Marine Policy 46 (2014) 113

Contents lists available at ScienceDirect

Marine Policy
journal homepage: www.elsevier.com/locate/marpol

Stakeholder perceptions of ecosystem service declines in Milne Bay,

Papua New Guinea: Is human population a more critical driver than
climate change?
J.R.A. Butler a,n, T. Skewes b, D. Mitchell c, M. Pontio c, T. Hills d
a

CSIRO Ecosystem Sciences and Climate Adaptation Flagship, EcoSciences Precinct, GPO Box 2583, Brisbane, QLD 4001, Australia
CSIRO Marine and Atmospheric Research and Climate Adaptation Flagship, EcoSciences Precinct, GPO Box 2583, Brisbane, QLD 4001, Australia
c
Asia-Pacic Field Division, Pacic Island Program, Conservation International, 211 Alotau, Papua New Guinea
d
Global Strategies Division, Conservation International, 2/204 Kent St. New Farm, QLD 4005, Australia
b

art ic l e i nf o

a b s t r a c t

Article history:
Received 27 July 2013
Received in revised form
8 December 2013
Accepted 20 December 2013
Available online 21 January 2014

Milne Bay Province (MBP) in Papua New Guinea is a priority seascape in the Coral Triangle marine
biodiversity hotspot. Goal 4 of the Coral Triangle Initiative on Coral Reefs, Fisheries and Food Security
promotes adaptation planning for small island ecosystems and communities threatened by climate
change, but information to identify vulnerable islands and priority interventions is limited. This study
adapted the Millennium Ecosystem Assessment (2005) framework in MBP with regional stakeholders to
project trends in harvested or cultivated provisioning ecosystem goods and services (EGS), human wellbeing, drivers of change and necessary management strategies, based on their tacit knowledge. In 2010
ve island subregions which are susceptible to food insecurity were assessed. Workshop participants
identied freshwater, garden food crops, coral, bche-de-mer, reef sh and sharks as the most important
EGS in all subregions. Terrestrial EGS contributed 43% of aggregated ecosystem-derived well-being, and
marine EGS 57%. By 2030 the overall condition of EGS was projected to decline by 450%. The primary
driver in all subregions was human population growth, and climate change impacts were predicted in
only two subregions. Improved garden and agricultural productivity and population control were the
highest ranked management strategies. Population relocation was also prioritised for two subregions
where human carrying capacities may soon be exceeded. Although none of the strategies addressed
climate change directly, all could yield climate adaptation and marine conservation co-benets by
enhancing ecosystem-based adaptation and community adaptive capacity. It is suggested that there is a
2030 year adaptation window in which to address population growth, which otherwise will continue
to erode the capacity of communities and ecosystems to cope with potentially extreme climate impacts
after mid-century.
& 2014 Elsevier Ltd. All rights reserved.

Keywords:
Livelihoods
Human population growth
Food security
Coral Triangle
Small islands
Climate adaptation

1. Introduction
Change, including climate change, is occurring at unprecedented rates, resulting in compounding impacts on ecosystems
[1]. Rural communities in less developed regions are particularly
sensitive because their livelihoods and well-being are highly
dependent upon local ecosystem goods and services (EGS) for
their subsistence and cash needs [2,3], exacerbated by their
limited adaptive capacity [4,5]. Adequate scientic data on the
status and trends of ecosystems and the EGS that they provide,
plus the nature of stressors on these EGS, are typically scant in
such regions, accentuating their vulnerability [2].

Corresponding author. Tel.: 61 7 47814205; fax: 61 7 47251570.

E-mail address: james.butler@csiro.au (J.R.A. Butler).

0308-597X/$ - see front matter & 2014 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.marpol.2013.12.011

In these situations stakeholders0 tacit knowledge, garnered

through rapid participatory assessments, may be an important
source of information with which to formulate policy and management interventions [3,6]. Engaging local stakeholders to characterise socialecological systems and design management strategies
can also foster adaptive capacity by enhancing social learning, the
co-production of knowledge, mitigating power asymmetries and
creating ownership of problems and solutions [2,711]. Conversely, the utility of local information may be constrained by
stakeholders0 world-views which are based on local experience,
and a lack of awareness of inuences from higher scales [12]. Also,
the integration of complex scientic information into participatory
processes, such as climate change modelling and its inherent
uncertainties, can be problematic. Unless information is contextualised in local terms [13] and stakeholders are sufciently
engaged in the production of knowledge it may lack salience,

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Fig. 1. Milne Bay Province in Papua New Guinea (PNG) and its major island groups and coral reefs.

credibility and legitimacy [14], resulting in an underestimation of

potential future impacts [10]. Furthermore, external scientists can
exercise considerable power in participatory processes, potentially
inuencing the results generated [11].
The Coral Triangle is one such region. It consists of the developing nations of Malaysia, Indonesia, the Philippines, Timor Leste, the
Solomon Islands and Papua New Guinea (PNG), and is a global
hotspot for marine biodiversity, containing over 75% of the world0 s
known hard coral species [15], over 3000 species of reef sh [16,17],
and 31% of the world0 s mangroves [18]. Approximately 120 million
people benet directly from marine-derived EGS [19], and many
inhabit the region0 s numerous remote, small islands [20,21]. In
response to escalating anthropogenic degradation of coastal and
marine habitats, exacerbated by climate change, and the imperative
to simultaneously protect biodiversity and livelihoods, the Coral
Triangle Initiative on Coral Reefs, Fisheries and Food Security (CTI)
was established in 2009. A partnership between the six Coral
Triangle countries, the CTI is funded by the Asian Development
Bank, USA and Australian Governments, the Global Environment
Fund and non-government organisations (NGOs) including Conservation International (CI). The CTI aims to achieve biodiversity
conservation, sustainable sheries and food security through the
implementation of marine protected areas (MPAs) and ecosystembased approaches to sheries management.
An additional focus of the CTI is climate change adaptation. Under
both worst case and best case scenarios of global carbon emissions,
climate projections suggest that until mid-century the region is
expected to experience increased sea surface temperatures, ocean
acidication, sea level rise and more variable rainfall and intense
tropical storms. This is likely to impact coral reefs through bleaching
and impeded calcication processes, and hence declines in the
productivity of reef sheries. Sea level rise is expected to impact
mangrove and sea grass habitat, and extreme rainfall and storms will
cause erosion and run-off from terrestrial catchments, further degrading inshore seagrass and coral habitats. After mid-century projections
diverge, and climate impacts may stabilise or escalate considerably
[22]. Small island ecosystems will be particularly exposed to sea level
rise [19] and increasingly variable rainfall [23].
Consequently Goal 4 of the CTI0 s Regional Plan of Action [19],
and the subsequent Region-wide Early Action Plan for Climate

Change Adaptation [24] is to plan adaptation for near-shore and

small island ecosystems. However, there is a lack of local-scale
scientic information and methodologies with which to prioritise
vulnerable communities and habitats, and to develop appropriate
adaptation strategies [19]. In addition, ne-scale analyses of
livelihood and food security issues are required to account for
the heterogeneity of social contexts typical of the CTI region [25].
Also, while there has been a recent focus on adaptation for marine
sheries [e.g. [2628]] and coastal ecosystems [e.g. [29]], there has
been little consideration of the relative importance of terrestrial
EGS to small island communities, or linkages to marine resource
use and conservation. Consequently, many adaptation studies have
been shery-focussed and overlook broader human development
and livelihood issues which may be of greater relevance for the
improvement of community well-being [30,31].
This paper presents the results of a pilot participatory exercise
designed to address these challenges in Milne Bay Province, PNG,
which is a priority seascape in the CTI and contains many small
islands of differing geomorphology, ecological and livelihood
characteristics. Our results show that regional stakeholders predict
acute declines in EGS and human well-being, and that human
population pressure is a more pre-eminent stressor than climate
change. Furthermore, some islands may be approaching carrying
capacity under current livelihood and technological systems,
requiring the relocation of inhabitants. Management strategies
identied to address population growth and its effects are likely to
generate co-benets for future climate adaptation by building
community and ecosystem adaptive capacity. We discuss the
relevance of these results for future adaptation planning in Milne
Bay, PNG and the CTI, and discuss shortcomings and potential
improvements for the method.

