacean; five species of

mollusc; 104 bird species, including both migrants and residents; and

several globally endangered vertebrates (Table S2; Stanley &

Broadhead, 2011).

Myanmar receives approximately 75% of its annual rainfall during

the summer monsoon months (June–September); Rakhine is the state

that receives the highest seasonal rainfall (>424 cm) during this period

(Sen Roy & Kaur, 2000). This high rainfall, combined with the flow of

low-salinity water from the Ganges–Brahmaputra River in the north,

leads to a low salinity of <18 parts per thousand (ppt) during the

summer monsoon season. Salinity increases to more than 34 ppt in

the dry seasons due to low rainfall and regular inundation of highly

saline waters flowing from the Andaman Sea in the south

(Ramaswamy & Rao, 2014). The coastline receives a small amount of

sediment discharge from the Ganges–Brahmaputra delta to the north

and the Ayeyarwady delta to the south. These salinity and

sedimentation regimes are key components of the abiotic

environment for this ecosystem, as mangroves occur along the

mesotidal coastal zone in areas where soft sediment is regularly

inundated throughout the tidal cycle (Ramaswamy & Rao, 2014;

Worthington et al., 2020).

2.1.1 | Threatening processes, drivers and

ecosystem decline

Rakhine mangroves are subject to anthropogenic, natural and climate

change-related threats (Figure 3). Mangrove loss and degradation in

Rakhine are driven predominantly by anthropogenic activities,

in particular their conversion to agricultural and aquacultural lands,

including oil palm plantations, rice paddies and shrimp farms

(De Alban et al., 2020; Goldberg et al., 2020; Storey, 2015). The

development of agriculture and aquaculture near mangroves can also

result indirectly in ecosystem degradation due to sea wall

construction, altering normal hydrology and changing the tidal

inundation dynamics, influencing the mangrove ecosystem. Farms can

also directly introduce pollutants, causing eutrophication of mangrove

LEE ET AL.3

FIGURE 1 (a) Mapped distribution of the

Rakhine mangrove forest on mud and the

location of the study region within Myanmar

(top right). The map also includes occupied

10-km

grid cells and a minimum convex

polygon encompassing all mangrove

occurrences within the study region.

(b) Photo of the ecosystem within the

Wunbaik reserved forest (photo credit: Don

Macintosh).

4LEE ET AL.

ecosystems (Friess et al., 2019). Logging and wood harvesting for

firewood and charcoal and bark peeling for dyes are other sources of

degradation to the Rakhine mangroves (Stanley & Broadhead, 2011).

Relative sea-level rise as a result of climate change is also expected to

negatively affect Rakhine mangroves by reducing land suitable for

mangroves, thus reducing their area (Alongi, 2015). Climate

change-driven increased frequency and intensity of storms, altered

precipitation and higher temperatures may also threaten the

ecosystem in the future (Alongi, 2015; Ward et al., 2016).

2.1.2 | Indicators and thresholds of ecosystem

degradation and collapse

Mangrove ecosystems are primarily characterized by mangrove trees

and shrubs acting as foundation species, with non-dominant animal

species playing smaller ecological roles (Geist et al., 2012; Marshall

et al., 2018). Therefore, mangrove vegetation loss or degradation was

used as an indicator of collapse risk, where an absence of true mangrove

species signifies the collapse of the ecosystem (Marshall et al., 2018).

When assessing the spatial Red List criteria (Criteria A and B) of

Rakhine mangroves, the ecosystem is considered to have collapsed

when pixel-specific mapped mangrove distribution is reduced to zero

as a result of the complete loss of any mangrove vegetation. For

Criterion C, a sufficient change in abiotic conditions, such as

sedimentation and/or salinity, can cause mangrove collapse when the

environment can no longer sustain mangrove vegetation (Krauss

et al., 2014; Peters et al., 2020). For Criterion D, the ecosystem is

considered collapsed when mangrove vegetation is degraded to the

point of complete loss of distribution (cf. function); see more details

below under Criterion D.

2.2 |Mapping distribution and degradation of

Rakhine mangroves

To assess Rakhine mangroves under the Red List criteria, various

maps of mangrove distribution at different times across the

assessment period are needed. To achieve this, multiple independent

sources of data to generate mangrove area estimates, each with their

FIGURE 2 Overall workflow of the methods, including (a) the data sources used and (b) distribution maps and estimates and (c) analysis done

for each criterion of the Red List of Ecosystems assessment.

LEE ET AL.5

respective accuracy assessments, were compiled and reviewed.

Including multiple sources of data and analytical methods can increase

confidence in the robustness of the Red List assessment outcomes

(Bland et al., 2017).

2.2.1 | Static mangrove distribution maps

Maps of Rakhine mangrove distribution were needed to assess

distribution change over time, as well as assess the extent of

ecosystem degradation. Four sources of static maps suitable for the

assessment over different time periods were used to investigate

changes in the distribution of Rakhine mangroves over time. Additional

steps were taken to ensure the comparability of the data from

different sources when they were used together in a single analysis:

1. the US Army Map Services (AMS) for Burma (Myanmar) created

from aerial photography and depicting mangrove distribution for

the period between 1943 and 1945 (AMS1944), which were

digitized manually (details in Supporting Information);

2. a freely available global mangrove map for 2000 produced by Giri

et al. (2011) from Landsat satellite data (Giri2000);

3. Global Mangrove Watch (GMW) version 2 maps with their

associated accuracy assessments, available for various years:

GMW1996,GMW2007,GMW2008,GMW2009,GMW2010,

GMW2015 and GMW2016 (Bunting et al., 2018). GMW was

included as it is one of the most recognized global datasets for

mangrove monitoring, providing time series information for global

mangrove areas, and is freely available to the public. GMW

provides straightforward mangrove area estimates to users

without further capacity to develop their own classifications. The

FIGURE 3 Mangrove ecosystem conceptual model highlighting the main threats and ecological processes. The red boxes indicate threats, the

blue ovals represent the abiotic processes, the blue hexagons represent the abiotic environment and the green hexagons represent the biotic

components of the mangrove ecosystem. Pointed arrowheads indicate positive effects, rounded arrowheads indicate negative effects and square

arrowheads indicate context-dependent effects.

6LEE ET AL.

overall accuracy of the dataset was 95.3%, with a 99% likelihood

that the confidence interval, using the Wilson score interval, was

4.5%–5.0%.

4. Two additional mangrove distribution maps for the years 2014 and

2019 were classified using a machine learning model for the study

region; 2014 was chosen as the earliest year with complete

Landsat 8 coverage, and 2019 was chosen as the latest year with

complete Landsat 8 coverage at the time of analysis. The 2019

map was used for the assessments of Criteria B, C and D. Manually

developing these two maps allowed accuracy assessments to be

conducted for these maps, reporting area estimates with

quantitative uncertainties. These maps were produced by applying

supervised random forest classifications to cloud-free composites

of Landsat 8 images acquired during the dry season (January to

March, October to December) for those 2 years using the Google

Earth Engine (Gorelick et al., 2017; details provided in the

Supporting Information) and are referred to as Classification2014

and Classification2019.

Classification2019 was used for any analysis where only one map

of mangrove distribution was required, while all other maps were

used in conjunction with others for trend analyses to estimate

mangrove area change through time (Figure 2b).

2.2.2 | Dense time series of mangrove distribution

In addition to estimating mangrove areas from static mangrove maps

that provide estimations of mangrove areas at specific times, a dense

Landsat time-series model was developed to estimate mangrove area

trends from 1988 to 2019 following the methods of Lee et al.

(2021). This included using random forest models to estimate

mangrove areas from every available Landsat image over the study

area. To train the random forest models, a spectral library suitable for

training and prediction was developed. The training points used to

generate the spectral library included target classes (mangrove;

water; cloud; others, which include all non-mangrove non-water land

cover such as non-mangrove forests, sand and bare ground) for each

of Landsats 5, 7 and 8 using high-resolution Google Earth imagery

(National Centre for Space Studies [CNES]/Airbus) for 2012 (Landsat

5) and 2018 (Landsats 7 and 8). Landsat bands were then extracted

from each of the training points from their respective Landsat

satellites to generate a set of explanatory covariates included in the

spectral library. To develop an image time series suitable for applying

the dense time-series classification model, 16-day repeat mosaics

were created for the entire study region for the period 1988 to

2019. Random forest models trained using the spectral libraries were

used to predict the land cover of each of these mosaics, creating

maps of mangroves, water, others and clouds. Estimates from

mosaics where there were any gaps in mapped distribution were

discarded, leaving 1132 estimates to enable analyses of mangrove

areal trends over a 31-year period. The accuracy of the dense time-

series classification model was assessed with 900 validation points,

including 200 validation points that were consistently mapped as

mangroves in AMS1944, Giri2000, GMW1996, GMW2016 and

Classification2019 and 700 validation points that were never

mapped as mangroves in the same maps. All 900 validation points

were then applied to the 1132 classified maps, and all points

classified as clouds were discarded, leaving 174,487 mangrove pixels

and 590,895 non-mangrove pixels that were used to assess the

entire dense time series. The results from these pixels were

subsequently used to estimate the accuracy of the dense time-series

classification model.

2.2.3 | Mapping mangrove degradation

To assess the extent of ecosystem degradation in the Rakhine

mangrove ecosystem, an additional supervised random forest

classification model, trained to classify mangrove pixels into degraded

and intact classes, was used (Lee et al., 2021). Occurrence points used

to train the classification model were collected following the criteria

below (for details, please refer to Lee et al., 2021).

For a pixel to be labelled as intact in the training set, it must meet

the following criteria:

•It contained part of a mangrove forest patch that was at least 5 ha

in area.

•It contained a closed canopy cover with no underlying substrate

observed from Google Earth imagery.

•It contained no obvious anthropogenic structures and disturbances

observed from Google Earth imagery.

•It maintained the above criteria for at least 5 years.

Pixels in the training set annotated as degraded met the following

criteria:

•Mangrove trees can be observed in Google Earth imagery (thus not

collapsed).

