Marine Pollution Bulletin 109 (2016) 700–715

Contents lists available at ScienceDirect

Marine Pollution Bulletin

journal homepage: www.elsevier.com/locate/marpolbul

Oil spill impacts on mangroves: Recommendations for operational

planning and action based on a global review
Norman C. Duke ⁎
James Cook University, TropWATER Centre, Townsville, QLD, Australia

a r t i c l e i n f o a b s t r a c t

Article history: Mangrove tidal wetland habitats are recognised as highly vulnerable to large and chronic oil spills. This review of
Received 6 January 2016 current literature and public databases covers the last 6 decades, summarising global data on oil spill incidents
Received in revised form 20 June 2016 affecting, or likely to have affected, mangrove habitat. Over this period, there have been at least 238 notable oil
Accepted 22 June 2016
spills along mangrove shorelines worldwide. In total, at least 5.5 million tonnes of oil has been released into man-
Available online 1 July 2016
grove-lined, coastal waters, oiling possibly up to around 1.94 million ha of mangrove habitat, and killing at least
Keywords:
126,000 ha of mangrove vegetation since 1958. However, there were assessment limitations with incomplete
Mangroves and unavailable data, as well as unequal coverage across world regions. To redress the gaps described here in
Oil spills reporting on oil spill impacts on mangroves and their recovery worldwide, a number of recommendations and
Impacts suggestions are made for refreshing and updating standard operational procedures for responders, managers
Pollution and researchers alike.
Monitoring © 2016 Elsevier Ltd. All rights reserved.
Risk management
Standard operational procedure
recommendations

Figure frontispiece. Shipping accidents cause most oil spills that threaten mangrove and tidal wetland habitats, as with the 2007 oil spill in Port Curtis, Queensland, Australia.

⁎ Corresponding author.
E-mail address: norman.duke@jcu.edu.au.

http://dx.doi.org/10.1016/j.marpolbul.2016.06.082
0025-326X/© 2016 Elsevier Ltd. All rights reserved.
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
1. Introduction
Mangroves are highly vulnerable to oil spills because oil deposits on
sensitive plant surfaces, affecting soils and dependant marine life caus-
ing death and sublethal impacts (Swan et al., 1994, Duke et al., 1999,
Duke and Burns, 2003; NOAA, 2014). This disruption affects ecosystem
services of mangroves, like fisheries production and shoreline protec-
tion worldwide. Oil spill impacts also persist for decades, and they can
occur at any time, at any location. So, for as long as oil is extracted and
transported around the world, the risks are ever present. Therefore, it
is essential to be prepared where possible with prior risk assessments
and surveys of baseline shoreline condition, along with database re-
cords of earlier impacts and instances of recovery. There is an urgent
need, for example, to quantify impact severity in a standard way,
starting with relatively simple parameters like: area of tree death; esti-
mated volume of oil reaching mangrove habitat; and, areas of potential
lethal and sublethal impacts. This article reviews current literature, and
lists a number of key recommendations and guidelines for improving
longer term management and monitoring of mangroves affected by
large oil spills.
2. Diverse and vulnerable mangrove assemblages
Tidal wetlands consist of forested mangrove woodlands, thick man-
grove and saltmarsh shrubbery, low dense samphire plains, and
microlagal covered saltpans (Tomlinson, 1994; Duke, 2011). In the tro-
pics, mangroves are often the dominant shoreline ecosystem comprised
chiefly of flowering trees and shrubs uniquely adapted to coastal and es-
tuarine tidal conditions (also see Spalding et al., 2010). They form dis-
tinctly dense structured habitat of verdant closed canopies fringing
coastal margins and tidal waterways of equatorial, tropical and subtrop-
ical regions of the world. Normally, but not exclusively, these vegetation
assemblages grow in soft sediments above mean sea level in the inter-
tidal zone of sheltered coastal environments and estuarine margins.
Mangroves are well-known for their morphological and physiological
adaptations, coping with salt, saturated anoxic soils and regular tidal in-
undation; notably with specialised attributes like: exposed air-breath-
ing roots above ground; extra, above-ground stem support structures;
salt-excreting leaves; low water potentials and high intracellular salt
concentrations to maintain favorable water relations in saline environ-
ments; and viviparous water-dispersed propagules. With such survival
traits, these habitats have key roles in coastal productivity and connec-
tivity, often supporting high biodiversity and biomass not possible in
upland vegetation, especially in more arid regions.
Mangroves are recognised as an economic and biological resource
with a number of highly beneficial ecosystem services (Mukherjee et
al., 2014, Duke and Schmitt, 2015). These include fisheries (nursery,
habitat and aquaculture), ecotourism and recreation, shoreline protec-
tion, coastal water quality buffering (‘coastal kidneys’), coastal primary
production and pollution abatement ranking highly. It is of concern that
these services have been seriously under-valued in recent decades
reflected in natural habitat replacement, alteration and general degra-
dation associated with inadequate management of coastal areas (Duke
et al., 2007; Richards and Friess, 2016). Mangroves collectively face a
number of primary threats with habitat loss caused by direct and indi-
rect human pressures coupled with global climate change. Along with
reclamation and changes to hydrological flows, pollution damage is spe-
cifically identified as a major threat to mangrove ecosystems worldwide
(Hensel et al., 2010). Such pressures have all contributed to an overall
deterioration of the adaptive resilience of mangrove plants, adversely
limiting their ability to respond and adapt to current human and natural
pressures (cs. Lovelock et al., 2015).
Mangroves and tidal wetlands are ancient ecosystems, having
evolved over at least 60 million years. During this time, the earth, sea
level and climate have changed dramatically. Mangroves today, com-
prised of plants and animals, are the survivors of changes through the
701
millennia. As such, these ecosystems have well-developed strategies
for survival in their capacity for dealing with change. Today, as tidal
wetlands ecosystems respond to change, they rely on their inherent
adaptive capacities (eg., Duke et al., 1998a). Where sea levels had
risen and fallen in the past, current occurrences of intact mangrove
stands, are tangible evidence of their resilience and abilities for success-
ful restoration, migration and re-establishment. However such abilities
are finite, and they cannot be taken for granted. In this context, it is co-
gent that mangrove communities around the world are currently in de-
cline both in area and function. The causes appear related in large part to
the combination of on-going ‘natural’ influences, coupled with an ever-
growing menace of direct and indirect human pressures (Duke, 2014a).
Despite such pressures and limitations, there exists a highly diverse
and thriving number of plant types within mangrove stands (Duke,
2014b, Duke and Schmitt, 2015). And, a well-recognised benefit of man-
grove habitats are their habitat role between shoreline and marine envi-
ronments where their biodiversity and structure uniquely provide
shelter, protection from erosion, as well as food for a number of terres-
trial and marine species, like juvenile fishes (e.g. Meynecke et al., 2008).
The forested stands are known also for their traditional forest products,
including timber for construction and local energy demands (e.g.
Walters, 2005). Despite these uses, and others, unique wetland forests
are known only in a few instances for being managed in a sustainable
way. Around the world, these habitats are considered wild and uncon-
trolled compared to upland plants used in urban, city, industrial park-
lands and forestry plantations. Never the less, unmanaged mangroves
occur along many heavily populated shorelines where they provide sig-
nificant, but often unrecognized services, like shoreline protection, nu-
trient plus carbon sinks, fisheries benefits, as well as providing distinct
aesthetic improvements to otherwise drab mud banks, flotsum rubbish
and constructed walls.
In summary, mangrove habitats are seriously influenced by a num-
ber of often deterministic ‘natural’ factors including: tidal inundation;
sea water and broad salinity fluctuations; exposure at times to severe
winds, waves and currents; variations over longer periods to overall
sea level changes; and, growth in unconsolidated sediments with fre-
quent instances of severe erosion and burial deposition. In recent centu-
ries, there are added significant pressures of severe physical damage by
human trampling, replacement with expanding human development,
and pollution (Duke, 2014a). All these have notable and profound influ-
ences on the distribution, functioning and well-being of mangrove hab-
itat. Few other natural ecosystems are subject to such an array of severe
abiotic influences, as listed above, including oil pollution.
3. The influence of oil spills on mangroves
Mangrove communities are particularly susceptible to damage from
large oil spills (also see: Proffitt, 1997; Hensel et al., 2010, Lewis et al.,
2011, Santos et al., 2011). When oil is released into coastal and estuarine
waters and washed ashore, it deposits on sensitive surfaces exposed
during the regular daily ebb and flow of tidal waters. This includes the
sediments that thinly cover the highly sensitive fine feeding roots of
mangrove trees. Once deposited, oil mostly adheres and rarely moves,
having adsorbed to oleophilic surfaces of plants and animals alike.
Only in incidences where oil volumes are very large does it refloat and
continue to spread in a significant way with subsequent tidal flushing.
In each case, oil coats breathing surfaces of mangrove roots, stems, seed-
lings, and surrounding sediments, as well as fauna present in burrows
and root hollows. When smothered with oil, shorter plants and animals,
die mostly within days. By contrast, taller mature trees and shrubbery,
oiled only on their exposed roots and sediments, might persist for six
or more months before dying. Plants are accordingly smothered,
poisoned and starved by oil spills; and the lighter the oils are, the
most damaging they are (Duke and Burns, 1999, 2003; Michel and
Rutherford, 2014).
702 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715

These differences between lethal responses and their quantification
are very informative, but there are also sublethal responses to consider.
The sublethal responses are rarely ever quantified (cs. Duke et al., 1997)
but they are often very important since they help us better understand
and define overall impacts and possible recovery trajectories. So, while
oil type and concentration levels have a primary influence on the im-
pacts observed, it is the responses of mangrove biota to all influential
factors that better define their recovery potential, and the longer term
fate of oil-impacted habitat. For these reasons, and that deposited oil
often becomes less visible after a few days and weeks, it is strongly rec-
ommended that deposited oil be surveyed and mapped accurately at
the time when the spill is taking place, or very soon afterwards.
There are also further influential variables affecting habitat
responses, like differences in the sensitivity of individual mangrove spe-
cies to oil and dispersed oil (Duke et al., 1998b; Duke and Burns, 2003;
Lewis et al., 2011). In Australian studies, mangrove species were tested
in planthouse trials where species showed a range of sensitivities from
highest in Aegiceras corniculatum and Avicennia marina, to lesser levels
in Rhizophora stylosa and Ceriops tagal. The more sensitive species
were those with greater salt excretion strategies as their likely co-related
physiological character. Duke and Burns (1999) further concluded that
sediment type was also important where those plants in more porous
sediments were more vulnerable than those growing in fine clay mud.
In addition, oil types were ranked by increasing impact on plants, from
least sensitive: Bunker C fuel oil; Arabian light crude; Gippsland light
crude; Thevenard crude; to Woodside condensate, the most harmful. In
general, impact level rankings corresponded to oil density, where light
oils were more harmful than dense heavy oils. Oil toxicity trials were
notably made during field trials and post-spill studies, and not during
spill incidents (Swan et al., 1994).
In field situations, oil-impacted mangrove ecosystems follow a rea-
sonably ordered set of condition states with primary, secondary and re-
sidual effects, leading to recovery and/or loss. The schematic model
presented in Fig. 1, shows key trajectories followed by oil-affected man-
grove plants, including: undisturbed (left side, green) to either recovery
(orange); and permanent loss from deterioration (lower right, pink).