2. Study site and methods

2.1. Study site
Milne Bay Province (MBP) is located at the south-eastern tip
of the PNG mainland (Fig. 1). MBP0 s total area is 256,804 km2, of
which 12,943 km2 is land, and 243,861 km2 is sea. There are

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Fig. 2. The 10 inhabited subregions of Milne Bay Province as dened by Skewes et al. [38].

Table 1
Areas and human population densities of the ve focal subregions in Milne Bay
Province. Human population data are sourced from the 2000 national census in
Skewes et al. [38].
Subregion Number
Total land
of islands area (km2)

Total
subregion
area (km2)

Population Land density

(people km  2)

Dawson
Woodlark
Misima
Trobriand
Louisiade

1027
16,631
11,088
19,738
15,084

2034
9120
14,849
57,344
11,243

9
13
14
62*
48

19
947
220
2003
1333

107.1
9.6
67.5
28.6
8.4

Includes parts of Goodenough, Fergusson and Normanby Islands (see Fig. 2).

approximately 210 islands in the province with continental, volcanic, atoll, raised limestone and coral cay geomorphologies and
related soil types [32]. It is a priority seascape in the Coral Triangle,
with 5355 km of reefs and shoals of less than 20 m deep [33], and
has one of the world0 s richest reef and shore sh faunas [34].
MBP0 s climate is monsoonal, with a pronounced dry (May
December) and wet (DecemberMay) season when tropical
cyclones also occur. The region is affected by the El Nio Southern
Oscillation, which can generate droughts (El Nio events) with an
average return rate of 10 years [35]. Severe El Nio events have
occurred on average every 30 years, with the most recent in 1997
1998, when there were widespread food and fresh water shortages
throughout PNG, inducing famine conditions [23]. In MBP, there
was no rainfall between March 1997 and January 1998, and island
communities were most acutely affected due to their naturally
limited water supplies and higher population densities relative to
the mainland. Communities maintained food security by substituting failed garden production with wild-harvested sh, shellsh,
fruit, coconuts, mammals and birds, receiving remittances from
mainland-based family or clan members to buy food, and the
provision of emergency food and water from the MBP, PNG and
Australian governments [36,37].
Based on variations in MBP0 s geomorphological, oceanographic
and climatic characteristics which determine underlying ecological
processes and the provision of EGS, Skewes et al. [38] identied 15

subregions, of which 10 are inhabited (Fig. 2). In this study we

focussed on the ve subregions (Table 1) which had inhabited islands
(rather than those with a combination of island and mainland
communities), because these were food insecure during the 1997
1998 drought. As well as having varied geomorphology and land areas,
population densities in 2000 ranged widely from 8.4 people km  2
(Louisiade) to 107.1 people km  2 (Dawson; Table 1).
The 2000 national census estimated MBP0 s human population
to be 196,044, of which 80% live on islands [32], with an average
growth rate of 2.7% annum  1 in 19902000 [39]. The Human
Development Index (HDI) for MBP in 2007 was 0.338, lower than
the national HDI of 0.44 [40]. In 2011 PNG0 s HDI was 0.466,
ranking it 153 out of 187 assessed nations [41], and the lowest
amongst the CTI partners.
Island livelihoods consist of a mixed subsistence and cash
economy, with almost total reliance on the consumption or sale
of primary produce [42,43]. The balance between these sectors is
largely determined by the spatial extent of the land and marine
resource base available to communities, and opportunities for
participation in the cash economy based on these resources
[39,44,45]. In general, subsistence food production is provided by
swidden-farmed gardening, which involves a rotation of forest
clearance, cultivation and fallowing before re-clearing. A variety of
produce is grown as subsistence or agricultural cash crops,
including yam, sweet potato, coconut, banana and taro, augmented by the raising of pigs and poultry and the hunting and
gathering of wild fruits, nuts, mammals and birds (wildlife) from
native forests. This is supplemented by the harvesting of a wide
range of marine resources. Destructive and indiscriminate shing
methods such as dynamite and traditional derris root are widely
used around reefs, and coral is harvested for building materials
and lime, and mangroves for timber [44].
The cash economy is primarily driven by commercial shing.
Sea cucumbers (bche-de-mer) have been the greatest source of
income for local communities. Catches are exported to the Asian
market, but stocks are overexploited [32,33,43], and in 2009 the
PNG National Fisheries Authority (NFA) enforced a national closure
of the shery. Other harvested invertebrates are trochus, clams,
lobsters and crabs, plus seaweed collection [32,46]. There is also an
expanding shery for shark ns which are exported to Asia, and

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

many shermen have transferred shing effort from bche-demer to sharks [32]. Other sources of income are nature-based
tourism, which is becoming established in MBP due to its globallyrecognised marine biodiversity [47], plus coconut and oil palm
plantation agriculture, gold mining and logging [44].

2.2. Stakeholder analysis

In 20092010 a research project was undertaken to develop an
adaptation planning process for MBP under the PNG CTI National
Plan of Action [48]. The project aimed to pilot a participatory, rapid
assessment through sequential engagement with stakeholders at
the regional and community level. The rst step was to consult
regional stakeholders to identify the relative vulnerability of
subregions, and to design priority adaptation strategies for each.
Subsequent community engagement in priority subregions was
then planned to validate the regional stakeholders0 perceptions
and recommendations.
In June 2010 a 2-day workshop was held in Alotau, the MBP
capital. Relevant regional stakeholder organisations were identied by the CI authors using Mitchell et al.0 s [49] criteria of power,
legitimacy and the urgency of their involvement. Twenty-two MBP
organisations were invited. A total of 26 people attended, representing the primary economic sectors (tourism, forestry, oil palm,
mining; n 7), MBP government divisions (forestry, health, sheries, mining, education, planning; n 10), media, NGOs and
church organisations (n 7) and other research organisations
(n 2).

2.3. Conceptual framework

The workshop process adapted the Millennium Ecosystem
Assessment (MA) [1] conceptual framework, which was selected
because of its proven effectiveness as a participatory tool capable
of engaging a range of stakeholders, enabling them to conceptualise social-ecological systems, and integrate scientic and local
knowledge [50]. Also, it facilitates a multi-scale systems view of
inuences on ecosystems and human well-being, and therefore
provided contextualisation of climate change relative to other
global, regional or local-scale issues.
The MA framework [1, p. 57] identies EGS as the benets
people obtain from functioning ecosystems. EGS are categorised
as provisioning (products obtained from ecosystems), cultural
(non-material benets), regulating (benets obtained from the
regulation of ecosystem processes) and supporting (services
necessary for the production of all other ecosystem services).
EGS contribute to human well-being, dened as basic material
for a good life, freedom of choice, health, good social relations, and
security, which is the antithesis of poverty. The condition and
trends of EGS are affected by drivers of change, dened (p. 87) as
any natural or human-induced factor that directly or indirectly
causes a change in an ecosystem, sub-divided into direct (e.g.
harvesting) and indirect (e.g. population growth, global economic
conditions). The framework also identies intervention options as
responses to drivers. Although it includes analysis of social
ecological system thresholds, the agency of decision-makers over
drivers, future scenarios and trade-offs in EGS, these elements
were not considered due to time constraints.
The adapted method focussed only on provisioning EGS, for
three reasons. First, a project scoping workshop held in April 2009,
where a prototype of the planning approach was tested with a
group of 10 regional stakeholders, indicated that there was some
confusion between the MA categories. Some provisioning EGS (e.g.
hunted wildlife) could also provide cultural EGS, and some

habitats (e.g. forest) could provide regulating EGS (e.g. carbon

sequestration) and provisioning EGS (e.g. harvested timber). This
potentially results in double-counting during EGS importance
ranking [1]. Second, the scoping workshop indicated that there
was insufcient time in a 2-day workshop to analyse all EGS
categories. Third, because the study focussed on food security and
livelihoods linked to the harvest or cultivation of natural resources,
provisioning EGS were the most relevant.