•Low canopy cover and/or isolated trees can be observed from

Google Earth imagery.

•Browning and/or tree death is observable from Google Earth

imagery.

Training points representing the two classes of ecosystem state

(157 degraded and 133 intact) were selected manually using Google

Earth imagery across the study region. The training points were

subsequently used to train a random forest classifier that included

five explanatory variables: annual normalized difference vegetation

index (NDVI) standard deviation and mean, annual normalized

difference moisture index (NDMI) standard deviation and mean and

annual normalized difference water index (NDWI) mean (Lee

et al., 2021; Table 1). The analysis yielded two maps of degradation

for Rakhine mangroves (2014 and 2019). The accuracy of these maps

and the area of each class were subsequently estimated following

good practice guidelines (Olofsson et al., 2014).

LEE ET AL.7

2.3 |Red List of Ecosystems assessment

2.3.1 | Criterion A—reduction in geographic

distribution

To obtain a comprehensive picture of the change in geographic

distribution of the Rakhine mangrove ecosystem and consider any

uncertainty that may exist, three independent analyses were

developed to investigate its change through time. Additionally, the

results from a previous study that estimated historical losses of

Rakhine mangroves are also reported (Storey, 2015).

To estimate reduction in mangrove extent over the past 50 years

(Criterion A1), linear and exponential models (after Bland et al., 2017)

were fitted to six static area estimates (AMS1944, GMW1996,

Giri2000, Classification2014, GMW2016, and Classification2019).

Only two out of the seven available GMW maps (1996 and 2016)

were included to prevent the model from being driven primarily by

them. To allow comparison between the different data sources, each

map was resampled to 200-m spatial resolution, which was larger

than the smallest patch of mangroves depicted in the maps from the

AMS (approximately 200 m 200 m), while all other maps were

originally at 30-m resolution. While accuracy metrics were not

available for AMS1944, this map was generated from aerial

photography by the US Army and included high levels of detail.

Uncertainty may have arisen due to misclassification of mangroves

by the cartographers, but there is unfortunately no method of

assessing this. Regardless, given their level of detail and inclusion of

mangroves as a specific class, this dataset provides a valuable

historical baseline to compare against (Murray et al., 2014). The

uncertainties within the maps are accounted for using statistical

models, and the use of multiple independent sources of area

estimates increases the robustness of the trend estimates (Bland

et al., 2017). A pixel counting approach was used to estimate

mangrove area from each map, as reference data were not available

for the older maps to allow for area estimation using error matrices

(Olofsson et al., 2014). The estimates and 95% confidence intervals

from the models were used as best-case and worst-case scenarios to

assess Criterion A1.

To estimate reduction in mangrove extent over any 50-year

period (Criterion A2b), two analyses were performed. First, seven area

estimates were derived from GMW, estimated at 30-m resolution

using pixel counting. This dataset included data over a 20-year time

frame, and in the absence of available socio-economic data to provide

information about the most likely trajectory over 50 years, a statistical

method was used to determine the most suitable model. Linear and

exponential models were fitted to the dataset (Bland et al., 2017).

These models were used to extrapolate the results into the future to

include the 50-year time frame (1996–2046; beginning from the first

GMW estimate), and the best and worst case scenarios were

estimated based on the estimates and their 95% confidence intervals,

the results representing the outcome of the assessment under

Criterion A2b. Second, the dense time-series model described in

Section 2.2.2 was used. A generalized additive mixed model (GAMM)

was fitted to the area estimate from the dense time-series model, and

the trend in extent was estimated using the R package mgcv

(Wood, 2017). The estimate and 95% confidence intervals were then

extrapolated to the required 50-year time frame (1988–2038;

beginning from the first available Landsat image used), using both the

absolute rate of decline and proportional rate of decline and the best

and worst case scenarios calculated to assess the ecosystem under

Criterion A2b.

To estimate reduction in mangrove extent when compared to a

historical baseline (approximately 1750; Criterion A3), two different

methods were used. First, previously published historical estimates

of mangrove areas, obtained from interviews conducted in five

villages in Northern Rakhine (Storey, 2015), were used. The estimate

from this method assumes that the rate of mangrove loss was the

same throughout the geographic extent of the ecosystem. Second,

the mangrove area estimates from the various maps collated

(AMS1944, GMW1996, Giri2000, Classification2014, GMW2016,

and Classification2019) were extrapolated backwards to 1750.

Assuming a constant rate of mangrove loss, both proportional loss and

absolute area loss were estimated to generate best and worst case

scenarios. The assumption that mangrove loss occurred at a constant

rate was unlikely to hold true, as development within Rakhine did not

occur at the same rate throughout the region (Storey, 2015), and a

TABLE 1 Satellite-derived covariates used to model mangrove degradation including the expected mechanism for each covariate to detect

mangrove degradation.

Covariate Proposed mechanism Reference

Annual NDVI mean Intact mangrove forests have higher mean NDVI as they are more photosynthetically

active and have higher canopy cover and LAI.

Kovacs et al. (2004)

Annual NDVI SD Intact mangroves have a more stable NDVI as they remain productive and have high

cover throughout the year as evergreen trees.

Verbesselt et al. (2016)

Annual NDMI mean Intact mangroves with higher canopy cover have higher average NDMI. Lucas et al. (2020)

Annual NDMI SD Intact mangroves have a more stable NDMI as they remain productive and have

high cover throughout the year as evergreen trees.

Verbesselt et al. (2016)

Annual NDWI mean Intact mangrove forests have lower average NDWI as they have higher canopy

cover and multi-spectral satellites cannot typically detect underlying water.

Abbreviations: LAI, leaf area index; NDMI, normalized difference moisture index; NDVI, normalized difference vegetation index; SD, standard deviation.

8LEE ET AL.

constant rate of loss over such a long time period is very unlikely.

Regardless, these results provide the only source of plausible

estimates possible, given data limitations, on historical mangrove

change required for the Red List assessment.

2.3.2 | Criterion B—restricted geographic

distribution

To assess Rakhine mangroves under Criteria B1 and B2, the Rakhine

mangrove map produced for 2019 (Classification2019, Section 2.2.1)

was used. The extent of occurrence (EOO, Criterion B1) was calculated

using a minimum convex polygon enclosing all mapped occurrences of

the ecosystem. The area of occupancy (AOO, Criterion B2)was

calculated by counting the number of 10 km 10 km grid cells that

contained the ecosystem while accounting for geometric uncertainty

(i.e., the placement of the cells used to assess AOO; Lee et al., 2019).

Both of these functions were calculated using the R package redlistr

(Lee et al., 2019). In addition to the EOO and AOO of the ecosystem,

one of three sub-criteria had to be met for an ecosystem to be listed

under Criteria B1 and B2: (a) an observed or inferred continuing decline,

(b) an observed or inferred threat and (c) the ecosystem existing at a

small number of threat-defined locations (Bland et al., 2017). As there is

already evidence that there is a continuing decline in the extent of the

Rakhine mangrove ecosystem based on results for Criterion A (i.e., Sub-

criteria a and b are met), there was no need to estimate the number of

threat-defined locations of the ecosystem (Sub-criterion c).

2.3.3 | Criterion C—environmental degradation

Environmental degradation of Rakhine mangroves can be caused by

various drivers, including altered hydrology and pollution as a result of

aquaculture, extreme weather phenomena and relative sea-level rise.

Data on aquaculture expansion in the area and the effects these farms

may have on surrounding mangroves would be required to assess

Criterion C for Rakhine mangroves due to the expansion of aquaculture,

none of which were available at the time of analysis. Data on future

risks from extreme weather phenomena were also unavailable. As a

result, only relative sea-level rise was assessed as the threat that can

cause environmental degradation leading to ecosystem collapse.

While the Sea Level Affecting Marsh Model (SLAMM; Clough

et al., 2016) is a commonly used model to predict future scales of

relative sea-level rise along coastlines, the data required to

parameterize and train such a model for the study region were not

available. Thus, a sea-level rise model for the Indo-Pacific developed by

Lovelock et al. (2015) was used, as reported by Murray et al. (2020).

This was trained using the surface elevation table-marker horizon

(SET-MH) method along with satellite-derived total suspended matter

data to predict the year of submergence of mangroves in 10-year time

steps. The relative severity of relative sea-level rise was estimated by

assuming any mangroves predicted to be submerged in 50 years will be

extensively degraded and likely to collapse due to drowning.

2.3.4 | Criterion D—disruption of biotic processes

and interactions

The relative severity of the mapped degradation, including a range of

plausible values, was estimated based on the training set that was

used to develop the degradation model. As all training points for the

degraded class were based on mangroves that were visibly degraded

in high-resolution satellite imagery, substantial disruption to normal

biotic processes must already have occurred, suggesting a high

relative severity of degradation. The results from these models were

used to estimate the extent of degradation for Rakhine mangroves for

2014 and 2019. The areas from each map and corresponding

uncertainties were estimated following Olofsson et al. (2014).

However, the Red List requires assessments to consider change over

50 years, and a 5-year period is not sufficient to extrapolate the results

to the required time frame. As a result, the ecosystem was assessed

under Criterion D3 by comparing the results for 2019 with a historical

baseline (1750s). In this scenario, it is assumed that no ecosystem

degradation was present in the 1750s due to low human population

density and no evidence of mangrove degradation in the region before

the 1800s (Storey, 2015). Additionally, the results from the assessment

by Murray et al. (2020), who assessed ecosystem degradation using a

model that evaluated trends in 12 vegetation indices for mangrove pixels

mapped by GMW (Worthington & Spalding, 2018), were also reported to

maintain comparability between the two assessments.

