Deforestation in The Ayeyarwady Delta (2014)

Global Environmental Change 24 (2014) 321–333
Contents lists available at ScienceDirect
Global Environmental Change

journal homepage: www.elsevier.com/locate/gloenvcha
Deforestation in the Ayeyarwady Delta and the conservation

implications of an internationally-engaged Myanmar
Edward L. Webb a,*, Nicholas R.A. Jachowski b, Jacob Phelps a, Daniel A. Friess b,c,
Maung Maung Than d, Alan D. Ziegler b
a
Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore, Singapore
b
Department of Geography, National University of Singapore, Singapore, Singapore
c
Singapore-Delft Water Alliance, National University of Singapore, Singapore, Singapore
d
Mangrove and Environmental Rehabilitation Network (MERN), Yangon, Myanmar
A R T I C L E I N F O A B S T R A C T
Article history: Myanmar is a country of huge biodiversity importance that is undergoing major political change,
Received 17 February 2013 bringing with it new international engagement. This includes access to international markets, which will
Received in revised form 18 October 2013 likely spur investment in export-oriented agriculture, leading to increased pressures on already
Accepted 21 October 2013
threatened ecosystems. This scenario is illustrated in the Ayeyarwady Delta, the country’s agricultural
heartland sustaining high deforestation rates. Using the Delta as a model system, we use an integrated
Keywords: approach to inquire about whether and how imminent agricultural reforms associated with an
Agriculture
internationally-engaged Myanmar could introduce new actors and incentives to invest in agricultural
Cropland
Expansion
expansion that could affect deforestation rates. We use a novel remote sensing analysis to quantify
Governance deforestation rates for the Delta from 1978 to 2011, develop business-as-usual deforestation scenarios,
Land use and contextualize those results with an analysis of contemporary policy changes within Myanmar that
Mangrove are expected to alter the principal drivers of land-cover change. We show that mangrove systems of
Policy Myanmar are under greater threat than previously recognized, and that agriculture has been the
principle driver of deforestation on the Delta. The centrality of agriculture to the Myanmar economy
indicates that emerging policies are likely to tip the scales towards agricultural expansion, agro-
industrial investment and potentially greater rates of deforestation due to the introduction of well-
funded investors, insufficient land tenure agreements, and low governance effectiveness. The broad
national challenge is to initiate environmental governance reforms (including safeguards) in the face of
significant pressures for land grabbing and opportunistic resource extraction.
ß 2013 Elsevier Ltd. All rights reserved.
1. Introduction mangroves in Myanmar, but which has sustained the highest

deforestation rate in the country (Leimgruber et al., 2005).
Myanmar is a heavily forested country (48% forest cover; FAO, Myanmar’s mangrove forests support critical natural
2010) that retains among the largest remaining tracts of forest in resources and rural livelihoods (e.g. Barbier et al., 2008).
Southeast Asia and some of the world’s ‘hottest hotspots’ for Mangroves support in- and offshore fishing industries that are
species diversity and endemism (Myers et al., 2000). Despite critical sources of protein and rural income for many communities
retaining expansive forests, there is concern over Myanmar’s long- (e.g. Loneragan et al., 2005) and have potential to support export-
term forest conservation: very little of its area is protected (6.3% of oriented coastal aquaculture and fisheries (Fabrikant, 2013). The
the total land area; World Bank, 2013a), and certain parts of the Ayeyarwady Delta system provides coastal protection and water
country have experienced high rates of deforestation (Leimgruber regulation services (e.g. Koch et al., 2009) to a region with poor
et al., 2005). Of particular concern is the Ayeyarwady Delta, an drainage/water infrastructure and high vulnerability to seasonal
expansive alluvial floodplain originally home to the largest tract of flooding (Than, 2001) and extreme weather events. Notably, the
Delta is also one of Myanmar’s key biodiversity areas (Tordoff
et al., 2005; World Wide Fund for Nature, 2008), hosting some of
the most floristically-diverse mangroves in the world (Spalding
* Corresponding author. Tel.: +65 6516 4184.
et al., 2010) and more than 30 species of ‘endangered’ fauna (IUCN,
E-mail addresses: [email protected], [email protected] (E.L. Webb), 2011) including the Ayeryawady dolphin (Orcaella brevirostris)
[email protected] (Nicholas R.A. Jachowski). (IUCN, 2011), estuarine crocodile (Crocodylus porosus), which
0959-3780/$ – see front matter ß 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.gloenvcha.2013.10.007
322 E.L. Webb et al. / Global Environmental Change 24 (2014) 321–333
numbers only 100 individuals in the lower Ayeryarwady Delta land-cover in the Ayeyarwady Delta, demonstrating that almost all
(Thorbjarnarson et al., 2008), mangrove terrapin (Batagur baska), deforestation in the Delta has been for rice agriculture and that the
sarus crane (Grus antigone) (Tordoff et al., 2005), and numerous Delta mangroves have been lost at a substantially greater rate than
migratory bird species, including the critically endangered spoon- previously thought. We then contextualize our results with an
billed sandpiper (Eurynorhynchus pygmeus) (World Wide Fund for analysis of recent policy changes within Myanmar. Our analysis
Nature, 2008). Moreover, the Delta’s mangroves represent a suggests that pressure on natural forests in Myanmar is expected
substantial store of belowground carbon of interest to global to increase as a result of policy developments favoring export-
climate change mitigation (Donato et al., 2011). oriented agriculture, improvements in technology and infrastruc-
Mangrove forests are recognized as one of the most threatened ture, and the introduction of actors (investors) who may benefit
ecosystems globally (Duke et al., 2007), and the Ayeyarwady Delta from land expropriation under insufficient land laws. We use this
deserves attention as one of the most threatened mangrove deltas discussion to frame the environmental policy and governance
in the world. The Delta is heavily populated (7.7 million people, challenges going forward in a rapidly changing context. This
13.7% of the nation’s population; Central Statistics Organization, integrated approach (Mattison and Norris, 2005) is instrumental to
2007) and a center of agricultural processing and production providing a framework to predict the salient, large-scale land use
(Than, 2001; Xiao et al., 2006), responsible for 35% of the change drivers that are expected on the Delta, and more broadly,
country’s rice production (FAO, 2013). As a result, the Delta has Myanmar.
been heavily impacted and most of its biodiversity is severely
threatened (World Wide Fund for Nature, 2008); mangrove 2. Methods
deforestation is recognized as a critical environmental issue for
the country (Phyu, 2012). Previous analyses have suggested 2.1. Data collection
agricultural expansion—largely for subsistence, limited domestic
markets or redistribution during the decades of military rule—and We quantified mangrove cover on the Ayeyarwady Delta at four
fuelwood extraction as the principle drivers of mangrove loss (Oo, times over a 33-year period: 1978, 1989, 2000 and 2011. Spectral
2002; Leimgruber et al., 2005; Giri et al., 2008). data were acquired from Landsat multispectral scanner (MSS),
Recent and unprecedented political reforms in Myanmar since thematic mapper (TM) and enhanced thematic mapper (ETM+)
2010 portend a potentially significant shift in the social, political imagery for 1978, 1989, 2000, 2008, 2009 and 2011 (Table 1,
and economic context of the country. These reforms, following 50 available from the USGS Earth Resources Observation and Science
years of economic and political isolation, are catalyzing Myanmar’s Center, http://eros.usgs.gov). A minimum of two images were used
return to the international community, and have prompted high- for each time period to fill data gaps in imagery from 2008, 2009
level foreign visits along with efforts to re-establish political ties and 2011, and to increase model accuracy (Table 1). Images from
(e.g. Baker, 2012), increase international aid (e.g., AusAID, 2013), 2008 and 2009 were used to test for an impact of Cyclone Nargis on
ease sanctions (e.g. Barta, 2012) and increase international trade mangrove cover. Elevation data were acquired from the ASTER
and investment (e.g, Japan Times, 2012). These initiatives have the GDEM Version 2 (available from the Earth Remote Sensing Data
potential to deliver profound social, humanitarian and economic Analysis Center, http://www.ersdac.or.jp), and were averaged
changes for Myanmar (e.g., Asian Development Bank, 2011), which using a 9 9 pixel neighborhood filter to correct any gross errors
has some of the lowest development indicators in the world in elevation.
(United Nations Development Program, 2011; World Bank, 2012), Four regions of the Delta, in proximity of the settlements of
and one of the lowest Purchasing Power Parity GDPs (US$1300, Saluzeik, Yegyawgyi, Mayan and Kyaiklat in the Ayeyarwady
ranking 200/225; Central Intelligence Agency, 2010). Yet these region, were selected for extensive ground-data collection (Fig. 1).
changes may also bring environmental conservation challenges The zones ranged in size from 270 to 460 km2. A total of 240 points
across Myanmar, as both small and industrial-scale producers were randomly selected within the four zones with the GRASS
within and outside Myanmar respond to new policies and
opportunities (Schmidt, 2012; Webb et al., 2012).
A large literature demonstrates an agricultural expansion- Table 1
deforestation link globally and within Myanmar (e.g., Geist and Summary of images and model traits.
Lambin, 2002; Lambin et al., 2003; Rudel et al., 2009; Songer et al., Model # Classes Factors Imagery Training dataset
2009), and Myanmar’s agricultural sector contributes one of the year dates (or ‘‘Model’’)
highest percentages of national GDP in the world (World Bank, 1978 3 MSS imagery 1974 Jan 6 H
2013a). Therefore, it is justifiable to inquire about whether and 1978 Nov 5
how imminent agricultural reforms associated with an interna- 1989 3 TM imagery 1989 Jan 16 H
tionally-engaged Myanmar could introduce new actors and 1989 Feb 1
2000 3 TM imagery 2000 Jan 31 H
incentives to invest in agricultural expansion that could affect
2000 Apr 4
the remaining natural systems. And further, whether current 2008 3 ETM+ imagery 2007 Dec 4 H
policy instruments are sufficient to both promote agricultural 2008 Jan 21
innovation while protecting local landholders and remaining 2008 Feb 6
2009 3 ETM+ imagery 2008 Dec 6 H
natural systems. Indeed, concerns over the potential environmen-
2009 Jan 7
tal and humanitarian impacts in Myanmar resulting from greater 2009 Mar 28
international engagement are already mounting (Burma Environ- 2011 5 ETM+ imagery 2010 Jan 10 C
mental Working Group, 2011; Webb et al., 2012; Chong, 2012; TM imagery 2011 Apr 3
Schmidt, 2012; Wang et al., 2013). MSS imagery 2000 Jan 31
Distance to 0-m 2000 Apr 4
In this study, we explore the historical trend of deforestation in
contour
the Ayeyarwady Delta, as a model system to Myanmar, and Distance to 5-m 1989 Jan 16
subsequently consider the potential ramifications of recent policy contour
developments designed to increase access to international Distance to large 1989 Feb 1
water body 1974 Jan 6
markets, on Myanmar’s broader environment. We first use a
1978 Nov 5
novel remote sensing analysis to document historical changes in
E.L. Webb et al. / Global Environmental Change 24 (2014) 321–333 323
Fig. 1. Map of study area showing the Ayeyarwady delta of Myanmar and the ground data collection points. Red points indicate random points, while yellow points indicate
half-way points where additional ground data were collected. 239 points were sampled in total with approximately 60 red and 60 yellow points in each zone. The areas of
each zone are: Zone 1 = 437 km2, Zone 2 = 460 km2, Zone 3 = 270 km2, Zone 4 = 270 km2.
(2011) Geographic Information System. During a field campaign The training dataset for Model C consisted of 647 points: 479
conducted between 16 March–19 April 2011, four teams of ground-collected points collected in 2011 (described previously),
researchers (including author MMT) travelled to the assigned and 168 remotely-collected water points (Table 2). The 168
points to determine the dominant land-cover type at a resolution remotely-collected water points were selected randomly from
of 30 m 30 m. Positions were recorded on a handheld geograph- large water bodies (>100 ha) determined from a normalized
ical positioning system (GPS) with an accuracy of <15 m; this difference vegetation index (NDVI) map derived from the Landsat
positional error was less than the size of the Landsat pixel of satellite imagery. Additional water points were collected in order
30 m 30 m (National Aeronautics and Space Administration, to better represent this class in the final training dataset.
2011). Four photographs, one in each cardinal direction, were Training dataset H consisted of 504 points distributed equally
taken for data screening and standardization (Fig. 2). A total of 239 (168 each) among the three classes: mangrove, non-mangrove and
additional points were collected at random while travelling to an water. The non-mangrove and water points were remotely
assigned GPS location, increasing the total number of ground collected and the mangrove data points were ground-collected
reference points to 479. Land cover was classified as mangrove, during the 2011 field excursion (Table 2). This dataset was deemed
degraded mangrove, agriculture, water or other (Fig. 2). Data were historically valid based on two key biological and environmental
screened and classifications standardized among the four teams. assumptions: (1) given the 33-year time period of this study, water
body and elevation variables do not change significantly and (2) if a
2.2. Model construction dense mangrove without evidence of human degradation exists
today it likely existed over the past 33 years. As we go further back
A 5-class contemporary model (Model C) was produced for in time the accuracy of these assumptions will decrease, so we
2011 and a 3-class historical model (Model H) was produced for would expect to see a similar decline in model prediction accuracy.
1978, 1989 and 2000. Model construction began with developing The water and non-mangrove training points were randomly
the necessary training datasets. To produce both contemporary as selected in the GIS from water bodies and non-tidal land areas that
well as historical mangrove models, two different training datasets were unlikely to support mangroves (Table 2, Fig. A1). Water
were required: a historically valid training data set for 3-class points were selected randomly from large water bodies (>100 ha)
analysis (Model H) and a contemporary training data set for 5-class determined from a normalized difference vegetation index (NDVI)
analysis (Model C). This distinction was necessary because 5-class map derived from the Landsat satellite imagery. Non-mangrove
data could not be accurately inferred for the historical time terrestrial points were selected randomly from land areas situated
periods. The training datasets were compiled from the ground- more than 10-km inland from the 5-m elevation contour
collected and remotely-collected data (described in Section 3.1). determined from the ASTER GDEM Version 2 digital elevation
model; thus ensuring that points were chosen in areas where it
would be physiologically impossible for mangroves to be present,
Table 2 to minimize class confusion (Table 2, Fig. A1). The mangrove
Descriptions of training datasets used for model derivations.
training points corresponded to the ground-collected points
Historical Training Dataset (H) Contemporary Training Dataset (C) classified as mangrove. Ground-collected points classified as
168 Historical mangrove pointsa 479 Ground collected points degraded mangrove were not included as it was uncertain whether
168 Historical non-mangrove points 168 Water points these areas would have been mangrove over the entire study
168 Historical water points period.
a
Subset of Model C ground collected points; the non-mangrove and water points An algorithmic approach to model development was chosen
were extracted from satellite images. over traditional data modeling approaches due to their superior
predictive ability (Breiman, 2001). The approach used three sequential minimal optimization method is a time-optimized
modeling methods—multilayer perception, rotation forest and way to train support vector classifiers (Platt, 1998).
sequential minimal optimization—which were combined using a Modeling was performed using Weka, a machine learning
majority-voting algorithm. The rotation forest method split the software (Hall et al., 2009). Modeling was conducted using a 10-
training set into N random subsets and then performed a principal fold cross-validation, which produced robust accuracy assess-
component analysis on each subset (Rodriguez et al., 2006). The ments involving confusion matrices, user accuracies, producer
principal components were then classified using decision trees. accuracies, overall classification accuracies and Cohen Kappa
The multilayer perceptron method is a type of neural network that coefficients (see Congalton, 1991). A unique model was trained for
uses back-propagation to model classes (Ware, 2000). The each time period using the input variables and training datasets
summarized in Tables 1 and 2.
Congruence between estimates derived using the two models
(validation) was accomplished by comparing the estimated
mangrove area (derived by applying the 3-class Model H to the
2011 image) with the area of dense mangrove estimated according
to Model C (also on the 2011 image). Because Model C was used to
report land cover for 2011 (owing to a higher number of classes
that were ground-referenced), the Model H estimation for 2011
was only used as a validation procedure and not included in the
land cover time-series.
2.3. Accuracy assessment
Accuracies were estimated using a 10-fold cross validation

