Content uploaded by Norman C Duke
Author content
All content in this area was uploaded by Norman C Duke on Feb 14, 2019
Content may be subject to copyright.
Large-scale dieback of mangroves in Australia’s Gulf of
Carpentaria: a severe ecosystem response, coincidental
with an unusually extreme weather event
Norman C. Duke
A
,
F
,John M. Kovacs
B
,Anthony D. Griffiths
C
,Luke Preece
D
,
Duncan J. E. Hill
B
,Penny van Oosterzee
D
,
E
,Jock Mackenzie
A
,
Hailey S. Morning
B
and Damien Burrows
A
A
TropWATER Centre, James Cook University, Townsville, Qld 4811, Australia.
B
Department of Geography, Nipissing University, 100 College Drive, North Bay, ON,
P1B 8L7, Canada.
C
Flora and Fauna Division, Department of Land Resource Management, CSIRO Complex,
564 Vanderlin Drive, Berrimah, PO Box 496, Palmerston, NT 0831, Australia.
D
TESS (Centre for Tropical Environment and Sustainability Sciences), James Cook University,
Cairns, Qld 4870, Australia.
E
BIOME5 Pty Ltd, PO Box 1200 Atherton, Qld 4883, Australia.
F
Corresponding author. Email: norman.duke@jcu.edu.au
Abstract. This study records and documents the most severe and notable instance ever reported of sudden and
widespread dieback of mangrove vegetation. Between late 2015 and early 2016, extensive areas of mangrove tidal wetland
vegetation died back along 1000 km of the shoreline of Australia’s remote Gulf of Carpentaria. The cause is not fully
explained, but the timing was coincident with an extreme weather event; notably one of high temperatures and low
precipitation lacking storm winds. The dieback was severe and widespread, affecting more than 7400 ha or 6% of
mangrove vegetation in the affected area from Roper River estuary in the Northern Territory, east to Karumba in
Queensland. At the time, there was an unusually lengthy period of severe drought conditions, unprecedented high
temperatures and a temporary drop in sea level. Although consequential moisture stress appears to have contributed to the
cause, this occurrence was further coincidental with heat-stressed coral bleaching. This article describes the effect and
diagnostic features of this severe dieback event in the Gulf, and considers potential causal factors.
Additional keywords: mangrove forests, plant–climate interactions, tidal wetlands.
Received 19 September 2016, accepted 19 December 2016, published online 14 March 2017
Introduction
During the summer of 2015–16, mangroves in the sparsely
populated Gulf of Carpentaria (the Gulf) area of northern
Australia suffered a particularly severe occurrence of dieback.
News of the incident was, at first, slow to emerge from this
remote, largely unmonitored region. A small number of reports
and images (e.g. Fig. 1) from concerned community members
along the Gulf coast were followed up by ad hoc scientific
surveys during 2016 (Fig. 2) to better define the affected area.
The total area affected extended from Roper River in the
Northern Territory to Karumba in Queensland, a distance of
more than 1000 km of shoreline.
At the time, there were no coincidental anthropogenic or
natural stochastic events likely to cause severe or large-scale
stresses on mangrove forests in the Gulf, such as a large oil spill
(Duke 2016), a severe storm event, a tropical cyclone (e.g.
Cahoon et al. 2003;Paling et al. 2008; Feller et al. 2015), river
flooding (Erftemeijer and Hamerlynck 2005), frost effects (Ross
et al. 2009), locust plagues or other severe herbivory (Reef et al.
2012).
In this first report, it was useful to briefly document key
circumstances surrounding the incident, along with the range
of factors associated with conditions before and during its
occurrence. Observations presented are based on available
evidence, including photographic images, maps, location
data and field survey observations showing areas of severe
loss in tidal wetlands along the southern shoreline of the
Gulf.
These initial observations of the dieback included its extent,
severity, affected habitat type, the species involved, patterns of
dieback, geophysical settings, prevailing hydrological settings,
recent weather conditions and timing of the incident. Such
considerations have been, and will be, used in the development
of hypotheses that will target more rigorous inquiries and assist
CSIRO PUBLISHING
Marine and Freshwater Research
http://dx.doi.org/10.1071/MF16322
Journal compilation CSIRO 2017 www.publish.csiro.au/journals/mfr
in identifying the cause. Based on current findings, an initial
hypothesis has been proposed concerning prevailing weather
conditions as well as other notable coincident factors.
Materials and methods
For these initial environmental surveys, a four-step rapid,
forensic assessment strategy was used, including: (1) an over-
view of the study area, with a general description, shoreline
features and notable biota; (2) validation of mapping methods,
linking satellite imagery and spatial views from available obli-
que photographs; (3) mapping of affected areas, linking satellite
imagery and spatial views from available oblique photographs;
and (4) brief (ad hoc) aerial and field surveys with continuous
image capture using the shoreline video assessment method
(S-VAM; Duke et al. 2010;Mackenzie et al. 2016), further
Fig. 1. Photograph of mangrove dieback taken during an early aerial
survey just east of Limmen Bight estuary mouth, Northern Territory on 29
February 2016 by Paul Barden (location: 1588051.5400S, 135847016.9400E).
0200 km
N
Roper
River
McArthur
River
Limmen
Bight &
Cox Rivers
Liechhardt
River
Norman
River
Centre
Island
Wellesley
Islands
Q U E E N S L A N D
N O R T H E R N
T E R R I T O R Y
G u l f
o f
C a r p e n t a r i a
Fig. 2. Photographs of mangrove dieback in five example locations (A–E) across the southern Gulf of Carpentaria
from the Northern Territory to Queensland. Locations shown in the map at the bottom right-hand corner. Location
A is the Roper River shoreline (14845026.7700S, 135823023.5500E); Location B is the Limmen Bight River shoreline
(1588050.4700S, 135847037.5900E; images taken by Tony Griffiths, 9 June 2016. Location C shows the Gangalidda
shoreline (16852040.5700S, 138855039.1900E; image taken by Roger Jaensch, 13–14 April 2016). Location D is at the
Leichhardt River mouth on 17 November 2016 (17834035.5500S, 139848056.7800E) and Location E is at Karumba,
Queensland, Norman River shoreline (1782602.1300S, 140850039.1800E) on 27 October 2016 (images from surveys
reported in this article).
BMarine and Freshwater Research N. C. Duke et al.
validation of spatial data with specific on-ground measurements
of conditions.
Study area: the Gulf
The Gulf of Carpentaria of tropical northern Australia (Fig. 3)is
a large, shallow (,70-m depth) coastal gulf of ,330 000 km
2
.
It almost equally straddles two political jurisdictions of
Australia, the Northern Territory to the west, and the state of
Queensland to the east. Industries in the region chiefly include
mining, grazing, ecotourism and fisheries.
The Gulf itself is the receiving water body of numerous large
and small river systems that drain ,92 000 GL of water into the
Gulf each year, mainly during the north-west monsoon between
January and March (Burford et al. 2009). The coastal areas
consist of low-lying swampy, chiefly level ground that is largely
inaccessible and comparatively little affected by direct human
activity, making it globally exceptional (Halpern et al. 2008).
However, there are notable effects of livestock trampling and
grazing, as well as feral animal damage. In addition, there are
several mining operations in the region. Pressures from such
activities may have added relevance in this semi-arid region
with relatively low annual rainfalls of ,950 mm on average
(Bureau of Meteorology, see www.bom.gov.au).
Although mangroves are abundant in the area, the dry climate
limits their diversity, height and extent. Owing to the dry
climate, large areas of high intertidal saltpans and saltmarsh
communities cover at least two- to threefold more area than
mangroves. These wide tidal wetland expanses are spread along
shorelines and upstream verges for tens of kilometres of
meandering estuaries of rivers, creeks and back beach systems
throughout the region. The seaward margin forms a low-lying,
continuous sweeping coastline often fringed with mangroves
along with occasional stretches of sandy beaches, fronted by
shallow, broad mud flats and large patches of seagrass meadows
(Poiner et al. 1987;Roelofs et al. 2005). These Gulf habitats
support thousands of marine turtles and dugongs (Bayliss and
Freeland 1989;Kennett et al. 2004), as well as numerous other
estuarine species (Long et al. 1995).