2.3.5 | Criterion E—quantitative risk analysis

Quantitative models that explicitly estimate the future risk of

ecosystem collapse are required to assess an ecosystem under Criterion

E(Blandetal.,2017). Such models should produce quantitative

estimates of ecosystem risk of collapse over a 50- to 100-year time

frame with explicit uncertainty. Existing quantitative mangrove models

that may be suitable for assessing Rakhine mangroves' risk of collapse

were reviewed. Process-based vegetation models may be suitable for

assessing mangrove ecosystems under Criterion E but are data and

computationally expensive when applied at the scale required for this

study. The most suitable potential model, the MANGRO model (Doyle

et al., 2003), was developed for mangroves in North America, where

mangrove species diversity is much lower, and the assumptions and

baseline data underpinning this model will not be appropriate to

Rakhine mangroves. Thus, Criterion E was not assessed.

3|RESULTS

3.1 |Criterion A

3.1.1 | Criterion A1

The analysis showed that Rakhine mangroves' historical distribution

declined from the initial estimate of 2771 km

in 1943 to 1566 km

LEE ET AL.9

2019 (43.8%, Figure 4). The best fitting model (linear), assessed with

root mean square error (RMSE), suggested that 33.3% (20.6%–44.1%)

of mangrove area was lost over the past 50 years (Figure 4). The

ecosystem therefore meets the category threshold for Vulnerable

(range: Least Concern–Vulnerable) under Criterion A1 (Table 2).

3.1.2 | Criterion A2b

Two approaches were used.

•GMW data (1996–2016): When estimating recent changes in

mangrove area, an exponential model returned a smaller RMSE

than a linear model (Figure S5), so the exponential model was used

to extrapolate predicted mangrove loss to 2046. The analysis

showed that the estimated mangrove area declined from 2230 km

in 1996 to 1857 km

in 2016 (Figure S5). This model estimated

that 36.5% of mangroves will be lost by 2046, with an upper

bound of 45.8% and a lower bound of 25.4%. With these

estimates, the ecosystem meets the criteria for Vulnerable (range:

Least Concern–Vulnerable) under Criterion A2b (Table 2).

•Landsat dense time series: The classification model applied to the

1132 Landsat mosaics achieved an overall accuracy of 95.8%, with

the mangrove class having a user's accuracy of 92.5% and

producer's accuracy of 72.3% (Table 3).

The GAMM with the best fit included cloud cover, time and day

of year all as non-linear explanatory variables. It also included the

satellite sensor as a random effect, a variance structure controlled by

the proportion of the mosaic that is cloud free and a temporal

correlation structure of 0.2 between each consecutive time step

(Figure 5; additional information on model selection in the Supporting

Information). When the model results are extrapolated to 2038, a best

estimate of 35.8% of mangrove extent will be lost using an absolute

rate of decline, while 33.4% of mangrove extent will be lost using a

proportional rate of decline (Figure 6a; Bland et al., 2017). Under the

worst case scenario, 73.3% of mangrove extent will be lost using an

absolute rate of decline, while 62.8% of mangrove extent will be lost

using a proportional rate of decline (Figure 6b). Based on these

results, the ecosystem meets the criteria for Vulnerable (Least

Concern–Endangered) under Criterion A2b (Table 2).

FIGURE 4 Estimated area of Rakhine mangroves estimated based

on (a) a fitted exponential model (n=6) and (b) a linear model (n=6),

with 95% confidence intervals shaded in dark grey. Area estimates are

based on the Army Map Services for Burma, Global Mangrove map

for 2000 from Giri et al. (2011), Global Mangrove Watch maps for

1996 and 2016 and a classified map based on Landsat 8 for the 2019

dry season. RMSE, root mean square error.

TABLE 2 Results for the Red List of Ecosystems assessment for all sub-criteria for Rakhine mangroves.

Criterion

Declining

distribution (A)

Restricted

distribution (B)

Environmental

degradation (C)

Biotic

disruption (D)

Quantitative risk

analysis (E)

Overall

ecosystem status

Sub-criterion 1 VU (LC–VU) LC NE DD NE CR (VU–CR)

Sub-criterion 2a NE LC LC

Sub-criterion 2b VU (LC–VU)

VU (LC–EN)

NE LC–VU

Sub-criterion 3 CR (VU–CR)

EN (VU–EN)

LC NE LC–VU

Note: Categories in brackets indicate plausible bounds. Two approaches were used to assess each of Criteria A2b and A3; both are reported.

Abbreviations: CR, critically endangered; DD, data deficient; EN, endangered; LC, least concern; NE, not evaluated; VU, vulnerable.

Sub-criterion 1 assesses a criterion over the past 50 years; Sub-criterion 2a assesses a criterion over the next 50 years; Sub-criterion 2b assesses a

criterion over any 50-year period, including the past, present and future; Sub-criterion 3 assesses a criterion's historical change since approximately 1750.

Results from Murray et al. (2020).

10 LEE ET AL.

3.1.3 | Criterion A3

Based on Storey (2015), only an estimated 6% of the historical

(1750s) mangrove extent at Rakhine remained in 2015. The second

analysis, combining the data from the collated maps (Section 2.2.1)

and assuming a linear rate of decline, estimated that 82.5% of

historical mangrove extent has been lost under the worst case

scenario, 62.7% has been lost under the best case scenario, with the

best estimate of 77.0% historical mangrove extent lost since 1750.

Without additional information that can reduce this uncertainty

further, these results were combined, following the precautionary

principle and keeping the best estimate from the previous assessment

(and a higher risk category), returning a status of Critically Endangered

under Criterion A3 (range: Vulnerable–Critically Endangered),

highlighting the high degree of uncertainty that remained (Table 2).

3.2 |Criterion B

The EOO of Rakhine mangroves was 54,874 km

. AOO was

estimated as 246 10 10 km grid cells. Additionally, there is evidence

of an ongoing decline in ecosystem extent, thus fulfilling Sub-criteria a

and b required to list the ecosystem under Criteria B1 and B2. Based

on these results, the ecosystem meets the criteria for Least Concern

under Criteria B1 and B2 (Table 2).

3.3 |Criterion C

Based on the sea-level rise model by Lovelock et al. (2015), under an

extreme scenario of 1.4-m relative sea-level rise in the region, 2.3% of

the mangrove area is predicted to be lost by 2100. However, it is

important to note the limitations of the model by Lovelock et al.

(2015). First, it did not account for mangrove plasticity or adaptation

and therefore may overestimate ecosystem risk as mangroves have

been shown to be resilient to sea-level rise (Duncan et al., 2018). On

the other hand, the low relative sea-level rise projected by the model

is due to a high concentration of total suspended matter in the water

column in the region, though river dams may reduce this in the future,

potentially leading to faster rates of sea-level rise.

Regardless of these uncertainties, the overall estimated mangrove

area loss is very low, and with this estimated level of relative sea-level

rise, along with the assumption that this leads to a relative severity of

>80% due to mangrove drowning, the ecosystem is assessed as Least

Concern under Criterion C2a (Table 2).

3.4 |Criterion D

Two results were used to assess Rakhine mangroves under Criterion

D. First, Worthington and Spalding (2018) concluded that 30.4% of

Rakhine mangroves will become degraded by 2050 when using the

TABLE 3 Error matrix of the dense

time-series classification model based on

stratified random sampling with

proportional allocation.

Mangrove Not mangrove Total (Wi) User's accuracy (%)

Mangrove 0.091 0.007 0.098 92.5

Not mangrove 0.035 0.867 0.902 96.1

Total 0.126 0.874 1

Producer's accuracy (%) 72.3 99.2 Overall accuracy: 95.8%

Note: Cell entries represent the proportion of the area. Mapped categories are in rows while the

reference categories are in columns.

FIGURE 5 Mangrove extent for Rakhine

mangroves estimated by the generalized

additive mixed model (GAMM), with 95%

confidence intervals shaded in grey. Symbols

represent the satellite that captured each

data point. Darker symbols represent images

with <20% cloud cover. The histogram at the

bottom shows the number of satellite images

available at 2-year intervals.

LEE ET AL.11

year 2000 as a baseline. The ecosystem was assessed as between

Least Concern and Vulnerable under Criterion D2b (Table 2),

depending on assumptions about the relative severity of degradation.

Second, based on the developed mangrove degradation model,

49.4% (Standard Error [SE] 1.2) of Rakhine mangroves were mapped

as degraded in 2014 and 49.6% (SE 1.1) in 2019 (Figure 7). Spatially,

degradation can be observed throughout the ecosystem. The

Wunbaik reserved forest was an exception within the ecosystem,

remaining as relatively intact mangroves though some degradation

can be observed along the edge of the forest reserve (Figure 7).

The 2014 and 2019 models had overall accuracies of 84.4% and

88.4%, respectively (Table 4). Based on the observed severity of

degradation of the training points, pixels classified as degraded were

assumed to have a plausible relative severity range from 50% to 99%.

If a historical baseline is used, assuming none of the ecosystem was

degraded in the 1750s, the ecosystem is classified as Least Concern if

the relative severity of the mapped degradation is less than 90%.

However, if a relative severity greater than 90% is used instead, the

ecosystem is classified as Vulnerable under Criterion D3 (Table 2;

Figure 8).

4|DISCUSSION

Mangrove loss in Myanmar has been well documented over the years

(De Alban et al., 2020; Murray et al., 2020; Veettil et al., 2018). While

the focused reassessment of Myanmar's coastal ecosystem, ‘Rakhine

mangrove forest on mud’, concluded that the ecosystem should

remain classified as Critically Endangered due to extensive historical

loss, the use of several new methods demonstrated a range of novel

uses for remote sensing for assessing risks to mangrove ecosystems.