Fig. 1. Schematic diagram depicting the effects of large oil spills on mangrove habitat.
Adapted from Duke et al. (1999) and NOAA (2014).
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
Note, the short term effects where some trees die while others survive
oiling.
The longer term effects occur over decades with habitat either recov-
ering via successful recruitment, or lost permanently in a cycle of pro-
gressive degradation and deterioration. Because of the propensity of
oil to clump, especially in cooler conditions, the areas oiled are often
patchy. The result can be a combination of both oiled and unoiled
patches occurring in close proximity. Accordingly, responses can be var-
iable and complicated. Some recovery of unoiled undisturbed patches
can be quite rapid (less than one year) with virtually undetected long-
term effects. The fate of oiled patches however may follow one of two
overall possible trajectories – one to recovery, and the other to deterio-
ration and loss. These latter trajectories of recovery or loss further de-
pend on the severity of oiling, and the severity of any subsequent
damaging events.
Where oiling results in sublethal damage only, then recovery can be
rapid (1–5 years) since affected trees only need to recover their canopy
foliage, and regain their associated fauna. But, where trees die, the re-
covery process will take much longer (5–25 years or more), involving
seedling recruitment, establishment, and the growth of replacement
plants to maturity with propagation. Such recovery processes are diffi-
cult to achieve in this physically dynamic and demanding environment.
By most accounts, tree growth to maturity can take at least 25–30 years.
Assessment protocols set out in the recommended operational guide-
lines below have been developed for assessments of both the extent
and severity of impacts, and the status of recovery/degradation
processes.
The aim has been to characterize and quantify the key characteristics
of any incident under consideration, with severity measures of impact
and condition status in rehabilitation, for any time during or after an
oil spill. For example, Da Silva et al. (1997) identified an apparent out-
come of continuing habitat deterioration where there was no progress
towards recovery, while mangrove recruitment after oiling became ex-
tremely vulnerable to subsequent episodic events like hurricanes and
other human-caused disturbances. By contrast, other authors (Lewis,
1982; Lamparelli et al., 1997; Kadafa, 2012a) when describing impact
trajectories have taken a positive view, where comparable stages and
timelines progress along a trajectory to full recovery of habitat. Unfortu-
nately, this is not always the case, and there appears to be a growing
number of impacted sites showing little or no recovery.
Fig. 2. Mangrove habitat zones (green shading) and shipping risk routes (blue tracks; Wikimedia Commons, 2016), describe major spill incidents (black astericks) only where mangroves
occur. This is representative of the 238 or so, reported oil spill incidents recorded between 1958 and 2016 (see Table 1, Supplementary Data Table 3; compiled from multiple sources). Note:
mangroves occur almost exclusively along intertidal shoreline margins, so their extent is restricted to enclosed margins of islands and continents within green shaded areas.
Adapted from Duke (2011).
703
4. Past incidences of larger oil spills affecting mangroves
Oil pollution incidents, as accidental or deliberate spillages, are the
chief way oil reaches mangrove habitat, especially those in high risk
coastal areas like ports, refineries and busy transport corridors. Since
1958, there have been at least 238 notable incidents of larger oil spills
reported as affecting or threatening mangrove habitats worldwide
(see Fig. 2; Table 1; Supplementary Data). Data on these spills have
been sourced from a number of incident reports, but particularly from
five exemplary online databases listing selections of marine spill inci-
dents (AMSA, CEDRE, NOAA, ITOPF and the Nigerian Oil Spill Monitor,
2016). Unfortunately, none of these sources list all spills affecting man-
grove habitat. And, as mentioned in the recommendations, their content
and access might also be usefully improved upon because many inci-
dents were in the vicinity of mangrove habitat but these were mostly
left unreported. Regardless, data gathered from these sources has
great value in this compilation, as compared and listed in the Supple-
mentary Data Tables 1 to 3. This first comprehensive review draws on
published research survey results, technical publications as well as
these online databases.
The distribution of reported incident locations (Fig. 2) generally
match the global distribution of mangrove habitat (green shaded
areas; see Duke, 2011) – demonstrating that oil spills occur very widely
affecting mangroves where ever this habitat occurs. Four types of oil
spills are represented including: pipeline ruptures; vessel incidents;
shore tank facility disruptions; and, well head damage. Sites of major
experimental field trials include those in Port Curtis, Australia in
1995–1998 (also see Fig. 5). Some open water incidents within the
mangrove area have not been included since they occurred appreciably
offshore and were considered unlikely to have impacted mangrove hab-
itats since they are restricted to intertidal shorelines. In each case, the
choice to include each incident was made based on local circumstances,
including distance offshore, when no observations were made about
impacts on specific shoreline mangrove habitats.
Understandably, there was a notable concentration of incidents in
areas of both high vessel traffic and extraction sources, like the Arabian
Sea, the Gulf of Mexico, South eastern Brazil, and the Niger Delta. How-
ever, the wider distributions dominate, and oil-affected mangroves
could be found in all 6 global sub-regions. This occurrence is explained
by the bulk of spill incidents having occurred during the transport of
704 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715

oil, with shipping accidents accounting for at least 54%. The other
causes, like leakages from pipelines, shore tanks and well heads,
accounted for the rest.
The region of least incidents was the West American sub region in
the Atlantic East Pacific. And, the area with the greatest concentration
of reported incidents was the East American region. While the reporting
from the latter region (cs. mostly NOAA, 2016) was especially inclusive
and detailed, this greater level of reporting provides a challenge to other
regions. Perhaps the region of most deficient information was West
Africa where public reporting of oil spills was far less rigorous (Odeyemi
and Ogunseitan, 1985; Kadafa, 2012a) despite a seemingly first rate
reporting service with the Nigerian Oil spill Monitor (2016). This meant
that it was harder to get a balanced and uniform global comparison, and
there are important observations stemming from this review. Firstly, it
was cogent that reported spill numbers worldwide would be significantly
greater should the same amount of reporting have been made worldwide
as those covered by NOAA and the other databases. And, having said that,
an international database that collated reports equally from all over the
world is a tangible goal. But this needs a greater commitment with
encouragement for global organizations to collaborate, either supporting
and contributing to a single database, or using shared regional databases
that apply and link comparable reporting standards and maintain public,
open accessibility.
While it may also be partly an artifact of progressively improved
reporting, it is more likely that the overall number of incidents had in-
creased much as shown, from 7 in the 1960s, to 59 in the 2000s. This
coincides with a growing global demand for oil. Over this period, the
number of incidents steadily increased in 4 of the 6 global sub regions
with apparent decreases in East Africa and Australasia, (Table 1; Fig.
3). The types of spill events were mostly vessel-related with collisions,
fires, holing and sinkings. Pipeline ruptures and damage to shore tank
facilities make up about 34% of incidents while accidents on offshore
well heads account for the remaining 9%. A lack of reporting and assess-
ment of the impact and death of mangrove vegetation greatly under-es-
timates the likely full extent of impacts on this habitat.
A number of reports have documented the death and/or damage of
mangrove vegetation associated with released oil (NOAA, 2014). Most
report overall oil volumes released into the environment. In total, over
the last 6 decades, N 5.5 million tones of oil were spilled into coastal wa-
ters where mangroves occur. This equates to around 23,216 tonnes per
spill. Volumes of oil released appear maintained at a high rate with little
change overall over half a century (see Fig. 4), as reflected especially in
subregions: East America, West Africa, Indo Malesia and Australasia.
Slight decreases were observed in the 2 other subregions of West Amer-
ica and East Africa.
The latter region of East Africa was affected by a number of war inci-
dents that damaged refineries and storage tanks, as those in Iran and
Kuwait affecting the Arabian Sea in the 1990s. These are some of the
largest releases of oil in modern times, with the damage in Kuwait caus-
ing the loss of around 1.5 m tonnes of crude oil during 1991 – (Böer,
1993; El-Nemr, 2006; CEDRE, 2016; Webecoist, 2016; NOAA, 2016).
Mangroves were affected, but the areas of oiling or tree death were

Table 1
A summary table of reported oil spill incidents having actual and likely impacts on mangrove habitats worldwide over the last 6 decades from the 1958 to 2016. Asterisks (*) mark the
measures of under-estimated impacts (representing b17% of incidents were observations were available). See Fig. 2, for global regions represented. Individual incidents and multiple
sources are listed in the Supplementary Data Table 3.

1958–1969 1970–1979 1980–1989 1990–1999 2000–2009 2010–2016 Total

Larger incident locations (number)

World total 7 28 39 56 59 49 238
East Africa 2 5 4 3 1 2 17
Indo-Malesia 0 1 0 2 5 8 16
Australasia 0 3 8 14 7 5 37
IWP global region 2 9 12 19 13 15 70
West America 0 1 1 0 2 0 4
East America 5 16 22 31 39 28 141
West Africa 0 2 4 6 5 6 23
AEP global region 5 19 27 37 46 34 168

Total oil spill threat

Spillage (tonnes) 83,832 1,685,234 852,043 1,735,425 393,205 775,773 5,525,512
Mean (tonnes/spill) 11,976 60,187 21,847 30,990 6664 15,832 23,216

Spill incident type (number)

Pipeline 0 2 6 11 14 16 49
Vessel 6 18 20 32 31 21 128
Shore Tank 1 4 9 8 6 5 33
Well Head 0 4 4 2 5 6 21
Other 0 0 0 3 3 1 7

Mangrove oiled* (ha)

World total 49 172 11,910 12,328 2962 665 28,086
Sites Measured * 1 9 9 11 7 6 43
Sites Observed, not measured 2 6 4 12 6 6 36
Likely, unconfirmed 4 13 26 33 46 37 159
% sites oiled 42.9 53.6 33.3 41.1 22.0 22.4 35.9
Studied incidents 1 8 8 7 4 1 29
Field trials 0 0 3 1 0 0 4

Mangrove oil-dead, reported* (ha)