method in which the training datasets were split randomly into 10
equal subsets; 90% of the data were used to build a model, which
was then tested on the remaining 10%. Ten models were built and
tested using this method, and the classification accuracies were
averaged to develop a robust estimate of the generalization
accuracy of each model (Kohavi, 1995). Classification accuracy for
the 3-class models for each year ranged between 89.9% and 95.6%
(Table A1). The non-mangrove class was most frequently
misclassified with an average producer accuracy over the four
time periods of 0.894, compared to 0.909 and 0.972 for the
mangrove and water classes, respectively. This is because the ‘non-
mangrove’ training points included multiple land uses, such as
bare ground, upland forest, urban, and agriculture. Thus, the
spectral variability was greatest in the non-mangrove class.
The 5-class model for 2011 had lower accuracy—as is expected
of higher order models—with an overall accuracy of 81.6% and a
kappa coefficient of 0.749 (Table A2). The mangrove, agriculture
and water classes had the highest accuracies, while the degraded
mangrove and other classes had significantly lower accuracies. The
degraded mangrove class likely had low accuracy, as its spectral
reflectance would be intermediate of mangrove and agriculture.
The ‘‘other’’ class likely had low accuracy because it was a ‘catch-
all’ class for a broad range of land-cover types that did not fit into
the other classes, and thus, did not have a constant spectral
Table 3
Mangrove area statistics for the Ayeyarwady Delta, Myanmar. The range was
calculated using producer and user accuracies. Model H for years 1978–2000 was
trained using 2011 dense mangroves from ground collected points. Mangrove
estimates for the year 2011 are dense mangroves only. The 2011 Mangrove areal
statistics using models trained on datasets H and C overlap, validating the use of
models based on dataset H for historical time periods for which more nuanced field
data were not available. The model based on Training Dataset H gives a slightly
higher predicted range of mangrove areas likely due the fact that it is less able to
distinguish dense and degraded mangrove, as degraded mangrove training points
were not available in the past.
Model year Area (km2) Range
1978 2623 2329–3030

1989 2138 1990–2228
2000 1198 1114–1256
2008 (pre-Nargis) 933 805–1060
2009 (post-Nargis) 918 820–973
2011 (Model H) 1096 937–1226
Fig. 2. Examples of ground data collection photos of the five land cover classes in
2011 (Model C) 938 705–1061
this study.
Table 4
Comparison of classification accuracy of the current study with other recent mangrove remote sensing studies.
Study Location Imagery type # of classes Classification accuracy % Kappa coefficient (3dp)
This study Myanmar Landsat + ASTER GDEM, Model H 3 89.9–95.6 0.848–0.935