The importance of mangrove ecosystems has been recognized as

a global priority for conservation (Friess et al., 2020). Despite this,

Myanmar remains one of the countries with the highest rates of

mangrove loss, with much of that degradation focused on the Rakhine

coastline. In Rakhine, mangroves provide essential ecosystem services

for the people living here by helping reduce the impacts of tropical

cyclones affecting the area; acting as the main source of fuel and

energy for the local people; providing vital fish, crab and shellfish

nurseries; and sequestering and storing carbon (Storey, 2015;

Zöckler & Aung, 2019). The assessment highlighted that the

ecosystem continues to be threatened and has a high risk of collapse

without conservation interventions to reduce ecosystem degradation

and land conversion as a result of expanding agriculture and

aquaculture and overexploitation of mangrove trees for firewood

and timber (Zöckler & Aung, 2019). Despite this high risk of

ecosystem collapse, the results also show that the rate of recent

mangrove area loss may be slowing in the past two to three decades,

with the results echoing a recent assessment conducted for the

neighbouring Indian Sundarbans mangroves where they also reported

reduced rates of loss (Sievers et al., 2020). Along with a reduced rate

of mangrove loss in recent years, the Wunbaik reserved forest

remained a region with relatively intact mangroves compared to the

rest of the ecosystem, though some degradation was still observed

here, particularly along the edge of the forest reserve. This suggests

that while the protected area managed to offer protection to the

mangroves by limiting direct mangrove deforestation, recent

developments have led to increased encroachment of degradation

into the forest reserve from surrounding areas. Stronger enforcement,

such as reducing the surrounding shrimp farms and restricting their

further expansion or reducing woodcutting by locals (Saw &

Kanzaki, 2015), will thus be required in the near future to ensure the

continued maintenance and survival of the mangrove ecosystem in

this area. It is important to acknowledge that ongoing political and

social unrest in Myanmar and Rakhine may force more people into

poverty and dependency on unsustainable harvesting and agricultural

practices (Ware, 2015). Mangrove conservation in Rakhine will

require a combination of socio-economic solutions aiding the people

here, sustainable use of mangroves, national land use policies that

take into account the increasing population in the region, increased

protection of the remaining intact mangrove forests by reducing and

restricting shrimp farm development and potential plans to restore

degraded mangroves through rehabilitation of abandoned shrimp

ponds (Maung & Sasaki, 2020; Oo, 2002; Veettil et al., 2018).

FIGURE 6 Predicted mangrove extent estimate in 2038 based on

the absolute rate of decline (dotted line) and proportion rate of

decline (dashed line), assuming (a) best estimate and (b) worst case

scenarios.

12 LEE ET AL.

Previous research has shown the importance of reassessments

for ecosystems and species (Cazalis et al., 2022); >750 species had

their Red List of Threatened Species status changed between 2019

and 2020 alone (IUCN, 2020). The reassessment of Rakhine

mangroves provided a more complete analysis of its risk of collapse

by incorporating additional lines of evidence (Bland et al., 2017)

and highlighted several sources of uncertainty that can be

minimized and quantitatively reported with satellite remote sensing

methods. Using a dense time-series of satellite imagery to classify

ecosystem areal change over 30 years allows non-linear trends and

quantitative uncertainty to be modelled that would otherwise be

impossible to present or be reported (Foody, 2010; Lee

et al., 2021), and using an explicit ecosystem conceptual model

developed for this ecosystem allows clear specification of the

potential drivers of ecosystem degradation here (Lee et al., 2021).

Despite this, uncertainty in the assessment outcome remained,

particularly with regard to the historical area change of the

ecosystem, as there are multiple estimates of the ecosystem's

FIGURE 7 Spatial summary of

degradation of Rakhine mangroves, showing

the percentage of 1 1 km grid cells

classified as degraded in (a) 2014 and (b)

2019.

TABLE 4 Error matrix of the Rakhine mangroves' degradation

models.

Degraded Intact User's accuracy

2014

Degraded 41.9% 8.1% 83.8%

Intact 7.5% 42.5% 85.0%

Producer's accuracy 84.9% 84.0% Overall accuracy:

84.4%

2019

Degraded 45.4% 7.4% 85.9%

Intact 4.2% 42.9% 91.1%

Producer's accuracy 91.5% 85.2% Overall accuracy:

88.4%

Note: Cell entries represent the percentage of the total area. Map

categories are in rows while the reference categories are in columns.

FIGURE 8 Plausible bounds for the extent of Rakhine mangroves'

degradation and relative severities under Criterion D3. CR, critically

endangered; EN, endangered; VU, vulnerable.

LEE ET AL.13

historical extent that are impossible to verify. Both estimates included

important assumptions that were unlikely to be met. The data from

Storey (2015) were estimates based on interviews with villagers in

Northern Rakhine, where the most extensive mangrove degradation

was observed, suggesting that relying on these estimates may

overestimate mangrove loss if extrapolated to the entire ecosystem.

The alternative estimate of historical mangrove extent was based on

linear back-extrapolation to more than 200 years ago from 80 years

of data. This required an assumption that there has been a constant

rate of deforestation over 250 years, as additional information on the

shape of the trajectory is not available, which is extremely unlikely

(Armenteras et al., 2017). Assessing trends in ecosystem areas over the

past 250 years, as required by the Red List guidelines, will likely be

challenging for many assessments. The AMS maps from the 1940s

were used as the earliest available data available to reduce uncertainty,

but a 170-year period without data still exists. To produce more reliable

estimates, physical and biological factors known to explain land use and

deforestation patterns can be used to parameterize deforestation

models (Brown et al., 2007), or distribution models with environmental

variables to estimate the expected distribution of the ecosystem in the

absence of anthropogenic effects can produce other points of

comparison (Keith et al., 2013). Ultimately, avenues to reduce the

uncertainty of estimating area trend over such a long period of time

was pursued within this study, and while some uncertainty remained

due to data paucity, an outcome of Critically Endangered remained as

the Red List suggests reporting the highest risk category reached based

on the precautionary principle (Bland et al., 2017).

With the continued development of satellite data and methods to

analyse data, there are still gaps that can be filled for more complete

Red List assessments. For example, Landsat satellites provided

continuous data beginning 30 years ago, allowing for statistical

modelling of ecosystem extent, which is otherwise impossible. As

time goes on, more satellite data will be collected to fill the 50-year

time frame required by the Red List, further reducing the need for

extrapolation and the uncertainty of the results. Additionally, data at

higher spatial and temporal resolution than Landsat, such as the

PlanetScope constellation of satellites (Planet Team, 2017), can be

added as additional sources of information as they become more

readily available. For example, PlanetScope data are able to accurately

characterize forest phenology at the tree-crown scale (Wu et al., 2021),

an important indicator signalling ecosystem change under climate

change (Ettinger et al., 2022). Historical satellite or aerial photography

that is being declassified offers another opportunity for providing long

time-series data to track historical ecosystem change (Nita et al., 2018),

similar to how the US AMS data were used in this study. However,

while historical imagery is an irreplaceable source of ecosystem

information, it is also difficult to assess the accuracy of these maps due

to their age, and it must be used with caution.

Using a dense time-series model to estimate ecosystem extent

trends enables the full use of the information available from regularly

collected satellite imagery. While attempts to assess the accuracy of

the dense time-series classification maps (Table 3) were made, these

relied on creating validation points where pixels were assumed to be

either always mangroves or never mangroves. Unfortunately, this

means that pixels where mangrove gain or loss occurred were not

assessed, and it is likely that these pixels are also more prone to

misclassification. Regardless, the best available data and methods

were used here to produce a quantitative estimate of ecosystem area

trends, which is invaluable for producing informative risk assessments

with significance testing.

While remote sensing offers unique opportunities to tracking

ecosystem change, it is important to note that the analyses that were

done for the reassessment did not include additional field data. When

estimating ecosystem degradation, the lack of field data limited the

results to only two possible classes of degradation. This method also

quantified mangrove degradation into a single, relatively simple

indicator despite the many contributing factors that can drive and

represent degradation (Yando et al., 2021; Figure 3). Quantitative field

measurements of environmental and biotic degradation will be

needed to estimate a continuous metric for the relative severity of

ecosystem degradation as required by the Red List. Furthermore,

combining the mangrove conceptual model with a more

conceptualized degradation framework based on field data will allow

us to paint a more holistic picture of the entire mangrove ecosystem,

its status and the threats that may affect it.

The results from this study highlight the benefits of incorporating

new data and methods that are released and developed to conduct

detailed, strategic reassessments of ecosystem risk. Not only will this

improve the accuracy of the risk assessments and fill previously

existing knowledge gaps, it can also identify further data and

methodological gaps to help guide future research. A template for

future mangrove Red Lists using satellite data is presented here. As

the Red List of Ecosystems continues to increase in prominence,

reassessments encourage assessors to consider the latest data and

methods available, ensuring the results are based on the most up-

to-date information available.

5|CONSERVATION IMPLICATIONS

By extending the assessment of the ‘Rakhine mangrove forest on

mud’with additional data and analyses, some of the uncertainties

involved in the first assessment were reduced. Uncertainty regarding

the historical change (since 1750) in ecosystem area remained, and

the ecosystem is still assessed as Critically Endangered (with a

plausible range between Vulnerable and Critically Endangered) to

follow the precautionary principle where the overall status of the

ecosystem is the highest risk category obtained (Bland et al., 2017).

While there is evidence that recent rates of mangrove losses are

slowing, coinciding with the reducing rates of mangrove deforestation

observed globally (Hamilton & Casey, 2016), this could be due to the

small amount of mangroves remaining, meaning there is less left to be

cleared. Successful recovery and restoration of the ecosystem to a

lower risk level will require increasing the current coverage of

mangrove area back towards the historical extent of the ecosystem

(at least 10% of its historical distribution in the 1750s).

14 LEE ET AL.

In addition to areal loss, there is also considerable mangrove

degradation observed, including around and within protected areas.

Within the existing remaining mangrove forests, degradation can lead

to declining ecosystem quality (Yando et al., 2021), and Rakhine

mangroves will still be classified as Vulnerable even in the absence of

the observed mangrove loss. This indicates that the establishment

of additional protected areas alone may not be sufficient to protect the

ecosystem. For example, the Wunbaik reserved forest is one of

the largest remaining protected mangrove forests in the region but is

suffering from degradation due to expansion of paddy fields and shrimp

ponds, along with illegal woodcutting by locals (Saw & Kanzaki, 2015), a

situation similar to the neighbouring Sundarbans (Roy, 2016). To

address the issue of increased mangrove exploitation, multiple key

points need to be addressed and solved. The main incentive identified

for locals to switch from less damaging subsistence fishing to more

destructive shrimp farming is to improve their financial situation, as no

alternatives are currently available to them (Saw & Kanzaki, 2015).