Preliminary validation of satellite image interpretation
Given the remoteness of the area, the extent and character of
mangroves and the recent dieback were mostly described and
mapped using satellite imagery. Field investigations were used
to validate and confirm interpretations made from satellite
imagery. Ten oblique aerial photographs were assessed in spe-
cific validation trials (e.g. Fig. 4). Dieback was readily identified
using available satellite imagery and all affected areas were
mapped.
Mapping of the affected area from satellite imagery
A preliminary scientific visualisation of historical Landsat
products (Fig. 5) was conducted for the southern coast of the
Gulf to determine whether the die-off of the mangroves was a
unique event. All imagery was acquired from the USGS Earth
Explorer website (https://earthexplorer.usgs.gov, last accessed
3 August 2016). Because the months of March, April and May
were found to have the least amount of cloud cover, a time
series based on Landsat 4, 5, 7 and 8 (30-m pixel resolution)
scenes collected from these three months were acquired from
1984 onwards for each of the six locations. For two of these
areas, namely the Nicholson River (path 100/row 72) and the
Limmen Bight (path 102/row 71), older Landsat Multispectral
Scanner (MSS) imagery (79-m pixel resolution) was also
examined dating back to 1972 and 1978 respectively. The
scientific visualisation of these historical records indicated
that the recent die-off was a unique event, occurring between
2015 and 2016. Consequently, a monthly series of Landsat 8
imagery was then created between 2015 and 2016 in order to
identify as much as possible the timing of the effects on the
mangroves.
To determine the extent of mangrove loss between 2015 and
2016, a quantitative binary change detection approach (Kovacs
et al. 2001) was conducted on recent Landsat 8 Level 1 products
(Table 1). Specifically, a near anniversary date comparison was
conducted between these years. Of the 12 Landsat 8 scenes used,
10 were acquired in April and, due to cloud cover, 1 each was
acquired from late March and early May.
The affected area occurs across two Universal Transverse
Mercator (UTM) zones. Therefore, the imagery covering
Queensland was left projected in UTM Zone 54, whereas the
Northern Territory imagery was left projected to UTM Zone 53.
Each image was first radiometrically corrected to surface
reflectance using PCI Geomatica’s (PCI Geomatics, Richmond
Hill, ON, Canada) atmospheric and terrain correction (ATCOR)
module. The surface reflectance images were then mosaicked
for each year and UTM zone, resulting in four separate mosaic
images.
0 500 km
Fig. 3. Catchment areas of the Gulf of Carpentaria (yellow shading)
between the Northern Territory and Queensland, Australia, where
,1000 km of the southern shoreline (grey shading) was affected by severe
dieback of mangroves in late 2015. Crosses mark the locations of five local
Australian Bureau of Meteorology Stations (http://www.bom.gov.au/)
referred to in the present study. Note also the locations of ad hoc aerial
surveys (red lines) conducted during 2016, including the western shoreline
from Roper River to Borroloola on 9 June (A), the eastern shoreline north of
the Norman River mouth and Karratha on 27 October (B) and the eastern
shoreline east and west of the Albert and Leichhardt Rivers near Burketown
on 17 November (C).
Severe mangrove dieback in the Gulf Marine and Freshwater Research C
To set the baseline extent of mangrove coverage for 2015, an
iterative per-pixel unsupervised classification approach using an
iterative self-organising data analysis technique (ISODATA)
algorithm based on all but the cirrus band of Landsat 8 was used
to classify the mangroves at a spatial resolution of 30 m. Using
ancillary data (e.g. Google Earth; https://www.google.com/earth/,
accessed 3 August 2016), on-screen manual editing was then
performed to remove any erroneouslyclassified mangrove pixels.
A normalised difference vegetation index (NDVI) was pro-
duced from each of the four surface reflectance mosaics. The
NDVI was calculated using the following formula:
NDVI ¼NIR redðÞ
NIR þredðÞ
For Landsat 8, the near-infrared (NIR) spectral band and the red
spectral band are collected within the wavelength range of 0.845–
0.885 mm and 0.630–0.680 mm respectively. For each UTM zone,
a NDVI difference image was then created by subtracting the
2016 NDVI image from the 2015 image. To identify areas of
135⬘40E 135⬘42E
Dead mangrove Limmen Bight River
02 km
Mangrove loss
Camera position
135⬘40E
15⬘6S
15⬘4S
15⬘6S
15⬘4S
Aerial oblique photo taken June 2016
135⬘42E 135⬘44E
135⬘44E
April 2015 April 2016
Fig. 4. An example of the current mangrove dieback recorded in satellite imagery of the Limmen Bight (Cox) River estuary mouth and shorelines
in the Northern Territory. The camera view (1584059.6300S, 135842012.7000E) was compared with change detection imagery comparing Landsat 8
April scenes for 2015 and 2016 (top panel) showing healthy vegetation in natural green colours and areas of dieback as dark purple–brown colours.
The bottom two images show yellow patches for the dieback (left), whereas the oblique aerial view shows the location with dieback patches in June
2016 (right).
0 200 km
103 / 70 102 / 70
102 / 71 101 / 71
100 / 72 99 / 72
A U S T R A L I A
UTM Zone 53
UTM Zone 54
UTM Zone 53
UTM Zone 54
N O R T H E R N
T E R R I T O R Y
Q U E E N S L A N D
G u l f
o f
C a r p e n t a r i a
Fig. 5. Southern coast of the Gulf of Carpentaria showing the coverage of
the Landsat scenes (Path/Row; see also Table 1). UTM, Universal Trans-
verse Mercator.
DMarine and Freshwater Research N. C. Duke et al.
mangrove loss, thresholding of the NDVI differenced image was
performed for areas identified as mangrove from the aforemen-
tioned 2015 classification procedure. The threshold value was
determined using visual inspection of the original imagery
(see Fig. 4). Full imagery and enlarged scenes presented in Fig.
6 are available in the Supplementary material.
Aerial and field survey of affected shorelines
Aerial and field surveys of shorelines and mangrove areas were
conducted on three occasions during 2016 (Fig. 3): (1) on 9 June,
of the shoreline from Roper River to McArthur River in the
Northern Territory (1484504200S, 13582303800Eto158440900S,
13683304800E) supported by Northern Territory Government
and JCU TropWATER Centre; (2) on 7 October, in the vicinity
of Karumba in Queensland (17826049.6500S, 14085004.6400E
to 17817042.7800S, 140854011.8900E) supported by WWF, The
Ocean Agency and Carpentaria Land Council Aboriginal
Corporation (CLCAC); and (3) on 17 November along the
shoreline around the Nicholson, Albert and Leichhardt
rivers (17822050.6000S, 139826013.2600Eto1783704.3000S,
14083205.9200E) supported by CLCAC.
In each case, aerial helicopter surveys were conducted along
significant portions of affected shorelines as both a fact-finding
mission and an opportunity to further validate interpretations made
from satellite imagery. The aircraft used were Robinson 44
helicopters, operated by either North Australian Helicopters
(www.northaustralianhelicopters.com.au) (1 above) or Cloncurry
Mustering Company (www.cloncurryhelicopterservices.com.au)
(2 and 3 above). Concurrent on-ground field inspections were
undertaken to evaluate selected mangrove sites during each flight.
In each case, a number of key features were observed, including the
species of trees and shrubs affected, the impact status (classified
primarily as lethal or sublethal), the presence of canopy foliage
(living and dead), the impact on saltmarsh plants, the presence of
leaf litter on the sediment surface and the presence of fauna.
Results
Mapping areas of mangrove dieback
Areas of shoreline and estuaries mapped extended from the
Walter and Roper rivers in the Northern Territory to the Norman
River and Brannigans Creek in Queensland (Fig. 5). Selected
scenes from along this shoreline are shown in Fig. 6 displaying
examples of the patterns of dieback observed, along with the
extent and severity of the dieback.
Three notable observations were apparent from the NDVI
difference images: (1) the dieback of mangroves was severe and
widespread; (2) the patches of mangrove dieback were often
quite large (100–200 m wide); and (3) there were distinctive
patterns associated with the dieback areas, for example man-
groves along estuarine water courses were less affected.