World total 49 37 6528 12,215 216 32 19,077
East Africa 0 0 0 0 0 0 0
Indo-Malesia 0 20 0 0 1 0 21
Australasia 0 5 2 8 15 0 30
IWP global region 0 25 2 8 16 0 52
West America 0 0 0 0 0 0 0
East America 49 12 88 13 0 0 161
West Africa 0 0 6438 12,195 200 32 18,665
AEP Global Region 49 12 6525 12,208 200 32 19,026
% dead of oiled 33.3 60.0 76.9 39.1 46.2 9.1 44.1
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
Fig. 3. Numbers of larger oil spill incidents affecting and likely to have affected mangrove
habitats worldwide for 6 global sub-regions during the last 6 decades (Table 1;
Supplementary Data).
unreported. Other large spills occurred in East America due to serious
accidents on two offshore well heads in the Gulf of Mexico, with the
Ixtoc 1 well releasing around 1 m tonnes in 1979 (Hooke (1997);
CEDRE, 2016; NOAA, 2016; ITOPF, 2016; Webecoist, 2016), and the
Deepwater Horizon well head releasing around 760,000 tonnes in
2010 (Khanna et al., 2013; CEDRE, 2016; Webecoist, 2016; NOAA,
2016; Nixon et al., 2016). Each of these very large spills oiled mangrove
habitat, but no reports describe the areas oiled or whether mangrove
vegetation died after oiling. This lack of assessment and reporting
from these very large spills makes it difficult to further evaluate the se-
verity of such spills on mangrove habitat.
As noted, reporting was generally lacking for the region of West Af-
rican shorelines of Angola, and particularly Nigeria where there were
very high concentrations of oil extraction infrastructure – as numerous
well heads and inter-connected piping (Odeyemi and Ogunseitan,
1985; Kadafa, 2012a, 2012b; Nigerian Oil Spill Monitor, 2016; see Sup-
plementary Data). While few prior reviews of world oil spill damage
to mangroves refer to these instances, a number of documents were
found describing the relatively large number of incidents reported in
this review. Mangroves have been affected by numerous oiling inci-
dents throughout the Niger Delta since at least 1979, until the present
day (Odeyemi and Ogunseitan, 1985; Seenreport, 2008; Twumasi and
Merem, 2006; Adeyemo et al., 2009; Kadafa, 2012a; FAO, 2010;
Eoearth, 2010; The Guardian, 2010, 2011; UNEP, 2011; Shell, 2015;
BBC, 2015; Oil Spill Intelligence Report, 2016). The amounts of oil re-
leased, although not always clearly quantified, were unusually large
compared to elsewhere in the world. For instance, where this global re-
view lists 238 incidents of oil spills since 1958 affecting mangroves
worldwide, in the Niger Delta there appears to have been that number
of incidents each year over the same time period (Odeyemi and
Ogunseitan, 1985; Kadafa, 2012a).
Fig. 4. Released oil spill volumes (tonnes) of larger oil spill incidents affecting and likely to have affected mangrove habitats worldwide for 6 global sub-regions during the last 6 decades
(Table 1; Supplementary Data).
705
A brief outline for each decade confirms the on-going situation in
Nigeria (see Nigerian Oil Spill Monitor, 2016, for overall information).
In the 1970s, mangroves were reportedly oiled in 783 incidents
(Odeyemi and Ogunseitan, 1985; Tolulope, 2004; Ukoli, 2005;
Twumasi and Merem, 2006; Kadafa, 2012a). In the 1980s, 11,210 ha of
mangroves were reportedly oiled in 2015 incidents (Ekekwe, 1983;
Snowden and Ekweozor, 1987; Thorhaug, 1992; Tolulope, 2004; Ukoli,
2005; Nwilo and Badejo, 2005; Twumasi and Merem, 2006; Aghalina
and Eyinla, 2009; Adeyemo et al., 2009; Kadafa, 2012a; NOAA, 2016).
In the 1990s around 12,195 ha were reportedly oiled in 4024 incidents
(Adeyemo et al., 2009; Twumasi and Merem, 2006; Kadafa, 2012a;
Eoearth, 2010; Human Rights Watch, 2016). In the 2000s mangroves
were reportedly oiled in 117 incidents (Adeyemo et al., 2009; FAO,
2010; Eoearth, 2010; The Guardian, 2010, 2011; Kadafa, 2012a; Shell,
2015; BBC, 2015). From 2010 to 2016, mangroves were reportedly
oiled in 5 larger incidents (Huffington Post, 2010; UNEP, 2011;
Sciencythoughts, 2013; Wikipedia, 2014; The Guardian, 2015; Oil Spill
Intelligence Report, 2016).
All together, these oil releases in the Niger Delta reportedly amount
to around 100,000 t each decade with likely severe damage caused to
local mangroves as well as to the human communities that depend on
them (Kadafa, 2012a). This amount of oil released equates to around
27% of that threatening mangroves worldwide. To this date, only the
Bodo West study of spills from pipeline sabotage (UNEP, 2011) has ad-
equately verified and described clear impacts on mangroves. But, more
is needed to document the apparently severe and chronic incidents
across the Niger Delta.
For all world regions (Fig. 2), the total amounts of oil released in each
of the 6 sub regions are shown in Fig. 4 for each decade since 1958. Note
that the amounts of oil released in West Africa were maintained, as de-
scribed above. As they stand, these levels matched those in the East
American with lesser amounts in Indo Malesia. There were small de-
clines in the Australiasian, East African and West American regions.
These differ from the trends with increasing number of incidents re-
gionally and globally over the last 60 years (Fig. 3). Total oil volumes re-
leased and likely to impact on mangrove habitat, range from 393,205 to
1,735,425 tonnes per decade. While these measures are generally useful
for overall quantification of incident severity, they have been less useful
in quantifying the incident impacts on mangrove habitat. More direct
measures are needed, like the areas of mangrove oiled, and the areas
of mangrove tree death.
5. Oiled mangrove habitat – a measure of lethal and sublethal impact
Of the total amounts of oil released, only a portion is likely to impact
on mangrove habitat. Much of the free oil may distribute elsewhere be-
fore reaching vulnerable mangrove shorelines. Released floating oil can
evaporate and degrade, be diverted using booms, or even recovered
using skimmers whilst at sea. A more direct measure of likely impacts
706 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
on mangrove habitat is the amount of oil reaching mangroves, and the
area of mangrove oiled. This measure best defines and describes the
subsequent responses of mangrove vegetation to both lethal and suble-
thal damage.
For the incidents reported, only 43 out of 238 (18.1%) were quanti-
fied for the amounts of mangrove habitat actually oiled (Table 1). This
total area of oil-affected mangrove was around 28,000 ha. But, because
this figure was derived from only a subset of incidents, the total amount
oiled is expected to have been around 5 times greater – up to around
155,000 ha.
To help quantify the measures of impact shown in the broadly
sourced data used in this review, it was convenient to categorise three
levels of mangrove oiling where: ‘Site Measured’ quantified those inci-
dents where oiled areas were measured (noted above); ‘Sites Observed,
not measured’ – were the sites with mangroves observed as oiled, but
not measured; and, ‘Likely, unconfirmed’ – were those sites likely to
be oiled since mangroves are known in that locality, but not reported
as oiled. The three categories of oiled mangrove incident reports were
tabulated across all 6 decades in Table 1.
The percentage ‘Likely, unconfirmed’ was overall around 67%. An-
other 14% of incidents had the briefest of comments about mangroves
being oiled. So, without other accompanying details, the presence of
oiled mangroves was deduced sometimes from photographs and
maps in the incident reports. This meant that b 20% of reports made de-
finitive mention of the presence of mangroves in the vicinity, despite
the habitat being unilaterally recognised as highly vulnerable.
A review of available reports shows the ten largest areas of man-
grove habitat damage (see Supplementary Data Table 2; noting source
references there in). These rank the reported single incident impacts
of oiled mangrove area for the countries represented with the largest
spills first:
• Nigeria with the Funiwa 5 Well blowout in 1980 oiling 5107 ha;
• Pakistan with the Tasman Spirit sinking in 2003 oiling around 1000 ha;
• Nigeria with the Pipeline rupture Bodo in 2008 oiling at least 1000 ha;
• The Philippines with the Solar 1 sinking in 2006 oiling 650 ha;
• Panama with the Texaco Refinery spill in 1986 oiling 377 ha;
• Nigeria Pipeline sabotage, Bodo West in 2011 oiling 366 ha;
• Brazil pipeline rupture near Sao Paulo in 1983 oiling around 300 ha;
• Micronesian islands of Yap with the sinking of the Kyowa Violet in
2002 oiling 300 ha; India with sinking of the MSC Chitra in 2010 oiling
around 200 ha; and
• Australia with the holing of the Era in 1992 oiling 100 ha.
6. Dead mangroves – the lethal response of mangrove habitat to
oiling
A specific and relatively easily deduced measure of impact on man-
grove habitat is the area of dead mangrove vegetation post oiling (see,
Duke and Burns, 2003). These data were available only from a relatively
small number of reported incidents, just 36 incidents, or 15.1% of all 238.
The total area of mangroves reportedly killed by oil spills was around
19,077 ha for the subset of sites, extrapolated to around 126,338 ha
for all incidents since 1958. There were a number of instances where a
clear record was made of no mangroves having died. These considered
observations of no lethal impact have been as important to this review,
as those reports with measured areas of dead mangroves. Unlike other
measures, the losses and damage to mangrove vegetation are both di-
rectly related to the overall vulnerability of this environmental receptor.
This severity measure is easily recognised, mapped and measured from
historical and current aerial/satellite imagery. In addition, change detec-
tion assessments with appropriate field validation have been useful in
quantifying losses and gains in other studies of vegetation cover, post
impact and post recovery (e.g., Duke et al., 1997).
The largest areas of mangrove habitat damage reported (see Supple-
mentary Data Table 2; noting source references there in) were in the
Niger Delta with: around 340 ha killed by the Funiwa 5 well head spill
in 1980; 200 ha killed by the Bodo pipeline rupture in 2008; and another
32 ha killed by the Bodo West pipeline sabotage in 2011. Elsewhere in
the World, the largest areas of oil-dead mangroves range generally
lower, with: 69 ha killed in Panama by the Texaco Refinery spill in
1986; 49 ha killed in Panama by the Witwater sinking spill in 1968;
20 ha killed in Indonesia with the sinking of the Showa Maru in 1975;
12 ha killed in Puerto Rico with the Jet Fuel tank spill in 1999; 10.5 ha
killed in Brazil by a Jet Fuel tank spill in 1999; 10 ha killed in Yap, Micro-
nesia with the sinking of the Kyowa Violet in 2002; and another 6 ha
killed in Puerto Rico by an earlier Jet Fuel tank spill in 1986.
These data of amounts (area) of mangrove death are perhaps the
most cogent, simple measures of impact severity, as the measures of le-
thal habitat damage. They provide the means to quantify and compare
relative impacts from which we might later derive realistic estimates
of habitat recovery. Despite the great value of these data, and their
ease of measurement, there have been relatively few reports recording
these useful measures. The inclusion of such key measured observations
in future reporting is one of the chief recommendations arising from
this review.
7. Relationship between oiling and the amount of mangrove damage
This review has drawn on a wide selection of marine oil spill inci-
dents to gain a much greater understanding of the variables and circum-
stances influencing oil-impacted mangrove habitat, and its possible
recovery. Further insights have been derived also from observations
and findings of experimental field trials (Supplementary Data Table
3). The overall aim has been to document key lessons learnt, and to de-
velop and support the revised list of recommendations, guidelines and
strategies for better informing and preparing responders, community
and managers faced with actual and potentially oil-damaged mangrove
habitat. Overall findings include the identification of knowledge gaps,
and other useful deductions considered to be lessons learnt. One of
these found that impact severity can be equated to the amount of lethal
damage. A further finding was an informative, emerging relationship
between the area of mangrove oiled (sublethal and lethal impacted)
and the area of lethal impact.
These relationships have been derived from 7 well-studied spill
sites, including (see Supplementary Data Table 3 for source references):
• the Texaco Refinery, Atlantic coast, Panama in 1986;
• Zoe Colocotronis, Cabo Rojo, Puerto Rico in 1973;
• World Encouragement, Botany Bay, New South Wales, Australia in
1979;
• the Sao Paulo pipeline, Brazil in 1983;
• Era, Spencer Gulf, South Australia in 1992;
• Solar I, Iliolo, The Philippines in 2006; and,
• the Bodo West pipeline sabotage, Niger Delta, Nigeria in 2011.
These more detailed findings show that for a total of around 1900 ha
of mangroves oiled, there were around 123 ha of mangrove death, or
around 6.5%. So, for any given area of mangrove death, there might be
a 15 times greater area of sublethal affected mangrove habitat. This
measure of sublethal damage (as decreased canopy densities) was fur-
ther observed in Panama with assessments of the Texaco Refinery spill
(Duke et al., 1997). In that incident, although there were no specific
measurements of the area of oiled mangrove, the proportion of man-
grove lethal damage was measured around 18% of the observed suble-
thal damage. In this case, the area of sublethal damage measured from
low density canopies was at least 5–6 times greater than the area of
dead mangroves. If the previous ratio of 6.5% were applied it means
that N1000 ha were oiled, and the canopy damage observed
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
(~377 ha) was the area of more severe sublethal damage. In the same
way, the previous measure of oiled mangroves above is likely also to
be much greater based on the area of dead mangroves measured. So
the total amount of oiled mangrove in all incident sites around the
world is more likely to have been around 1900,000 ha since 1958.
Replicated field trials (Fig. 5) were conducted in central Queensland
to further test dispersant use and bioremediation treatments (Duke et
al., 2000). Oils were weathered and applied at a pre-determined rate
of 5 l·m− 2, based on field observations from hydrocarbon sampling
from actual oil spill sediments in Panama (Duke et al., 1997). The im-
pacts of Gippsland light crude and Bunker C fuel oil on mature stands
of Rhizophora stylosa were assessed over at least 18 months post oiling
in each case. With the application of light crude oil, around half of the
trees in each plot died within 6–12 months, compared to control plots.
By contrast, there were no significant differences with the application
of the heavier fuel oil between plots.
These findings suggest that tree death occurred only after exceeding
a defined threshold concentration of oil present in sediments. And, that
there was a relational link between the area of dead and dying man-
grove trees and the oil type and volume (concentration) required to
kill mature vegetation. Although precise dosage levels can be refined,
this current estimate provides a tangible measure by which to extrapo-
late first order estimates of released oil volume based on observed areas
of tree death – or vice versa. This has already been usefully applied in fo-
rensic post-incident investigations and legal considerations with a case
in Nigeria (BBC, 2015; Shell, 2015; The Guardian, 2015; The Maritime
Executive, 2015).
Of further critical importance, there was also a notable order of mag-
nitude difference between recovery periods of lethal impacted
(deforested) habitat (Fig. 1) taking multiple decades for re-establish-
ment (with seedling establishment, growth and development, plus re-
establishment of animals), compared with months and years for recov-
ery of sublethal impacted habitat (as mostly for re-establishment of an-
imals). This section outlines such vulnerabilities of mangrove habitat in
relation to large oil spills, taking into account important geophysical and
temporal variables. These are important for predicting the likely impact
of spills on mangroves in any given circumstance.
The loss of keystone individual plants of mangrove communities is
likely to have substantial consequences for tidal wetland habitats. Mi-
croclimate is altered with the absence of canopy, resulting in increased
temperatures and lower humidity. Physical habitat is lost for plants and
animals that attach to mangrove stems and roots. Sediment can also be
lost when mangrove roots decay and become detached over large areas.
Any loss of sediment, especially at the seaward edge of a mangrove, will
represent a significant impediment to the re-establishment of the hab-
itat (Duke, 2001).
Fig. 5. Field trials tested the resilience mature mangrove stands to oil and dispersant
during 1995–1998 in Port Curtis, near Gladstone (Fig. 2).
707
These observations describe two distinct levels of impact, with one
being easily quantifiable and available. These are considered essential
to both quantify such variables with each incident, and to learn more
about how mangroves respond to the presence of petroleum on their
roots, and coating surrounding sediment surfaces. While oil on exposed
foliage is known to be lethal, the impacts on taller trees with oiling
around their trunks were more cryptic and previously uncertain. The
observed dosage to sediments surrounding mature trees (noted above
for the central Queensland field trials) was checked and found compara-
ble to concentrations in sediments measured during accidental large
spill events where trees died (Duke and Burns, 2003). Lower trial dos-
age levels of 0.1 to 0.5 l per m2 resulted in smaller patchy impacts on
vegetation and fauna, with no tree death (Duke et al., 2000).
8. Retention of oil in mangrove sediments
Oil can at times be retained within sediments of the intertidal zone
for decades following an oil spill. And, there may not always be harmful
effects on plants, their growth or reproduction. Sediments deposited in
intertidal habitats, especially mangroves, are often fine-grained, water-
logged and rich in organic matter. These mostly oleophilic properties fa-
vour retention of oil and other contaminants via physical and chemical
processes (Lewis et al., 2011, Santos et al., 2011).
The abundance of burrows made by crabs and other fauna has been
shown to enhance the retention of oil at depth within intertidal sedi-
ment (Duke et al., 2000; Culbertson et al., 2007, Hensel et al., 2010).
Mudflats are less likely to accumulate oil relative to mangrove plants
given the lack of vegetation and debris to trap material, and their loca-
tion lower in the intertidal zone where tidal flushing is regular. Also,
as noted previously, oil in mangrove sediments degrades more rapidly
in places where there is greater tidal range and exposure (Burns et al.,
1993, Burns et al., 2000).
Chronic impacts on mangroves have been observed across long pe-
riods (up to decades) following oil spills, with the symptoms most com-
monly observed being: death of trees with seedling regeneration;
defoliation and canopy thinning; leaf yellowing; reduced height growth
of surviving trees and poor seedling establishment (Duke et al., 1997,
Hensel et al., 2010, Lewis et al., 2011); plus, likely toxic response defor-
mities and morphological changes such as pneumatophore branching
(Duke et al., 2005), reduced lenticel numbers (Böer, 1993), and genetic
mutations like variegated leaves and chlorophyll-deficient propagules
(Duke and Watkinson, 2002).
Distinguishing both acute and chronic effects, as well as the different
responses from toxic or smothering impacts, are all important for
predicting the ecological consequences of any spill and for modeling
subsequent impact and recovery. Because of the often complex circum-
stances surrounding any spill incident where the dynamic between
long-lived trees and short lived fauna are responding to an altered, com-
plex and heterogeneous environment, it is understandably challenging.
Additional controlled and replicated studies undertaken in realistic field
settings (see Duke et al., 2000) are needed to address the important data
gaps identified above. For instance, it is important to derive realistic
dose-response data, and to follow longer term processes associated
with the accumulation and breakdown of oil in sediment.
There are complex inter-relationships also amongst plants and ani-
mals. And, it is clear that each depends on the other in significant
ways. So, a decline in the condition of either plants or animals will affect
the well-being of the other biota types (Cannicci et al., 2008). One ready
example is the influence of crabs as ecosystem engineers, with their
burrowing and soil bioturbation affecting sediment aeration, nutrient
turnover, water exchange and leaching of salt within sediments. Crabs
also have a major influence on mangrove forest structure with their se-
lective predation of propagules. Other examples include the beneficial
role of sessile sponges on mangrove roots preventing borer attack, stim-
ulating rootlet growth, and enhancing nutrient uptake; and heat-
708 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715