Thus, alternative financial incentives and sources of income and

livelihoods for the locals will need to be developed and enhanced as a

means of mangrove conservation. Moreover, improving local

participation in mangrove management has been shown to be effective

in reducing long-term conflicts between local communities and the

government agencies in charge of management, improving local

resilience to sudden disasters and participants' livelihoods and

imparting a sense of security and community (Islam et al., 2013). Lastly,

restoration of abandoned shrimp ponds also provides an opportunity to

improve the overall status of the ecosystem, reducing degradation. This

is already occurring naturally in some areas (Maung & Sasaki, 2020),

though further investments can potentially lead to further, more

targeted mangrove recovery and become a source of income for the

local communities (Damastuti & de Groot, 2017). Successful mangrove

restoration is an ongoing challenge that requires more than just

planting mangroves, and the proper allocation of resources is essential

to ensure local support. Mechanisms that are fair and equitable,

providing a potential source of income to the local communities

through sustainable mangrove use, are needed to ensure the long-term

sustainability of the mangrove ecosystem and the people that reside

next to it (Lovelock & Brown, 2019).

AUTHOR CONTRIBUTIONS

Calvin K. F. Lee: Conceptualization; methodology; visualization;

writing—review and editing; writing—original draft; validation; formal

analysis. Emily Nicholson: Conceptualization; supervision; project

administration; writing—review and editing; funding acquisition. Clare

Duncan: Conceptualization; writing—review and editing; supervision;

methodology. Hedley S. Grantham: Writing—review and editing.

David A. Keith: Writing—review and editing; conceptualization. Rob

Tizard: Writing—review and editing. Nicholas J. Murray: Writing—

review and editing; methodology; conceptualization; supervision.

ACKNOWLEDGEMENTS

The authors acknowledge the contributions of Adam Duncan, Win

Thuya Htut, Nyan Hlaing, Aung Htat Oo and Kyaw Zay Ya for their

work in building the foundation for the Red List of Ecosystems in

Myanmar. CKFL was in part supported by the Australian Government

Research Training Program Scholarship, HKU seed fund for basic

research (#202011159154) and the HKU 45th-round PDF scheme.

We acknowledge funding from the Australian Research Council

(FT190100234 to EN; LP170101143 to EN, HSG, DAK and NJM).

The maps from the Army Map Service (AMS) were downloaded from

the University of Texas Libraries.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the

corresponding author upon reasonable request.

ORCID

Calvin K. F. Lee https://orcid.org/0000-0001-8277-8614

Rob Tizard https://orcid.org/0000-0002-9815-4117

REFERENCES

Alongi, D.M. (2015). The impact of climate change on mangrove forests.

Current Climate Change Reports, 1(1), 30–39. https://doi.org/10.1007/

s40641-015-0002-x

Armenteras, D., Espelta, J.M., Rodríguez, N. & Retana, J. (2017).

Deforestation dynamics and drivers in different forest types in Latin

America: three decades of studies (1980–2010). Global Environmental

Change, 46, 139–147. https://doi.org/10.1016/j.gloenvcha.2017.

09.002

Aung, U. (2007). Policy and practice in Myanmar's protected area system.

Journal of Environmental Management, 84(2), 188–203. https://doi.org/

10.1016/j.jenvman.2006.05.016

Bland, L.M., Keith, D.A., Miller, R.M., Murray, N.J. & Rodríguez, J.P. (2017).

Guidelines for the application of IUCN Red List of Ecosystems categories

and criteria, version 1.1. Gland, Switzerland: IUCN.

Bland, L.M., Nicholson, E., Miller, R.M., Andrade, A., Carré, A., Etter, A.

et al. (2019). Impacts of the IUCN Red List of Ecosystems on

conservation policy and practice. Conservation Letters, 40(5), e12665.

https://doi.org/10.1111/conl.12666

Bland, L.M., Regan, T.J., Dinh, M.N., Ferrari, R., Keith, D.A., Lester, R. et al.

(2017). Using multiple lines of evidence to assess the risk of

ecosystem collapse. Proceedings of the Royal Society B, 284(1863),

20170660. https://doi.org/10.1098/rspb.2017.0660

Brown, S., Hall, M., Andrasko, K., Ruiz, F., Marzoli, W., Guerrero, G. et al.

(2007). Baselines for land-use change in the tropics: application to

avoided deforestation projects. Mitigation and Adaptation Strategies for

Global Change, 12(6), 1001–1026. https://doi.org/10.1007/s11027-

006-9062-5

Bunting, P., Rosenqvist, A., Lucas, R., Rebelo, L.-M., Hilarides, L.,

Thomas, N. et al. (2018). The Global Mangrove Watch—a new 2010

global baseline of mangrove extent. Remote Sensing, 10(10), 1669.

https://doi.org/10.3390/rs10101669

Cardinale, B.J., Duffy, J.E., Gonzalez, A., Hooper, D.U., Perrings, C.,

Venail, P. et al. (2012). Biodiversity loss and its impact on humanity.

Nature, 486(7401), 59–67. https://doi.org/10.1038/nature11148

Cazalis, V., Di Marco, M., Butchart, S.H.M., Akçakaya, H.R.,

González-Suárez, M., Meyer, C. et al. (2022). Bridging the research-

implementation gap in IUCN Red List assessments. Trends in

Ecology & Evolution, 37(4), 359–370. https://doi.org/10.1016/j.tree.

2021.12.002

LEE ET AL.15

Clough, J., Park, R., Propato, M., Polaczyk, A., Brennan, M., Behrens, D.,

et al. (2016). SLAMM 6.7 technical documentation.

Damastuti, E. & de Groot, R. (2017). Effectiveness of community-based

mangrove management for sustainable resource use and livelihood

support: a case study of four villages in Central Java, Indonesia.

Journal of Environmental Management, 203, 510–521. https://doi.org/

10.1016/j.jenvman.2017.07.025

De Alban, J.D.T., Jamaludin, J., Wong de Wen, D., Than, M.M. &

Webb, E.L. (2020). Improved estimates of mangrove cover and change

reveal catastrophic deforestation in Myanmar. Environmental Research

Letters, 15(3), 034034. https://doi.org/10.1088/1748-9326/ab666d

Doyle, T.W., Girod, G.F. & Books, M.A. (2003). Modeling mangrove forest

migration along the southwest coast of Florida under climate change.

In: Ning, Z., Turner, R.E., Doyle, T.W., Abdollahi, K., Thornton, A.,

Reyes, E. et al. (Eds.) Integrated assessment of the climate change

impacts on the Gulf Coast region. Lafayette, LA: USGS, pp. 211–222.

Duncan, C., Owen, H.J.F., Thompson, J.R., Koldewey, H.J.,

Primavera, J.H. & Pettorelli, N. (2018). Satellite remote sensing to

monitor mangrove forest resilience and resistance to sea level rise.

Methods in Ecology and Evolution, 9(8), 1837–1852. https://doi.org/10.

1111/2041-210X.12923

Estoque, R.C., Myint, S.W., Wang, C., Ishtiaque, A., Aung, T.T., Emerton, L.

et al. (2018). Assessing environmental impacts and change in

Myanmar's mangrove ecosystem service value due to deforestation

(2000–2014). Global Change Biology, 24(11), 5391–5410. https://doi.

org/10.1111/gcb.14409

Ettinger, A.K., Chamberlain, C.J. & Wolkovich, E.M. (2022). The increasing

relevance of phenology to conservation. Nature Climate Change, 12(4),

305–307. https://doi.org/10.1038/s41558-022-01330-8

Foody, G.M. (2010). Assessing the accuracy of land cover change with

imperfect ground reference data. Remote Sensing of Environment,

114(10), 2271–2285. https://doi.org/10.1016/j.rse.2010.05.003

Friess, D.A., Rogers, K., Lovelock, C.E., Krauss, K.W., Hamilton, S.E.,

Lee, S.Y. et al. (2019). The state of the world's mangrove forests: past,

present, and future. Annual Review of Environment and Resources, 44(1),

89–115. https://doi.org/10.1146/annurev-environ-101718-033302

Friess, D.A., Yando, E.S., Abuchahla, G.M.O., Adams, J.B., Cannicci, S.,

Canty, S.W.J. et al. (2020). Mangroves give cause for conservation

optimism, for now. Current Biology, 30(4), R153–R154. https://doi.org/

10.1016/j.cub.2019.12.054

Geist, S.J., Nordhaus, I. & Hinrichs, S. (2012). Occurrence of species-rich

crab fauna in a human-impacted mangrove forest questions the

application of community analysis as an environmental assessment

tool. Estuarine, Coastal and Shelf Science, 96, 69–80. https://doi.org/

10.1016/j.ecss.2011.10.002

Giri, C., Ochieng, E., Tieszen, L.L., Zhu, Z., Singh, A., Loveland, T. et al.

(2011). Status and distribution of mangrove forests of the world using

earth observation satellite data. Global Ecology and Biogeography,

20(1), 154–159. https://doi.org/10.1111/j.1466-8238.2010.00584.x

Goldberg, L., Lagomasino, D., Thomas, N. & Fatoyinbo, T. (2020). Global

declines in human-driven mangrove loss. Global Change Biology,

26(10), 5844–5855. https://doi.org/10.1111/gcb.15275

Gorelick, N., Hancher, M., Dixon, M., Ilyushchenko, S., Thau, D. &

Moore, R. (2017). Google Earth Engine: planetary-scale geospatial

analysis for everyone. Remote Sensing of Environment, 202, 18–27.

https://doi.org/10.1016/j.rse.2017.06.031

Hamilton, S.E. & Casey, D. (2016). Creation of a high spatio-temporal

resolution global database of continuous mangrove forest cover for

the 21st century (CGMFC-21). Global Ecology and Biogeography, 25(6),

729–738. https://doi.org/10.1111/geb.12449

IPBES. (2019). Summary for policymakers of the global assessment report

on biodiversity and ecosystem services. Intergovernmental Science-

Policy Platform on Biodiversity and Ecosystem Services.