This study Myanmar Landsat + ASTER GDEM, Model C 5 81.6 0.749
Fatoyinbo et al. (2008) Mozambique Landsat + satellite altimetry 8 93.9 0.870
Fatoyinbo et al. (2008) Mozambique Landsat 8 71.0 N/R
Ruiz-Luna et al. (2010) NW Mexico Landsat 2 >85.0 >0.810
Alatorre et al. (2011) NW Mexico Landsat 7 84.0 N/R
Rakotomavo and Fromard (2010) Madagascar Landsat 11 84.4 N/R
N/R, not reported.
identity. Also, there were relatively few ‘‘other’’ points in the of 2000–2011. The calculations excluded the Meinmahla Kyun
training dataset used to train the model, decreasing the ability of Wildlife Sanctuary and remote offshore islands (137 km2)
the model to accurately predict that class. because our analysis indicated that the protection of those areas
A high congruence was observed between Model H and Model C has been effective in preventing deforestation over the 33-year
for the 2011 estimates (Table 3). Whereas the mangrove area period, which was assumed to continue going forward. Thus, the
predicted by the 5-class and 3-class 2011 models differed by scenarios relate only to unprotected mangroves—i.e. 95% of the
185 km2, the range of possible mangrove areas overlapped 1978 mangrove cover. The BAU scenarios had three assumptions.
substantially, providing validation for the models using Training First, it was assumed that the remaining mangroves were situated
Dataset H (Table 3). Overall, the level of accuracy obtained through on land suitable for rice cultivation. Second, future mangrove
this novel classification method was comparable to, or greater conversion to agriculture would not be prohibited by costs of site
than, other studies utilizing Landsat imagery for mangrove treatment related to the prevention of soil salinization. Lastly,
classification (Table 4). underlying drivers of deforestation remained constant.
2.4. Deforestation scenarios 2.5. Contextualizing the case study
We developed deforestation projections that were based on the We undertook a literature review to contextualize this case
quantified deforestation rates: i.e. business-as-usual (BAU) con- study. We focused a review on the deforestation literature,
ditions (Sloan and Pelletier, 2012). As our models produced value specifically in terms of linking market access and agricultural
ranges for each year of analysis, we produced four deforestation production with forest transitions. Given the assumption that the
scenarios: (1) a best-case scenario (lowest possible 33-year political transition in Myanmar is corresponding with the
deforestation rate), where the deforestation rate was calculated development of policies to facilitate access to international
based on the lowest value of the 1979 mangrove cover range and agricultural markets, we evaluated how deforestation pressures
the highest value of the 2011 mangrove cover range; (2) a mean may be shaped by greater investment in export-oriented agricul-
deforestation rate, calculated as the regression line through ture. We also inquired as to who might be the agents of change,
estimates of 1978, 1989, 2000 and 2011; (3) a worst-case based given that the (gradual) integration of Myanmar into the
on the highest 1979 mangrove cover range and lowest 2011 international enterprise networks will introduce new actors into
mangrove cover range; and (4) the most recent deforestation trend the agricultural sector.
Fig. 3. Map showing mangrove land cover in the Ayeyarwady Delta, Myanmar, in 1978, 1989, 2000 and 2011. The large island that has remained completely forested is the
Meinmahla Kyun Wildlife Sanctuary.
Table 5 The vast majority (81%) of dense mangrove loss was caused by
Mangrove transitions in the Ayeyarwady Delta, Myanmar. Land cover classes in
conversion to rain-fed rice paddy, some of which has likely been
2011 that were mangrove in 1978.
abandoned and subsequently regenerated to degraded mangrove
Land class 2011 Area (km2) Area (%) (13%) (Table 5, Fig. 4). Mangrove patches >100 km2, comprising
Degraded mangrove 277.8 13.1 nearly 60% of the total mangrove area in 1978, accounted for only
Agriculture 1727.6 81.2 25% of total area by 2011 (Fig. 5). Moreover, by 2011 the mangrove
Water 3.4 0.2 landscape was highly fragmented and consisted of no patches
Other 118.9 5.6
>300 km2. Most of these remaining fragmented mangroves are
Total 2127.7 100
expected to support significantly reduced marine and terrestrial
resources (Barbier, 2003; Polidoro et al., 2010) as well as reduced
coastal defense functions (Barbier et al., 2011). Indeed, deforesta-
3. Results and discussion tion in the Delta has substantially impacted biodiversity (World
Wide Fund for Nature, 2008; IUCN, 2011) and potentially the well-
3.1. Rapid, agriculture-driven deforestation in the Delta being of rural communities. For example, anecdotal evidence
(described by Feagin et al., 2010) suggests that the loss and
Mangrove area cover has changed dramatically in the degradation of the Ayeyarwady mangroves may have increased
Ayeyarwady Delta since 1978 (Fig. 3). In 1978, mangroves covered population risk to the impacts of Cyclone Nargis.
2623 km2 (range of estimate 2329–3030 km2, Table 1). Over the The mean 33-year deforestation rate BAU scenario suggested that
33-year study period, mangrove cover declined by 64.2%, at an Ayeyarwady mangroves (outside Meinmahla Kyun) could be
average rate of 51 km2 (3.1%) per year, to an area of 938 km2 (range completely deforested by about 2026 (Fig. 6). Using worst- and
705–1061 km2) in 2011. Deforestation rate varied substantially best-case BAU scenarios, the forecast to depletion ranged from 2019
over the three periods, from 44.1 km2 per yr in 1978–1989, to to 2035, respectively, and the 2000–2011 deforestation rates
85.5 km2 per year in 1989–2000, to 23.6 km2 per year from 2000 to projected mangrove depletion by 2044. Business-as-usual defores-
2011 (Table 3). During the 1989–2000 period, 44% of remaining tation scenarios suggest that most, if not all of the unprotected
mangrove area was lost. The only area of the mangrove not Ayeyarwady mangroves will be lost in the next few decades at a rate
sustaining a precipitous decline was the 137 km2 Meinmahla Kyun faster than other mangrove deforestation hotspots in the region, such
Wildlife Sanctuary and small, remote offshore islands (Fig. 3). It is as the Mekong Delta from 1965–1995 (prior to major investment in
notable that we report higher levels of confidence in estimates for shrimp aquaculture, Thu and Populus, 2007). However, the BAU
both historical and present day land-cover estimates, than with scenarios should be viewed with caution because deforestation may
previous studies that used traditional land-cover analyses decelerate as the system vanishes; yet it is noteworthy that our
(Leimgruber et al., 2005; Giri et al., 2008). The results also analysis shows a fairly rapid deforestation rate from 2000 to 2011.
returned high classification accuracies (cf. Leimgruber et al., 2005), Rigorous, quantitative historical analysis of deforestation
and which included error estimates (range per land cover type). drivers from 1978 to 2011 was not possible given the lack of
Notably, our estimate of mangrove loss between 1989 and 2000 robust, reliable socioeconomic data for Ayeyarwady (see Woods
(44% of remaining forest cover) is more than double that of the only and Canby, 2011; Bissinger, 2012), limiting our ability to
other previous study (20% loss; Leimgruber et al., 2005). quantitatively evaluate specific drivers (cf. Geist and Lambin,
Cyclone Nargis did not result in immediate reduction of 2002) or to make spatially-explicit predictive deforestation
mangrove areal extent (Table 3). Extensive flooding occurred scenarios (cf. Sloan and Pelletier, 2012). Despite this caveat, our
during Cyclone Nargis, and on-the-ground observations and analysis clearly indicates that rice agriculture has been the
permanent plots noted extensive damage to some mangrove dominant proximal driver of deforestation in the Delta, likely
species, particularly in low-lying sites (Aung et al., 2011) or areas alongside firewood extraction (see Oo, 2002; Fritz et al., 2009). This
that had already been severely degraded owing to fuelwood assessment concurs with all previous assessments of land
harvest and conversion to rice paddies (Fritz et al., 2009). In transformations in the Delta, which highlight the role of rice
addition, some mangrove species were able to recover rapidly after agriculture and associated agrarian policies in shaping historical
the disturbance event (Aung et al., 2011, 2013). land-use transformations the Delta (e.g. Than, 2001; Kurosaki et al.,
Fig. 4. Mangrove land cover change map showing 2011 mangrove land cover (green) and transitions into other land uses by 2011 (‘‘Mangrove Change Classes’’). Grey and light
blue areas were never mangrove over the time periods modeled in this study.
potential for growth (Dapice et al., 2010) and is a development