restricted weevil borers that severely limit propagule establishment
that might also influence mature forest structure (Duke, 2001).
9. Recovery potential of oil-damaged habitat
Post spill assessments suggest that structural recovery of oil-dam-
aged mangrove forests takes place over a period of at least 3 decades.
For instance, an estimated time of recovery of 30–36 years was deduced
from 16 oil spills affecting various mangrove locations (Duke and Burns,
1999). Based on the ranked recovery state phases described in Table 2,
the previous incident data were re-assessed as part of this treatment.
The recovery period relationship is re-affirmed, as displayed in Fig. 6.
It must be emphasized however that this recovery only refers to forest
structure, and not the dependant fauna. From other studies, it seems
that recovery of fauna and full habitat recovery may take longer periods
than recovery of forest structure alone (Salmo and Duke, 2010).
It seems that once the toxic components of spilled oil have dissipat-
ed and broken down to more benign residual products, mangrove
stands may re-establish in similar ways to non oil-impacted forests
(see Duke, 2001). Erosion may also follow deforestation with tree deg-
radation and subsequent dislodgement of root systems and soil binding.
Elevation within the intertidal zone governs inundation regime, which
is critical for establishment and growth of mangroves with particular re-
quirements that vary between species. Therefore, any loss of soil and el-
evation via erosion will influence mangrove re-establishment (Lewis,
2005). Instances have also been reported where dead trees and

Table 2
Mangrove forest recovery goes through 6 phases and one negative state, based on structural regrowth from sublethal (marginal) and lethal (central, gap) impacts (also Fig. 1), respectively.
Indicators of gap condition status have been adapted from Duke (2001). Faunal status, while undoubtably important also, has not been included in this treatment. Actual times taken for
forest recovery vary due to the different trajectories of sublethal and lethal impacts.

Recovery
Recovery phase of status % Sub lethal Lethal state representation
gap percent trajectory Lethal trajectory in gap in gap

Natural pre-damage Reference Foliage dense with yellowing leaf numbers b10%. Trees mostly alive throughout stand; occasional dead
state condition Seedling bank under closed mature canopy. trees and up to ~10% light gaps in ambient conditions.

1. Recently oiled Positive Yellowing and loss of foliage in affected areas, and Tree death (within 6–12 months after spill), dead
1–10 presence of dead, low-placed seedlings. Some seedlings and saplings. Trees with dead yellow leaves
surviving seedlings. and small twigs present. Mostly dead seedlings.

2. Recovery Positive Loss of foliage in affected areas, and presence of dead, Deterioration of dead trees missing small branches
preliminary 11–30 low-placed seedlings. and twigs. No appreciable recruitment, some
seedlings.

3. Recovery Positive Foliage density in recovery with new growth. Deterioration of dead trees missing large branches and
established 31–50 Re-establishment seedling bank under re-established upper stems. Establishment of additional seedling
canopies. recruits in open areas.

4. Recovery Positive Foliage density in recovery with new growth. Notable large stumps remain with some exposed
progressed 51–70 Re-establishment seedling bank under re-established roots. Saplings dominate in dense stands, in the forest
canopies. gaps. Immature, low level canopy closure.

5. Recovery Positive Foliage density in recovery with new growth. Reduced remnant dead stumps & wood sections.
advanced 71–90 Re-establishment seedling bank under re-established Canopy closure advanced. Notable thinning of saplings
canopies. and young seedlings present.

6. Structural Positive Normal foliage density with canopy closed. Site None or occasional remnant mature-sized stumps.
recovery in final 91–100 Maximal Canopy Height unaffected. Presence of Canopy closed. Damaged area Site Maximal Canopy
stages of seedling bank of 3–6 year old young plants, and a Height restored. Formation of seedling bank of 3–6
completion. notable gap between mature canopy trees. year old recruits, notable class gap to mature canopy.