2.4. Workshop process

The workshop was facilitated by two CI authors (DM and MP)
in English and Tok Pisin, and divided into four sessions of 34 h
each. In Session 1 the above elements of the MA conceptual
framework were introduced, explained and discussed. In Session
2 potential drivers of change were considered. Representatives of
the major economic sectors and MBP government gave 510 min
oral presentations on the current and potential future trends in
each economic sector. The MBP government health division presented the 2000 census data, and used the 2.7% annum  1 growth
rate to illustrate that the MBP population could increase by 78%
to 350,000 by 2030, and double by 2040. To avoid biasing
participants0 perceptions of changes in EGS condition linked to
population pressure (see Section 2.5 below), population densities
(Table 1) and projections were not presented for each subregion.
The CSIRO authors presented climate change projections for
MBP and PNG, collated from contemporary modelling and downscaling (Table 2). Data presented were based on the A1B SRES
global emission scenario, which represents a mid- to high-range of
surface warming [57]. However, as for the CTI regional projections
[22], there are unlikely to be major differences between emissions
scenarios and their inuence on climate until after mid-century.
These data were then applied to a threat-asset interaction model
to illustrate potential impacts on EGS [38]. Results showed that
widespread impacts on marine ecosystems such as coral bleaching
are most likely to occur after mid-century, and sea level rise may
reach 10 cm by 2030 and 40 cm by 2100 (Table 2). Although
annual rainfall may increase, it will become more erratic, and El
Nio events may become more frequent, with deleterious impacts
on freshwater hydrology and supplies, garden and agricultural
production. By 2100 more intense tropical storms will elevate
destructive effects on reefs and mangroves.
In Session 3 stakeholders self-selected into a group assigned to
each island subregion (Table 1), based on their familiarity with
that subregion. This resulted in ve groups of four to six people.

Table 2
Climate change parameters and projected changes from 1990 levels, derived from
the A1B SRES global emission scenario. These were presented at the workshop and
applied in the threat-asset interaction model to illustrate potential impacts on
ecosystem goods and services.
Parameter

2030

2100

Mean annual sea surface temperature (1C)a,b,c

Mean annual ambient air temperature (1C)a,b
Sea level rise (cm)d
Mean annual rainfall (% change)c,e,f
Ocean acidication (pH)b

1.0
0.51.0
10
10
 0.1

3.0
1.52.0
40
30
 0.3

[51].
[52].
c
[53].
d
[54].
e
[55].
f
[56].
b

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Each group was then given an A0 map of the subregion including

basic biophysical (islands, reefs, rivers) and socio-economic (settlements, schools) features, ip chart paper and pens, and they
tabulated answers to these questions for the subregion:

Table 3
Provisioning ecosystem goods and services (EGS) and their contribution to human
well-being scored from 1 (low) to 5 (high). Abbreviations for subregions are
DDawson; W Woodlark; M Misima; T Trobriand; L Louisiade.
EGS

1. Which provisioning EGS contribute to human well-being, and

what is the relative importance of each, scored from 1 (low) to
5 (high)?
2. What is the predicted change in condition of each EGS by 2030
from its current state, scored from  5 (absent) to 5 (greatly
improved)?
3. What are the direct and indirect drivers responsible for the
predicted condition of each EGS in 2030?
4. What management strategies are required to address the
drivers for each EGS?
The projected year of 2030 was selected by participants
because it is in the foreseeable future and therefore reduced the
degree of uncertainty in the trends of drivers and EGS condition,
but allowed a sufciently large temporal window for the consideration of strategic management responses. The scoring scale of
15 was selected by the authors to minimise potential discrepancies in valuations between the groups.
Session 3 concluded with each group presenting their results,
and any points requiring clarication or debate were discussed,
and scores were adjusted if necessary. Over lunch the authors
entered the nal data into pre-prepared computer spreadsheets.

(a) Terrestrial
Freshwater
Garden food crops
Forest timber
Agricultural crops
Wildlife
Fruit and nuts
Minerals (land)
Sub-total

5
5

5
5
5

5
5
5
5
5
5
5
35

5
5
5
5
3
5

5
5
5

28

19

5
5
5
5
5
5
4

5
5
5
5
4
5
3

5
5
5
5

36
71
49
51

34
62
45
55

5
2

15

17

5
5
5
5
3

5
5
5
5
4

(b) Marine
Bche-de-mer
Coral
Reef sh
Sharks
Turtles
Mangrove timber
Seaweed
Crabs
Shellsh (currency)
Dugong
Minerals (seabed)
Sub-total
Total
Terrestrial EGS contribution (%)
Marine EGS contribution (%)

Total

5
5
2
23
38
39
61

31
48
35
65

25
44
43
57

25
25
20
15
14
10
5
114
25
25
25
25
16
10
7
5
5
4
2
149
263
43
57

2.5. Data analysis

In Session 4 the authors presented the collated data back to the
participants for further discussion, in the following sequence.
Results of the statistical analyses were considered signicant
where po 0.05:
1. The contribution of different EGS to human well-being in each
subregion and across all subregions. A 2 contingency table was
applied to compare amongst subregions the relative contribution to aggregated well-being of terrestrial and marine EGS.
2. The predicted change in EGS0 condition, and average changes
for all EGS in each subregion and across all subregions. To
assess the relative changes for terrestrial versus marine EGS,
changes for each category were averaged across all subregions
and compared statistically using a t-test.
3. The occurrence of direct and indirect drivers of change linked
to changes in EGS condition for each subregion and across all
subregions. The relative occurrence of drivers was compared
using 2 contingency tables.
4. The human well-being impact (HWI) of predicted changes for
each EGS, calculated from the formula:

2.6. Subregion synthesis

To synthesise this information for each subregion, the DriverPressure-Impact-State-Response (DPSIR) framework [58] was
applied after the workshop. Drivers describe large scale socioeconomic or ecosystem conditions and trends, pressures refer to
patterns of ecological system alterations, state includes observable changes in system dynamics, impacts are measurable changes
in social costs and/or benets linked to a resource, and responses
are policy or management actions required to address changes in
the system [59]. Although the MA [1] framework also identies
strategies and intervention options, the DPSIR approach more
explicitly depicts social responses as a part of the system. It was
considered that the MA0 s indirect drivers are equivalent to
drivers, direct drivers equivalent to pressures, the HWI combines
the state and impact components, and the management strategies are equivalent to responses. To standardise the presentation
of the issues in each component they were listed in descending
rank order.

HWIEGS1 REGS1  C EGS1

where REGS1 is the contribution to well-being of EGS1, and CEGS1
is the predicted change in condition of EGS1 by 2030. The total
HWI for each EGS across all subregions, and the mean HWI for
each subregion was also presented.
5. The correlations between each subregion0 s predicted changes
in EGS condition, HWI and population density in 2000, using a
Pearson0 s correlation coefcient (r).
6. The importance of each management strategy (MSI) in each
subregion, calculated from the formula:
MSI1 HWIEGS1 HWIEGS2 HWIEGSn
where a strategy (MS1) was identied to address the drivers
impacting on EGS1, EGS2 and EGSn. MSI1 was calculated by
summing the HWI for EGS1, EGS2 and EGSn.

3. Results
3.1. EGS and human well-being
A total of 18 EGS were identied, comprising seven terrestrial
and 11 marine (Table 3). Freshwater, garden food crops, bche-demer, harvested coral, reef sh and sharks occurred in all subregions and contributed highly to human well-being throughout.
The remaining EGS occurred variably and had lesser overall
importance. However, in some locations they were of high
importance (e.g. wildlife, agricultural crops, fruit and nuts in
Misima; mangrove timber in Misima and Trobriand).
For all subregions the contributions to aggregated well-being
from marine EGS were greater than terrestrial EGS. Although this

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Table 4
The provisioning ecosystem goods and services (EGS) in each subregion and their
predicted change in condition by 2030. Abbreviations for subregions are
D Dawson; W Woodlark; M Misima; T Trobriand; L Louisiade.
EGS

Table 5
Occurrence of indirect and direct drivers of change inuencing the condition of
ecosystem goods and services within each subregion by 2030. Abbreviations for
subregions are D Dawson; W Woodlark; M Misima; T Trobriand;
L Louisiade.