priority (see President U Thein Sein’s inaugural address at http://
www.burmalibrary.org/docs11/NLM2011-03-31.pdf). The govern-
ment of Myanmar recognizes the need to reshape agricultural
production towards export-oriented markets. The Ministry of
Agriculture is promoting expansion of rubber, jute, cotton,
sugarcane and palm oil industries (Ministry of Agriculture and
Irrigation, 2010), which are attracting foreign investment (Burma
Environmental Working Group, 2011; Montlake, 2013). The Delta
is experiencing expansion of higher-value and export-oriented
agricultural products such as oilseed crops, pulses and vegetables
(see Aung, 2011; Zaw et al., 2011). Moreover, there is emerging
interest in developing Myanmar’s coastal aquaculture industry,
particularly cultured shrimp for export (Fabrikant, 2013). Although
our results show that less than 2% of Delta mangroves were
supplanted by active aquaculture since 1978 (see also Giri et al.,
2008), pond aquaculture represents major future potential for both
small-scale and commercial income generation, as has occurred in
neighboring Thailand (Flaherty et al., 1999; Goss et al., 2000) and
Fig. 5. Bar chart showing the relative area comprising mangrove patches of different
areas, across a 33-year study period in the Ayeyarwady Delta, Myanmar.
Vietnam (Thu and Populus, 2007). These developments strongly
suggest that as Myanmar increases its access to international
markets, it will emphasize export-oriented crops, potentially
2004; Leimgruber et al., 2005; Okamoto, 2005; Giri et al., 2008; leading to ‘‘crop booms’’ (Hall, 2011), which have important
Matsuda, 2009; Okamoto, 2009). The literature suggests that rice implications for an understanding of the actors driving land use
agricultural expansion in the Delta region—and in other parts of change and potentially deforestation.
Myanmar—was largely for smallholder paddy, facilitated by the Specific policies reflect efforts to improve access to international
State through assistance with land preparation, irrigation and markets, including in the agricultural sector. For example, Myanmar
agricultural extension services (Matsuda, 2009). State-facilitated is signatory to the China-ASEAN Free Trade Agreement that will
agricultural expansion has been seen in other tropical countries, remove tariffs on 90% of goods by 2015 (http://fta.mofcom.gov.cn/
where deforestation prior to the 1990s was driven largely by state- topic/chinaasean.shtml). Myanmar has also increased cooperation
supported activities (Rudel, 2007; Rudel et al., 2009). In addition, with Thailand, which plans to triple bilateral trade by 2015
farmers in the Delta and across Myanmar experience severe land (Pratruangkrai, 2012). The government has broadened the banking
tenure insecurity, which has important implications for agricul- sector, increasing the number of sanctioned private banks, increas-
tural development in an environment that may facilitate large- ing local credit access, and introducing swift international monetary
scale agro-industrial investment (see below). transfers that will facilitate remittances and investment (Asian
Development Bank, 2011, 2012; Kyaw, 2012a). Facilitated by these
3.2. Agriculture expansion and land use change: A look forward for policies, Myanmar reported a 28% increase in foreign trade during
Myanmar the first 9 months of the 2011–2012 fiscal year (Xinhua, 2012), as
well as a substantial increase in Asian investment (Kate and Kubota,
Given that the Delta is one of Myanmar’s key rice growing 2012) including plans for increased trade with India, Bangladesh and
regions (Xiao et al., 2006), future agricultural investment and China (Mirdha, 2011; Te Te, 2011).
expansion in response to new access to international markets, are The agriculture sector of Myanmar currently employs 70% of
likely to affect the Delta. At the national scale, agriculture is the labor force and contributes >40% of GDP (Central Intelligence
considered central to Myanmar’s economic reform, has great Agency, 2010), yet is currently characterized by inefficiency, low
quality and high levels of farmer debt. Investments into updated
farming technologies, improved varieties, crop diversification,
education, and increased credit access could revitalize the sector
(Kurosaki et al., 2004; Aung, 2011; Dapice et al., 2011, 2012). The
Myanmar Agricultural Bank recently doubled credit available to
farmers (Asian Development Bank, 2011), and a new agricultural
bank was established in 2012 (Szep, 2012). Access to credit along
with increased efficiencies and technology could help stimulate
agricultural production, potentially doubling rice exports by 2017
(Chanjaroen, 2012).
Both small- and industrial-scale agriculture and aquaculture
remains heavily limited by infrastructure (Dapice et al., 2010;
Ministry of Agriculture and Irrigation, 2010; Asian Development
Bank, 2012), particularly in the Delta following Cyclone Nargis (Fritz
et al., 2009) but there is evidence that infrastructure development is
a domestic priority. Thailand and Myanmar are developing a major
seaport and industrial complex (see http://www.daweiport.com/),
Fig. 6. Area cover of mangroves in the Ayeyarwady Delta, Myanmar from 1978 to and are improving a transportation network that will extend to
2011 (with error bars), and four deforestation projections. Scenarios a–c are based neighboring countries (Boot, 2010). The Japanese government has
on the 33-year deforestation rate, with (a) being the best-case, (b) the regression
line through all points (i.e., the mean) and (c) being the worst-case. Scenario (d) is a
recently agreed to absolve approximately $3.36 billion in Myanmar
BAU scenario based on the 2000–2011 deforestation rate. The calculations exclude debt, coupled with major loans and investment in infrastructure
the Meinmahla Kyun Wildlife Sanctuary and remote offshore islands (137 km2). (Fuller, 2012; Kyaw, 2013), and the World Bank is providing a $165
million interest-free loan to support priority needs, including the banking reforms, farmer access to credit and capital for
infrastructure (World Bank, 2013b). investment will be greatly overshadowed by foreign investment
The trends in Myanmar we detail above—improved market given the low levels of smallholder capital (low incomes) and
access, technologies and infrastructure—mirror those in other collateral (insecure land tenure for leveraging). Further, foreign
tropical countries since the 1990s, with important effects on land capital is a ‘‘necessary condition’’ for economic growth of
use and forest cover (Angelsen and Kaimowitz, 1999; Geist and developing countries (Zoomers, 2010)—especially in countries
Lambin, 2002; Abdullah and Hezri, 2008; Brando et al., 2013). such as Myanmar where FDI is seen as a critical element of kick-
Notably, export-oriented agricultural expansion such as planta- starting the economy. Finally, the Foreign Investment Law passed
tions, and agricultural trade are the principal forces underlying in November 2012 provides numerous incentives to increase FDI,
deforestation in Asia since the 1990s, having usurped smallholder including tax breaks, a guarantee against nationalization of the
expansion as the principle driver in the decades prior (Rudel, 2007; investment, ability to repatriate investment gains, and land leases
Rudel et al., 2009; Hall, 2011; DeFries et al., 2013). Because of 50 years with two 10 year extensions (which contravenes the
Myanmar has been largely sidelined from the international Vacant, Fallow and Virgin Lands Management Law of 2012
markets since the 1950s, improved market access suggests that stipulating 30 year leases) (Mayer Brown JSM, 2012; Buchanan
an expansion of export-oriented agriculture (e.g., rubber and oil et al., 2013). Thus, foreign investors are expected to be important
palm) is likely to hold going forward in Myanmar. actors in the ‘‘control and use of land’’ (Zoomers, 2010).
Thus, the expectation is that the policy developments designed to Second, although land tenure and land reform are considered
improve international market access, combined with improvements critical to equitable economic growth, agricultural output, improved
to technology and infrastructure, will result in greater pressure food security and environmental sustainability in Myanmar (Food
being placed on natural systems across Myanmar, potentially Security Working Group, 2012; Kyaw, 2012b), rural and farmland
leading to greater deforestation rates (Angelsen and Kaimowitz, tenure in Myanmar remains insufficient. Two relevant laws recently
1999; Geist and Lambin, 2002). Neighboring countries across passed in Myanmar, the Farmland Law (2012) and the Vacant, Fallow
continental Southeast Asia also offer salient examples. For example, and Virgin Lands Management Law (2012), have made strides
conversion of upland forest to oil palm is a major driver of toward improved land tenure security (English translations of both
deforestation in both Malaysia and Indonesia (Koh and Wilcove, laws available online at http://www.burmalibrary.org/show.php?-
2008). Vietnamese aquaculture increased production from cat=1200). These laws have established a land use certificate and
168,000 tons in 1990 to 2,700,000 tons in 2010 (FAO, 2006-2013) registration system for cultivated land (the Farmland Law), and an
in response to greater market access; aquaculture expansion has application system to lease State land for lawful purposes, including
been implicated as the principle driving force of increased mangrove agricultural development (The VFV Law; Oberndorf, 2012). Never-
deforestation rates (De Graaf and Xuan, 1998), especially in the theless, these laws lack sufficient provisions for secure tenure to
northeast Mekong delta from 1995 to 2001 (687 ha/year or 13.1% rural farmers: the State retains ownership of all land (any changes in
annual rate; Thu and Populus, 2007). In addition, improved cropping require permission, and the State can confiscate land if
infrastructure has been associated with road-facilitated deforesta- ‘‘conditions of use are not met’’); the definition of ‘famer’ is so broad
tion in the highlands of northern Thailand (Delang, 2005), and as to include investors and management personnel, who can receive
expansion of rubber at the expense of forest in Xishuangbanna, confiscated land or apply for land leases under the VFV Law; disputes
southern China (Li et al., 2007). However, roads combined with involving allocation or use are not heard by a court of law; women do
capital investment in urban areas, may also facilitate rural-to-urban not enjoy explicit equal landholding rights; customary law is not
migration, which may reduce pressure on forests as marginal lands sufficiently recognized; and there is a lack of free, prior and informed
are abandoned (e.g. Rudel et al., 2005; Izquierdo et al., 2011). consent for leasing of VFV lands that might interfere with existing
land claims of any groups (Human Rights Foundation of Monland,
3.3. Agents of change transforming or protecting the landscape 2012; Oberndorf, 2012).
The combination of rapidly expanding export-oriented agricul-
Based on the case study of the Ayeyarwady Delta, and in line ture (‘‘crop booms’’, Hall, 2011), increasing FDI and insecure land
with previous findings across the region, we anticipate that new tenure are likely to create conditions that facilitate land grabbing by
market incentives for export-oriented agriculture will likely result elites, both national and domestic (Zoomers, 2010; Borras and
in broad and increased future deforestation (Rudel et al., 2009; Franco, 2010). Indeed, more than 750 cases of land disputes and
DeFries et al., 2010). However the mechanisms and actors confiscations were reported in the first year of the Farmland and VFV
responsible for future transformations are evolving. While laws (New Light of Myanmar, 2013), indicating significant loopholes
historically deforestation in the Delta was principally driven by in the law that facilitate land disputes and open the door for land
small-holder rice farmers receiving government support, these grabbing, with significant advantages to well-funded and well-
patterns will not necessarily hold true into the future. Based on connected elites (Osborne, 2013). Certainly, land seizures in
experiences across the region and emerging evidence from within Myanmar have been and continue to be commonly associated with
Myanmar, it is likely that powerful interest groups, both domestic the military (Zaw and Khaing, 2013); however there is evidence of
and foreign, will play an increasingly significant role in enterprise- non-state actors obtaining land through extralegal means, e.g. agro-
driven deforestation. The behavior of these elites, in conjunction industrial plantation development without regard for terms of lease