Possible Negative Foliage absent in impacted gap area Dependent on state of gap degradation. Absence of
deterioration condition integrated roots, living seedlings, saplings and young
state post trees. Evidence of scouring.
recovery, likely
from phases 1 to 5
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
Fig. 6. Recovery of mangrove forest structure killed during various oil spills (see
Supplementary Data Table 3). Recovery was quantified in terms of total above ground
biomass of mangrove stands before and after deforestation (see Table 2). The linear
regression of ‘best fit’ shows a trend towards recovery of after around 3 decades
(dashed line: r = 0.878; N = 7; P b 0.005). Note, this trend does not necessarily
represent full habitat recovery since faunal biota were not considered.
dislodged roots have been carried away by waves and tides, causing a
second wave of mortality through scouring of newly-established plants
around 6 years after the initial oil spill impact (Duke et al., 1997, Duke
and Burns, 1999). While oil-impacted, fringing mangroves on high en-
ergy coastlines are at greatest risk (Duke and Burns, 1999), it is such ex-
posed stands that maintain and support the integrity of the structurally-
delicate inner stands. Re-establishment of these stands is critical but
there is little evidence demonstrating how this might happen. In any
case, it means that all mangrove stands are highly vulnerable to oil
spills, and that lethal damage should be avoided at all costs!
Biotic and abiotic interactions with mangrove ecosystems are com-
plex, and each spill has a different set of circumstances. Hence, some
generalisations about the long-term term consequences of oil spills in
mangroves are difficult to make. In summary, the primary factors
most likely to influence recovery time are oil type (heavy versus
light), quantity, and oil condition (fresh and concentrated versus
weathered and dispersed). High initial impacts are most likely where
fresh, light oils are involved, but also possible also with concentrated,
heavy oils where they smother exposed surfaces (Duke et al., 1998b).
Broadly applied dense oiling is also more likely to result in delayed re-
covery than small patches of oiling over several hectares or less. Recov-
ery clearly depends on biotic factors also, like the supply of propagules
for colonization, and variations in species tolerances.
Furthermore, recovery may proceed more rapidly in areas of high
tidal ranges where the effects of more flushing appears to break down
deposited oil faster (Burns et al., 1993, 2000). This is shown from the
currently considered data in Fig. 7, which shows the relationship be-
tween tidal range for particular incidents, and the estimated annual re-
covery rates (Supplementary Data Table 3). The recovery rates vary
from 3 to 6% per year for a tidal ranges varying from 0.6 to 3.9 m. The
implications of these preliminary findings are that the application of ar-
tificial flushing may be beneficially applied to accelerate the recovery of
at least the structural components of mangrove stands.
A number of management options exist to assist recovery of man-
groves following an oil spill. These focus on accelerating the rate of
breakdown of oil (via degradation and bioremediation) and planting
seedlings to assist natural regeneration. As a means for restoring man-
grove cover, facilitating natural regeneration is generally preferred
over planting as it is more cost-effective, and it can have a better out-
come for structure and composition (Lewis, 2005). For oil spills, seed-
ling re-establishment may only proceed once toxic compounds in
709
Fig. 7. Rates of annual recovery of mangrove forest structure influenced by different tidal
ranges (see Fig. 6; Supplementary Data Table 3). The linear regression of ‘best fit’ shows a
trend where greater tidal range equates to greater rates of recovery (dashed line: r =
0.827; N = 7; P b 0.005). Note, this trend does not necessarily represent complete
habitat recovery since faunal biota were not considered.
sediments have sufficiently diminished, as the agents causing harm. Fi-
nally, if changes in elevation and alterations in tidal flow have occurred
following severe or prolonged damage, re-engineering may be needed
to create suitable conditions for establishment and growth (Schmitt
and Duke, 2015). For areas that have been denuded of cover and debris,
strategic planting or the positioning of artificial structures may assist to
capture and stabilise sediment and create nodes of advanced recovery
(also see Gedan and Silliman, 2009).
In conclusion, on-going improvements for better managing recovery
and rehabilitation of oil-impacted mangrove tidal wetland habitat, de-
pends on the successful implementation of robust assessment and mon-
itoring strategies and measures for gauging the success of intervention
and mitigation efforts. These measures need to be ecologically sound
while providing quantifiable, longer term comparisons of reliable, prac-
tical and useful evaluations of habitat condition and health. While a
great deal of progress has been made in some aspects, the recommenda-
tions that have arisen from these considerations are considered practi-
cal and useful next level improvements in the sustainable
management of valuable but vulnerable tidal wetland habitats
worldwide.
10. Operational response guidelines for oil-damaged mangroves
As noted, there is a great need for standard reporting of each spill in-
cident affecting mangroves. This includes sharing these findings in a
public forum. Standard descriptors start with: the area of oiled man-
grove and surrounding sediments; as well as areas of oil impacted de-
forestation. Finally, there is a need to promote rehabilitation of
damaged areas, especially for areas deforested by oiling. With these
measures, it is understandably considered best practice to support nat-
ural processes of natural seedling recruitment, and to promote plant
growth generally. For instance, only if absolutely necessary, should sup-
plemental planting be used to aid recovery and seedling re-establish-
ment when stand integrity has not yet reached self-reliance.
A more complete set of suggested actions are listed in Table 3, cover-
ing the four key threat phase periods that make up a ‘Pre, During, Im-
pact, Post’ (PDIP) Spill Response plan. It is notable, that this strategy
reflects the nomenclature to be used by AMSA and CSIRO in the new
Australian National Oil Spill Monitoring Handbook. By way of further
explicit explanation, nine (9) key recommendations are also given for
improving the management of oil-damaged mangrove and tidal wet-
land habitat. Further included are brief guidelines for responders to
710 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715

Table 3
Actions and response strategies for monitoring mangrove habitats impacted, or likely to be impacted, by large oil spills, as a set of recommended Standard Operational Procedures. The
types of responses, actions and their timing depend on the Threat Phase of the spill incident. The four Phases include: “Pre-Spill”, “During Spill Pre-Impact” and “During Spill Post-Impact”
and “Post Spill”. Modified from Duke and Burns (1999). This describes the “Pre, During, Impact, Post“(PDIP) Spill Response Plan for the four (4) critical phases when oil spills threaten, or
impact on, mangrove and tidal wetland habitat.

Oil spill threat phase Operational monitoring actions Response measures Justification

Pre-spill Proactive preparedness – Training of personnel in
Before Spill Incident high risk areas, equipment preparations, plus the
compilation of Reports, Data and the approval of
current best practice Strategies for Monitoring
and Assessment of Oiled Mangrove Habitat.
• Feedback from prior incidents of monitoring,
recovery and mitigation measures used.
• Risk & Resource Assessments Minimization of
risk at possible spill point sources, including
safe ship handling practices.
• Contingency Planning with the Compilation of
best advice, equipment and personnel training
for effective rapid response.
• Publication of a coastal resource atlas,
identifying risk levels and habitat vulnerability.
• Research into fate and effects of oils and
dispersed oils on mangrove communities.
• Regular re-assessment and updating of
response, cleanup and management protocols.
• Maintenance of a contact list of mangrove
monitoring specialists.
• Test likely products and strategies used in
mitigation and cleanup.
• Supply & Equipment Stockpiles developed.
During Spill pre-impact. Oil threatening mangroves – reactive phase
Spill Incident Taking improved greatly by any prepared response
Place - oil threatens measures.
mangrove habitat. • Urgent need for monitoring of mangrove
habitat to start with the Rapid Response Units
during the spill incident.
• Prediction of Threat - Spatial & Temporal -
Predict movement of oil and evaluate threats to
potential stranding sites of oil along the
shoreline.
• Apply Contaminant Delay and Diversion
Strategies - evaluate the range of response
options, including “do nothing but monitor the
incident”.
• Dispersant Application - review cleanup and
containment techniques subject to habitat, oil
type, etc.
• Recovery of Oil - Assess removal, disposal, and
assisted degradation of oil.
During Spill post-impact. Newly oiled mangrove - Reactive Phase improved
Spill Incident Taking greatly by any prepared response measures.
Place - oil deposited • Urgent need for monitoring of mangrove
amongst mangrove habitat to start with the Rapid Response Units
plants and surrounding during the spill incident. There is only a very
sediments. short period of days and weeks before vital
evidence is lost.
• Habitat Affected - Shoreline Assessments -
evaluation of affected habitat.
• Remote Imagery & Mapping - before damage -
evidence of canopy damage.
• Infauna Assessment - assessment of infauna
condition.
• Stranded Oil Assessment - map oil ‘hot spots -
assessment of deposited oil
Post spill. Previously oiled mangrove – opportunities to
Oil Long Settled, evaluate mitigation strategies, as well as the
Mangrove notably quantification of past impacts and any recovery 1) Standard reporting. Document what is known
Impacted, and/or processes. of the primary oil spill incident – along with any
Recovery Taking Place. intervening severe events, like cyclones, or
• Habitat Damage & Recovery Assessments - additional oil spills.
evaluation of damage, response and recovery of 2) Standard assessment methods. Specifically for
Specific responses include:
1) Oil spill database. Develop and support a
national and/or regional database and atlas of
oil spill incidents affecting mangroves.
2) Current contacts. A network of current
stakeholders, like the Australian Mangrove &
Saltmarsh Network (www.amsn.net.au).
3) Shoreline risk evaluation. Shoreline mapping
using latest satellite imagery; in combination
with shoreline condition monitoring and
assessments (like the Shoreline Video
Assessment Method) for the entire shoreline
area at risk.
4) Standard operational procedures. An
approved and agreed set of
procedures/strategies for the monitoring and
rehabilitation of oil damaged mangrove habitat,
based on past incident experience, reports and
research.
5) Standard methods. Production of a series of
approved Standard Assessment Methods, like:
mangrove cleanup techniques; how to
recognise mangrove recovery status, and aging
dead tree decomposition; species identification
material that are regionally relevant for plants
and animals.
Specific responses include:
1) Standard operational guidelines. Follow any
agreed and approved Monitoring Assessment
Strategies and Field Procedural Check Sheets.
2) Standard reporting. Document Incident
circumstances as recommended in publically
available reports.
3) Standard assessment methods. Specifically for
identification of: oil type (collect fresh samples
from source); oil volume released; oil volume
salvaged; locations likely to be oiled; type of
corrective chemicals used, like dispersants;
installation of booms; use of skimmers;
document impacts on mobile marine and bird
fauna; collect water quality data.
4) Oil spill database. Add new and revised
information to the national oil spill database
and atlas.
Specific responses include:
1) Standard operational guidelines. Follow any
agreed and approved Monitoring Assessment
Strategies and Field Procedural Check Sheets.
2) Standard reporting. Document ongoing
incident circumstances.
3) Standard assessment methods. Specifically for
identification of: oil volume released (collect oil
sample, floating/deposited on trees and on
mud); map locations where oil deposited; note
oiled ‘hot spots’; document impacts on fauna;
collect data and samples of each species
present; collect water quality data.
4) Oil spill database. Add new and revised
information to the national oil spill database
and atlas.
Specific responses include:
This applies to unspecified threats from large oil
spill incidents that may affect sensitive mangrove
habitat.
It is essential best practice to be prepared using
the best available resources and people.
Preparedness need be applied at a National and
regional level since major oil spills can occur at
anytime and at any location.
This ‘In between’ Phase is the best time to
review, discuss and approve a series of
dedicated, current best practice guides along
with current best practice monitoring
strategies. It is also the time to prepare contact
lists for particular specialists who can assist
both during the Rapid Response Phases, and the
more considered post incident assessments.
Oil spill current – or within 2 weeks.
Oil not in mangroves but it threatens.
The primary aim at this phase is to protect sites
of sensitive mangrove habitat. To be effective,
this requires a good knowledge of the
whereabouts of any particularly vulnerable
habitats, as well as the current climatic
conditions (like wind direction, waves, storms,
etc), and current movements.
Responders also need to have an eye on the
future with additional, targeted monitoring to
further improve future responses (like the
trialing of cleaning techniques, innovative
products, etc).
This includes the consideration and
establishment of incident specific influential
parameters – as listed in Response Items. These
particularly concern the locations of oil
deposition, as well as reasonably accurate
measures of habitat likely to be affected by oil.
Oil spill current – or within 2 weeks.
Oil in mangroves, coating exposed roots, soil &
wildlife.
At this critical phase, it is essential to determine
the extent of likely and actual oil spill impacts
related to both lethal and sublethal responses
by affected mangrove habit. This requires a
good knowledge of the indicators by those
undertaking the monitoring.
It is essential during this phase to establish key
criteria that will frame all post spill assessment
reporting. For example, it is important to know
about the extent of likely sublethal damage.
This can only be determined if the extent of
oiled mangrove is mapped during this short
phase. Much of the visual evidence of broader
oiling disappears shortly after this phase.
Oil spill much earlier – maybe last year or decades
since.
For this phase, it is essential to document what
can be discovered about the earlier oil spill
incident.
Important parameters to add at this time
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715 711

Table 3 (continued)

Oil spill threat phase Operational monitoring actions Response measures Justification

oiled habitat.
• Remote Imagery & Mapping - after damage -
evidence of longer term canopy damage.
• Assessment of Associated Biota - condition of
mangrove infauna.
• Oil in Sediment - analysis of residual oil in
sediments.
• Assist in restoration of habitat where and if
necessary.
• Assessment of prior Restoration Interventions
identification of; visual presence of oil; collect
water samples; collect sediment samples for oil
assessment; assess status of mangrove habitat –
condition; structure, biodiversity; record dead
trees and status of decomposition;
measure/sample presence of fauna, biomass,
species, abundance.
4) Oil spill database. Add new and revised
information to the national oil spill database
and atlas. Especially reporting on recovery
status, and how affected sites responded to
applied mitigation strategies.
include: the full extent of structural (lethal)
damage (this would not have been evident at
the time of the spill); the extent of sublethal
damage to reproductive success & productivity;
impacts on dependent animals, including local
fisheries; review lessons learnt during the
incident for reducing impacts of future spills.
It is important at this phase to establish
amounts of progress, if any, towards recovery.
This requires those monitoring to have a
standard set of robust indicators for ranking
levels of damage, and recovery. Such objective
measures are needed when comparing
incidents, and to get more measured
evaluations of the effectiveness of prior
mitigation strategies.