Mangrove extent mapping and NDVI difference images were
used to estimate the total area of dieback. This was then
compared with our estimate of the total area of mangroves in
each of the 14 catchment areas. The total areas of mangrove
dieback and total mangrove cover observed in 2015–16 satellite
imagery are listed in Table 2. These data show that most losses,
up to ,5500 ha, were in the Northern Territory, amounting to
,9% of mangroves in the area. Losses in Queensland were less,
but still amounted to more than 3% of affected mangroves in that
area of the state. Overall, this amounted to more than 7400 ha of
mangrove loss across the southern Gulf. This figure represents
,6% of mangroves across the entire region.
Patterns and features of mangrove dieback
At least three primary response patterns were observed in veg-
etation affected by dieback based on mapping and aerial surveys
(see Fig. 7), including: (1) more or less complete loss of the
shoreline fringing zone (Fig. 7a); (2) upper margin loss of
mangroves bordering saltpans and saltmarsh flats (Fig. 7b); and
(3) mostly intact, healthy fringes along estuarine channel mar-
gins and upstream riverine stands (Fig. 7c), even where these
enter the sea.
A key notable feature of this recent mangrove dieback
incident was its synchronous occurrence along more than
1000 km of shoreline. At smaller scales, there are several
alternative drivers also likely to cause dieback (Duke 2014).
These include disturbances, dieback and loss of mangrove
vegetation like: the common presence of small circular light
gaps (50–100 m wide) attributed to lightning strikes (Duke
2001;Amir 2012); shoreline erosion from severe cyclonic
storms (Baldwin et al. 2001); shoreline erosion associated with
Table 1. Landsat 8 image products used in the mangrove change detection procedure (see Fig. 5)
WRS-2, Worldwide Reference System; UTM, Universal Transverse Mercator
Scene ID WRS-2 path/row Acquisition date Projected UTM zone
LC81010712015112 LGN00 101/71 22 April 2015 53
LC81020702015119 LGN00 102/70 29 April 2015 53
LC81020712015119 LGN00 102/71 29 April 2015 53
LC81030702015094 LGN00 103/70 15 April 2015 53
LC81030702016081 LGN00 103/70 21 March 2016 53
LC80990722015114 LGN00 99/72 24 April 2015 54
LC81000722015121 LGN00 100/72 1 May 2015 54
LC81010712016115 LGN00 101/71 24 April 2016 53
LC81020702016106 LGN00 102/70 15 April 2016 53
LC81020712016106 LGN00 102/71 15 April 2016 53
LC80990722016117 LGN00 99/72 26 April 2016 54
LC81000722016108 LGN00 100/72 17 April 2016 54
Severe mangrove dieback in the Gulf Marine and Freshwater Research E
0 250 km
Back
g
round ima
g
er
y
: Landsat 8 OLI B and 4 (Red) collected April 2015
Fig. 6. Seven example scenes (A–G) distributed across the southern Gulf, as indicated on the map at the bottom right,
showing areas of mangrove loss (red) detected in Normalised Difference Vegetation Index (NDVI) difference images of
Landsat 8 scenes from April 2015 to April 2016. Also shown are the surviving mangrove areas (green). Central locations
for respective views were as follows: Location A, 14820017.0800S, 135842012.6100E; Location B, 14842046.0000S,
135822055.9200E; Location C, 1586053.9500S, 135844011.0300E; Location D, 15844036.0900S, 136833019.5000E; Location E,
15854047.2300S, 136852034.3300E; Location F, 17831010.3300S, 139829038.4600E; and Location G, 17828050.9700S,
140847017.2200E. Operational Land Imager (OLI). Note: enlarged versions of these scenes (Figs S1–S7), and difference
images of the entire affected shoreline (GIS shape files), are available in Supplementary material to this article.
FMarine and Freshwater Research N. C. Duke et al.
sea level rise (Lovelock et al. 2015; and, depositional gain with
mangrove islands and mud banks forming at river mouths
(Asbridge et al. 2016). Still other drivers include impoundment
with blockages to tidal drainage resulting in altered hydrology
and root burial from rapid sediment deposition.
An understanding of such indicators and their associated
drivers of change underpinned our interpretation of the current
dieback. One key pattern was ubiquitous across the affected
area, shown in the example image (Fig. 7b) of dieback along an
inner stand margin. This appears to be an instance of ecotone
shift (Duke 2014). In such a case, mangrove loss appears to have
progressed in a unidirectional contraction of the inner mangrove
zone margin, matched by a corresponding expansion of the
lower elevation margin of saltmarsh and saltpan. This was
consistent with the progressive loss of mangroves attributed to
unusually low moisture levels associated with tidal inundation.
There were also other distinctive features of this incident. For
example, there was a concurrent loss of vegetation from two
species zone ecotones at the same time. Two dominant species
zones within several mangrove stands in the Gulf each showed
signs of ecotone shift with the dieback (Fig. 8). This was shown
further by an unusual co-occurrence of dieback along two
parallel fronts: one for Avicennia marina and another for
Rhizophora stylosa. Judging by the zone of sublethal damage
(yellowed and partially defoliated canopies) at the lower profile
margins, there appeared to be a unidirectional loss (contraction)
towards the water’s edge margin.
These observations show that multiple mangrove species
were affected by the dieback and each had retreated from its
respective intertidal profile position (Fig. 9). In this way,
mangrove losses followed distinct patterns consistent with the
contraction of species zones away from their inner and higher
elevation margins.
In more extreme instances, notable sections of seaward
shoreline mangrove stands were completely lost. This explains
the instances (such as Fig. 7aand almost the case in Fig. 8b)
where either the zone or the entire shoreline margin had died
back.
Furthermore, although mangroves are predominantly
restricted to coastal areas, they are known to be influenced by
conditions within upstream catchment areas (cf. Duke et al.
1998). It was considered useful for ongoing assessments to
quantify mangrove dieback within individual catchments
across the affected region. These regional variations in man-
grove canopy loss are shown in 14 catchment areas across the
assessment region (Fig. 10). A minimal level of 0.4% was
measured for Groot Eylandt (Catchment area 1) compared with
a maximal level of ,26% within the Robinson River (Catch-
ment area 8). Dieback severity was greatest (,8–25%) in
mangrove areas of the catchments of the Rosie (Catchment
area 6), McArthur (Catchment area 7), Robinson (Catchment
area 8) and Calvert (Catchment area 9) rivers. It is an important
question as to why the severity levels were highest in these
systems.
Table 2. Summary of mangrove vegetation cover within the Gulf of Carpentaria and along its southern shoreline
The Gulf shoreline extends across two regional jurisdictions of Australia, namely the Northern Territory (NT) and the state of Queensland (Qld). The National
Vegetation Information System (NVIS) was accessed in June 2016 (National Vegetation Information System 2016). Note, although there were notable
differences between NVIS mapping estimates and those made in the present study for the southern Gulf areas, these are likely to be methodological errors
(Rogers et al. 2016) rather than actual measures of change in mangrove vegetation cover. UTM, Universal Transverse Mercator
Gulf section Feature and image source Mangrove/saltmarsh vegetation cover (ha)
UTM Zone 53 (NT) UTM Zone 54 (Qld) Total zones (NT and Qld)
Total Gulf Mangrove: extrapolated/NVIS mapping 74 405/49 203 131 780/172 441 206 185/221 644
Saltmarsh/saltpan: NVIS mapping 233 002 182 731 415 733
Tidal wetland: extrapolated/NVIS mapping 282 205/308 399 355 172/354 931 637 377/663 330
Southern Gulf Mangrove: Landsat 8 April 2015/NVIS mapping 58 139/38 452 64 364/85 263 122 503/123 715
Dieback loss: Landsat 8 April 2015–16 5493 1912 7405
Percentage loss 9.4 3.0 6.0
(a)(b)(c)
Fig. 7. Indicative patterns in vegetation zones relate to the severity of mangrove dieback, including (a) the more-or-less entire dieback of the shoreline
zone, (b) the dieback of the inner edge bordering saltmarsh and saltpans and (c) dense normal fringing vegetation along tidal channels. Images were taken
during the aerial survey on 9 June 2016 between Centre Island and the mouth of Limmen Bight River, Northern Territory.