Mean change
Drivers of change

(a) Terrestrial
Freshwater
Garden food crops
Forest timber
Wildlife
Agricultural crops
Fruit and nuts
Minerals (land)
Mean change
(b) Marine
Bche-de-mer
Coral
Reef sh
Sharks
Turtles
Mangrove timber
Dugong
Seaweed
Minerals (seabed)
Crabs
Shellsh (currency)
Mean change
Mean change all EGS

4
4

2
3
2
2

4

 4.0

 2.3

4
2
2
4
3

4
1
2
2
4

3
3
3
3
3
3
3
 3.0

3
4
4
4
4
4

1
2
2
2

 3.8

 1.8

4
3
3
4
2
3
2
2

4
3
4
4
3
3
3
3

3
1
2
2

2

 3.8
 2.0
 2.6
 3.2
 3.0
 3.0
 2.5
 2.5
 1.0
 3.0
 2.0

 2.0
 1.9

 2.8
 2.9

1
3
 3.0
 3.4

 2.4
 2.4

 2.9
 2.9

 3.4
 3.6

 2.6
 3.2
 2.8
 2.8
 2.8
 3.5
 3.0
 3.0

ranged in a marineterrestrial split of 5149% in Misima and 65

35% in Woodlark (Table 3), there was no statistically signicant
difference amongst subregions (2 2.568, d.f. 4, p 0.632). Over
all subregions, marine EGS contributed 57% and terrestrial EGS
contributed 43% to aggregated well-being.
3.2. Predicted changes in EGS condition
The condition of all EGS was predicted to decline by 2030
(Table 4). Across all subregions, bche-de-mer was predicted to
decline the most (  3.8), followed by fruit and nuts (  3.5), garden
food crops and sharks ( 3.2). Participants highlighted that in the
case of bche-de-mer, garden food crops and sharks, their current
status was already poor, and consequently projected declines
could result in extreme degradation.
There was variation between subregions in mean EGS decline,
with the highest for Trobriand (  3.6) and the lowest for Louisiade
(  1.9). The mean predicted decline ( 7s.e.) for terrestrial EGS
(  3.0 70.18, n 24) was greater than for marine EGS (  2.8 7 0.17,
n 33), but the difference was not statistically signicant (t 0.86,
d.f. 51, p 0.393). The mean condition of all EGS was predicted to
decline by 2.9 (70.12, n 57), or 450%.
3.3. Drivers of change
Five indirect and 10 direct drivers of change were identied
(Table 5). Amongst the indirect drivers, human population growth
was identied most frequently (41%), followed by storms (16%) and
climate change (12%), and this difference was signicant (2 21.82, d.
f.4, po0.001). Discussion claried that storms referred to the
continued occurrence of tropical cyclones that are currently experienced in MBP, rather than the intensication of storms linked to
climate change. Population growth was the only indirect driver
recorded for all ve subregions, while the remaining drivers occurred
in only three or two subregions each. Climate change was only
mentioned for Misima (causing coral bleaching, sea level rise impacting mangroves and eroding turtle nesting beaches) and Trobriand
(droughts limiting freshwater supplies and impacting forests). When

(a) Indirect
Human population growth
Declining soil fertility
Tourism development
Storms
Climate change
(b) Direct
Overshing
Pollution from gardens/mining/logging
Forest clearing for gardens/mining/logging
Destructive shing (reef sh)
Destructive shing (seagrass)
Mining overexploitation
Mangrove overharvesting
Overhunting wildlife
Destructive coral harvesting
Shipping pollution

Total (%)

4
2
2
4
3a

5
1

5
1
1

23
4
4
7
5

(41)
(9)
(9)
(16)
(12)

4
1
2

4
1
2
1

4
1
2
1

18
9
9
5
2
2
2
2
2
1

(35)
(17)
(17)
(10)
(4)
(4)
(4)
(4)
(4)
(2)

1
1

3
1
1
1

3
5
2
2
1
1
1

2
2b

2
1
1
1

1
1
1

a
Coral bleaching, sea level rise impacting mangroves and eroding turtle
nesting beaches.
b
Droughts limiting freshwater supplies and impacting forests.

discussed, participants conrmed that they were aware of some

possible effects (e.g. sea level rise, coral bleaching) already occurring
in MBP, but they understood from the climate projection data (Table 2)
and the asset-threat interaction model that extreme effects on EGS
were most likely to occur after mid-century. None of the groups which
had not identied climate change impacts wished to change their
assessment after discussion.
The most frequently identied direct driver was overshing (36%),
which occurred in all subregions (Table 5), followed by pollution from
gardens, mining or logging and forest clearing for gardens, mining or
logging (16%) and destructive shing for reef sh (10%), and this
difference was signicant (2 43.12, d.f.11, po0.001). The remaining direct drivers occurred infrequently across the subregions, and
were specic to certain activities occurring in those subregions (e.g.
mining in Woodlark and Misima).
3.4. Human well-being impacts (HWI)
The greatest total HWI of predicted declines in EGS condition
was for bche-de-mer (  95), followed by garden food crops and
sharks (  80), freshwater and reef sh ( 65), forest timber and
agricultural crops ( 55) (Table 6). The greatest mean HWI was for
Dawson (  16.1) and the lowest was for Louisiade (  9.2).
3.5. EGS condition change, HWI and human population density
correlations
There was a positive correlation between population density in
2000 and mean change in EGS condition by 2030 (r 0.62), and a
stronger correlation between population density and HWI
(r 0.73), but neither were statistically signicant.
3.6. Management strategies
Twenty-four management strategies were identied (Table 7).
Across all subregions the most important strategy was improved
garden/agricultural productivity (MSI 180), followed by human
population control (MSI 140) and community education for
sheries management (MSI 126). Improved garden/agricultural
productivity, community education for sheries management, and

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Table 6
The human well-being impacts (HWI) of projected changes in provisioning
ecosystem goods and services0 (EGS) condition by 2030. Abbreviations for subregions are D Dawson; W Woodlark; M Misima; T Trobriand; L Louisiade.
EGS
(a) Terrestrial
Freshwater
Garden food crops
Forest timber
Wildlife
Agricultural crops
Fruit and nuts
Minerals (land)
(b) Marine
Bche-de-mer
Coral
Reef sh
Sharks
Turtles
Mangrove timber
Dugong
Seaweed
Minerals (seabed)
Crabs
Shellsh (currency)
Mean HWI

 20
 20

 15
 15
 15
 15
 15
 15
 15

 15
 20
 20
 12
 20
 20

5
 10
 10
8

 65
 80
 55
 39
 55
 35
 15

 20
5
 10
 10
 16

 20
 15
 15
 20
 10
 15
4
8

 20
 15
 20
 20
 12
 15
6
9

 15
5
 10
 10

 10.6

 14.1

 95
 50
 65
 80
 47
 30
 10
 17
2
 15
 10
 9.2

2
 15
 16.1

 10
 16.0

Table 7
Management strategy importance (MSI) for each subregion. Note that because all
human well-being impacts were negative (see Table 6), all scores are negative, but
for ease of presentation negative signs are omitted. Abbreviations for subregions
are D Dawson; W Woodlark; M Misima; T Trobriand; L Louisiade.
Management strategy

Total MSI

Improved garden/agricultural productivitya

Human population control
Community education for sheriesb
Fisheries researchc
Improved national BDMd management
Improved national shark management
MPAs for turtles and dugong
MPAs for BDM
MPAs for coral reefs
Human population relocation
Re-forestation
Aquaculturee
Community education for agriculturef
Improved regulation of logging
De-salination plants for water supplies
Mangrove MPAs and restoration
Forest and wildlife research
Improved pollution regulationg
Import building materials
Improved national turtle legislation
Environmentally-sustainable mining
International climate change mitigation
Wildlife harvest controls
Sustainable coral harvesting

20
20
19
9
20
20
9
20

35
35
20

45

60
60
27
46
20
20
15
20
15
15

20
35
15

180
140
126
97
95
80
62
60
60
55
49
46
45
43
35
30
27
25
20
19
17
10
8
5

20
10
16
15

45
42
20
20
22
20
15

45
24
16

15
20

15
10

15
10
10

45
15
20
15
15

20
15
15
12

15

10
20

9
2

10
15
10
8

international climate change mitigation, explicitly addressed climate

impacts, and this was only identied by the Misima working group.
3.7. Subregion synthesis

Total HWI

 10
 15
 10
4

 20

 20
 10
 10
 20
9

Includes fruit and nut production.