large oil spills. The following points and recommendations are made
based on the prior studies and observations considered in this article.
11. Specific recommendations for monitoring oil-damaged
mangroves
The history of oil spills affecting mangroves suggests that recovery
and rehabilitation generally takes at least 3 decades (as noted above),
depending on climate, tidal range and geographic circumstances
(Lewis et al., 2011; Duke, 2001; NOAA, 2014). However, sufficiently de-
tailed studies of long term impacts are relatively few in number. What is
required is more explicit reporting using a standard selection of key de-
fining parameters. While an array of factors influence the outcome of an
oil spill incident on mangrove stands, only a few parameters are needed
to quantify impact severity and on-going changes towards recovery, or
degradation. By measuring these parameters, it should be possible to
quantify overall impacts and timeframes for recovery.
This understanding would greatly improve the management of oil-
damaged mangrove ecosystems. For example, oil type and the the
amount of habitat oiled are key determinants of impact severity, and
the longer term responses. With this knowledge, responders might be
able to more confidently apply targeted intervention with a reasonable
expectation of achieving certain outcomes and in reducing impacts. This
could include interventions like: the application of dispersants; the re-
direction of oil deposits; the reduction to oil deposited within mangrove
stands; enhanced flushing; and bioremediation. The following nine (9)
recommendations outline a series of strategies intended to guide better
and more targeted responses that promote and accelerate recovery of
mangrove habitat affected by larger oil spills.
For relevant spill incidents, both past and present background infor-
mation should be made available, collected and archived in a central
public repository. Full access to all relevant documents is essential if
there are to be further improvements in future response actions. This
database could build on prior reviews of large oil spill incidents, like
those from around Australia (see Duke and Burns, 2003). It is important
to piece together, capture and link all relevant historical facts and obser-
vations that might otherwise be lost.
Exemplary examples of existing databases include those mentioned
in this review, including: AMSA (2016); CEDRE (2016); ITOPF (2016);
NOAA (2016) and the Nigerian Oil Spill Monitor (2016). These model
resources could be improved upon with: more defined content (cs. spe-
cific geographic areas, incident types, etc), better search capabilities
(like, mangroves affected), a wider range of standardized entry fields,
and multiple choice selections to help responders upload their observa-
tions in a standard format.
Recommendation 3: Where possible, either as part of pre-spill con-
tingency planning, or as part of rapid assessment pre-impact, it is
highly beneficial that baseline condition of the oil- threatened man-
grove shorelines be measured and evaluated using accepted
methodologies, to support future ‘before and after impact’ com-
parisons. Where this is not possible, appropriate comparative sites
need to be established in suitable locations known to reflect as
well as can be determined, the pre-impact state of the impacted
mangroves.

The ideal baseline monitoring of shorelines threatened by oil spills

Recommendation 1: Follow a Pre, During, Impact, Post (PDIP) Re-
sponse Plan (Table 3) to minimise impacts and costs associated
with oil-damaged mangrove and tidal wetland habitat.
For overall management of large oil spill impacts on mangrove and
tidal wetland habitat it is necessary to follow a plan like the all
encompassing strategy of Pre, During, Impact, Post (PDIP) Response
Plan, adapted from Duke and Burns (2003) and described in Table 3.
Recommendation 2: Report, record and make available publically,
all relevant information about oil spill incidents where they impact
on mangroves and tidal wetland habitat.
needs to be undertaken prior to an oil spill incident. But, since oil spills
are known to occur anywhere, as demonstrated in this review, it re-
quires regional shorelines be fully surveyed beforehand. This appears
to present a difficult challenge, but there are solutions found in recent
innovations in shoreline surveillance. One is critical evidence captured
in high definition satellite imagery and data. Remote satellite mapping
is extremely valuable in providing overall quantifications of habitat
types and locations, but there are limitations to its application which be-
comes obvious only when ground verification is found lacking.
An important recent compliment to satellite remote sensing is the
Shoreline Video Assessment Method (S-VAM) as used from either
boat-based (Mackenzie et al., 2016) or low-level, aerial platforms
(Duke et al., 2010). The S-VAM approach provides a unique, improved
perspective that allows quantification of shoreline processes as well as
likely impacting agents, like severe climatic events, storms, and large
712 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
oil spills. For instance, S-VAM aerial survey data was used to establish
shoreline baseline condition, including mangroves, threatened during
the offshore Montara oil spill incident off the north-west coast of Aus-
tralia in 2009 (CEDRE, 2016; AMSA, 2016). A notable feature of this
low level aerial monitoring method is that it allowed not only establish-
ment of baseline condition, but it also provided a permanent reference
database from which to evaluate future change along these shorelines.
The future application of this method depends on two key contribu-
tions: continued image acquisition through space and time; and, the de-
velopment of a dedicated data management and display portal, called
ShoreView. The former requires its adoption by coastal surveillance
agencies at a national scale, and the latter is under development for
the Port Curtis Port Alma Coastal Habitat Archive and Monitoring Pro-
gram (Duke and Mackenzie, 2015). For those interested, the prototype
PTTEP Montara ShoreView display portal is available for viewing at:
http://203.101.224.239/.
Recommendation 4: Collect all descriptive data on large oil spills in
a standardized, expanded format, especially where spills might im-
pact on mangrove and tropical saltmarsh habitat.
It is recommended that some additional fields be considered for
inclusion in future evaluation and monitoring guides. The value of
reports on oil spill incidents affecting mangroves could be greatly en-
hanced by adding the suggested fields (see Duke and Burns, 2003), as
fully detailed in Table 4. And, these would be most useful if they were
freely available as per Recommendation 1. The information fields
listed expand slightly, but importantly, on those listed by NOAA
and other organizations. The new observations made in this review
support further useful post-spill deductions, for example, regards
extrapolated volumes of oil lost based on the area of oil-related
dieback.
Recommendation 5: Determine, using accepted methods, the
type, extent and concentrations of oil within the impacted areas
of mangrove (cs., water surface, water column, sediment and
mangrove surfaces and buried sediment) at both the time of oiling
and at some period after initial oiling (i.e. 3–6 months) when lethal
impacts are apparent, and the type, extent and severity of man-
grove damage (e.g. severity of defoliation or deforestation) is
known.
The severity of damage caused by oiling of mangrove habitat, should
be measured in terms of deforestation and habitat loss (the lethal im-
pact), as well as sampling and measuring oil concentrations and oil
type and condition from contaminated sediments. These and other
influencing factors are needed in the making of a baseline assessment.
There is further value in using the reported lethal dosage of 5 l.m−2 of
mature Rhizophora stands in Panama (Duke et al., 2000), to estimate ei-
ther the area of mangroves likely killed by the known volume lost, or
vice versa, to deduce the likely volume of oil from the area of dead man-
groves. It follows that the area of mangrove death would not be the only
a measure of impact, but this measure is a useful proxy for the amount
of oil deposited in mangrove habitat. This allows impact/response asses-
sors a means to estimate the minimal volume of oil lost in mangroves, as
the amount it takes to cause an observed area of tree death. Clearly, such
an estimate will be approximate, since there are other factors to consid-
er, like oil type. The accuracy is expected to be improved as we learn
more from better and more specific reporting, as noted in these Recom-
mendations (2, 3 & 4).
Table 4
A proposed standard list of information fields for reporting on large oil spill incidents, es-
pecially those resulting in mangrove dieback and habitat loss. Modified and adapted from
Duke and Burns, 1999; NOAA, 2014.
Standard reporting oil spill incident database fields
Primary information
Date, or period, of spill
Date/time of landings of oil by location
Location, name and site coordinates
Tidal range, tidal phase during oil landings
Site exposure condition – aspect, direction
Types of oil
Quantities of oil released
Weather conditions at the time and subsequently
Description of spill, details of incident, vessels involved
Observations/measurements/samples, slick observations
Spill response
Description of response action
Volume of oil recovered
Remediation/mitigation strategies used
Type and quantity of dispersant used
Sediment condition, physical site parameters
Sediment type
Hydrocarbons in sediments
Coastal features and water movements
Biotic condition, impact assessment
Area and extent of mangrove oiled – time of landing
Extent of mangrove deforestation – over one year
Mangrove species affected – oil contacted
Other mangrove biota affected – death of animals
Status of plant condition, regrowth, recruitment
Remediation applied
Natural restoration, no intervention, location areas
Assisted restoration, types and location of interventions
Habitat recovery assessment
Monitoring strategies conducted
Research investigations undertaken
Maps and remote sensing imagery
Field sketch maps and notes
Satellite imagery, highest resolution, GIS layers
Aerial imagery flown
Historical imagery, background status, change detection
Image records and documentation
Photographic imagery archive
Official reports, publications
Popular media, newspapers, websites
Contact person(s)/organisation(s), email, web sites
Spill management and clean-up
Impact assessment/monitoring/research
Recovery monitoring and current status
Site access restrictions, authorities and delegations
Recommendation 6: Post-spill monitoring of mangrove reforesta-
tion and the presence of hydrocarbons in sediments needs to be
conducted over 3–4 decades to assess habitat recovery where
trees have died.
Habitat rehabilitation from recruitment to canopy closure and mat-
uration of previously deforested areas from oiling may take around
35–50 years from the initial incident of first impact. Note that habitat
condition refers to the diversity and health of both plants and animals.
Recovery rates are expected to be difficult to predict because there are
a number of influential factors like, tidal range, temperature and rainfall,
coupled with site conditions of species assemblages, sediment type, site
exposure. There are also different residual oil effects, with their differing
rates of degradation. A further notable factor often overlooked is the
prior condition and resilience of mangroves before they had been
impacted by a current incident. While it is essential where possible to
establish baseline condition, this recommendation requires the identifi-
cation of adequate longer term funding and support through
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715 713

appropriate mechanisms starting with insurance and company clean-
up obligations.
Recommendation 7: Mangroves along exposed foreshores and
fringing the waters' edge are more vulnerable to damage and re-
quire the highest level of protection.
Rates of recovery are largely influenced by site exposure where
those in sheltered locations recover slowly while those in exposed loca-
tions may have initially recovered quickly only to be subsequently killed
by “driftwood scouring”. This degradation occurs after the roots of dead
trees have decomposed and become mobilised, after around 5–6 years,
as observed in Panama (Duke et al., 1997). This occurrence appears to
sensitive habitats like coral reefs, while burning can produce solid resi-
due that can wash ashore and also contaminate the intertidal zone.
Should oil enter a mangrove stand then a decision must be made at
the time to act, or not, based on best available evidence of all extenuat-
ing circumstances. Information reviewed by Hensel et al. (2010) pro-
vides some well-founded suggestions for the cleanup of oil spilled
within mangroves and associated intertidal habitats. A summary is pre-
sented (Table 5) as an additional guide for making informed decisions
about cleanup intervention. Of these, the most likely beneficial inter-
vention appears to be gentle flushing (Fig. 7), since this simulates
Table 5
Once oil is within a mangrove stand, the following is a brief list of options ranging from no
action, to varying forms of intervention (adapted from Hensel et al., 2010; NOAA, 2014).