with government agencies and often in contest with other local agreement under the VFV Law (Oberndorf, 2012), and land
actors, will be key in decisions over how Myanmar’s landscape is confiscation for Chinese investment in agribusiness, particularly
transformed and/or protected. in northern Myanmar (Buchanan et al., 2013). In the Ayeyarwady
First, it is likely that a significant portion of capital investment Delta, farmers are seeking reparations or the return of land that may
for agriculture will come internationally. Although foreign direct have been seized by the government (Zaw and Khaing, 2013).
investment (FDI) to agriculture in Myanmar is low (Bissinger, Nevertheless, potentially massive influxes of investment for
2012), the lifting of sanctions has already increased investment, agroindustry, such as Wilmar International’s interest in expanding
with Myanmar now seen as one of the world’s biggest ‘‘agricultural the sugar industry to the Ayeyarwady Delta (Montlake, 2013), will
investment frontiers’’ (Osborne, 2013). This suggests that well- likely serve as a test of whether private industry will successfully
financed international elites will be entering Myanmar, with levels exploit current land tenure insecurity to gain government approval
of capital much greater than available locally. For example, despite to acquire and develop large-scale agroindustry on contested land.
Thus, as seen in other developing countries, Myanmar’s policy highlights a critical need for significantly improved data collection
objectives to promote private enterprise and increase FDI may also and analysis to capture existing, on-the-ground land use patterns,
facilitate the transfer of control and use of land, leading to a as well as increased consideration for how contemporary policy
‘‘foreignisation of space’’ (Zoomers, 2010) to investors from China, reforms may shape the future environment.
Thailand, and Malaysia. Moreover, the need for an improved protected area system in
Third, Myanmar scores extremely low in all aspects of Myanmar has been highlighted for over a decade (Rao et al., 2002,
governance—including in rule of law and the control of gover- 2013). In the Ayeyarwady Delta, for example, the Meinmahla Kyun
nance—and in terms of providing an open and secure business and Wildlife Sanctuary has successfully protected both forest cover
investment environment (www.govindicators.org; www.herita- (our analysis) and the endangered estuarine crocodile (Thorbjar-
ge.org). In emerging economies, corporate wealth tends to be narson et al., 2008). Especially given the comparatively small area
highly concentrated (La Porta et al., 1999; Claessens et al., 2000), currently under conservation, more financing could be provided to
suggesting that elite actors are likely to work closely with the significantly expand the protected areas network (Rao et al., 2002).
government to secure land leases, possibly exploiting loopholes in In the Delta, there is equally scope for community-based
existing land laws. Given the recent analyses that strongly reforestation and forest management programs, which could
implicate industrial-scale, export-oriented agriculture in tropical rehabilitate mangroves and help to fulfill demand for fuelwood in
deforestation (Rudel et al., 2009; DeFries et al., 2010), it therefore the Delta (see Macintosh et al., 2012), and substantial investment
seems entirely possible that an Indonesian-style system of crony is expected for direct conservation funding to conserve coastal
capitalism could emerge in Myanmar, with well-connected elite species and habitats (e.g. Critical Ecosystems Partnership Fund,
agro-industrial investors (palm oil, rubber) making significant land 2012; Wildlife Conservation Society, 2013).
deals in the absence of a well-articulated, and enforced, land law To date, however, Myanmar’s protected areas network has
and tenure system (Oberndorf, 2012). remained small, and environmental regulations in Myanmar have
Substantial on-the-ground work needs to be done to more been largely absent (e.g. Li, 2008), and rarely or arbitrarily enforced
deeply evaluate the behavior of actors, on-the-ground incentives (Burma Environmental Working Group, 2011; Rao et al., 2002).
and investment in agriculture and agro-industry, how actors are There are considerable challenges to developing and implementing
acquiring land and navigating/exploiting the current land tenure environmental safeguards and environmental impact assessments
system, and the consequences for forest cover dynamics in in the context of Southeast Asia, where they are often viewed as
Myanmar. Our study on the physical dimensions of land use disincentives to development, investment and short-term eco-
change in the Delta incorporated on-the-ground data, however, we nomic gains (Li, 2008). These challenges may be aggravated by the
lack critical socioeconomic data that could advance our under- rapid arrival of foreign financial interests and new waves of
standing of the political economy of increased focus on export- agricultural land-grabbing that may conflict with efforts to
oriented agriculture, introduction of foreign actors and their increase environmental regulation; this may be particularly acute
interaction with the existing power structure in Myanmar, and the within Myanmar, given its complex socio-political landscape and
dynamics of emerging land law; all of which will affect forest cover extremely low governance capacity (World Bank, 2012). Moreover,
in the future. A nuanced approach will be essential for well- Myanmar’s largest foreign investors—China and Thailand—may
informed policy (Borras and Franco, 2010). However, adequate and view economic liberalization as an opportunity to export polluting
verifiable data on these critical components needed for a complete industries (Boehler, 2012), extract raw materials (e.g., Yap, 2010;
understanding of the current and potential drivers of land use Asian Development Bank, 2012), and supply agricultural exports
change in Myanmar are largely absent, as noted by a recent study (e.g., Yunfei and Lingling, 2010). The fear is that these interests
on Myanmar’s forest law and policy (Woods and Canby, 2011). have little regard for environmental impacts (Casey, 2007; Burma
Environmental Working Group, 2011; Fuller, 2011) and may
3.4. Policy and governance challenges going forward influence both policymaking and enforcement. Moreover, leverag-
ing foreign aid and technical support to encourage improved
In this paper we have identified trends of forest loss in the natural resource governance is largely considered ineffective
Ayeyarwady Delta and the critical role of rice agriculture to (Carbonnier, 2011), especially in a region dominated by emerging
deforestation. The opening up of Myanmar to international donors (Woods, 2008; Reilly, 2012).
investment and enterprise promises to bring new actors, pressure A high priority needs to be placed on land tenure reform, in
on forest, and governance challenges. High rates of historical order to promote equitable economic growth and improved
mangrove deforestation, and the growing tension between rapid agricultural output, mitigate against land grabbing, and promote
economic development and mangrove conservation in the environmental sustainability. Community-based conservation
Ayeyarwady Delta illustrate the broader need for significant mechanisms and responsible local-level land management, for
attention to be paid to numerous aspects of environmental example, often depend heavily on secure land tenure, recognition
governance to ensure ‘‘environment-friendly reform’’ that con- of customary land use practices and rights to local stewardship
siders multiple objectives (Webb et al., 2012). The quality of (Mendelsohn, 1994; Angelsen and Kaimowitz, 1999). Policy
environmental policy and governance may ultimately determine analysis has already put forth a set of guidelines for improving
whether business-as-usual scenarios are maintained, improve, or the current Farmland Law and the Vacant, Fallow and Virgin Lands
worsen. Going forward, policymakers need to consider several Management Law, including freedom to farm crops of farmers’
policy dimensions to prevent significant increases in deforestation choice, recognition of equal rights for women, recognition of
while focusing on greater agricultural production (e.g., Angelsen, customary laws governing land use, creation of independent
2010; Sayer et al., 2013). bodies to hear and adjudicate land disputes, prioritization of land
There are mounting calls within Myanmar for unambiguous allocation for marginalized citizens, and adhering to the concepts
environmental safeguards and mainstreamed environmental and of free, prior and informed consent (Oberndorf, 2012). Beyond
social impact assessments, improved land-use planning, a amending current law, however, recommendations for a compre-
strengthened environment agency and training for government hensive Land Law have been made to explicitly safeguard
on how to deal with environmental issues, and mechanisms for smallholder rights, create secure land tenure (e.g. prevention of
public engagement and consultations (Rao et al., 2002; Burma land foreclosures), and clarify both the land registration procedure
Environmental Working Group, 2011). And, this analysis further and the land use classification system (Oberndorf, 2012). Land
tenure reform would provide opportunity for smallholders to drivers requires an integrated approach (Mattison and Norris,
improve their long-term agricultural objectives, allow them to 2005). Our quantitative analysis demonstrates that Myanmar’s
better access international markets, and prevent land seizures Ayeyarwady Delta mangroves are under considerably greater
(Song, 2013), while at the same time clearly demarcating State threat than previously documented. Further, there is an
managed lands such as production forests and protected areas. urgent need to revisit business-as-usual scenarios such as the
Such transparency could help reduce contestation over land and one we present for the Ayeyarwady Delta, and to reconsider
promote better land stewardship. conservation priority-setting and threat analyses, and to
Environmental governance policy needs to advance alongside anticipate future drivers of change. As Myanmar’s government
the development of policies designed to enhance international takes steps toward political and economic reform and is
enterprise and agricultural output. Given that previous research rewarded with increased international engagement, private
has strongly implicated large, enterprise-oriented agroindustrial investment and overseas aid, it also potentially faces increased
development in recent tropical deforestation in Asia, policy environmental pressures. These necessitate proactive environ-
mechanisms to protect forests, both in the Delta and in upland mental safeguarding and precautionary management. However,
areas, need emphasis. Given clear land use planning, agrotechno- given the current state of environmental governance and
logical changes could simultaneously facilitate intensification and regional precedents, recent policy developments seem poised
land sparing for forest conservation (Lambin and Meyfroidt, 2011, to deeply and negatively affect remaining natural ecosystems
but see Morton et al., 2006). Policy could require conservation set- across Myanmar.
aside programs funded by large-scale agroindustry, to not only
enhance the current protected area system but also to require Acknowledgments
large-scale farming to occur on limited land (Rudel, 2007). Such
policies will likely only occur through efforts by coalitions of local, ELW and DAF were supported by the Singapore-Delft Water
national and international conservation interests that ‘‘pressure for Alliance, grants R303-001-001-272 and R303-001-024-414. ELW
institutional safeguards for natural forests and contest growth was supported by Ministry of Education, Government of
coalition plans for expanding agricultural enterprises at the Singapore grant R-154-000-400-133. ADZ and NRAJ were
expense of tropical forests’’ (Rudel, 2007). Such grassroots farmer supported by NUS Faculty of Arts and Sciences grant R109-000-
coalitions appear to be forming (e.g. Lwin and Ei, 2013). However, 109-646 and Singapore-Delft Water Alliance grant 303-001-020-
such a protectionist agenda needs to be undertaken carefully so 414. JP was supported by the Harry S. Truman Foundation and
that it does not result in ‘‘green grabbing’’, which could end up National University of Singapore President’s Graduate Fellowship.
excluding local people from forests (e.g. Fairhead et al., 2012). The authors thank D. Dapice for early conceptual discussions, and
D. Dapice and K.W. Krauss for helpful comments on the
4. Conclusion manuscript.
Understanding future prospects for conservation in a dynamic