Reef sh, coral reefs, turtles and mangroves.
c
Includes sharks, turtles, dugong, bche-de-mer, seaweed.
d
BDM bche-de-mer.
e
Reef sh, turtles, shellsh.
f
Includes gardens and fruit and nuts.
g
Gardens, mining and logging.
b

improved national bche-de-mer and shark management were the

only strategies identied for all subregions. Human population
control was identied for all subregions except Misima, and
population relocation was identied for Dawson and Trobriand,
and was the most important strategy for Dawson. Only one strategy,

The DPSIR models for each subregion synthesised these results

(Figs. 37). For all subregions human population growth was the
highest or equal-highest ranked driver. Climate change was of
lesser importance and was only identied in Misima and Trobriand. Storms were of slightly higher importance, and identied
in three subregions. The highest ranked pressures were overshing (four subregions) or pollution from gardens and logging
(one subregion). Although the principal state/impacts (i.e. HWI)
and responses (i.e. management strategies) were similar across the
subregions, there were differences in the relative rankings. Also, in
some subregions specic EGS of local importance were addressed,
such as turtles and turtle marine protected areas (MPAs) in
Woodlark (Fig. 4) and Misima (Fig. 5), mangrove timber and
mangrove MPAs and restoration in Trobriand (Fig. 6), and wildlife
hunting and regulation in Louisiade (Fig. 7).

4. Discussion
The participants gave the following broad prognosis for the ve
subregions. Although marine EGS contribute 57% of aggregated
human well-being, terrestrial EGS were also important, contributing 43%. Overall there will be a decline in EGS0 condition of 4 50%
by 2030, and this will be of similar magnitude for terrestrial and
marine EGS. The projected declines in bche-de-mer, garden food
crops and sharks is of concern because their status is already
degraded. This decline will primarily be driven by population
growth, which was noted in all subregions and was the most
frequently identied indirect driver. This inuences the most
prevalent direct driver, overshing, plus pollution and forest
clearing from gardens and logging, which occurred in all subregions. The greatest impact on well-being will be through further
declines in bche-de-mer, garden food crops, sharks, freshwater
and reef sh. The priority management strategies were improved
garden production, human population control and community
education on sheries management. However, the DPSIR models
illustrated that there was some variation in the rankings of
pressures, impacts and responses between subregions, and there
were specic responses for EGS of localised importance.
Participants0 assessments were not completely comprehensive,
however. Fishery species such as trochus, lobster and molluscs are
known to be of commercial importance throughout the islands
[32], but were not listed as EGS. The commercial bche-de-mer
and shark n sheries are inuenced by the growing Asian
economy and related export markets [32], but this was not
identied as a driver. Similarly, monetisation of the local economy
is driving materialism, erosion of traditional norms, institutions
and leadership, plus drug, alcohol and debt problems amongst
younger generations, and disputes over land and sea tenure
[42,43], but these were not mentioned. One explanation for these
omissions could be the lack of time available in the workshop to
explore and discuss these issues in detail. Another may be that
participants were from regional organisations, and therefore
lacked specic knowledge of global or community-scale drivers,
which is a weakness of analysing social-ecological systems from
the perspective of stakeholders at only one level [1,12].
In spite of this, more detailed studies in other island subregions
of MBP mirror the results, suggesting that participants0 perceptions were largely representative. A household survey carried out
by the NFA [43] in the Northern Owen-Stanley and Samarai
subregions revealed that farming and gardening is the most

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Drivers
1. Human population growth
2. Tourism development
2. Storms

Pressures
1.
2.
3.
4.

Overfishing
Destructive fishing (reef fish)
Pollution from gardens/logging
Forest clearing for
gardens/logging

Responses
1.
2.
2.
2.
2.
2.
2.
3.
4.
5.
5.
5.

Human population relocation

Human population control
Improved garden/agricultural productivity
Improved national BDM management
Improved national shark management
MPAs for BDM
De-salination plants for water supplies
Community education for fisheries
International climate change mitigation
MPAs for turtles and dugong
Fisheries research
Improved national turtle legislation

State/Impact
1.
1.
1.
1.
1.
2.
2.
3.

BDM
Sharks
Freshwater
Garden food crops
Agricultural crops
Coral
Reef fish
Turtles

Fig. 3. Summary DPSIR system model for the Dawson subregion. See Table 5 for indirect drivers (drivers) and direct drivers (pressures), Table 6 for human well-being
impacts (state/impact), and Table 7 for management strategies (responses). Issues are listed by descending or equal rank. Abbreviations are BDM (bche-de-mer) and MPA
(marine protected areas).

Drivers
1. Human population growth

Pressures
1. Pollution from gardens/logging
2. Overfishing
3. Forest clearing for
gardens/logging/mining
3. Destructive fishing (reef fish)
4. Overhunting wildlife
4. Mangrove overharvesting
4. Mining overexploitation

Responses
1.
1.
2.
3.
3.
4.
4.
5.
5.
5.
5.
6.
7.
8.

Human population control

Improved garden/agricultural productivity
Re-forestation
Community education for fisheries
Improved national BDM management
MPAs for turtles and dugong
Aquaculture
MPAs for coral reefs
Improved regulation of logging
Mangrove MPAs and restoration
Improved pollution regulation
Improved national shark management
Sustainable coral harvesting
Environmentally-sustainable mining

State/Impact
1.
2.
3.
3.
4.
4.
4.
4.
5.
6.
7.

BDM
Turtles
Crabs
Garden food crops
Freshwater
Forest timber
Reef fish
Sharks
Coral
Wildlife
Minerals

Fig. 4. Summary DPSIR system model for the Woodlark subregion. See Table 5 for indirect drivers (drivers) and direct drivers (pressures), Table 6 for human well-being
impacts (state/impact), and Table 7 for management strategies (responses). Issues are listed by descending or equal rank. Abbreviations are BDM (bche-de-mer) and MPA
(marine protected areas).

common household occupation, followed closely by shing and

collecting marine resources. This reects participants0 ranking of
garden food production, bche-de-mer, reef sh and sharks as
being universally important EGS. The survey also showed that
communities consistently identied population growth as the
cause of overshing, and that as a consequence the status of
marine resources is predicted to decline further, and bche-demer and reef sh in particular. In addition, destructive shing,
pollution and tourism development were identied as impacts,
and improved community-based sheries management and education were the highest ranking responses. Declining soil fertility

has also been identied in the Samarai subregion, where fallow

periods for swidden-cultivated gardens have fallen due to population growth and increasing demand for gardening land [42].
The positive but statistically insignicant correlations between
subregions0 population densities in 2000 and projected changes in
EGS and HWI may cast doubt on the validity of the participants0
assessments. However, this result is not surprising considering the
small sample size of subregions (n 5), and likely error and
variation in the data. First, each stakeholder group may have
applied different scores for the same EGS contributions to wellbeing and changes in EGS condition. Although this was minimised

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Drivers
1.
1.
2.
3.
3.

Human population growth

Storms
Climate change
Declining soil fertility
Tourism development

Pressures
1. Overfishing
2. Forest clearing for
gardens/mining/logging
2. Destructive fishing (seagrass)
3. Pollution from
gardens/mining/logging
3. Mining overexploitation

Responses
1.
1.
1.
2.
3.
4.
4.
4.
5.
5.
5.
5.
6.