apply generally where trees die in exposed stands for various reasons Intervention action Comment
(like severe storm winds), and resulting in stand replacement and turn-
No action This is recommended for most affected areas.
over (Duke, 2001). Booms and other barriers, Can be effective but generally only for low
including adsorbant booms and energy settings and where advanced
pillows preparation enables structures to be
positioned in advance of the spill coming
Recommendation 8: Post-spill monitoring of animals needs to con- ashore.
tinue as part of habitat rehabilitation assessments for more than Manual removal Most applicable for heavy oils as they can be
two years to adequately assess their recovery. persistent and cause long term risks;
however, trampling and other forms of
physical damage are most likely to be
unreasonably destructive.
While mangrove trees take several months to die following a large Vacuuming Suitable only for the outer fringe of a
mangrove and/or in open stands, coarse
spill (as for Recommendation 6), this contrasts with the impact on ani- rather than fine sediment and heavier rather
mals which occurs within hours of the oil first settling after tide levels than light oil. Access by the necessary heavy
first drop. During the following days, the habitat can smell of equipment is a major limitation.
decomposing fauna, like those in Port Curtis after experimental oiling Flooding and flushing using Designed to remove oil from sediment
seawater surface and from mangrove parts; most
(Duke and Burns, 1999). The animals chiefly comprised small fish, shell-
effective for light and medium oils, on a
fish and crustaceans, including sesarmid crabs and pistol shrimps. How- receding tide when oil is moving out of the
ever, unlike the process for replacing trees taking decades, the crabs and mangrove and if sediment will not be
other animals reported dead in the trials (Duke et al., 2000), had re- disturbed. This type of intervention shows
established in plots after 1–2 months. However, it was notable that fau- great promise since this method concurs with
observed more rapid oil degradation rates in
nal activity levels remained relatively low compared to those in un- areas of higher tidal ranges (see Burns et al.,
oiled plots after 2 years. 2000; Duke et al., 2000; Fig. 7).
Longer term funding support is essential also for research and mon- Cleaners & dispersants To assist tidal flow or flushing treatments to
itoring of animals affected by oil spills. Animals are essential for the detach and dissolve oil from sediment and
mangrove parts; however, some cleaners
health and well-being of mangrove vegetation, as shown in the complex
may be toxic.
trophic and functional relationships observed (cs., McIvor and Smith, Bioremediation Promoting microbial breakdown of oil
1995). through nutrient addition and aeration;
noted that this approach is yet to be fully
evaluated and nutrient addition can
negatively affect mangrove health. But see
Recommendation 9: Once oil enters a mangrove tidal wetland, Duke et al. (2000) for a successful trial of
very serious consideration must be made for ‘no action’ where just bioremediation in Central Queensland and
about any intervention will result in further habitat damage, like Santos et al. (2011) for a further
trampling of sediments and seedlings, cutting away dead wood, comprehensive account.
Removal of oiled debris Appropriate when no further oil is likely to
sweeping up dead vegetation, planting seedlings, burning off oil, reach the shore (debris acts as a barrier),
scraping up oiled surface sediments, and applying dispersants or physical damage to mangroves can be
other industrial chemicals. Cleanup access and procedures must minimised and the activity does not result in
only be applied where these intervention actions are proven not significant habitat alteration (also see Duke et
al., 1999).
to cause harm to recovery processes.
Floating loose sorbents This includes dust or fibre sorbents, including
persistent polymers (i.e. plastics banned in
Australia and other countries, for their
persistence and negative influences), floating
Overall, the treatment of oil offshore is regarded as the preferred op- organic materials of little use (e.g. wool, hair
tion over the treatment of oil deposited within intertidal habitat. This coconut fibre, straw), or inorganic materials
applies especially to mangroves where clean up is extremely difficult (e.g. zeolites, cat litter, plasma silicates).
These may well sink and they are generally
since stands are often dense, remote and inaccessible. And, that the
not well researched. However, new products
clean up intervention is most likely damaging in itself; for example, and processes are imminent. For example, see
trampling, root breakage and other forms of physical impacts. Options the AMSA (2016) Oil Spill Control Agents
for treating oil offshore include mechanical containment and skimmer Register list. The product called OilZorb, as an
recovery, agitation and mixing, dispersant use, and burning. However, example with potential for better response
applications in and about mangrove areas.
chemical additives such as dispersants may have an impact on other
714 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
greater tidal flushing with reported correlative benefits for more rapid
longer term recovery (Burns et al., 2000).
If there are any doubts about making an intervention, then such un-
certainty must emphasis the need for more research and testing. It fol-
lows that in making a decision to intervene, then that action should be
rigorously reviewed, monitored and tested to properly evaluate its
value and future merit in subsequent responses – compared to doing
nothing. During a spill, is arguably the best time for making selective tri-
als and evaluations since there is both opportunity with experimental
conditions, and public support. Justification is readily made in the inter-
ests of improving future responses, and in learning from past experi-
ences and mistakes – all of which needs to be publically documented
and published.
Acknowledgements
I acknowledge assistance from the sub editor and an anonymous
reviewer.
Appendix A. Supplementary data
Supplementary data to this article can be found online at http://dx.
doi.org/10.1016/j.marpolbul.2016.06.082.
References
Adeyemo, O.K., Ubiogoro, O.E., Adedeji, O.B., 2009. Oil exploitation, fisheries resources and
sustainable livelihood in the Niger delta, Nigeria. Nature & Fauna, FAO Regional Office
for Africa Vol. 24, pp. 56–61.
Aghalina, S.A., Eyinla, E., 2009. Oil exploration and marine pollution: evidence from the
Niger Delta, Nigeria. J. Hum. Ecol. 28 (3), 177–182.
AMSA, 2016. Australian Maritime Safety Authority. Major historical oil spill incidents in
Australian waters (https://www.amsa.gov.au/environment/major-historical-
incidents/ Accessed March 2016).
BBC, 2015. Shell agrees $84 m deal over Niger Delta oil spill. (Posted: 7 January 2015)
http://www.bbc.com/news/world-30699787.
Böer, B., 1993. Anomalous pneumatophores and adventitious roots of Avicennia marina
(Forssk.) Vierh. Mangroves two years after the 1991 Gulf War oil spill in Saudi Arabia.
Mar. Pollut. Bull. 27, 207–211.
Burns, K.A., Codi, S., Duke, N.C., 2000. Gladstone, Australia field studies: weathering and
degradation of hydrocarbons in oiled mangrove and salt marsh sediments with and
without the application of an experimental bioremediation protocol. Mar. Pollut.
Bull. 41, 392–402.
Burns, K.A., Garrity, S.D., Levings, S.C., 1993. How many years until mangrove ecosystems
recover from catastrophic oil spills? Mar. Pollut. Bull. 26, 239–248.
Cannicci, S., Burrows, D., Fratini, S., Smith III, T.J., Offenberg, J., Dahdouh-Guebas, F., 2008.
Faunal impact on vegetation structure and ecosystem function in mangrove forests: a
review. Aquat. Bot. 89, 186–200.
CEDRE, 2016. Centre of Documentation, Research and Experimentation on Accidental
Water Pollution, Database of spill incidents and threats in waters around the world.
http://wwz.cedre.fr/en/Our-resources (Accessed March 2016).
Culbertson, J.B., Valiela, I., Peacock, E.E., Reddy, C.M., Carter, A., VanderKruik, R., 2007.
Long-term biological effects of petroleum residues on fiddler crabs in salt marshes.
Mar. Pollut. Bull. 54, 955–962.
Da Silva, E.M., Peso-Aguiar, M.C., Navarro, M.D.F.T., Chastinet, C.d.B.E.A., 1997. Impact of
petroleum pollution on aquatic coastal ecosystems in Brazil. Environ. Toxicol.
Chem. 16 (1), 112–118.
Duke, N.C., 2001. Gap creation and regenerative processes driving diversity and structure
of mangrove ecosystems. Wetl. Ecol. Manag. 9, 267–279.
Duke, N.C., 2011. Mangroves. In: Hopley, D. (Ed.), Encyclopedia of Modern Coral Reefs.
Structure, Form and Process. Springer, Dordrecht, The Netherlands, pp. 655–663.
Duke, N.C., 2014a. Mangrove Coast. In: Harff, J., Meschede, M., Petersen, S., Thiede, J.
(Eds.), Encyclopedia of Marine Geosciences. Springer, Dordrecht, Netherlands,
pp. 1–17.
Duke, N.C., 2014b. World Mangrove iD: expert information at your fingertips. Smart
Phone App, Apple & Android versions. MangroveWatch Ltd and TropWATER James
Cook University, Townsville, Australia.
Duke, N.C., Burns, K.A., 1999. Fate and effects of oil and dispersed oil on mangrove ecosys-
tems in Australia. Final Report to the Australian Petroleum Production Exploration
Association. Australian Institute of Marine Science and CRC Reef Research Centre.
Duke, N.C., Burns, K.A., 2003. Fate and effects of oil and dispersed oil on mangrove ecosys-
tems in Australia. Environmental Implications of Offshore Oil and Gas Development
in Australia: Further Research. A Compilation of Three Scientific Marine Studies. Aus-
tralian Petroleum Production and Exploration Association (APPEA), Canberra,
pp. 232–363 (521 pp.).
Duke, N.C., Mackenzie, J., 2015. Annual Report: Port Curtis and Port Alma Coastal Habitat
Archive and Monitoring Program (PCPA CHAMP). Report Produced for the Ecosystem
Research and Monitoring Program Advisory Panel as Part of Gladstone Ports
Corporation's Ecosystem Research and Monitoring ProgramCentre for Tropical
Water and Aquatic Ecosystem Research (TropWATER) Publication 15/37. James
Cook University, Townsville (50 pp.).
Duke, N.C., Schmitt, K., 2015. Mangroves: Unusual Forests at the Seas Edge. In: Pancel, L.,
Köhl, M. (Eds.), Tropical Forest Handbook. Springer-Verlag, Berlin, Heidelberg.
Duke, N.C., Watkinson, A.J., 2002. Chlorophyll-deficient propagules of Avicennia marina
and apparent longer term deterioration of mangrove fitness in oil-polluted sedi-
ments. Mar. Pollut. Bull. 44, 1269–1276.
Duke, N.C., Ball, M.C., Ellison, J.C., 1998a. Factors influencing biodiversity and distribution-
al gradients in mangroves. Glob. Ecol. Biogeogr. Lett. Mangrove Spec. Issue 7, 27–47.
Duke, N.C., Bell, A.M., Pedersen, D.K., Roelfsema, C.M., Bengtson Nash, S., 2005. Herbicides
implicated as the cause of severe mangrove dieback in the Mackay region, NE Austra-
lia: consequences for marine plant habitats of the GBR World Heritage Area. Mar.
Pollut. Bull. 51, 308–324.
Duke, N.C., Burns, K.A., Dalhaus, O., 1998b. Effects of oils and dispersed-oils on mangrove
seedlings in planthouse experiments: a preliminary assessment of results two
months after oil treatments. In: Beck, C. (Ed.)APPEA Journal. APPEA, Canberra,
pp. 631–636.