Severe mangrove dieback in the Gulf Marine and Freshwater Research G
Timing of this occurrence of severe dieback
The timing of the mangrove dieback event in the Gulf appears to
have been more or less synchronous along the entire affected
shoreline. Evidence of the timing was first observed by con-
cerned locals around mid to late November 2015. Just north of
Karumba in Queensland, ,10-km section of mangrove seaward
fringing shoreline was observed to be undergoing dieback. This
was confirmed as a widespread response from satellite images
taken over the period (e.g. see Table 3). It seems that mangrove
areas were notably affected during November–December 2015,
the end of the unusually long dry season. Based on anecdotal
observations from community members, there may also have
been concurrent losses of seagrass during the month earlier in
November. However, this needs to be confirmed.
Discussion
The extent and severity of mangrove dieback in Australia’s
southern Gulf of Carpentaria appears to be unprecedented.
There had been no previous reports of mangrove dieback at this
scale or within such a short time frame. The respective factors,
along with observations of spatial patterns described at local
(Figs 8,9) and regional (Fig. 10) scales, combined with temporal
Elevation Profile, Zonation and Patterns of Mangrove Dieback
Upland–supratidal
Tidal wetland
Saltmarsh–saltpan Mudflat
Highest water
spring
Mangrove Mangrove-all species
Avicennia
Rhizophora
Av
DiebackDieback
Dieback
Mean sea
level
High tide
Mean
sea level
Fig. 9. Cross-section view of mangrove zonation with two species dominants (Avicennia marina and Rhizophora stylosa) having different fronts, or
ecotones, of dieback patterns observed in the Gulf of Carpentaria.
Avicennia
(a)(b)
Avicennia
Rhizophora Rhizophora
Fig. 8. Two views of mangrove zonation showing the sea edge Avicennia marina, then Rhizophora stylosa, and backed by
another A. marina zone bordering the saltpan with saltmarsh (a) on the mainland shoreline west of Centre Island (Northern
Territory) and (b) near the Leichhardt River estuary mouth (Queensland). The two mangrove species show separate dieback
fronts at their respective higher elevation margins in several locations along the southern Gulf of Carpentaria coast.
Percentage mangrove loss
1. Groot Eylandt 8. Robinson River
9. Calvert River
10. Settlement Creek
11. Mornington Island
12. Nicholson–Leichhardt Rivers
13. Morning Inlet
14. Flinders–Norman Rivers
500 km
2. Walker River
3. Roper River
4. Towns River
5. Limmen Bight River
6. Rosie River
7. McArthur River
0.4
0.5–3.3
3.4–7.1
7.2–15.8
15.9–25.6
Fig. 10. Proportional losses of mangroves for 14 catchment areas border-
ing the southern shorelines of the Gulf of Carpentaria. For the loss groupings
(interval classifications), we used the Jenks natural breaks approach (Jenks
1967). The question raised by these differences is why were Catchments 8
and 9 more severely affected than the others?
HMarine and Freshwater Research N. C. Duke et al.
observations (Table 3), are each considered potentially influ-
ential and indicative of the likely cause.
Although this incident of mangrove dieback in the Gulf also
coincided with effects of hot water on corals in north-eastern
Australia (Normille 2016;Pratchett and Lough 2016), we
suspect additional factors were responsible for the severe
dieback of mangroves in the Gulf. The most likely factors
appeared to be those consistent with severe moisture deficit
and the extreme weather conditions at the time.
Monthly satellite imagery was used to show changes to
mangrove areas before November 2015 and after mangroves
had died back significantly. The timing of peak dieback
appeared to be synchronous with the end of the extreme weather
conditions (e.g. National Academies of Sciences, Engineering,
and Medicine 2016). Significantly, the region was affected by
above-record air temperatures, as well as high sea temperatures.
Although there may have been some link with higher sea
temperatures on Australia’s north-east coast (Wolanski 2016),
this was not the only factor likely to have caused mangrove
dieback. In addition, the bulk of mangrove dieback occurred in
relative unison towards the end of the unusually extended dry
season that affected southern parts of the Gulf of Carpentaria
(Australian Broadcasting Corporation 2015). Based on such
observations, a working hypothesis is that mangroves died from
localised moisture stress.
Was this severe dieback the first instance of climate change-
induced ‘longer, hotter droughts’ affecting mangrove forests by
moisture stress? The likelihood of stress-related dieback from
such weather conditions leading to extended water deficit within
mangroves is very concerning, because the frequency of severe
drought reportedly increased worldwide (Dai 2013;Trenberth
et al. 2014;Moise et al. 2015;Clark et al. 2016). In addition,
these droughts have also become hotter (Intergovernmental
Panel on Climate Change 2014). Concerns remain about anthro-
pogenic effects where climate change-induced drought has
reportedly caused increased forest tree mortality (McDowell
and Allen 2015), as well as degradation of forest structure and
function generally (Clark et al. 2016).
Large-scale mangrove retreat at high intertidal saltmarsh–
mangrove ecotones, referred to as ecotone shift (Duke 2014),
has been reported for longer periods of drying in eastern
Australia (e.g. Eslami-Andargoli et al. 2013). However, these
have occurred over decadal periods (N. C. Duke, A. Basile,
C. Field, J. R. Mackenzie, J.-O. Meynecke and A. L. Wood,
unpubl. data), rather than the rapid, sudden loss reported for the
current incident.
Comparable incidents and functional processes
To further understand the reasoning behind our conclusions, it is
useful to briefly evaluate the various types of condition states in
mangrove and tidal wetland ecosystems around the world.
Overall, the dominant vegetation type in temperate tidal
wetlands is saltmarsh (Duke et al. 1998). In contrast, mangroves
dominate wet tropical and subtropical tidal wetlands where
saltmarsh plants and saltpan microphytes are also present
coexisting in varying proportions with mangroves, depending
on longer-term rainfall conditions (N. C. Duke et al., unpubl.
data).
In semi-arid tropical locations, like Australia’s Gulf of
Carpentaria, the habitat types co-occur, with mangroves occu-
pying less than 50% of the tidal wetland niche, and tropical
saltmarsh and saltpans occupying the rest. In such drier tropical
and subtropical situations, these habitat states appear con-
strained by moisture (rainfall) conditions, rather than by tem-
perature (N. C. Duke et al., unpubl. data). This is a globally
significant distinction (Osland et al. 2016;Ward et al. 2016).
Overall, the case studies of changes of mangrove dieback
observed can be grouped into three types of incidents coinci-
dental with extreme weather events, other than severe storms or
large waves. These incidents result in often notable responses as
shifts between alternative tidal wetland vegetation state types.
However, not all have been consistent with reported longer-term
trends in weather conditions (see Alongi 2015). Overall, these
incident types include the following.
1. Temperature-dominated changes to mangrove v. saltmarsh
vegetation at mangrove higher latitude limits, because man-
grove expansion is linked to warmer conditions (Gilman
et al. 2008;Osland et al. 2013,2016;Saintilan et al. 2014)or
the alternative, mangrove dieback or retreat, is linked to
severe frosts (Kao et al. 2004;Ross et al. 2009;Feller et al.
2010;Saintilan et al. 2014).
2. Rainfall-dominated changes to mangrove v. saltmarsh vege-
tation within the latitudinal limits of mangroves, because
Table 3. Observations of the onset and timing of severe mangrove dieback
Satellite images were used to quantify the timing of impacts on mangrove dieback along the shoreline north of the Norman River mouth near Karumba, Gulf of
Carpentaria. Reference data include regional rainfall and the duration of the 2015 dry season. nv, not visible
Date Tide level Seagrass intertidal Mangrove seaward zone Mangrove landward
zone
Regional
rainfall (mm)
Number of dry
season months
30 August 2015 Mid nv Intact, full and dense Intact, sparse, patchy 0.7 7
1 October 2015 Low Intact, full and dense Intact, full and dense Intact, sparse, patchy 0.1 8
2 November 2015 Very
low
Mostly absent, some patches Intact, full and dense Intact, sparse, patchy 0.4 9
4 December 2015 Low Mostly absent, some patches Reduced canopies, sparse,
patchy
Sparse, patchy 28.4 10
5 January 2016 Low Mostly absent, some patches Mostly absent, sparse, patchy Mostly absent, sparse,
patchy
257.9 Wet season
Severe mangrove dieback in the Gulf Marine and Freshwater Research I
mangrove expansion is linked to wetter conditions (Duke
et al. 2003;Gilman et al. 2008;Eslami-Andargoli et al. 2013;
Duke 2014;Osland et al. 2014,2016) or the alternative,
mangrove losses or retreat, are linked to drought and
decreased precipitation (Duke et al. 2003;Gilman et al.