Improved garden/agricultural productivity

Community education for agriculture
Community education for fisheries
Fisheries research
MPAs for turtles and dugong
Improved national BDM management
Improved national shark management
MPAs for BDM
MPAs for coral reefs
Re-forestation
Forest and wildlife research
Environmentally-sustainable mining
Improved national turtle legislation

State/Impact
1.
1.
2.
2.
2.
2.
2.
2.
2.
2.
2.
2.
3.
4.
5.

BDM
Sharks
Freshwater
Garden food crops
Forest timber
Wildlife
Agricultural crops
Fruit and nuts
Minerals
Coral
Reef fish
Mangrove timber
Turtles
Seaweed
Dugong

Fig. 5. Summary DPSIR system model for the Misima subregion. See Table 5 for indirect drivers (drivers) and direct drivers (pressures), Table 6 for human well-being impacts
(state/impact), and Table 7 for management strategies (responses). Issues are listed by descending or equal rank. Abbreviations are BDM (bche-de-mer) and MPA (marine
protected areas).

Drivers
1.
2.
2.
3.

Human population growth

Storms
Climate change
Declining soil fertility

Pressures
1. Overfishing
2. Forest clearing for
gardens/logging
3. Pollution from gardens/logging
3. Destructive fishing (reef fish)
3. Destructive fishing (coral)
3. Overharvesting mangroves

Responses
1.
1.
2.
3.
4.
4.
4.
4.
4.
4.
5.
5.
5.
5.
5.
6.

Human population control

Improved garden/agricultural productivity
Fisheries research
Community education for fisheries
Improved national BDM management
Improved national shark management
MPAs for BDM
Aquaculture
Improved regulation of logging
Import building materials
MPAs for turtles and dugong
MPAs for coral reefs
Human population relocation
De-salination plants for water supplies
Mangrove MPAs and restoration
Forest and wildlife research

State/Impact
1.
1.
1.
1.
1.
1.
1.
2.
2.
2.
3.
3.
4.
5.

Garden food crops

Forest timber
Agricultural crops
Fruit and nuts
BDM
Reef fish
Sharks
Freshwater
Coral
Mangrove timber
Wildlife
Turtles
Seaweed
Dugong

Fig. 6. Summary DPSIR system model for the Trobriand subregion. See Table 5 for indirect drivers (drivers) and direct drivers (pressures), Table 6 for human well-being
impacts (state/impact), and Table 7 for management strategies (responses). Issues are listed by descending or equal rank. Abbreviations are BDM (bche-de-mer) and MPA
(marine protected areas).

by applying a narrow range for scoring (i.e. 15), and by reviewing

data with participants to address any inconsistencies, such error
will have been amplied through the multiplicative calculation of

HWI. Second, there are likely to be inherent variations in the

relationships between population density and the productive
capacity of terrestrial and marine ecosystems. Soil types vary with

10

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

Drivers
1. Human population growth
2. Declining soil fertility
2. Tourism development

Pressures
1. Overfishing
2. Forest clearing for
gardens/logging
3. Pollution from gardens/logging
3. Destructive fishing (reef fish)
3. Destructive fishing (coral)
3. Overhunting wildlife
3. Shipping pollution

Responses
1.
2.
3.
3.
3.
4.
4.
4.
4.
5.
5.

Human population control

Improved garden/agricultural productivity
Community education for fisheries
Improved national BDM management
Improved national shark management
MPAs for coral reefs
Re-forestation
Aquaculture
Improved regulation of logging
Improved regulation of pollution
Wildlife harvest controls

State/Impact
1.
2.
2.
2.
2.
3.
4.
5.
5.

BDM
Garden food crops
Forest timber
Reef fish
Sharks
Shellfish
Wildlife
Freshwater
Coral

Fig. 7. Summary DPSIR system model for the Louisiade subregion. See Table 5 for indirect drivers (drivers) and direct drivers (pressures), Table 6 for human well-being
impacts (state/impact), and Table 7 for management strategies (responses). Issues are listed by descending or equal rank. Abbreviations are BDM (bche-de-mer) and MPA
(marine protected areas).

islands0 geomorphology, and some are less productive and prone

to loss of fertility and drought than others [32,42]. Also, some
islands have coastal lagoons which provide accessible supplies of
marine resources which are valuable during periods of reduced
garden production, while others do not [23]. Finally, the population data could be inaccurate, a common problem in PNG national
censuses [60].
Assuming that the results were representative, it appears that
Dawson may soon exceed its carrying capacity under current
livelihood and technological systems. With the highest population
density of 107.1 people km  2 in 2000, and the highest projected
decline in human well-being by 2030, participants recommended
population relocation, and this was the most important strategy.
Based on their HWI, Misima and Trobriand may also be approaching
carrying capacity with densities of 67.5 km  2 and 28.6 km  2,
respectively. The assessment for Dawson supports the hypothesis
of Foale [42] that above a density of 100 people km  2 food security
cannot be guaranteed for islands in MBP, and therefore this may be
a threshold beyond which human well-being rapidly declines.
However, the results suggest that for some subregions such as
Misima and Trobriand, this threshold may be o100 people km  2.
More generally, these results support the contention by Bourke [61]
and Allen and Bourke [35] that in contrast to mainland PNG where
population densities are lower, small islands are most food insecure,
and vulnerable communities must be identied for policy intervention.
Consequently, until 2030 population growth rather than climate change was the greatest perceived driver of human wellbeing. This was not because participants ignored climate change:
the working groups for Misima and Trobriand identied climate
impacts, and in discussion participants claried that they had
already observed climate impacts in MBP, and these were only
considered to be a potentially major driver of EGS condition after
mid-century. However, in spite of this apparent awareness, some
groups may have underestimated its potential future impacts. The
presentation of complex climate data and projections by science
experts can lack salience, credibility and legitimacy for lay audiences [14] unless contextualised [10], for example through visualisation [13]. Although the method aimed to achieve this by
applying a threat-asset interaction model [38], which illustrated

potential impacts on EGS through regional climate projections

under the worst case A1B SRES global emission scenario, the
results may not have been fully understood or assimilated by all
participants.
In addition, population effects may have been over-estimated,
diminishing the relative potential inuence of climate change.
Although population projections were deliberately not illustrated
for the subregions, projections for MBP were shown using the
19902000 annual average growth rate of 2.7% annum  1, which
yielded a 78% increase by 2030. However, preliminary results from
the 2011 national census show that the MBP population growth
rate in 20002011 had declined to 2.3% annum  1 [60], and
consequently population effects may have been over-anticipated
by participants.
These results have important implications for the development
of climate adaptation planning in MBP. Rather than focus on
climate-specic issues, all but one of the strategies identied
(international climate change mitigation by the Misima working
group) were motivated by the need to control human population
growth and concomitant ecosystem degradation. Hence the question arises whether these strategies can deliver co-benets for
climate adaptation, and whether any will be mal-adaptive. It seems
that many will be complimentary rather than mal-adaptive, and
also no regrets (i.e. they will yield benets under any future
conditions of change [62]). Reducing human population impacts
would enhance the health and resilience of ecosystems to climate
change, and many of the measures identied such as MPAs,
mangrove restoration and re-forestation represent ecosystembased adaptation [63,64]. For example, mangroves can provide
coastal protection against sea level rise erosion and inundation
[65], and mangrove planting has been prioritised as a cost effective
strategy for combating coastal inundation by the PNG government
[66]. Coral reefs buffer coasts against storms [67], and re-forestation
mitigates against landslide risk and facilitates multi-use agricultural
production systems which maintain soil fertility [68]. Community
education and training on sheries and agricultural sustainability
will also potentially build human and social capital, and hence
adaptive capacity, addressing a recognised lack of understanding
amongst MBP communities of the nite extent of natural resources
[42,43].