environment characterized by shifting policies and economic Appendix A
Fig. A1. Distribution of historical training points for Model H. There were 168 points of each class: mangrove (green dots, dense mangrove points taken from the ground
surveys), non-mangrove (brown dots) and water (white dots). Water points were randomly selected from within the blue areas and non-mangrove terrestrial points were
selected within the area delineated by the black line.
Table A1
Ten-fold cross-validation accuracy assessments of historical land-cover models based on Training Dataset H (historical).
Model 2009 2008 2000 1989 1978
PA UA PA UA PA UA PA UA PA UA
Mangrove 0.940 0.893 0.863 0.863 0.952 0.930 0.958 0.931 0.845 0.888
Non-mangrove 0.988 0.965 0.994 0.994 0.917 0.951 0.929 0.929 0.899 0.839
Water 0.863 0.935 0.845 0.845 1.000 0.988 0.958 0.988 0.952 0.976
Overall accuracy 93.1% 90.1% 95.6% 94.8% 89.9%
Cohen’s Kappa 0.896 0.851 0.935 0.923 0.848
PA, producer’s accuracy; UA, user’s accuracy.
Table A2 Boehler, P., 2012. Dawei power plant cancelled. The Irrawaddy, January 10, In:
Ten-fold cross-validation accuracy assessments of 2011 Ayeyarwady land cover http://www.irrawaddy.org/article.php?art_id=22804.
model based on Training Dataset C (contemporary). Boot, W., 2010. Abhisit will seek business advantages in Burma. The Irrawaddy,
October 7, In: http://www2.irrawaddy.org/article.php?art_id=19668.
PA UA Borras, S.M., Franco, J.C., 2010. Contemporary discourses and contestations around
pro-poor land policies and land governance. Journal of Agrarian Change 10,
Mangrove 0.869 0.752 1–32.
Degraded mangrove 0.333 0.500 Brando, P.M., Coe, M.T., DeFries, R., Azevedo, A.A., 2013. Ecology, economy and
Agriculture 0.840 0.866 management of an agroindustrial frontier landscape in the southeast Amazon.
Water 0.972 0.972 Philosophical Transactions of the Royal Society B 368 , http://dx.doi.org/
Other 0.529 0.563 10.1098/rstb.2012.0152.
Overall accuracy 81.6% Breiman, L., 2001. Statistical modeling: the two cultures. Statistical Science 16,
Cohen’s Kappa 0.749 199–231.
Buchanan, J., Kramer, T., Franco, J., 2013. Developing Disparity: Regional Investment
PA, producer’s accuracy; UA, user’s accuracy. in Burma’s Borderlands. Transnational Institute, Netherlands. http://
www.tni.org/work-area/burma-project. (accessed 07.10.13).
Burma Environmental Working Group, 2011. Burma’s Environment: People, Pro-
References blems, Policies. Wanida Press, Chiang Mai. , In: http://www.bewg.org.
Carbonnier, G., 2011. Introduction: the global and local governance of extractive
Abdullah, S.A., Hezri, A.A., 2008. From forest landscape to agricultural landscape in resources. Global Governance 17, 135–147.
the developing tropical country of Malaysia: pattern, process and their signifi- Casey, M., 2007. Myanmar’s Other Disaster: Its Environment. Associated Press,
cance on policy. Environmental Management 42, 907–917. October 16, In: http://www.msnbc.msn.com/id/21306082/ns/world_news-worl-
Alatorre, L.C., Sánchez-Andrés, R., Cirujano, S., Beguerı́a, S., Sánchez-Carrillo, S., d_environment/t/myanmars-other-disaster-its-environment/#.Ty9Kv-P9NeI.
2011. Identification of mangrove areas by remote sensing: the ROC curve Central Intelligence Agency, 2010. Burma. The World Factbook. https://www.
technique applied to the Northwestern Mexico coastal zone using Landsat cia.gov/library/publications/the-world-factbook/geos/bm.html (accessed
imagery. Remote Sensing 3, 1568–1583. 20.03.12).
Angelsen, A., Kaimowitz, D., 1999. Rethinking the causes of deforestation: lessons Central Statistics Organization (2007) Available at http://www.myanmararchives.-
from economic models. The World Bank Research Observer 14, 73–98. com/myanmardata2007/default1.htm. Accessed 16 February 2013.
Angelsen, A., 2010. Policies for reduced deforestation and their impact on agricul- Chanjaroen, C., 2012. Myanmar Seeks to Restore Status among Top Rice Exporters.
tural production. Proceedings of the National Academy of Sciences USA 107, Bloomberg, October 29. http://www.bloomberg.com/news/2012-10-28/myan-
19639–19644. mar-seeks-to-regain-top-spot-in-rice-exports-southeast-asia.html (accessed
Asian Development Bank, 2011. Asian Development Outlook 2011: South–South 12.02.13).
Economic Links: Myanmar.. http://beta.adb.org/publications/asian-develop- Chong, K.Y., 2012. Religiously protecting Myanmar’s environment. Science 337, 1604.
ment-outlook-2011-south-south-economic-links (accessed 16.02.13). Claessens, S., Djankov, S., Lang, L.H.P., 2000. The separation of ownership and
Asian Development Bank, 2012. Myanmar in Trasnition: Opportunities and Chal- control in East Asian Corporations. Journal of Financial Economics 58, 81–112.
lenges., In: http://www.adb.org/sites/default/files/pub/2012/myanmar-in- Congalton, R.G., 1991. A review of assessing the accuracy of classification of
transition.pdf. remotely sensed data. Remote Sensing of Environment 37, 35–46.
Aung, N.M., 2011.In: Agricultural Efficiency of Rice Farmers in Myanmar: A Critical Ecosystems Partnership Fund, 2012. Indo-Burma Biodiversity Hotspot:
Case Study in Selected Areas, Institute of Developing Economies Discussion 2011 Update.. http://www.cepf.net/SiteCollectionDocuments/indo_burma/
Paper 306, In: http://www.ide.go.jp/English/Publish/Download/Dp/pdf/ IndoBurma_ecosystemprofile_2011_update.pdf (accessed 12.02.13).
306.pdf. Dapice, D.O., Vallely, T.J., Wilkinson, B., 2010.In: Revitalizing Agriculture in Myan-
Aung, T.T., Than, M.M., Katsuhiro, O., Yukira, M., 2011. Assessing the status of three mar: Breaking Down Barriers, Building a Framework for Growth. Ash Center for
mangrove species restored by the local community in the cyclone-affected area Democratic Governance and Innovation Occasional Paper. http://ash.harvar-
of the Ayeyarwady Delta, Myanmar. Wetlands Ecology and Management 19, d.edu/Home/Programs/Institute-for-Asia/Publications/Occasional-Papers
195–208. (accessed 20.03.12).
Aung, T.T., Mochida, Y., Than, M.M., 2013. Prediction of recovery pathways of Dapice, D.O., Vallely, T.J., Wilkinson, B., McPherson, M., 2011.In: Myanmar Agricul-
cyclone-disturbed mangroves in the mega delta of Myanmar. Forest Ecology ture in 2011: Old Problems and New Challenges. Ash Center for Democratic
and Management 293, 103–113. Governance and Innovation Occasional Paper. http://ash.harvard.edu/
AusAID, 2013. Australia – Myanmar Aid Program Strategy (2012–14), PUB18.. Home/Programs/Institute-for-Asia/Publications/Occasional-Papers (accessed
http://www.ausaid.gov.au/publications/Pages/myanmar-country-strategy- 20.03.12).
2012-14.aspx (accessed 15.02.13). Dapice, D.O., Saich, A.J., Vallely, T.J., 2012.In: Appraising the Post-sanctions Pro-
Baker, P., 2012. Obama, in an emerging Myanmar, vows support. The New York spects for Myanmar’s Economy: Choosing the Right Path. Ash Center for
Times, November 18. http://www.nytimes.com/2012/11/19/world/asia/ Democratic Governance and Innovation Occasional Paper. http://ash.harvar-
obama-heads-to-myanmar-as-it-promises-more-reforms.html?pagewante- d.edu/Home/Programs/Institute-for-Asia/Publications/Occasional-Papers
d=all (accessed 20.01.13). (accessed 20.03.12).
Barbier, E.B., 2003. Habitat–fishery linkages and mangrove loss in Thailand. Con- DeFries, R., Rudel, T., Uriarte, M., Hansen, M., 2010. Deforestation driven by urban
temporary Economic Policy 21, 59–77. population growth and agricultural trade in the twenty-first century. Nature
Barbier, E.B., Koch, E.W., Silliman, B.R., Hacker, S.D., Wolanski, E., Primavera, J., Geoscience 3, 178–181.
Granek, E.F., Polasky, S., Aswani, S., Cramer, L.A., Stoms, D.M., Kennedy, C.J., Bael, DeFries, R., Herold, M., Verchot, L., Macedo, M.N., Shimabukuro, Y., 2013. Export-
D., Kappel, C.V., Perillo, G.M.E., Reed, D.J., 2008. Coastal ecosystem-based oriented deforestation in Mato Grosso: harbinger or exception for other tropical
management with nonlinear ecological functions and values. Science 319, forests? Philosophical Transactions of the Royal Society B 368, 20120173,
321–323. http://dx.doi.org/10.1098/rstb.2012.0173.
Barbier, E.B., Hacker, S.D., Kennedy, C., Koch, E.W., Stier, A.C., Silliman, B.R., 2011. De Graaf, G.J., Xuan, T.T., 1998. Extensive shrimp farming, mangrove clearance and
The value of estuarine and coastal ecosystem services. Ecological Monographs marine fisheries in the southern provinces of Vietnam. Mangroves and Salt
81, 169–193. Marshes 2, 159–166.
Barta, P., 2012. Myanmar wins over more old foes. The Wall Street Journal, January Delang, C.O., 2005. The political ecology of deforestation in Thailand. Geography 90,
16, In: http://online.wsj.com/article/SB100014240529702045559045771644- 225–237.
92935463460.html. Donato, D.C., Kauffman, J.B., Murdiyarso, D., Kurnianto, S., Stidham, M., Kanninen,
Bissinger, J., 2012. Foreign investment in Myanmar: a resource boom but a devel- M., 2011. Mangroves among the most carbon-rich forests in the tropics. Nature
opment bust? Contemporary Southeast Asia 34, 23–52. Geoscience 4, 293–297.
Duke, N.C., Meynecke, J.-O., Dittmann, S., Ellison, A.M., Anger, K., Berger, U., Cannicci, Kyaw, K., 2012 August 25. Land reform key to Burma’s future. The Diplomat. http://
S., Diele, K., Ewel, K.C., Field, C.D., Koedam, N., Lee, S.Y., Marchand, C., Nordhaus, thediplomat.com/2012/08/25/land-reform-key-to-burmas-future/. (accessed
I., Dahdouh-Guebas, F., 2007. A world without mangroves? Science 6, 41–42. 02.10.13).
Fabrikant, M., 2013. Huge potential for Myanmar Seafood Sector. The Paramus Post, Kyaw, A.T., 2013 January 14. Japanese ODA to fund Myanmar’s infrastructure. The
January 30. http://www.paramuspost.com/article.php/201301300718385 Myamar Times. http://www.mmtimes.com/index.php/business/3756-japa-
(accessed 12.02.13). nese-oda-to-fund-nation-s-infrastructure.html. (accessed 12.02.13).