2008;Duke 2014).
3. Rainfall-dominated saltmarsh dieback at, or beyond, higher
latitudinal limits of mangroves linked to drought, decreasing
rainfall and sea level changes (McKee et al. 2004;Silliman
et al. 2005).
Although mangrove dieback in the Gulf had seemingly
similar associated factors to those described for saltmarsh in
Incident type 3 above, they were distinctive and different
because the dieback affected saltmarsh plants only, and not
mangroves. This was not the case in the first two incident types.
In contrast, these types showed a shift between mangrove-
dominated and saltmarsh-dominated vegetation communities
within the same tidal wetland niche. Each involved the expan-
sion of one state–type at the expense of the other. This created a
seemingly common feature of ecotone shift between the alter-
native vegetation states. The shift direction appeared primarily
dependent on trends within the chief influencing factors of
temperature and rainfall.
Accordingly, where there had been mangrove expansion at
mangrove poleward limits, like the eastern coast of North
America (Osland et al. 2013), this involved a corresponding
replacement or loss of saltmarsh habitat. It was significant that
the Gulf dieback occurred well within the latitudinal range of the
species affected. Therefore, it seems more likely that the Gulf
dieback corresponds to a Type 2 incident, influenced mostly by
rainfall. In addition, this implies that there may have been a
severe moisture deficit.
Extreme weather conditions associated with mangrove
dieback in the Gulf
It was significant that the Gulf instance of severe mangrove
dieback was more or less coincidental with unusually hot water
temperatures observed off the Australian north-east coast
(Normille 2016;Wolanski 2016). Although mangroves are
known to be reasonably heat tolerant (Medina 1999), there
were exceptionally high temperatures recorded at the time
when mangrove dieback peaked (Bureau of Meteorology 2015;
Hope et al. 2016). Record high temperatures were recorded
during the preceding 5–6 months of the extended dry season
(Fig. 11a).
But also present, there were at least two additional severe
weather and hydrological conditions likely to affect mangrove
vegetation. These included an unusually prolonged severe
drought, and a temporary drop in sea level during the latter
months of that drought period. The low rainfall conditions
experienced in November 2015 are shown in Fig. 11b. These
dry season conditions prevailed for an unusually long period of
,10–11 months.
A third factor, sea level, was observed to have dropped
temporarily by up to 20 cm in the Gulf at the time of the severe
dieback. This is shown in the map of Pacific Ocean seawater
levels during October 2015 (Fig. 11c). Although there remain
questions about how the mean drop would have interacted with
tidal levels, it is likely that upper areas of mangroves would have
been less inundated during normal high tide periods of the
critical months when these upper zone plant habitats were both
heat stressed and in moisture deficit.
It is highly likely that the combination of these three factors
at least contributed to the severe dieback observed. It is further
significant that all these parameters were each correlated with
the Southern Oscillation Index (SOI) and the El Nin
˜o–Southern
Oscillation (ENSO) cycle for this region (Becker et al. 2012;
Hilbert et al. 2014). Although it is difficult to be sure about the
exact trigger causing this severe mangrove dieback, it is under-
stood that it had not occurred on such a scale before. In addition,
although it was also synchronous with the late 2015 period, it is
also apparent there was more than one variable involved.
All things considered, it appears that stresses on mangrove
plants were not only present, but also would likely have been
cumulative. For instance, we have looked closely at the duration
of contiguous days where monthly rainfall had not reached
recognised ‘wet season’ levels. Such climate variables were
analysed for each of the five local Bureau of Meteorology
stations (www.bom.gov.au) with averaged data extending back
up to ,30 years. This assessment has shown the period up to
October–November 2015 as unusual.
In a further line of inquiry, the patterns and distribution of
dieback were assessed from the mapping. The occurrences of
severe dieback along mostly upper elevation contour zones
(Figs 8,9) were consistent with circumstances of severe mois-
ture stress along vegetation zone margins. Given the hot, dry
weather conditions at the time, coupled with lower sea levels,
each of these factors would have contributed to a decrease in soil
moisture, with stresses on the plants extending down the tidal
profile from higher and inner mangrove zones. In addition,
given the synchronicity of drought conditions with the timing of
peak dieback, this implies that a lack of moisture was a likely
trigger, occurring at the end of an unusually prolonged drought
period.
These longer-term changes describe the unusual congruence
of key detrimental growth factors towards the end of the long
2015 dry season (Fig. 12). Key evidence for the likely cause of
this occurrence of severe mangrove dieback in the Gulf, are
summarised as follows:
unprecedented large patches of dead and defoliated mangrove
vegetation adding up to more than 7400 ha spread across
1000 km of shoreline
dieback areas matched zonation contours, extending down
profile from mostly higher elevation contour levels
dieback involved key dominant species present (A. marina,
R. stylosa and Ceriops tagal) and at their respective higher
zonation ecotones along upper tidal profile elevations
dieback was synchronous with the end of the unusually long
dry season of ,9–10 months in November–December 2015, a
period of unusually extended high evapotranspiration
dieback occurred when regional annual rainfall levels were
low, temperatures were high (see also Bureau of Meteorology
2015) and sea levels were notably lower at the time.
the severity of dieback was variable across coastal catchment
areas, with up to 25% mangrove losses in the more-or-less
central subcatchment area (Fig. 10).
JMarine and Freshwater Research N. C. Duke et al.
Based on the present findings, it is concluded that the most
likely cause of this unusual and unprecedented incidence of
severe mangrove dieback was the unusually long duration of
arid, hot conditions in the southern Gulf of Carpentaria region
towards the end of the 2015 dry season. These conditions
(Bureau of Meteorology 2015) were synchronous with local
high temperatures, low rainfall, extended drought period, and a
temporary fall in sea level. A notable correlate of these weather
and sea level variables was the SOI.
Implications of severe dieback in mangrove ecosystems
This incident of severe and extensive dieback has immediate
implications with likely consequences for the ecosystem ser-
vices previously provided by the damaged mangrove habitat.
The areas most likely affected by the loss of mangrove area and
health are those associated with fisheries and fishing, with the
Gulf also being an extremely popular area for recreational
fishing. The total value of the Gulf fishery is worth ,A$30
million per annum (Dambacher et al. 2015;QDEEDI 2011). The
Gulf is also the hub of the northern prawn fishery, one of Aus-
tralia’s most valuable and iconic commercial fisheries, targeting
banana prawns and tiger prawns. The long-term average annual
catch is valued at ,A$12 million (Dambacher et al. 2015). In
addition, the total commercial harvest of mud crabs, Scylla
serata,is,1190 tonnes (Mg), worth ,A$19 million, with
recreational fishers harvesting at least 50% of the total catch
(Queensland Department of Employment Economic Develop-
ment and Innovation 2011).
Commercial fisheries in the Queensland section alone
grossed a total value of A$14 million in 2006 with ,2300 Mg
(Greiner and Gregg 2010). The largest of the commercial
fisheries, the Queensland Inshore Fin Fish Fishery, had a total
harvest of 2365 Mg, worth ,A$15.5 million in 2012, with
recreational fishers taking in less than 10% of the catch
(Queensland Department of Agriculture Fisheries and Forestry
2014). Popular finfish in the fishery include barramundi Lates
calcarifer, blue salmon Eleutheronema tetradactylum, grunter
⫺20
⫺20 ⫺10 10 200
100 150 200
Monthly Msla without cycles referenced ot 1993–2013 (cm)
250
400%
Percentage of Mean
Rainfall percentages (AWAP HiRes)
Max. Temp. Deciles (AWA grds 1911-pres.)