Islam, K.K., Rahman, G.M., Fujiwara, T. & Sato, N. (2013). People's

participation in forest conservation and livelihoods improvement:

experience from a forestry project in Bangladesh. International Journal

of Biodiversity Science, Ecosystem Services & Management, 9(1), 30–43.

https://doi.org/10.1080/21513732.2012.748692

IUCN. (2012). Habitats classification scheme (version 3.1).

IUCN. (2020). The IUCN Red List of threatened species.

Keith, D.A., Ferrer-Paris, J.R., Nicholson, E. & Kingsford, R.T. (2020). IUCN

global ecosystem typology 2.0.

Keith, D.A., Rodríguez, J.P., Brooks, T.M., Burgman, M.A., Barrow, E.G.,

Bland, L. et al. (2015). The IUCN Red List of Ecosystems: motivations,

challenges, and applications. Conservation Letters, 8(3), 214–226.

https://doi.org/10.1111/conl.12167

Keith, D.A., Rodríguez, J.P., Rodríguez-Clark, K.M., Nicholson, E.,

Aapala, K., Alonso, A. et al. (2013). Scientific foundations for an IUCN

Red List of Ecosystems. PLoS ONE, 8(5), e62111. https://doi.org/10.

1371/journal.pone.0062111

Kovacs, J.M., Flores-Verdugo, F., Wang, J. & Aspden, L.P. (2004).

Estimating leaf area index of a degraded mangrove forest using high

spatial resolution satellite data. Aquatic Botany, 80(1), 13–22. https://

doi.org/10.1016/j.aquabot.2004.06.001

Krauss, K.W., McKee, K.L., Lovelock, C.E., Cahoon, D.R., Saintilan, N.,

Reef, R. et al. (2014). How mangrove forests adjust to rising sea level.

New Phytologist, 202(1), 19–34. https://doi.org/10.1111/nph.12605

Lee, C.K.F., Duncan, C., Nicholson, E., Fatoyinbo, T.E., Lagomasino, D.,

Thomas, N. et al. (2021). Mapping the extent of mangrove ecosystem

degradation by integrating an ecological conceptual model with

satellite data. Remote Sensing, 13(11), 2047. https://doi.org/10.3390/

rs13112047

Lee, C.K.F., Keith, D.A., Nicholson, E. & Murray, N.J. (2019). Redlistr: tools

for the IUCN Red Lists of Ecosystems and Threatened Species in R.

Ecography, 42(5), 1050–1055. https://doi.org/10.1111/ecog.04143

Lee, C.K.F., Nicholson, E., Duncan, C. & Murray, N.J. (2021). Estimating

changes and trends in ecosystem extent with dense time-series

satellite remote sensing. Conservation Biology, 35(1), 325–335. https://

doi.org/10.1111/cobi.13520

Lee, S.Y., Primavera, J.H., Dahdouh-Guebas, F., McKee, K., Bosire, J.O.,

Cannicci, S. et al. (2014). Ecological role and services of tropical

mangrove ecosystems: a reassessment. Global Ecology and

Biogeography, 23(7), 726–743. https://doi.org/10.1111/geb.12155

Leimgruber, P., Kelly, D.S., Steininger, M.K., Brunner, J., Müller, T. &

Songer, M. (2005). Forest cover change patterns in Myanmar (Burma)

1990–2000. Environmental Conservation, 32(4), 356–364. https://doi.

org/10.1017/S0376892905002493

Lovelock, C.E. & Brown, B.M. (2019). Land tenure considerations are key

to successful mangrove restoration. Nature Ecology & Evolution, 3(8),

1135. https://doi.org/10.1038/s41559-019-0942-y

Lovelock, C.E., Cahoon, D.R., Friess, D.A., Guntenspergen, G.R.,

Krauss, K.W., Reef, R. et al. (2015). The vulnerability of Indo-Pacific

mangrove forests to sea-level rise. Nature, 526(7574), 559–563.

https://doi.org/10.1038/nature15538

Lucas, R., Van De Kerchove, R., Otero, V., Lagomasino, D., Fatoyinbo, L.,

Omar, H. et al. (2020). Structural characterisation of mangrove forests

achieved through combining multiple sources of remote sensing data.

Remote Sensing of Environment, 237, 111543. https://doi.org/10.

1016/j.rse.2019.111543

Marshall, A., Schulte to Bühne, H., Bland, L. & Pettorelli, N. (2018).

Assessing ecosystem collapse risk in ecosystems dominated by

foundation species: the case of fringe mangroves. Ecological Indicators,

91, 128–137. https://doi.org/10.1016/j.ecolind.2018.03.076

Maung, W.S. & Sasaki, J. (2020). Assessing the natural recovery of

mangroves after human disturbance using neural network

classification and Sentinel-2 imagery in Wunbaik Mangrove Forest,

Myanmar. Remote Sensing, 13(1), 52. https://doi.org/10.3390/

rs13010052

Murray, N.J., Clemens, R.S., Phinn, S.R., Possingham, H.P. & Fuller, R.A.

(2014). Tracking the rapid loss of tidal wetlands in the Yellow Sea.

16 LEE ET AL.

Frontiers in Ecology and the Environment, 12(5), 267–272. https://doi.

org/10.1890/130260

Murray, N.J., Keith, D.A., Duncan, A., Tizard, R., Ferrer-Paris, J.R.,

Worthington, T.A. et al. (2020). Myanmar's terrestrial ecosystems:

status, threats and conservation opportunities. Biological Conservation,

252, 108834. https://doi.org/10.1016/j.biocon.2020.108834

Myint, W. & Stanley, D.O. (2011). The mangrove vegetation of Wunbaik

Reserved Forest. Yangon, Myanmar: FAO-UN.

Nita, M.D., Munteanu, C., Gutman, G., Abrudan, I.V. & Radeloff, V.C.

(2018). Widespread forest cutting in the aftermath of World War II

captured by broad-scale historical Corona spy satellite photography.

Remote Sensing of Environment, 204, 322–332. https://doi.org/10.

1016/j.rse.2017.10.021

Olofsson, P., Foody, G.M., Herold, M., Stehman, S.V., Woodcock, C.E. &

Wulder, M.A. (2014). Good practices for estimating area and assessing

accuracy of land change. Remote Sensing of Environment, 148, 42–57.

https://doi.org/10.1016/j.rse.2014.02.015

Oo, N. (2002). Present state and problems of mangrove management in

Myanmar. Trees, 16(2–3), 218–223. https://doi.org/10.1007/s00468-

001-0150-6

Peters, R., Walther, M., Lovelock, C., Jiang, J. & Berger, U. (2020). The

interplay between vegetation and water in mangroves: new

perspectives for mangrove stand modelling and ecological research.

Wetlands Ecology and Management, 28(4), 697–712. https://doi.org/

10.1007/s11273-020-09733-0

Planet Team. (2017). Planet application program interface: in space for life

on Earth.

Ramaswamy, V. & Rao, P.S. (2014). Chapter 17: the Myanmar continental

shelf. Geological Society, London, Memoirs, 41(1), 231–240. https://doi.

org/10.1144/M41.17

Richards, D.R. & Friess, D.A. (2016). Rates and drivers of mangrove

deforestation in Southeast Asia, 2000–2012. Proceedings of the

National Academy of Sciences, 113(2), 344–349. https://doi.org/10.

1073/pnas.1510272113

Rodríguez, J.P., Keith, D.A., Rodríguez-Clark, K.M., Murray, N.J.,

Nicholson, E., Regan, T.J. et al. (2015). A practical guide to the

application of the IUCN Red List of Ecosystems criteria. Philosophical

Transactions of the Royal Society B: Biological Sciences, 370(1662),

20140003. https://doi.org/10.1098/rstb.2014.0003

Roy, A.K.D. (2016). Local community attitudes towards mangrove forest

conservation: lessons from Bangladesh. Marine Policy, 74, 186–194.

https://doi.org/10.1016/j.marpol.2016.09.021

Saw, A.A. & Kanzaki, M. (2015). Local livelihoods and encroachment into a

mangrove forest reserve: a case study of the Wunbaik Reserved

Mangrove Forest, Myanmar. Procedia Environmental Sciences, 28, 483–

492. https://doi.org/10.1016/j.proenv.2015.07.058

Sen Roy, N. & Kaur, S. (2000). Climatology of monsoon rains of Myanmar

(Burma). International Journal of Climatology, 20(8), 913–928. https://

doi.org/10.1002/1097-0088(20000630)20:8%3C913::AID-JOC485%

3E3.0.CO;2-U

Sievers, M., Chowdhury, M.R., Adame, M.F., Bhadury, P., Bhargava, R.,

Buelow, C. et al. (2020). Indian Sundarbans mangrove forest

considered endangered under Red List of Ecosystems, but there is

cause for optimism. Biological Conservation, 251, 108751. https://doi.

org/10.1016/j.biocon.2020.108751

Spalding, M.D., Fox, H.E., Allen, G.R., Davidson, N., Ferdaña, Z.A.,

Finlayson, M. et al. (2007). Marine ecoregions of the world: a

bioregionalization of coastal and shelf areas. Bioscience, 57(7), 573–

583. https://doi.org/10.1641/B570707

Stanley, D.O. & Broadhead, J.S. (2011). Integrated mangrove management

plan for Wunbaik Reserved Forest, Myanmar.

Storey, D. (2015). A socio-economic assessment of mangroves areas in

North Rakhine State.