Fairhead, J., Leach, M., Scoones, I., 2012. Green grabbing: a new appropriateion of La Porta, R., Lopez-de-Silanes, F., Shleifer, A., 1999. Corporate ownership around the
nature? Journal of Peasant Studies 39, 237–261. world. The Journal of Finance 54, 471–517.
FAO, 2010.In: Global Forest Resources Assessment 2010. Rome. http:// Lambin, E.F., Geist, H.J., Lepers, E., 2003. Dynamics of land-use and land-cover
www.fao.org/forestry/fra/fra2010/en/ (accessed 15.02.13). change in tropical regions. Annual Review of Environment and Resources 28,
FAO, 2006–2013.In: National Aquaculture Sector Overview. Viet Nam. National 205–241.
Aquaculture Sector Overview Fact Sheets. Text by Nguyen, T.P. and Truong, H.M. Lambin, E.F., Meyfroidt, P., 2011. Global land use change, economic globalization,
In: FAO Fisheries and Aquaculture Department [online]. Rome. Updated 10 and the looming land scarcity. Proceedings of the National Academy of Sciences
October 2005. http://www.fao.org/fishery/countrysector/naso_vietnam/en USA 108, 3465–3472.
(cited 22.01.13). Leimgruber, P., Kelly, D.S., Steininger, M.K., Brunner, J., Müller, Songer, M., 2005.
FAO, 2013. Agro-Maps: Global Spatial Database of Agricultural Land-use Statistics.. Forest cover change patterns in Myanmar (Burma) 1990–2000. Environmental
http://kids.fao.org/agromaps (accessed 16.02.13). Conservation 32, 356–364.
Fatoyinbo, T.E., Simard, M., Washington-Allen, R.A., Shugart, H.H., 2008. Landscape- Li, J.C. (2008) Environmental Impact Assessments in Developing Countries: An
scale extent, height, biomass and carbon estimation of Mozambique’s man- Opportunity for Greater Environmental Security? Working Paper No. 4. United
grove forests with Landsat ETM+ and Shuttle Radar Topography Mission States Agency for International Development (USAID) and Foundation of Envi-
elevation data. Journal of Geophysical Research 113, G02–S06. ronmental Security & Sustainability (FESS). Available from http://research.fi-
Feagin, R.A., Mukherjee, N., Shanker, K., Baird, A.H., Cinner, J., Kerr, A.M., Koedam, N., t.edu/sealevelriselibrary/documents/doc_mgr/338/Global_EIAs_in_Develo-
Sridhar, A., Arthur, R., Jayatissa, L.P., Seen, D.L., Menon, M., Rodriguez, S., ping_Countries_-_Li_2008.pdf.
Shamsuddoha, Md., Dahdouh-Guebas, F., 2010. Shelter from the storm? Use Li, L., Aide, T.M., Ma, Y., Liu, W., Cao, M., 2007. Demand for rubber is causing the loss
and misuse of coastal vegetation bioshields for managing natural disasters. of high diversity rain forest in SW China. Biodiversity and Conservation 16,
Conservation Letters 3, 1–11. 1731–1745.
Flaherty, M., Vandergeest, P., Miller, P., 1999. Rice paddy or shrimp pond: tough Loneragan, N.R., Adnan, N.A., Connolly, R.M., Manson, F.J., 2005. Prawn landings and
decisions in rural Thailand. World Development 27, 2045–2060. their relationship with the extent of mangroves and shallow waters in western
Food Security Working Group, 2012.In: The role of land tenure security for small- peninsular Malaysia. Estuarine, Coastal and Shelf Science 63, 187–200.
holder farmers in national development, Policy Discussion brief, Food Security Lwin, S.T.A., Ei, K.K., 2013 August 20. Myanmar farmers call for amendments to Land
Working Group of Myanmar. Law. RFA. http://www.rfa.org/english/news/myanmar/farmers-082020131-
Fritz, H.M., Blount, C.D., Thwin, S., Thu, M.K., Chan, N., 2009. Cyclone Nargis storm 80804.html. (accessed 07.10.13).
surge in Myanmar. Nature Geoscience 2, 448–449. Macintosh, D.J., Mahindapala, R., Markopoulos, M. (Eds.), 2012. Sharing Lessons on
Fuller, T., 2011. Myanmar backs down, suspending dam project. The New York Mangrove Restoration. Mangroves for the Future and Gland, Switzerland: IUCN,
Times, September 30. http://www.nytimes.com/2011/10/01/world/asia/myan- Bangkok, Thailand.
mar-suspends-construction-of-controversial-dam.html (accessed 20.03.12). Mattison, E.H.A., Norris, K., 2005. Bridging the gap between agricultural policy,
Fuller, T., 2012. Long reliant on China, Myanmar now turns to Japan. The New York land-use and biodiversity. Trends in Ecology and Evolution 20, 610–616.
Times, October 10. http://www.nytimes.com/2012/10/11/world/asia/long-re- Matsuda, M., 2009. Dynamics of rice production development in Myanmar: growth
liant-on-china-myanmar-now-turns-to-japan-for-help.html?pagewanted=all centers, technological changes, and driving forces. Tropical Agriculture and
(accessed 20.02.13). Development 53, 14–27.
Geist, H.J., Lambin, E.F., 2002. Proximate causes and underlying driving forces of Mayer Brown, J.S.M., 2012. Details of Myanmar’s new Foreign Investment Law.
tropical deforestation. BioScience 52, 143–150. http://www.mayerbrown.com/Details-of-Myanmars-New-Foreign-Invest-
Giri, C., Zhu, Z., Tieszen, L.L., Singh, A., Gillette, S., Kelmelis, J.A., 2008. Mangrove ment-Law-11-19-2012/. (accessed 07.10.13).
forest distributions and dynamics (1975–2005) of the tsunami-affected region Mendelsohn, R., 1994. Property rights and tropical deforestation. Oxford Economic
of Asia. Journal of Biogeography 35, 519–528. Papers 46, 750–756.
Goss, J., Burch, D., Rickson, R.E., 2000. Agri-food restructuring and third world Ministry of Agriculture and Irrigation, 2010. Business opportunities in Myanmar
transnationals: Thailand, the CP Group and the global shrimp industry. World agriculture.. http://www.moai.gov.mm/index.php?option=com_content&-
Development 28, 513–530. view=category&layout=blog&id=15&Itemid=9. (accessed 20.03.12).
GRASS Development Team, 2011.In: Geographic Resources Analysis Support Sys- Mirdha, R.U., 2011 August 3. Bangladesh, Myanmar Agree to Increase Trade. The
tem (GRASS) Software, Open Source Geospatial Foundation Project, In: http:// Daily Star.. http://www.thedailystar.net/newDesign/news-details.php?-
grass.osgeo.org. nid=196943. (accessed 20.03.12).
Hall, D., 2011. Land grabs, land control, and Southeast Asian crop booms. Journal of Montlake, S., 2013 August 8. Robert Kuok’s plantation firm eyes Myanmar as next
Peasant Studies 38, 837–857. frontier. Forbes. http://www.forbes.com/sites/simonmontlake/2013/08/08/
Hall, M., Frank, E., Holmes, G., Pfahringer, B., Reutemann, P., Witten, I.H., 2009. The robert-kuoks-plantation-firm-eye-myanmar-as-next-sugar-frontier/.
WEKA data mining software: an update. SIGKDD Explorations 11, 10–18. (accessed 07.10.13).
Human Rights Foundation of Monland, 2012. Examination and critique of the 2012 Morton, D.C., DeFries, R.S., Shimabukuro, Y.E., Anderson, L.O., Arai, E., Espirito-Santo,
Farmland Bill, May 29, 2012.. http://rehmonnya.org/archives/2238. (accessed F.B., Freitas, R., Morisette, J., 2006. Cropland expansion changes deforestation
02.10.13). dynamics in the southern Brazilian Amazon. Proceedings of the National
IUCN (International Union for the Conservation of Nature), 2011. IUCN Red List of Academy of Sciences USA 103, 14637–14641.
Threatened Species. Version 2011.2. , In: http://www.iucnredlist.org. Myers, N., Mittermeier, R.A., Meittermeier, C.G., da Fonseca, G.A.B., Kent, J., 2000.
Izquierdo, A.E., Grau, H.R., Aide, T.M., 2011. Implications of rural-urban migration Biodiversity hotspots for conservation priorities. Nature 405, 853–858.
for conservation of Atlantic forest and urban growth in Misiones, Argentina National Aeronautics and Space Administration, 2011. The Landsat Program., In:
(1970–2030). Ambio 40, 298–309. http://landsat.gsfc.nasa.gov.
Japan Times, 2012. Edano meets Suu Kyi, vows support for Myanmar economy to New Light of Myanmar, 2013. Farmland disputes to be solved in accord with laws
aid reform. The Japan Times, January 13. http://www.japantimes.co.jp/text/ and rules in consideration of rights and opportunity of each people for correct
nn20120113a9.html (accessed 20.03.12). results. The New Light of Myanmar, 28 September. http://myanmargeneva.org/
Kate, P.T., Kubota, Y., 2012. Asian firms in pole as Myanmar beckons. Reuters, newseng2013.htm. (accessed 02.10.13).
January 9. http://in.reuters.com/article/2012/01/09/myanmar-business-idIN- Oberndorf, R.B., 2012. Legal Review of Recently Enacted Farmland Law and Vacant,
DEE80804H20120109 (accessed 20.03.12). Fallow and Virgin Lands Management law. Food Security Working Group,
Koch, E.W., Barbier, E.B., Silliman, B.R., Reed, D.J., Perillo, G.M.E., Hacker, S.D., Myanmar. http://www.forest-trends.org/publication_details.php?publicatio-
Granek, E.F., Primavera, J.H., Muthiga, N., Polasky, S., Halpern, B.S., Kennedy, nID=3274. (accessed 02.10.13).
C.J., Kappel, C.V., Wolanski, E., 2009. Non-linearity in ecosystem services: Okamoto, I., 2005.In: Transformation of the Rice Marketing System and Myanmar’s
temporal and spatial variability in coastal protection. Frontiers in Ecology Transition to a Market Economy, Discussion Paper No. 43, Institute of Devel-
and the Environment 7, 29–37. oping Economies, Japan External Trade Organization, Chiba.
Koh, L.P., Wilcove, D.S., 2008. Is oil palm agriculture really destroying tropical Okamoto, I., 2009. Transformation of the rice marketing system after market
biodiversity? Conservation Letters 1, 60–64. liberalization in Myanmar. In: Fujita, K., Mieno, F., Okamoto, I. (Eds.), The
Kohavi, R., 1995. A study of cross-validation and bootstrap for accuracy estimation Economic Transition in Myanmar after 1988: Market Economy Versus State
and model selection. In: Mellish, C.S. (Ed.), Proceedings of the 14th International Control. NUS Press, pp. 216–245.
Joint Conference on Artificial Intelligence. Morgan Kaufmann Publishers, pp. Oo, N.W., 2002. Present state and problems of mangrove management in Myanmar.
1137–1143. Trees 16, 218–223.
Kurosaki, T., Okamoto, I., Kurita, K., Fujita, K., 2004. Rich Periphery, Poor Center: Osborne, L., 2013. As Myanmar opens up, forests face new risks. Deutsche Welle, 6
Myamar’s Rural Economy under Partial Transition to a Market Economy. Discus- August 2013. http://dw.de/p/19KUI. (accessed 02.10.13).