(a)(b)
(c)
1 February to 30 September 2015
Distribution Based on Gridded Data
Australian Bureau of Meteorology
October 2015
Australian Bureau of Meteorology
300%
200%
150%
125%
100%
80%
60%
40%
20%
0%
Temp. Decile Ranges
10
8–9
4–7
2–3
1
Highest on
record
Lowest on
record
Very much
above average
Very much
below average
Above average
Below average
Average
0
20
October 2015
Issued: 14/07/2016
© Commonwealth of Australia 2016,
Australian Bureau of Meteorology
ID code: IGMapAWAPPercentagesHighRes
Issued: 14/07/2016
© Commonwealth of Australia 2016,
Australian Bureau of Meteorology
ID code: IGMapAWAPDeciles
Australian Government
Bureau of Meteorology
Australian Government
Bureau of Meteorology
http://www.born.gov.au
http://www.born.gov.au
Fig. 11. Three notable factors (Bureau of Meteorology 2015) concurrent with the severe mangrove dieback in the Gulf include (a) maximum temperature
levels (February–September 2015 in Australia), coincident with the hottest March days on record, (b) low rainfall percentages for October 2015 in
Australia, coincident with an unusually prolonged dry season, and (c) the temporary sea level anomaly for October 2015 in the Pacific region, coincident
with a temporary drop of 20 cm in sea level that was consistent with the El Nin
˜o–Southern Oscillation (ENSO) cycle affecting subsurface heat across the
Pacific Basin. The top images were supplied by Andrew Watkins with the Australian Bureau of Meteorology (BOM). The lower image was found on the
Aviso website (www.aviso.altimetry.fr). AWAP HiRes, Australian Water Availability (BOM) Project High Resolution imagery; AWA grds, Australian
Water Availability grids; CLS/CNES, Collecte Localisation Satellites Group, Centre National d’Etudes Spatiales (French Space Agency).
Severe mangrove dieback in the Gulf Marine and Freshwater Research K
Pomadasys kaakan, king salmon Polydactylus macrochir and
mangrove jack Lutjanus argentimaculatus. All these species
have close associations with, if not specific dependencies on,
mangrove habitat, making the extent of dieback of mangroves of
immense concern to commercial and recreational fisheries of the
Gulf region.
The impacts on mangrove ecosystem services in the Gulf
region also affect shoreline protection and carbon capture. There
is a high risk of further losses of carbon to the atmosphere should
there be more dieback or there is interrupted recovery of the
7400 ha of currently affected mangroves. For example, damage
would be greatly exacerbated within the next 10 years should the
area be struck by a tropical cyclone (Paling et al. 2008) while
shoreline vegetation re-establishes exposed shoreline fringing
stands. These points raise serious concerns for the future of
mangroves and other habitats across the region, especially
because they co-occur with the threats of sea level rise contrib-
uting to ongoing accelerated shoreline erosion and retreat.
Climate projections over the next several decades show
higher temperatures, increased evaporation rates and warmer
oceans (Moise et al. 2015). Accordingly, there is an increased
likelihood of future severe and extended droughts across parts of
Northern Australia (Dai 2013). This dieback incident revealed a
previously unrecognised sensitivity and vulnerability of man-
grove tidal wetland ecosystems to fluctuations in climate. A
greater understanding of specific catchment-scale drivers of
drought and other dieback effects is needed urgently to assist
targeted management strategies for enhancing the resilience of
mangrove shorelines faced with climate change. Such well-
informed strategies are essential to government agencies at all
levels needing to effectively manage longer-term adaptation
goals. This incident is a wake-up call for the improvement of
social and economic resilience of remote communities, along
with the health of natural environments along Australia’s
northern shorelines.
Author contributions
N. C. Duke led the investigation, the compilation of data and
writing the document. J. M. Kovacs, D. J. E. Hill and H. S.
Morning were responsible for the satellite image change detec-
tion component of the investigation. A. D. Griffiths assisted with
the aerial field survey and the writing. L. Preece assisted with the
writing and searching for resources. P. van Oosterzee assisted
with the writing and searching for resources. J. Mackenzie
assisted with the writing and searchingfor resources. D. Burrows
assisted with the writing and research support.
Supplementary material
Detailed imagery are available as Supplementary material to
this paper, including enlarged versions of the seven image plates
displayed in Fig. 6, plus GIS files of mangrove loss areas along
the entire affected coast.
Acknowledgements
Early imagery and observations of the severe mangrove dieback in the Gulf
were generously provided by Paul Barden (29 February 2016), Gavin
(‘Groover’) New (Carpentaria Barra and Sport Fishing Charters; 1–3 March
2016) and Roger Jaensch (Jaensch Ornithology & Conservation; 13–14
April 2016). Project works were supported by TropWATER Centre, James
Cook University. Support for aerial surveys are acknowledged in the article,
but especially include that provided by the Parks and Wildlife Commission
1996
5.0
5.5
6.0
6.5
7.0
7.5
8.0
8.5
9.0
9.5
1998 2000
Dry season duration
(number of months of high
evapotranspiration)
Annual rainfall (mm)
Mean air
temperature (⬚C)
SST (⬚C)Sea level rise (m)SOI
2002 2004 2006 2008 2010 2012 2014 2016 1996
27.2
27.4
27.6
27.8
28.0
28.2
28.4
28.6
28.8
29.0
29.2
29.4
A-Rain
Rain3yr
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
1996
1.1
1.1
1.2
1.2
1.3
1.3
1.4
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
1996
⫺15.0
⫺10.0
⫺5.0
0
5.0
10.0
15.0
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
1996
400
600
800
1000
1200
1400
1600
1800
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
1996
26.0
26.5
27.0
27.6
28.0
28.5
1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
Fig. 12. Six factors likely to have influenced mangrove condition at the time include dry season duration, annual rainfall, mean air temperature, sea surface
temperature (SST), sea level and the Southern Oscillation Index (SOI). Values were averaged for the region (where possible) for the 18-year period from 1998
to 2016. Data were sourced from the Australian Bureau of Meteorology, including stations (www.bom.gov.au) at Normanton, Burketown, Borroloola,
McArthur River and Centre Island (also see Fig. 3).
LMarine and Freshwater Research N. C. Duke et al.
of the Northern Territory with the Department of Land Resource Manage-
ment, and the Carpentaria Land Council Aboriginal Corporation. J. M.
Kovacs acknowledges financial support through the Natural Sciences and
Engineering Research Council of Canada (Grant #RGPIN-2014-06188).
References
Alongi, D. M. (2015). The impact of climate change on mangrove forests.
Current Climate Change Reports 1, 30–39. doi:10.1007/S40641-015-
0002-X
Amir, A. A. (2012). Canopy gaps and the natural regeneration of Matang
mangroves. Forest Ecology and Management 269, 60–67. doi:10.1016/
J.FORECO.2011.12.040
Asbridge, E., Lucas, R., Ticehurst, C., and Bunting, P. (2016). Mangrove
response to environmental change in Australia’s Gulf of Carpentaria.
Ecology and Evolution 6(11), 3523–3539. doi:10.1002/ECE3.2140
Australian Broadcasting Corporation (2015). Back to square one: Gulf shires
lobby for drought declaration. Available at http://www.abc.net.au/
news/2015-05-12/queensland-gulf-shires-lobby-for-drought-declaration/
6459722 [Verified 12 January 2017].