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

The prognosis for the ve subregions presents a major challenge

for the CTI in MBP. As a priority seascape the projected declines in
reef sh, reef coral, bche-de-mer and sharks across all subregions is
of conservation concern. Equally, widespread declines in garden
food production and freshwater and hence the carrying capacity of
islands are also critical because, as illustrated by the 19971998
drought, shortages of cultivated terrestrial foods result in the
transferral of effort to wild-harvested marine resources to maintain
food availability (through subsistence) and access (through sale for
cash or exchange for other food). Although many of the management
strategies mirror those advocated by the PNG CTI National Plan of
Action (e.g. MPAs, national shery management plans, community
education in sheries management [48]), it is clear that to achieve
both marine biodiversity conservation and food security human
population control should be the priority. Improving the condition
of terrestrial EGS, and garden food production in particular, is also
essential to maintain food security and reduce harvesting pressure
on marine resources. So far the relevance of terrestrial EGS for
marine biodiversity conservation have not been fully recognised by
the PNG CTI National Plan of Action, or more widely in the Pacic.
The methodology has several limitations which should be
addressed in the design of similar exercises. As discussed above,
making comparisons between subregions may be compromised by
the differing scorings applied by the workshop groups, and
detailed analysis of community and global-scale drivers and some
EGS was lacking. Furthermore, different working groups may have
had varying perceptions of appropriate management strategies
required to address the same drivers. Replication of the process at
the community scale, particularly in the priority subregions of
Dawson, Misima and Trobriand would provide deeper insights,
and triangulate and compliment the regional stakeholders0 perceptions. Local analyses would also enable disaggregation of those
communities, households or individuals who may be disproportionately impacted by declines in specic EGS [69,70]. This is
important because population densities vary between and within
islands in MBP [42], and some communities may be more
dependent on terrestrial or marine EGS than others [37,42,43].
If other subregions which include island and mainland communities are to be assessed (e.g. Collingwood Bay, Northern OwenStanley, Samarai; Fig. 2), a ner scale of analysis will also be
necessary to differentiate amongst population densities within
them. Local-scale analyses would also differentiate between sociocultural attributes of communities, which vary across MBP [36,42],
potentially inuencing their adaptive capacity.
Moreover, the method only considered local community beneciaries of EGS. Hence it did not account for off-site beneciaries
such as tourists and global society, who derive cultural EGS from
the regions0 unique marine biodiversity, exemplied by the international donor, NGO and government support for the CTI. Understanding trade-offs between off-site and local beneciaries is
fundamental for the design of effective governance and marketbased mechanisms required to manage endangered and iconic
marine species [69,7174]. Furthermore, due to risks of confusion
amongst participants and double-counting, which is an inherent
challenge in the valuation of EGS [1,75], we did not assess cultural,
regulating and supporting EGS.
Also, the MA framework only captures the direct contribution of
ecosystems to communities0 well-being, and ignores income
through other employment or remittances. Although remittances
have contributed 33% of household income on Ware Island in the
Samarai subregion [76], the NFA survey of a wider area found that
remittances plus non-agricultural and non-sheries related employment provided only 13% of household income [43]. Consequently,
island livelihoods remain directly dependent on ecosystems for cash,
subsistence and cultural needs [43], and wild-harvested marine and
terrestrial EGS provide an important backstop for food security

11

[36,37,42]. Moreover, this dependence is likely to persist, given the

limited opportunities for alternative, formal waged employment in
such remote locations of PNG [42]. Thus the MA framework is an
appropriate tool for assessing socialecological systems such as the
islands in MBP.
While there is evidence that participatory processes can foster
adaptive capacity by enhancing social learning [2,11], the coproduction of knowledge, mitigation of power asymmetries, and
creating ownership of problems and solutions [710], the impacts
of the process were not evaluated in these terms. The presence of
researchers may also affect the outcomes of participatory processes due to complex and subtle power relationships [11] but this
was not assessed, either. Hence future exercises should carry out
such evaluations. In particular, an assessment of the effectiveness
of the threat-asset interaction model as a tool for visualising
climate impacts on EGS is necessary, and the utility of the DPSIR
models which were only developed after the workshop.

5. Conclusions
When combined with studies in other subregions [42,43], the
results presented here indicate that island communities in MBP
are becoming increasingly vulnerable to change. Population
growth and resultant declining condition of terrestrial and marine
EGS is evident in many islands, with the consequence that their
carrying capacity under current livelihood and technological
systems may soon be exceeded, and food security cannot be
guaranteed. The ubiquitous over-population problem, combined
with weakening cultural ties between islands has nullied any
safety valves that allowed out-migration during historical periods
of population pressure [42]. With the exception of aquaculture
[43] and nature-based tourism, which is regarded as a potential
panacea for simultaneously achieving biodiversity conservation
and economic development in the region [25,47,77], there are few
alternative employment opportunities for island communities
which might relieve pressure on EGS. In addition, the rise of
materialism through the expanding cash economy is weakening
traditional institutions, norms and governance and causing social
dysfunction, which exacerbates resource overexploitation [42,43].
These interlinked issues are undermining communities0 capacity
to cope with potential future shocks such as droughts, severe
storms, inundation or shery collapses.
Similar trends are evident in remote areas of the Solomon Islands
[78]. Consequently there is a growing awareness that if Melanesian
communities are to adapt to the increasing rates and magnitude of
change, development efforts should focus on building self-reliance
through grass-roots institutions [31], and managing the fundamental
drivers of population growth and materialistic expectations [78].
Relocation of island inhabitants will also be necessary, as already
experienced in PNG0 s Cartaret Islands [79,80] and Bali-Witu Islands
[81,82]. If the aims of the CTI are to be achieved in MBP, similar
strategies are required, augmented by improved garden production
and community education in natural resource management. It is
suggested that there is a 2030 year adaptation window in which to
control population growth, which will otherwise continue to erode
the adaptive capacity of communities and ecosystems to cope with
potentially extreme climate impacts after mid-century [3].
In spite of methodological limitations and necessary evaluation
and improvement, including the possible inclusion of more recent
and sophisticated approaches to classifying and evaluating EGS
(e.g. [8386]), we believe that the participatory process and DPSIR
models generated a representative assessment of the status of
socialecological systems within ve island subregions in MBP.
The approach identied the relative importance of different EGS
for human well-being, pre-eminent drivers and their impacts on

12

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

EGS condition, the subregions requiring priority intervention and

specic adaptation strategies necessary for each. However, these
data are only a preliminary diagnosis which should form the rst
stage of a wider multi-scale, multi-stakeholder adaptive comanagement process to develop adaptation pathways for livelihoods [3]. Critical next steps are to engage local communities
within the subregions to further rene this assessment, and
integrate their world-views with those of the regional stakeholders. In the absence of more detailed scientic data, such
information is clearly vital for conserving marine biodiversity
and managing food security in the Coral Triangle. This may
become increasingly valuable as the effects of population pressure
and climate change compound, requiring more rapid assessments
and feedback for policy and management.