Duke, N.C., Burns, K.A., Swannell, R.P.J., 1999. Research into the bioremediation of oil spills
in tropical Australia: with particular emphasis on oiled mangrove and salt marsh hab-
itat. Report to the Australian Maritime Safety Authority and the Great Barrier Reef
Marine Park Authority.
Duke, N.C., Burns, K.A., Swannell, R.P.J., Dalhaus, O., Rupp, R.J., 2000. Dispersant use and a
bioremediation strategy as alternate means of reducing impacts of large oil spills on
mangroves: the Gladstone field trials. Mar. Pollut. Bull. 41, 403–412.
Duke, N.C., Meynecke, J.O., Dittmann, S., Ellison, A.M., Anger, K., Berger, U., Cannicci, S.,
Diele, K., Ewel, K.C., Field, C.D., Koedam, N., Lee, S.Y., Marchand, C., Nordhaus, I.,
Dahdouh-Guebas, F., 2007. A world without mangroves. Science 317, 41–42.
Duke, N.C., Pinzón, Z.S., Prada, M.C., 1997. Large-scale damage to mangrove forests follow-
ing two large oil spills in Panama1. Biotropica 29, 2–14.
Duke, N.C., Wood, A., Hunnam, K., Mackenzie, J., Haller, A., Christiansen, N., Zahmel, K.N.,
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 Monitor-
ing Plan (244 pp.) Report to PTTEP Australasia by Uniquest University of Queensland,
School of Biological Sciences, Brisbane (See the ShoreView prototype: http://203.101.
224.239/ see the report: http://www.environment.gov.au/coasts/oilspill/
publications/pubs/shoreline-ecological-assessment.pdf).
Ekekwe, E., 1983. The Funiwa-5 oil well blowout. The Petroleum Industry and the Nigeri-
an Environment, Proceedings of an International Seminar, pp. 64–68.
El-Nemr, A., 2006. Petroleum Contamination in Warm and Cold Marine Environments.
Novinka Books, New York (140 pp.).
Eoearth, 2010. The encyclopedia of earth. Updated Dec 2010 http://www.eoearth.org/
view/article/159877/ (Accessed in March 2016).
FAO, 2010. Food and Agriculture Organization of the United Nations. Global Forest Re-
sources Assessment: Main Report Friends of Mangroves (28 November 2011. Report
online. http://www.fao.org/docrep/013/i1757e/i1757e.pdf. Accessed March 2016).
Gedan, K.B., Silliman, B.R., 2009. Using facilitation theory to enhance mangrove restora-
tion. Ambio 38, 109.
Hensel, P., Proffitt, E.C., Delgado, P., Shigenaka, G., Yender, R., Hoff, R., Mearns, A.J., 2010.
In: Hoff, R. (Ed.), Oil Spills in Mangroves Planning and Response Considerations.
Hooke, N., 1997. Maritime Casualties, 1963–1996. second ed. LLP Limited, London.
Huffington Post, 2010. Niger Delta ExxonMobil Spill, Nigeria - May 2010. http://www.
huffingtonpost.com.au/entry/porter-ranch-gas-leak-worry_us_
569fdcdae4b0fca5ba76528f?section=australia&slideshow=true#gallery/
5600991ce4b0fde8b0cf879e/0 (Accessed March 2016. Pipeline spill, Nigeria).
Human Rights Watch, 2016. The Impact of Oil Operations on the Environment. Oil Spills
and Hydrocarbon Pollution. https://www.hrw.org/reports/1999/nigeria/Nigew991-
05.htm (Accessed April 2016).
ITOPF, 2016. Case studies. Database of major oil spills from 1967 to the present day.
http://www.itopf.com/knowledge-resources/data-statistics/gis/gis-case-studies/
(Accessed March 2016).
Kadafa, A.A., 2012a. Environmental Impacts of Oil Exploration and Exploitation in the
Niger Delta of Nigeria. Glob. J. Sci. Front. Res. Environ. Earth Sci. 12 (3) (on line).
Kadafa, A.A., 2012b. Oil Exploration and Spillage in the Niger Delta of Nigeria. Civ. Environ.
Res. 2 (3), 38–51.
Khanna, S., Santos, M.J., Ustin, S.L., Koltunov, A., Kokaly, R.F., Roberts, D.A., 2013. Detection
of Salt Marsh Vegetation Stress and Recovery after the Deepwater Horizon Oil Spill in
Barataria Bay, Gulf of Mexico Using AVIRIS Data. PLoS One 8 (11), e78989.
Lamparelli, C.C., Rodeigues, F.O., Orgler de Moura, D., 1997. Long term assessment of an oil
spill in a mangrove forest in Sao Paulo, Brazil. In: Kjerfve, B., Drude de Lacerda, L., Salif
Diop, W.H. (Eds.), Mangrove Ecosystem Studies in Latin America and Africa. UNESCO,
Paris, France, pp. 191–203.
Lewis, M., Pryor, R., Wilking, L., 2011. Fate and effects of anthropogenic chemicals in man-
grove ecosystems: a review. Environ. Pollut. 159, 2328–2346.
Lewis III, R.R., 1982. Mangrove forests. In: Lewis, R.R. (Ed.), Creation and Restoration of
Coastal Plant Communities. CRC Press, Florida, pp. 153–171.
Lewis III, R.R., 2005. Ecological engineering for successful management and restoration of
mangrove forests. Ecol. Eng. 24, 403–418.
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., Triet, T.,
2015. The vulnerability of Indo-Pacific mangrove forests to sea-level rise. Nature
526, 559–563.
Mackenzie, J.R., Duke, N.C., Wood, A.L., 2016. The Shoreline Video Assessment Meth-
od (S-VAM): using dynamic hyperlapse image acquisition to evaluate shoreline
mangrove forest structure, values, degradation and threats. Mar. Pollut. Bull.
http://dx.doi.org/10.1016/j.marpolbul.2016.05.069.
 N.C. Duke / Marine Pollution Bulletin 109 (2016) 700–715
McIvor, C.C., Smith III, T.J., 1995. Difference in crab fauna of mangrove areas at a south-
western Florida and a northeastern Australian location: implications for leaf litter
processing. Estuaries 18, 591–597.
Meynecke, J.-O., Lee, S.Y., Duke, N.C., 2008. Linking spatial metrics and fish catch reveals
the importance of coastal wetland connectivity to inshore fisheries in Queensland,
Australia. Biol. Conserv. 141, 981–996.
Michel, J., Rutherford, N., 2014. Impacts, recovery rates, and treatment options for spilled
oil in marshes. Mar. Pollut. Bull. 82, 19–25.
Mukherjee, N., Sutherland, W.J., Dicks, L., Hugé, J., Koedam, N., Dahdouh-Guebas, F., 2014.
Ecosystem service valuations of mangrove ecosystems to inform decision making and
future valuation exercises. PLoS One 9, 1–9.
Nigerian Oil Spill Monitor, 2016. Incident Archive. Oil spill Data Provided by the National
Oil Spill Detection and Response Agency of Nigeria. https://oilspillmonitor.ng
(Accessed May–June 2016).
Nixon, Z., Zengel, S., Baker, M., Steinhoff, M., Fricano, G., 2016. Shoreline oiling from the
Deepwater Horizon oil spill. Mar. Pollut. Bull. 107, 170–178.
NOAA, 2014. Oil Spills in Mangroves. Planning & Response Considerations. US Depart-
ment of Commerce, National Oceanic and Atmospheric Administration (NOAA), Na-
tional Ocean Service, Office of Response and Restoration, Seattle, Washington (96
pp.).
NOAA, 2016. Incident Archive. The Complete List of all Publicly Released Information on
Any Oil Spill Incident. US Department of Commerce, National Oceanic and Atmo-
spheric Administrationhttp://www.incidentnews.noaa.gov/browse/date?page=1
(Accessed March–April 2016).
Nwilo, C.P., Badejo, T.O., 2005. Oil spill problems and management in the Niger Delta. In-
ternational Oil Spill Conference, Miami, Florida, USA.
Odeyemi, O., Ogunseitan, O.A., 1985. Petroleum industry and its pollution potential in Ni-
geria. Oil Petrochem. Pollut. 2 (3), 223–229.
Oil Spill Intelligence Report, 2016. Spill Contaminates Coastal Communities in Nigeria. 8
Mar 2016 http://search.proquest.com.elibrary.jcu.edu.au/docview/1774548853/
E48DEB1BCD6E4D1DPQ/1?accountid=16285 (Accessed March 2016).
Proffitt, C.E. (Ed.), 1997. Managing Oil Spills In Mangrove Ecosystems: Effects, Remedia-
tion, Restoration, and Modeling. OCS Study MMS 97-0003. U.S. Department of the In-
terior, Minerals Management Service, Gulf of Mexico OCS Region, New Orleans (76
pp.).
Richards, D.R., Friess, D.A., 2016. Rates and drivers of mangrove deforestation in Southeast
Asia, 2000–2012. PNAS 113 (2), 344–349.
Salmo III, S.G., Duke, N.C., 2010. Establishing mollusk colonization and assemblage pat-
terns in planted mangrove stands of different ages in Lingayen Gulf, Philippines.
Wetl. Ecol. Manag. 18, 745–754.
Santos, H., Carmo, F., Paes, J.S., Rosado, A., Peixoto, R., 2011. Bioremediation of mangroves
impacted by petroleum. Water Air Soil Pollut. 216, 329–350.
Schmitt, K., Duke, N.C., 2015. Mangrove management, assessment and monitoring. In:
Pancel, L., Köhl, M. (Eds.), Tropical Forest Handbook. Springer-Verlag, Berlin Heidel-
berg (30 pp.).
Sciencythoughts, 2013. Oil Spill in Bayelsa State, Nigeria. Shell well 62, Nigeria. 15 April
2013 http://sciencythoughts.blogspot.com.au/2013/04/oil-spill-in-bayelsa-state-
nigeria.html (Accessed March 2016).
715
Seenreport, 2008. Oil Spill in Angola. 11 April 2008. Takula Drill Rig, Angola http://
seenreport.com/2008-04-11-oil-spill-in-angola/ (Accessed March 2016).
Shell, 2015. Bodo oil spills and litigation. Posted: 7 January 2015 http://www.shell.com.
ng/environment-society/our-response/bodo-spills.html.
Snowden, R.J., Ekweozor, I.K.E., 1987. The impact of a minor oil spillage in the estuarine
Niger delta. Mar. Pollut. Bull. 18, 595–599.
Spalding, M., Kainuma, M., Collins, L., 2010. World Atlas of Mangroves. Earthscan, London.
Swan, J.M., Neff, J.M., Young, P.C., 1994. Energy Research and Development Corporation
(Australia), Australian Petroleum Production Exploration Association. Environmental
Implications of Offshore Oil and Gas Development in Australia: the Findings of an In-
dependent Scientific Review. Sydney.
The Guardian, 2010. Nigeria's Agony Dwarfs the Gulf Oil Spill. Posted 30 May 2010 http://
www.theguardian.com/world/2010/may/30/oil-spills-nigeria-niger-delta-shell
(Accessed March 2016. Pipeline spill, Nigeria).
The Guardian, 2011. Shell Accepts Liability for Two Oil Spills in Nigeria. 3 Aug 2011
http://www.theguardian.com/environment/2011/aug/03/shell-liability-oil-spills-
nigeria.
The Guardian, 2015. Shell Announces £55 m Payout for Nigeria Oil Spills. Posted: 7 January
2015 http://www.theguardian.com/environment/2015/jan/07/shell-announces-55m-
payout-for-nigeria-oil-spills.
The Maritime Executive, 2015. Shell Reaches Nigerian Oil Spill Agreement. Posted: 4 April
2015 http://www.maritime-executive.com/article/shell-reaches-nigerian-oil-spill-
agreement.
Thorhaug, A., 1992. Oil spills in the tropics and subtropics. In: Connell, D.W., Hawker, D.W.
(Eds.), Pollution in Tropical Aquatic Systems. CRC Press Inc., Boca Raton, Florida,
pp. 99–128.
Tolulope, A.O., 2004. Oil exploration and environmental degradation: the Nigerian expe-
rience. International Information Archives, International Society for Environmental
Information Science. EIA04–039 Vol. 2, pp. 387–393.
Tomlinson, P.B., 1994. The Botany of Mangroves. Cambridge University Press, Cambridge,
U.K.
Twumasi, Y.A., Merem, E.C., 2006. GIS and Remote Sensing Applications in the Assessment
of Change within a Coastal Environment in the Niger Delta Region of Nigeria. Int.
J. Environ. Res. Public Health 3 (1), 98–106.
Ukoli, M.K., 2005. Environmental factors in the management of the oil and gas industry in
Nigeria. www.cenbank.org.
UNEP, 2011. Environmental Assessment of Ogoniland. Job No.: DEP/1337/GE. United Na-
tions Environment Programme, Nairobi, Kenya. 262 pages www.unep.org/nigeria.
Walters, B.B., 2005. Ecological effects of small-scale cutting of Philippine mangrove for-
ests. For. Ecol. Manag. 206, 331–348.
Webecoist, 2016. Sick Slicks: Counting Down The 13 Worst Oil Spills. Web Ecoist,
Momtastic http://webecoist.momtastic.com/2010/05/04/sick-slicks-counting-down-
the-13-worst-oil-spills/ (Accessed March 2016).
Wikimedia Commons, 2016. World Shipping Routes. sourced March 2016 https://
commons.wikimedia.org/wiki/File:Shipping_routes_red_black.png.
Wikipedia, 2014. Petroleum Industry in Nigeria. Environmental Impact. Oil spills and
water contamination, 2014.