Baldwin, A., Egnotovich, M., Ford, M., and Platt, W. (2001). Regeneration
in fringe mangrove forests damaged by Hurricane Andrew. Plant
Ecology 157(2), 151–164. doi:10.1023/A:1013941304875
Bayliss, P., and Freeland, W. J. (1989). Seasonal distribution and abundance
of dugongs in the western Gulf of Carpentaria. Wildlife Research 16,
141–149. doi:10.1071/WR9890141
Becker, M., Meyssignac, B., Letetrel, C., Llovel, W., Cazenave, A., and
Delcroix, T. (2012). Sea level variations at tropical Pacific islands since
1950. Global and Planetary Change 80–81, 85–98. doi:10.1016/
j.gloplacha.2011.09.004
Bureau of Meteorology (2015). Special Climate Statement 51 – an excep-
tional autumn hot spell in northern and central Australia. Available at
http://www.bom.gov.au/climate/current/statements/scs51.pdf [Verified
12 January 2017].
Burford, M. A., Rothlisberg, P. C., and Revill, A. T. (2009). Sources of
nutrients driving production in the Gulf of Carpentaria, Australia: a
shallow tropical shelf system. Marine and Freshwater Research 60,
1044–1053. doi:10.1071/MF08291
Cahoon, D. R., Hensel, P., Rybczyk, J., McKee, K. L., Proffitt, C. E., and
Perez, B. C. (2003). Mass tree mortality leads to mangrove peat collapse
at Bay Islands, Honduras, after Hurricane Mitch. Journal of Ecology 91,
1093–1105. doi:10.1046/J.1365-2745.2003.00841.X
Clark, J. S., Iverson, L., Woodall, C. W., Allen, C. D., Bell, D. M., Bragg,
D. C., D’Amato, A. W., Davis, F. W., Hersh, M. H., and Ibanez, I. (2016).
The impacts of increasing drought on forest dynamics, structure, and
biodiversity in the United States. Global Change Biology 22,
2329–2352. doi:10.1111/GCB.13160
Dai, A. (2013). Increasing drought under global warming in observa-
tions and models. Nature Climate Change 3, 52–58. doi:10.1038/
NCLIMATE1633
Dambacher, J. M., Rothlisberg, P. C., and Loneragan, N. R. (2015).
Qualitative mathematical models to support ecosystem-based manage-
ment of Australia’s Northern Prawn Fishery. Ecological Applications
25, 278–298. doi:10.1890/13-2030.1
Duke, N. C. (2001). Gap creation and regenerative processes driving
diversity and structure of mangrove ecosystems. Wetlands Ecology
and Management 9(3), 267–279. doi:10.1023/A:1011121109886
Duke, N. C. (2014). Mangrove coast. In ‘Encyclopedia of Marine
Geosciences’. (Eds J. Harff, M. Meschede, S. Petersen, and J. Thiede.)
pp. 1–17. (Springer: Dordrecht, Netherlands.) Available at http://link.
springer.com/referenceworkentry/10.1007%2F978-94-007-6644-0_186-1
[Verified 12 January 2017].
Duke, N. C. (2016). Oil spill impacts on mangroves: recommendations for
operational planning and action based on a global review. Marine
Pollution Bulletin 109, 700–715. doi:10.1016/J.MARPOLBUL.2016.
06.082
Duke, N. C., Ball, M. C., and Ellison, J. C. (1998). Factors influencing
biodiversity and distributional gradients in mangroves. Global Ecology
and Biogeography Letters 7, 27–47. doi:10.2307/2997695
Duke, N. C., Lawn, P., Roelfsema, C. M., Phinn,S., Zahmel, K. N., Pedersen,
D., Harris, C., Steggles, N., and Tack, C. (2003). ‘Assessing Historical
Change in Coastal Environments. Port Curtis, Fitzroy River Estuary and
Moreton Bay Regions.’ (Historical Coastlines Project, Marine Botany
Group, Centre for Marine Studies, The University of Queensland:
Brisbane, Qld, Australia.)
Duke, N. C., Wood, A., Hunnam, K., Mackenzie, J., Haller, A., Christiansen,
N., Zahmel, K., and Green, T. (2010). Shoreline ecological assessment
aerial and ground surveys, 7–19 November 2009. As part of the
Scientific Monitoring Study of the West Atlas Monitoring Plan. Report
to PTTEP Australasia by Uniquest University of Queensland, School of
Biological Sciences, Brisbane. Available at http://www.environment.
gov.au/marine/marine-pollution/montara-oil-spill/scientific-monitoring-
studies#two [Verified 12 January 2017].
Erftemeijer, P. L., and Hamerlynck, O. (2005). Die-back of the mangrove
Heritiera littoralis dryand, in the Rufiji Delta (Tanzania) following El
Nin
˜o floods. Journal of Coastal Research 42, 228–235.
Eslami-Andargoli, L., Dale, P., and Sipe, N. (2013). Does spatial scale affect
the pattern of mangrove change under different rainfall regimes? An
example in southeast Queensland, Australia. Austral Ecology 38,
208–218. doi:10.1111/J.1442-9993.2012.02393.X
Feller, I. C., Lovelock, C. E., Berger, U., McKee, K. L., Joye, S. B., and Ball,
M. C. (2010). Biocomplexity in mangrove ecosystems. Annual Review
of Marine Science 2, 395–417. doi:10.1146/ANNUREV.MARINE.
010908.163809
Feller, I. C., Dangremond, E. M., Devlin, D. J., Lovelock, C. E., Proffitt, E.,
and Rodriguez, W. (2015). Nutrient enrichment intensifies hurricane
impact in scrub mangrove ecosystems in the Indian River Lagoon, FL,
USA. Ecology 96(11), 2960–2972. doi:10.1890/14-1853.1
Gilman, E. L., Ellison, J., Duke, N. C., and Field, C. (2008). Threats to
mangrovesfrom climate change and adaptation options: a review.Aquatic
Botany 89, 237–250. doi:10.1016/J.AQUABOT.2007.12.009
Greiner, R., and Gregg, D. (2010). Considering recreational catch and
harvest in fisheries management at the bio-regional scale. Fisheries
Management and Ecology 17, 336–345.
Halpern, B. S., Walbridge, S., Selkoe, K. A., Kappel, C. V., Micheli, F.,
D’Agrosa, C., Bruno, J. F., Casey, K. S., Ebert, C., Fox, H. E., Fujita, R.,
Heinemann, D., Lenihan, H. S., Madin, E. M. P., Perry, M. T., Selig,
E. R., Spalding, M., Steneck, R., and Watson, R. (2008). A global map
of human impact on marine ecosystems. Science 319, 948–952.
doi:10.1126/SCIENCE.1149345
Hilbert, D. W., Hill, R., Moran, C., Turton, S. M., Bohnet, I., Marshall,N. A.,
Pert, P. L., Stoeckl, N., Murphy, H. T., Reside, A. E., Laurance, S. G. W.,
Alamgir, M., Coles, R., Crowley, G., Curnock, M., Dale, A., Duke, N. C.,
Esparon, M., Farr, M., Gillet, S., Gooch, M., Fuentes, M., Hamman, M.,
James, C. S., Kroon, F. J., Larson, S., Lyons, P., Marsh, H., Meyer
Steiger, D., Sheaves, M., and Westcott, D. A. (2014) ‘Climate Change
Issues and Impacts in the Wet Tropics NRM Cluster Region.’ (James
Cook University: Cairns, Qld, Australia.)
Hope, P., Wang, G., Lim, E.-P., Hendon, H. H., and Arblaster, J. M. (2016).
What caused the record-breaking heat across Australia in October 2015?
Bulletin of the American Meteorological Society 97(12), S122–S126.
doi:10.1175/BAMS-D-16-0141.1
Intergovernmental Panel on Climate Change (2014). Synthesis report.
Summary for policymakers. Available at http://www.ipcc.ch/report/
ar5/syr/ [Verified 12 January 2017].
Jenks, G. F. (1967). The data model concept in statistical mapping. In
‘International Yearbook of Cartography, volume 7’. (Ed. K. Frenzel.)
pp. 186–190. (George Philip: Bonn-Bad, Godensberg, Germany.)