Veettil, B.K., Pereira, S.F.R. & Quang, N.X. (2018). Rapidly diminishing

mangrove forests in Myanmar (Burma): a review. Hydrobiologia, 822(1),

19–35. https://doi.org/10.1007/s10750-018-3673-1

Verbesselt, J., Umlauf, N., Hirota, M., Holmgren, M., Van Nes, E.H.,

Herold, M. et al. (2016). Remotely sensed resilience of tropical forests.

Nature Climate Change, 6(11), 1028–1031. https://doi.org/10.1038/

nclimate3108

Ward, R.D., Friess, D.A., Day, R.H. & MacKenzie, R.A. (2016). Impacts of

climate change on mangrove ecosystems: a region by region overview.

Ecosystem Health and Sustainability, 2(4). https://doi.org/10.1002/

ehs2.1211

Ware, A. (2015). Secessionist aspects to the Buddhist–Muslim conflict in

Rakhine State, Myanmar. In: Kingsbury, D. & Laoutides, C. (Eds.)

Territorial separatism in global politics: Causes, outcomes and resolution.

London: Routledge.

Webb, E.L., Jachowski, N.R.A., Phelps, J., Friess, D.A., Than, M.M. &

Ziegler, A.D. (2014). Deforestation in the Ayeyarwady Delta and the

conservation implications of an internationally-engaged Myanmar.

Global Environmental Change, 24, 321–333. https://doi.org/10.1016/j.

gloenvcha.2013.10.007

Win, S., Towprayoon, S. & Chidthisong, A. (2020). Mangrove status, its

ecosystem, and climate change in Myanmar: a study in Ayeyarwaddy

Delta coastal zone. IOP Conference Series: Earth and Environmental

Science, 496, 012007. https://doi.org/10.1088/1755-1315/496/1/

012007

Wood, S.N. (2017). Generalized additive models: an introduction with R, 2nd

edition. Boca Raton, Florida, USA: CRC Publishing.

Worthington, T. & Spalding, M.D. (2018). Mangrove restoration potential: a

global map highlighting a critical opportunity. Gland, Switzerland: IUCN.

Worthington, T.A., zu Ermgassen, P.S.E., Friess, D.A., Krauss, K.W.,

Lovelock, C.E., Thorley, J. et al. (2020). A global biophysical typology

of mangroves and its relevance for ecosystem structure and

deforestation. Scientific Reports, 10(1). https://doi.org/10.1038/

s41598-020-71194-5

Wu, S., Wang, J., Yan, Z., Song, G., Chen, Y., Ma, Q. et al. (2021).

Monitoring tree-crown scale autumn leaf phenology in a temperate

forest with an integration of PlanetScope and drone remote sensing

observations. ISPRS Journal of Photogrammetry and Remote Sensing,

171, 36–48. https://doi.org/10.1016/j.isprsjprs.2020.10.017

Yando, E.S., Sloey, T.M., Dahdouh-Guebas, F., Rogers, K.,

Abuchahla, G.M.O., Cannicci, S. et al. (2021). Conceptualizing

ecosystem degradation using mangrove forests as a model system.

Biological Conservation, 263, 109355. https://doi.org/10.1016/j.

biocon.2021.109355

Zöckler, C. & Aung, C. (2019). The mangroves of Myanmar. In: Gul, B.,

Böer, B., Khan, M.A., Clüsener-Godt, M. & Hameed, A. (Eds.) Sabkha

ecosystems. Cham: Springer International Publishing, pp. 253–268.

SUPPORTING INFORMATION

Additional supporting information can be found online in the

Supporting Information section at the end of this article.

How to cite this article: Lee, C.K.F., Nicholson, E., Duncan, C.,

Grantham, H.S., Keith, D.A., Tizard, R. et al. (2023). Assessing

the conservation status of mangroves in Rakhine, Myanmar.

Aquatic Conservation: Marine and Freshwater Ecosystems,1–17.

https://doi.org/10.1002/aqc.4058

LEE ET AL.17

Tracking mangrove condition changes using dense Landsat time series

Article

Full-text available

Dec 2024
REMOTE SENS ENVIRON

Scalable Mangrove Monitoring with Limited Field Data: Integrating MREDT and DACN-M

Article

Full-text available

Sep 2024

Mangroves play a crucial ecological and economic role but face significant threats, particularly on Hainan Island, which has the highest mangrove species diversity in China. Remote sensing and AI techniques offer potential solutions for monitoring these ecosystems, but challenges persist due to difficult access for field sampling. To address these issues, we propose a novel model combining a Mangrove Rough Extraction Decision Tree (MREDT) and a Dynamic Attention Convolutional Network (DACN-M). Initially, we used drones and field surveys to conduct multiple observations in Dongzhaigang Nature Reserve, identifying the boundaries of the mangroves. Based on these features, we constructed the MREDT model to mitigate model failure caused by light instability, simplifying transfer to other study areas without requiring annotated samples or extensive field surveys. Next, we developed the DACN-M model, which refines the rough extraction features from MREDT and incorporates contextual information for more accurate detection. Experimental results demonstrate that our proposed method effectively differentiates mangroves from other vegetation, achieving F1 Scores above 75% and IoU values greater than 60% across six study areas. In conclusion, our proposed method not only accurately identifies and monitors mangrove distribution but also offers the significant advantage of being transferable to other study areas without the need for annotated samples or field surveys. This provides a robust and scalable solution for protecting and preserving critical mangrove ecosystems and supports effective conservation efforts in various regions.

Mapping the Extent of Mangrove Ecosystem Degradation by Integrating an Ecological Conceptual Model with Satellite Data

Article

Full-text available

May 2021

Anthropogenic and natural disturbances can cause degradation of ecosystems, reducing their capacity to sustain biodiversity and provide ecosystem services. Understanding the extent of ecosystem degradation is critical for estimating risks to ecosystems, yet there are few existing methods to map degradation at the ecosystem scale and none using freely available satellite data for mangrove ecosystems. In this study, we developed a quantitative classification model of mangrove ecosystem degradation using freely available earth observation data. Crucially, a conceptual model of mangrove ecosystem degradation was established to identify suitable remote sensing variables that support the quantitative classification model, bridging the gap between satellite-derived variables and ecosystem degradation with explicit ecological links. We applied our degradation model to two case-studies, the mangroves of Rakhine State, Myanmar, which are severely threatened by anthropogenic disturbances, and Shark River within the Everglades National Park, USA, which is periodically disturbed by severe tropical storms. Our model suggested that 40% (597 km2) of the extent of mangroves in Rakhine showed evidence of degradation. In the Everglades, the model suggested that the extent of degraded mangrove forest increased from 5.1% to 97.4% following the Category 4 Hurricane Irma in 2017. Quantitative accuracy assessments indicated the model achieved overall accuracies of 77.6% and 79.1% for the Rakhine and the Everglades, respectively. We highlight that using an ecological conceptual model as the basis for building quantitative classification models to estimate the extent of ecosystem degradation ensures the ecological relevance of the classification models. Our developed method enables researchers to move beyond only mapping ecosystem distribution to condition and degradation as well. These results can help support ecosystem risk assessments, natural capital accounting, and restoration planning and provide quantitative estimates of ecosystem degradation for new global biodiversity targets.

Assessing the Natural Recovery of Mangroves after Human Disturbance Using Neural Network Classification and Sentinel-2 Imagery in Wunbaik Mangrove Forest, Myanmar

Article

Full-text available

Dec 2020

In this study, we examined the natural recovery of mangroves in abandoned shrimp ponds located in the Wunbaik Mangrove Forest (WMF) in Myanmar using artificial neural network (ANN) classification and a change detection approach with Sentinel-2 satellite images. In 2020, we conducted various experiments related to mangrove classification by tuning input features and hyper-parameters. The selected ANN model was used with a transfer learning approach to predict the mangrove distribution in 2015. Changes were detected using classification results from 2015 and 2020. Naturally recovering mangroves were identified by extracting the change detection results of three abandoned shrimp ponds selected during field investigation. The proposed method yielded an overall accuracy of 95.98%, a kappa coefficient of 0.92, mangrove and non-mangrove precisions of 0.95 and 0.98, respectively, recalls of 0.96, and F1 scores of 0.96 for the 2020 classification. For the 2015 prediction, transfer learning improved model performance, resulting in an overall accuracy of 97.20%, a kappa coefficient of 0.94, mangrove and non-mangrove precisions of 0.98 and 0.96, respectively, recalls of 0.98 and 0.97, and F1 scores of 0.96. The change detection results showed that mangrove forests in the WMF slightly decreased between 2015 and 2020. Naturally recovering mangroves were detected at approximately 50% of each abandoned site within a short abandonment period. This study demonstrates that the ANN method using Sentinel-2 imagery and topographic and canopy height data can produce reliable results for mangrove classification. The natural recovery of mangroves presents a valuable opportunity for mangrove rehabilitation at human-disturbed sites in the WMF.

Monitoring tree-crown scale autumn leaf phenology in a temperate forest with an integration of PlanetScope and drone remote sensing observations

Article

Full-text available

Jan 2021
ISPRS J PHOTOGRAMM

In temperate forests, autumn leaf phenology signals the end of leaf growing season and shows large variability across tree-crowns, which importantly mediates photosynthetic seasonality, hydrological regulation, and nutrient cycling of forest ecosystems. However, critical challenges remain with the monitoring of autumn leaf phenology at the tree-crown scale due to the lack of spatially explicit information for individual tree-crowns and high (spatial and temporal) resolution observations with nadir view. Recent availability of the PlanetScope constellation with a 3 m spatial resolution and near-daily nadir view coverage might help address these observational challenges, but remains underexplored. Here we developed an integration of PlanetScope with drone observations for improved monitoring of crown-scale autumn leaf phenology in a temperate forest in Northeast China. This integration includes: 1) visual identification of individual tree-crowns (and species) from drone observations; 2) extraction of time series of PlanetScope vegetation indices (VIs) for each identified tree-crown; 3) derivation of three metrics of autumn leaf phenology from the extracted VI time series, including the start of fall (SOF), middle of fall (MOF), and end of fall (EOF); and 4) accuracy assessments of the PlanetScope-derived phenology metrics with reference from local phenocams. Our results show that (1) the PlanetScope-drone integration captures large inter-crown phenological variations, with a range of 28 days, 25 days, and 30 days for SOF, MOF, and EOF, respectively, (2) the extracted crown-level phenology metrics strongly agree with those derived from local phenocams, with a root-mean-square-error (RMSE) of 4.1 days, 3.0 days and 5.4 days for SOF, MOF, and EOF, respectively, and (3) PlanetScope maps large variations in autumn leaf phenology over the entire forest landscape with spatially explicit information. These results demonstrate the ability of our proposed method in monitoring the large spatial heterogeneity of crown-scale autumn leaf phenology in the temperate forest, suggesting the potential of using high-resolution satellites to advance crown-scale phenology studies over large geographical areas.