sion Paper 23, Institute of Economic Research, Hitotsubashi University, Tokyo. Phyu, A.S., 2012. Scientists fear Myanmar’s disappearing mangroves. Myanmar
Kyaw, A.T., 2012 January 13. Banking industry awakens from slumber. The Myan- Times, November 9. http://www.mmtimes.com/index.php/national-news/
mar Times. http://www.mmtimes.com/2012/business/608/biz60801.html. 3009-scientists-fear-myanmar-s-disappearing-mangroves.html. (accessed
(accessed 16.02.13). 08.10.13).
Platt, J., 1998. Fast training of support vector machines using sequential minimal 6080-india-wants-to-invest-in-burmas-agricultural-and-energy-sectors.html.
optimization. In: Schoelkopf, B., Burges, C., Smola, A. (Eds.), Advances in Kernel (accessed 20.03.12).
Methods – Support Vector Learning. The MIT Press, pp. 185–208. Than, M., 2001. Changing faces of the Ayeyarwady (Irrawaddy) Delta (1850–2000).
Polidoro, B.A., Carpenter, K.E., Collins, L., Duke, N.C., Ellison, A.M., Ellison, J.C., Institute of Southeast Asian Studies, Singapore.
Farnsworth, E.J., Fernando, E.S., Kathiresan, K., Koedam, N.E., Livingstone, Thorbjarnarson, J., Platt, S.G., Khaing, U.S.R., 2008. A population survey of the
S.R., Miyagi, T., Moore, G.E., Nam, V.N., Ong, J.E., Primavera, J.H., Salmo III, estuarine crocodile in the Ayeyarwady Delta. Myanmar, Oryx 4, 317–324.
S.G., Sanciangco, J.C., Sukardjo, S., Wang, Y., Yong, J.W.H., 2010. The loss of Thu, P.M., Populus, J., 2007. Status and changes of mangrove forest in Mekong Delta:
species: mangrove extinction risk and geographic areas of concern. PLoS One 5, case study in Tra Vinh, Vietnam. Estuarine, Coastal and Shelf Science 71,
e10095, http://dx.doi.org/10.1371/journal.pone.0010095. 98–109.
Pratruangkrai, P., 2012. Thailand, Myamar aim to triple trade by 2015. The Nation, Tordoff, A.W., Eames, J.C., Eberhardt, K., Baltzer, M.C., Davidson, P., Leimgruber, J.,
November 3. http://www.nationmultimedia.com/business/Thailand-Myan- Uga, U., Than, U.A., 2005. Myanmar: Investment Opportunities in Biodiversity
mar-aim-to-triple-trade-by-2015-30193589.html. (accessed 16.02.13). Conservation. Birdlife International, In: http://birdlifeindochina.org/content/
Rakotomavo, A., Fromard, F., 2010. Dynamics of mangrove forests in the Mangoky myanmar-investment-opportunities-biodiversity-conservation.
River delta, Madagascar, under the influence of natural and human factors. United Nations Development Program, 2011. Human Development Report 2011.,
Forest Ecology and Management 259, 1161–1169. In: http://hdr.undp.org/en/.
Rao, M., Htun, S., Platt, S.G., Tizard, R., Poole, C., Myint, T., Watson, J.E.M., 2013. Wang, C., Wang, F., Wang, Q., Yang, D., Li, L., Zhang, X., 2013. Preparing for
Biodiversity conservation in a changing climate: a review of threats and Myanmar’s environment-friendly reform. Environmental Science and Policy
implications for conservation planning in Myanmar. Ambio., http:// 25, 229–233.
dx.doi.org/10.1007/s13280-013-0423-5. Ware, M., 2000. WEKA Documentation. http://weka.sourceforge.net/doc.stable/
Rao, M., Rabinowitz, A., Khaing, S.T., 2002. Status review of the protected-area Webb, E.L., Phelps, J., Friess, D.A., Rao, M., Ziegler, A., 2012. Environment-friendly
system in Myanmar, with recommendations for conservation planning. Con- reform in Myanmar. Science 336, 295.
servation Biology 16, 360–368. Wildlife Conservation Society, 2013. Myanmar Biodiversity: Conservation Invest-
Reilly, J., 2012. A Norm-Taker or a Norm-Maker? Chinese aid in Southeast Asia. ment Vision. Wildlife Conservation Society, Yangon, Myanmar.
Journal of Contemporary China 21, 71–91. World Wide Fund for Nature, 2008. Irrawaddy freshwater swamp forests. In:
Rodriguez, J.J., Kuncheva, L.I., Alonso, C.J., 2006. Rotation forest: a new classifier Cleveland, C.J. (Ed.), Encyclopedia of Earth, In: http://www.eoearth.org/arti-
ensemble method. IEEE Transactions on Pattern Analysis and Machine Intelli- cle/Irrawaddy_freshwater_swamp_forests.
gence 28, 1619–1630. Woods, K., Canby, K., 2011. Baseline study 4, Myanmar: overview of forest law
Rudel, T., 2007. Changing agents of deforestation: from state-initiated to enterprise enforcement, governance and trade. EU FLEGT Facility. http://www.euflegt.e-
driven processes, 1970–2000. Land Use Policy 24, 35–41. fi.int. (accessed 07.10.13).
Rudel, T.K., Coomes, O.T., Moran, E., Achard, F., Angelsen, A., Xu, J., Lambin, E., 2005. Woods, N., 2008. Whose aid? Whose influence? China, emerging donors and
Forest transitions: towards a global understanding of land use change. Global the silent revolution in development assistance. International Affairs 84,
Environmental Change 15, 23–31. 1205–1221.
Rudel, T.K., Schneider, L., Uriarte, M., Turner III, B.L., DeFries, R., Lawrence, D., World Bank (2012) The World Bank’s Worldwide Governance Indicators. Available
Geoghegan, J., Hecht, S., Ickowitz, A., Lambin, E., Birkenholtz, T., Baptista, S., at http://info.worldbank.org/governance/wgi/index.asp. Accessed 14 February
Grau, R., 2009. Agricultural intensification and changes in cultivated areas, 2013.
1970–2005. Proceedings of the National Academy of Sciences USA 106, 20675– World Bank (2013a) World Development Indicators. Available at http://data.world-
20680. bank.org/indicator. Accessed 15 February 2013.
Ruiz-Luna, A., Escobar, A.C., Berlanga-Robles, C., 2010. Assessing distribution pat- World Bank (2013b) World Bank group to support Myanmar’s plan to improve
terns, extent, and current condition of Northwest Mexico mangroves. Wetlands people’s access to electricity [press release]. Available at http://www.world-
30, 717–723. bank.org/en/news/press-release/2013/02/05/World-Bank-Group-to-Support-
Sayer, J., Sunderland, T., Ghazoul, J., Pfund, J.-L., Sheil, D., Meijaard, E., Venter, M., Myanmar-8217-s-Plan-to-Improve-People-8217-s-Access-to-Electricity.
Boedhihartono, A.K., Day, M., Garcia, C., van Oosten, C., Buck, L.E., 2013. Ten Accessed 12 February 2013.
principles for a landscape approach to reconciling agriculture, conservation, Xiao, X., Boles, S., Frolking, S., Li, C., Babu, J., Salas, W., Moore III, B., 2006. Mapping
and other competing land uses. Proceedings of the National Academy of paddy rice agriculture in South and Southeast Asia using multi-temporal MODIS
Sciences USA 110, 8349–8356. images. Remote Sensing of Environment 100, 95–113.
Schmidt, C., 2012. As isolation ends, Myanmar faces new ecological risks. Science Xinhua, 2012. Myanmar’s foreign trade up 28% in nine months. Xinhua, January 12.
337, 796–797. http://www.asean-cn.org/Item/4396.aspx. (accessed 20.03.12).
Sloan, S., Pelletier, J., 2012. How accurately may we project tropical forest-cover Yap, C.-W., 2010. Chinese weapons maker signs Myanmar Deal. The Wall Street
change? A validation of a forward-looking baseline for REDD. Global Environ- Journal, June 23. http://online.wsj.com/article/SB1000142405274870390000-
mental Change 22, 440–453. 4575324420133952674.html?mod=fox_australian#articleTabs%3Darticle.
Song, S., 2013. Land seizure in Myanmar continues despite reforms, government (accessed 20.03.12).
faces internal and external pressure to develop land policies. International Yunfei, Z., Lingling, D., 2010. Myanmar business leaders stress China-Myanmar
Business Times, August 27, In: http://www.ibtimes.com/land-seizure-myan- cooperation in face of FTA challenges. Xinhua, January 1. http://news.xinhua-
mar-continues-despite-reforms-government-faces-internal-external-pres- net.com/english/2010-01/01/content_12741034.htm. (accessed 20.03.12).
sure-develop. Zaw, H.N., Khaing, A.K., 2013. Military involved in massive land grabs: Parliamen-
Songer, M., Aung, M., Senior, B., DeFries, R., Leimgruber, P., 2009. Spatial and tary report. The Irrawaddy, March 5. http://www.irrawaddy.org/human-rights/
temporal deforestation dynamics in protected and unprotected dry forests: a military-involved-in-massive-land-grabs-parliamentary-report.html.
case study from Myanmar (Burma). Biodiversity and Conservation 1001–1018. (accessed 07.10.13).
Spalding, M., Kainuma, M., Collins, L., 2010. World Atlas of Mangroves. Earthscan, Zaw, K., Lwin, N.N., Nyein, K.T., Thandar, M., 2011. Agricultural transformation,
London. institutional changes, and rural development in Ayeyarwady Delta, Myan-
Szep, J., 2012. Myanmar seeks to rejoin top rice exporters. Reuters, August 9. http:// mar. In: Intal, Jr., P.S., Oum, S., Simorangkir, M.J.O. (Eds.), Agricultural Devel-
in.reuters.com/article/2012/08/09/myanmar-rice-idINL4E8IJ21O20120809. opment, Trade and Regional Cooperation in Developing East Asia. ERIA,
(accessed 14.02.13). Jakarta, pp. 307–355.
Te Te, 2011. India wants to invest in Burma’s agricultural and energy sectors. Zoomers, A., 2010. Globalisation and the foreignisation of space: seven processes
Mizima News, October 19. http://www.mizzima.com/edop/interview/ driving the current global land grab. The Journal of Peasant Studies 37, 429–447.

Deforestation in The Ayeyarwady Delta (2014)

Uploaded by

Copyright:

Available Formats

Deforestation in The Ayeyarwady Delta (2014)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Deforestation in The Ayeyarwady Delta (2014)

Uploaded by

Copyright:

Available Formats

Global Environmental Change 24 (2014) 321–333

Contents lists available at ScienceDirect

Global Environmental Change

Deforestation in the Ayeyarwady Delta and the conservation

1. Introduction mangroves in Myanmar, but which has sustained the highest

2.3. Accuracy assessment

Accuracies were estimated using a 10-fold cross validation

Model year Area (km2) Range

1978 2623 2329–3030

This study Myanmar Landsat + ASTER GDEM, Model H 3 89.9–95.6 0.848–0.935

N/R, not reported.

2.4. Deforestation scenarios 2.5. Contextualizing the case study

potential for growth (Dapice et al., 2010) and is a development

Understanding future prospects for conservation in a dynamic

Model 2009 2008 2000 1989 1978

PA, producer’s accuracy; UA, user’s accuracy.

You might also like