Severe mangrove dieback in the Gulf Marine and Freshwater Research M
Kao, W. Y., Shih, C. N., and Tsai, T. T. (2004). Sensitivity to chilling
temperatures and distribution differ in the mangrove species Kandelia
candel and Avicennia marina. Tree Physiology 24, 859–864.
doi:10.1093/TREEPHYS/24.7.859
Kennett, R., Munungurritj, N., and Yunupingu, D. (2004). Migration
patterns of marine turtles in the Gulf of Carpentaria, northern Australia:
implications for Aboriginal management. Wildlife Research 31, 241–
248. doi:10.1071/WR03002
Kovacs, J. M., Wang, J., and Blanco-Correa, M. (2001). Mapping mangrove
disturbances using multi-date Landsat TM imagery. Environmental
Management 27, 763–776. doi:10.1007/S002670010186
Long, B. G., Poiner, I. R., and Wassenberg, T. J. (1995). Distribution,
biomass and community structure of megabenthos of the Gulf of
Carpentaria, Australia. Marine Ecology Progress Series 129, 127–139.
doi:10.3354/MEPS129127
Lovelock, C. E., Cahoon, D. R., Friess, D. A., Guntenspergen, G. R., Krauss,
K. W., Reef, R., Rogers, K., Saunders, M. L., Sidik, F., Swales, A.,
Saintilan, N., Thuyen, L. X., and Triet, T. (2015). The vulnerability of
Indo-Pacific mangrove forests to sea-level rise. Nature 526, 559–563.
doi:10.1038/NATURE15538
Mackenzie, J. R., Duke, N. C., and Wood, A. L. (2016). The Shoreline Video
Assessment Method (S-VAM): using dynamic hyperlapse image acqui-
sition to evaluate shoreline mangrove forest structure, values, degrada-
tion and threats. Marine Pollution Bulletin 109, 751–763. doi:10.1016/
J.MARPOLBUL.2016.05.069
McDowell, N. G., and Allen, C. D. (2015). Darcy’s law predicts widespread
forest mortality under climate warming. Nature Climate Change 5,
669–672. doi:10.1038/NCLIMATE2641
McKee, K., Mendelssohn, I. A., and Materne, M. D. (2004). Acute saltmarsh
dieback in the Mississippi River deltaic plain: a drought-induced
phenomenon? Global Ecology and Biogeography 13, 65–73.
doi:10.1111/J.1466-882X.2004.00075.X
Medina, E. (1999). Mangrove physiology: the challenge of salt, heat, and
light stress under recurrent flooding. In ‘Mangrove Ecosystems in
Tropical America’. (Eds A. Ya´n
˜ez-Arancibia and A. L. Lara-
Domı´nguez.) pp. 109–126. (Instituto de Ecologı´a, Universidad Nacional
Autonoma de Me´xico: Mexico DF, Mexico; International Union for the
Conservation of Nature and Natural Resources, Regional Office for
Mesoamericas (UICN/ORMA): San Jose, Costa Rica; National Oceanic
and Atmospheric Administration, National Marine Fisheries Service
(NOAA/NMFS): Silver Spring, MD, USA.)
Moise, A., Abbs, D., Bhend, J., Chiew, F., Church, J., Ekstro¨m, M., Kirono,
D., Lenton, A., Lucas, C., McInnes, K., Monselesan, D., Mpelasoka, F.,
Webb, L., and Whetton, P. (2015). Monsoonal north cluster report,
climate change in Australia projections for Australia’s natural resource
management regions: cluster reports. CSIRO and Bureau of Meteorolo-
gy, Canberra, ACT, Australia.
National Academies of Sciences, Engineering, and Medicine (2016). ‘Attri-
bution of Extreme Weather Events in the Context of Climate Change.’
(The National Academies Press: Washington, DC, USA.)
National Vegetation Information System (2016). Information on the
extent and distribution of vegetation types in Australian landscapes.
Available at https://www.environment.gov.au/land/native-vegetation/
national-vegetation-information-system [Verified 12 January 2017].
Normille, N. (2016). Survey confirms worst-ever coral bleaching at Great
Barrier Reef. In ScienceInsider, 19 April 2016. Available at http://
www.sciencemag.org/news/2016/04/survey-confirms-worst-ever-coral-
bleaching-great-barrier-reef [Verified 12 January 2017].
Osland, M. J., Enwright, N. M., Day, R. H., and Doyle, T. W. (2013). Winter
climate change and coastal wetland foundation species: salt marshes vs.
mangrove forests in the southeastern United States. Global Change
Biology 19, 1482–1494. doi:10.1111/GCB.12126
Osland, M. J., Enwright, N. M., and Stagg, C. L. (2014). Freshwater
availability and coastal wetland foundation species: ecological transi-
tions along a rainfall gradient. Ecology 95, 2789–2802. doi:10.1890/13-
1269.1
Osland, M. J., Enwright, N. M., Day, R. H., Gabler, C. A., Stagg, C. L., and
Grace, J. B. (2016). Beyond just sea-level rise: considering macrocli-
matic drivers within coastal wetland vulnerability assessments to climate
change. Global Change Biology 22, 1–11. doi:10.1111/GCB.13084
Paling, E., Kobryn, H., and Humphreys, G. (2008). Assessing the extent of
mangrove change caused by Cyclone Vance in the eastern Exmouth
Gulf, northwestern Australia. Estuarine, Coastal and Shelf Science 77,
603–613. doi:10.1016/J.ECSS.2007.10.019
Poiner, I., Staples, D., and Kenyon, R. (1987). Seagrass communities of the
Gulf of Carpentaria, Australia. Marine and Freshwater Research 38,
121–131. doi:10.1071/MF9870121
Pratchett M., Lough J. (2016). Coral Bleaching Taskforce: more than
1000 km of the Great Barrier Reef has bleached. In The Conversation,
7 April 2016. Available at http://theconversation.com/coral-bleaching-
taskforce-more-than-1-000-km-of-the-great-barrier-reef-has-bleached-
57282 [Verified 18 May 2016].
Queensland Department of Agriculture Fisheries and Forestry (2014). Gulf
of Carpentaria Inshore Fin Fish Fishery: 2012 fishing year report.
QDAFF, Brisbane Qld, Australia.
Queensland Department of Employment Economic Development and Inno-
vation (2011). Annual status report 2011: Mud Crab Fishery. QDEEDI,
Brisbane, Qld, Australia.
Reef, R., Ball, M. C., and Lovelock, C. E. (2012). The impact of a locust
plague on mangroves of the arid Western Australia coast. Journal of
Tropical Ecology 28, 307–311. doi:10.1017/S0266467412000041
Roelofs, A., Coles, R., and Smit, N. (2005). ‘A Survey of Intertidal Seagrass
from Van Diemen Gulf to Castlereagh Bay, Northern Territory, and
From Gove to Horn Island, Queensland.’ (Queensland Department of
Primary Industries and Fisheries: Brisbane, Qld, Australia; CRC Reef
Research Centre: Canberra, ACT, Australia; Biodiversity Conservation,
NT Department of Infrastructure, Planning and Environment: Darwin,
NT, Australia.)
Rogers, K., Boon, P., Branigan, S., Duke, N. C., Field, C. D., Fitzsimons, J.,
Kirkman, H., Mackenzie, J., and Saintilan, N. (2016). Can Australian
legislation and policy protect mangrove and saltmarsh and their ecosys-
tem services? Marine Policy 72, 139–155. doi:10.1016/J.MARPOL.
2016.06.025
Ross, M. S., Ruiz, P. L., Sah, J. P., and Hanan, E. J. (2009). Chilling damage
in a changing climate in coastal landscapes of the subtropical zone: a
case study from south Florida. Global Change Biology 15, 1817–1832.
doi:10.1111/J.1365-2486.2009.01900.X
Saintilan, N., Wilson, N. C., Rogers, K., Rajkaran, A., and Krauss, K. W.
(2014).Mangrove expansionand salt marsh declineat mangrove poleward
limits. Global Change Biology 20, 147–157. doi:10.1111/GCB.12341
Silliman, B. R., Koppel, J. d., Bertness, M. D., Stanton, L. E., and
Mendelssohn, I. A. (2005). Drought, snails, and large-scale die-off of
southern US salt marshes. Science 310, 1803–1806. doi:10.1126/SCI
ENCE.1118229
Trenberth, K. E., Dai, A., van der Schrier, G., Jones, P. D., Barichivich, J.,
Briffa, K. R., and Sheffield, J. (2014). Global warming and changes in
drought.Nature Climate Change4, 1 7–22. doi:10.1038/NCLIMATE2067
Ward, R. D., Friess, D. A., Day, R. H., and MacKenzie, R. A. (2016). Impacts
of climate change on mangrove ecosystems: a region by region over-
view. Ecosystem Health and Sustainability 2(4), e01211. doi:10.1002/
EHS2.1211
Wolanski, E. (2016). Oceanography and coral bleaching. Available at
https://research.jcu.edu.au/tropwater/news-and-events/oceanography-
and-coral-bleaching [Verified 12 January 2017].
www.publish.csiro.au/journals/mfr
NMarine and Freshwater Research N. C. Duke et al.