Indian Sundarbans mangrove forest considered endangered under Red List of Ecosystems, but there is cause for optimism

Article

Full-text available

Sep 2020
BIOL CONSERV

Accurately evaluating ecosystem status is vital for effective conservation. The Red List of Ecosystems (RLE) from the International Union for the Conservation of Nature (IUCN) is the global standard for assessing the risk of ecosystem collapse. Such tools are particularly needed for large, dynamic ecosystem complexes, such as the Indian Sundarbans mangrove forest. This ecosystem supports unique biodiversity and the livelihoods of millions, but like many mangrove forests around the world is facing substantial pressure from a range of human activities. Holistic, standardised and quantitative environment risk assessment frameworks are essential here, because previous assessments have either been qualitative in nature, or have generally considered single threats in isolation. We review these threats and utilise the RLE framework to quantitatively assess the risk of ecosystem collapse. Historical clearing and diminishing fish populations drove a status of Endangered (range: Vulnerable to Endangered), and ongoing threats including climate change and reduced freshwater supply may further impact this ecosystem. However, considering recent change, the outlook is more optimistic. Mangrove extent has stabilised, and analysis of mangrove condition highlights that only a small proportion of the forest is degraded. Using the RLE provides an authoritative avenue for further protection and recognition of the issues facing this UNESCO World Heritage Site. We also identify knowledge and data gaps in the Sundarbans that are likely common to coastal systems globally. By articulating these and presenting opportunities and recommendations, we aim to further the conservation goals of the IUCN and the implementation of its new assessment framework.

A global biophysical typology of mangroves and its relevance for ecosystem structure and deforestation

Article

Full-text available

Sep 2020

Mangrove forests provide many ecosystem services but are among the world's most threatened ecosystems. Mangroves vary substantially according to their geomorphic and sedimentary setting; while several conceptual frameworks describe these settings, their spatial distribution has not been quantified. Here, we present a new global mangrove biophysical typology and show that, based on their 2016 extent, 40.5% (54,972 km 2) of mangrove systems were deltaic, 27.5% (37,411 km 2) were estuarine and 21.0% (28,493 km 2) were open coast, with lagoonal mangroves the least abundant (11.0%, 14,993 km 2). Mangroves were also classified based on their sedimentary setting, with carbonate mangroves being less abundant than terrigenous, representing just 9.6% of global coverage. our typology provides a basis for future research to incorporate geomorphic and sedimentary setting in analyses. We present two examples of such applications. Firstly, based on change in extent between 1996 and 2016, we show while all types exhibited considerable declines in area, losses of lagoonal mangroves (− 6.9%) were nearly twice that of other types. Secondly, we quantify differences in aboveground biomass between mangroves of different types, with it being significantly lower in lagoonal mangroves. Overall, our biophysical typology provides a baseline for assessing restoration potential and for quantifying mangrove ecosystem service provision. Mangrove forests provide valuable ecosystem functions and services including carbon storage, coastal protection, fisheries enhancement and tourism 1-4 ; however, large declines in global mangrove area have historically been estimated 5. Recent high-resolution assessments of mangrove change suggest considerable slowing of losses 6 , likely driven by growing wealth, increasing clarity of ownership, national efforts to sustainably manage forest estates and increased awareness of the ecosystem services they provide 7 .

Global declines in human‐driven mangrove loss

Article

Full-text available

Jul 2020

Global mangrove loss has been attributed primarily to human activity. Anthropogenic loss hotspots across Southeast Asia and around the world have characterized the ecosystem as highly threatened, though natural processes such as erosion can also play a significant role in forest vulnerability. However, the extent of human and natural threats has not been fully quantified at the global scale. Here, using a Random Forest‐based analysis of over one million Landsat images, we present the first 30‐meter resolution global maps of the drivers of mangrove loss from 2000‐2016, capturing both human‐driven and natural stressors. We estimate that 62% of global losses between 2000‐2016 resulted from land‐use change, primarily through conversion to aquaculture and agriculture. Up to 80% of these human‐driven losses occurred within six Southeast Asian nations, reflecting the regional emphasis on enhancing aquaculture for export to support economic development. Both anthropogenic and natural losses declined between 2000‐2016, though slower declines in natural loss caused an increase in their relative contribution to total global loss area. We attribute the decline in anthropogenic losses to the regionally‐dependent combination of increased emphasis on conservation efforts and a lack of remaining mangroves viable for conversion. While efforts to restore and protect mangroves appear to be effective over decadal time scales, the emergence of natural drivers of loss presents an immediate challenge for coastal adaptation. We anticipate that our results will inform decision making within conservation and restoration initiatives by providing a locally‐relevant understanding of the causes of mangrove loss.

The increasing relevance of phenology to conservation

Article

Apr 2022

Shifts in phenology can impact organism fitness, ecosystem function, and goods and services from nature. Climate change management must better integrate phenology to optimize conservation outcomes as these impacts increase.

Bridging the research-implementation gap in IUCN Red List assessments

Article

Jan 2022
TRENDS ECOL EVOL

The International Union for Conservation of Nature (IUCN) Red List of Threatened Species is central in biodiversity conservation, but insufficient resources hamper its long-term growth, updating, and consistency. Models or automated calculations can alleviate those challenges by providing standardised estimates required for assessments, or prioritising species for (re-)assessments. However, while numerous scientific papers have proposed such methods, few have been integrated into assessment practice, highlighting a critical research–implementation gap. We believe this gap can be bridged by fostering communication and collaboration between academic researchers and Red List practitioners, and by developing and maintaining user-friendly platforms to automate application of the methods. We propose that developing methods better encompassing Red List criteria, systems, and drivers is the next priority to support the Red List.

Conceptualizing ecosystem degradation using mangrove forests as a model system

Article

Oct 2021
BIOL CONSERV

The status and potential degradation of an ecosystem is often difficult to identify, quantify, and characterize. Multiple, concurrent drivers of degradation may interact and have cumulative and confounding effects, making mitigation and rehabilitation actions challenging to achieve. Ecosystem status assessments generally emphasize areal change (gains/losses) as a primary indicator; however, this over-simplifies complex ecosystem dynamics and ignores metrics that would better assess ecosystem quality. Consideration of multiple indicators is necessary to characterize and/or anticipate ecosystem degradation and appropriately identify factors causing changes. We utilize mangrove forests as a model system due to their distribution across a wide range of geographic settings, their position in the inherently dynamic coastal zone, and the multiple natural and anthropogenic pressures they face. We present a conceptual framework to: i) examine drivers of ecosystem degradation and characterize system status, and ii) delineate the roles of biogeographic and geomorphic variability, site history and typology, and references. A complementary workflow is proposed for implementing the conceptual framework. We demonstrate the universal applicability of our conceptual framework through a series of case studies that represent locations with differing drivers of degradation and biogeographic and geomorphic conditions. Our conceptual framework facilitates scientists, conservation practitioners, and other stakeholders in considering multiple aspects of ecosystems to better assess system status and holistically evaluate degradation. This is achieved by critically evaluating suitable comparisons and relevant elements in assessing a site to understand potential actions or the outcome of previously implemented management strategies.

Myanmar's terrestrial ecosystems: Status, threats and conservation opportunities

Article

Dec 2020
BIOL CONSERV

Myanmar is highly biodiverse, with more than 16,000 plant, 314 mammal, 1131 bird, 293 reptile, and 139 amphibian species. Supporting this biodiversity is a variety of natural ecosystems—mostly undescribed—including tropical and subtropical forests, savannas, seasonally inundated wetlands, extensive shoreline and tidal systems, and alpine ecosystems. Although Myanmar contains some of the largest intact natural ecosystems in Southeast Asia, remaining ecosystems are under threat from accelerating land use intensification and over-exploitation. In this period of rapid change, a systematic risk assessment is urgently needed to estimate the extent and magnitude of human impacts and identify ecosystems most at risk to help guide strategic conservation action. Here we provide the first comprehensive conservation assessment of Myanmar's natural terrestrial ecosystems using the IUCN Red List of Ecosystems categories and criteria. We identified 64 ecosystem types for the assessment, and used models of ecosystem distributions and syntheses of existing data to estimate declines in distribution, range size, and functioning of each ecosystem. We found that more than a third (36.9%) of Myanmar's area has been converted to anthropogenic ecosystems over the last 2–3 centuries, leaving nearly half of Myanmar's ecosystems threatened (29 of 64 ecosystems). A quarter of Myanmar's ecosystems were identified as Data Deficient, reflecting a paucity of studies and an urgency for future research. Our results show that, with nearly two-thirds of Myanmar still covered in natural ecosystems, there is a crucial opportunity to develop a comprehensive protected area network that sufficiently represents Myanmar's terrestrial ecosystem diversity.

Assessing the conservation status of mangroves in Rakhine, Myanmar

Abstract

Recommended publications

Mapping the Extent of Mangrove Ecosystem Degradation by Integrating an Ecological Conceptual Model w...

Threatened Ecosystems of Myanmar. An IUCN Red List of Ecosystems Assessment. Version 1.0

Estimating changes and trends in ecosystem extent with dense time‐series satellite remote sensing

Myanmar's terrestrial ecosystems: Status, threats and conservation opportunities