Acknowledgements
The authors thank the relevant stakeholders who attended the
workshop in Alotau. The research was supported by grants from CI
and AusAID. Toni Darbas and Ingrid van Putten provided helpful
comments on earlier drafts.
References
[1] Millennium Ecosystem Assessment. Ecosystems and human well-being: a
framework for assessment. Washington DC: Island Press; 2005.
[2] Ensor J. Uncertain futures: adapting development to a changing climate.
Rugby: Practical Action Publishing; 2011.
[3] Butler JRA, Suadnya W, Puspadi K, Sutaryono Y, Wise RM, Skewes TD, et al.
Framing the application of adaptation pathways for rural livelihoods and
global change in Eastern Indonesian islands. Global Environ Change, http://dx.
doi.org/10.1016/j.gloenvcha.2013.12.004, in press.
[4] Adger WN, Huq S, Brown K, Conway D, Hulme M. Adaptation to changing
climate in the developing world. Prog Dev Stud 2003;3:17995.
[5] Adger WN. Vulnerability. Global Environ Change 2006;16:26881.
[6] Smit B, Wandel J. Adaptation, adaptive capacity and vulnerability. Global
Environ Change 2006;16(3):28292.
[7] Cash DW, Clark WC, Alcock F, Dickson NM, Eckley N, Guston DH, et al.
Knowledge systems for sustainable development. Proc Natl Acad Sci USA
2003;100(14):808691.
[8] Ballard D. Using learning processes to promote change for sustainable
development. Action Res 2005;3:13556.
[9] Brown VA. Leonardo0 s vision: a guide to collective thinking and action.
Rotterdam: Sense Publishers; 2008.
[10] Gidley JM, Fien J, Smith J-A, Thomsen DC, Smith TF. Participatory futures
methods: towards adaptability and resilience in climate vulnerable communities. Environ Policy Gov 2009;19:42740.
[11] Fazey I, Kesby M, Evely A, Latham I, Wagatora D, Hagasua J-E, et al. A threetiered approach to participatory vulnerability assessment in the Solomon
Islands. Global Environ Change 2010;20:71328.
[12] Enfors EI, Gordon LJ, Peterson GD, Bossio D. Making investments in dryland
development work: participatory scenario planning in the Makanya catchment, Tanzania. Ecol Soc 2008;13:42.
[13] Shaw A, Sheppard S, Burch S, Flanders D, Wiek A, Carmichael J, et al. Making
local futures tangible - synthesizing, downscaling, and visualizing climate
change scenarios for participatory capacity building. Global Environ Change
2009;19:44763.
[14] Cash DW, Borck JC, Patt AG. Countering the loading-dock approach to linking
science and decision making: comparative analysis of El Nio/Southern
Oscillation (ENSO) forecasting systems. Sci Technol Hum Values 2006;31:
46594.
[15] Veron JEN, DeVantier LM, Turak E, Green AL, Kininmonth SJ, Stafford-Smith
MG, et al. Delineating the Coral Triangle. Galaxea 2009;11:91100.
[16] Carpenter KE, Springer VG. The center of the center of marine shore sh
biodiversity: the Philippine Islands. Environ Biol Fishes 2005;72:46780.
[17] Allen GR. Conservation hotspots of biodiversity and endemism for Indo-Pacic
coral reef shes. Aquat Conserv: Mar Freshw Ecosyst 2007;18:54156.
[18] Spalding MD, Kainuma M, Collins L. World mangrove atlas. London: Earthscan,
with International Society for Mangrove Ecosystems, Food and Agriculture
Organization of the United Nations, The Nature Conservancy, UNEP World
Conservation Monitoring Centre, United Nations Scientic and Cultural
Organisation, United Nations University; 2010.
[19] Coral Triangle (CTI) Secretariat. Coral Triangle Initiative on coral reefs, sheries
and food security: regional plan of action. Jakarta, Indonesia: Interim CTI
Secretariat; 2009.
[20] Christensen AE, Mertz O. Researching Pacic island livelihoods: mobility,
natural resource management and nissology. Asia Pac Viewp 2010;51:27887.

[21] Gough KV, Bayliss-Smith T, Connell J, Mertz O. Small island sustainability in

the Pacic: introduction to the special issue. Singap J Trop Geogr 2010;31:19.
[22] Hoegh-Guldberg O, Hoegh-Guldberg H, Veron JEN, Green A, Gomez ED, Lough
J, et al. The Coral Triangle and climate change: ecosystems, people and
societies at risk. Brisbane: WWF Australia; 2009.
[23] Bourke RM. How Papua New Guinea villagers survived the 1997 drought and
frosts. Dev Bull 2005;67:279.
[24] Coral Triangle (CTI) Secretariat. Region-wide early action plan for climate
change adaptation for the nearshore marine and coastal environment and
small island ecosystems (REAP-CCA). Jakarta, Indonesia: Interim CTI Secretariat; 2011.
[25] Foale S, Adhuri D, Alino P, Allison EH, Andrew N, Cohen P, et al. Food security
and the Coral Triangle Initiative. Mar Policy 2013;38:17483.
[26] Badjeck MC, Allison EH, Halls AS, Dulvy NK. Impacts of climate variability and
change on shery-based livelihoods. Mar Policy 2010;34:37583.
[27] Bell JD, Kronen M, Vunisea A, Nash WJ, Keeble G, Demmke A, et al. Planning
the use of sh for food security in the Pacic. Mar Policy 2009;33:6476.
[28] Bell J, Johnson J, Hobday A. Vulnerability of tropical Pacic sheries and
aquaculture to climate change. Noumea, New Caledonia: Secretariat of the
Pacic Community; 2010.
[29] Grantham HS, McLeod E, Brooks A, Jupiter SD, Hardcastle J, Richardson AJ,
et al. Ecosystem-based adaptation in marine ecosystems of tropical Oceania
in response to climate change. Pac Conserv Biol 2011;17:24158.
[30] Carneiro C. Marine management for human development: a review of two
decades of scholarly evidence. Mar Policy 2011;35:35162.
[31] Schwarz A-M, Bn C, Bennett G, Boso D, Hilly Z, Paul C, et al. Vulnerability
and resilience of remote rural communities to shocks and global changes:
empirical analysis from Solomon Islands. Global Environ Change 2011;21:
112840.
[32] Kinch J. A review of sheries and marine resources in Milne Bay Province,
Papua New Guinea. Kavieng: National Fisheries Authority and Coastal Fisheries Management and Development Project; 2007.
[33] Skewes T, Kinch J, Polon P, Dennis D, Seeto P, Taranto T, et al. Research for
sustainable use of beche-de-mer resources in Milne Bay Province, Papua New
Guinea. Cleveland: CSIRO Division of Marine Research Final Report; 2002.
[34] Allen GR, Kinch JP, McKenna SA, Seeto P, editors. A rapid marine biodiversity
assessment of Milne Bay Province, Papua New Guinea Survey II (2000).
Washington DC: Center for Applied Biodiversity Science (CABS); 2003.
[35] Allen BJ, Bourke RM. The 1997 drought and frost in PNG: overview and policy
implications. In: Bourke RM, Allen MG, Salisbury JG, editors. Food security for
Papua New Guinea. Proceedings of the Papua New Guinea food and nutrition
2000 conference. ACIAR proceedings no. 99. Canberra: Australian Centre for
International Agricultural Research; 2001, p. 15563.
[36] Jonathon A. The El Nio drought: an overview of the Milne Bay experience. In:
Bourke RM, Allen MG, Salisbury JG, editors. Food security for Papua New
Guinea. Proceedings of the Papua New Guinea food and nutrition 2000
conference. ACIAR proceedings no. 99. Canberra: Australian Centre for International Agricultural Research; 2001, p. 20913.
[37] Mogina J. Food aid and traditional strategies for coping with drought:
observations of responses by villagers to the 1997 drought in Milne Bay
Province. In: Bourke RM, Allen MG, Salisbury JG, editors. Food security for
Papua New Guinea. Proceedings of the Papua New Guinea food and nutrition
2000 conference. ACIAR proceedings no. 99. Canberra: Australian Centre for
International Agricultural Research; 2001, p. 2018.
[38] Skewes T, Lyne V, Butler JRA, Mitchell D, Poloczanska E, Williams K, et al.
Melanesian coastal and marine ecosystem assets: assessment framework and
Milne Bay case study [CSIRO nal report to the CSIRO AusAID Alliance].
Brisbane: CSIRO Marine and Atmospheric Research; 2011.
[39] Kinch J. Social evaluation study for the Milne Bay community-based coastal
and marine conservation program. PNG/99/G41 [UNOPS contract for services
report ref: C00-1076]; 2001.
[40] McGillivray, M.; 2012. http://www.deakin.edu.au/news/2012/23042012
PNGMarkIndexresults.php [accessed 1.12.12].
[41] UNDP. Human development report 2011: sustainability and equity, a better
future for all. New York: United Nations Development Program; 2011.
[42] Foale S. Sharks, sea slugs and skirmishes: managing marine and agricultural
resources on small, overpopulated islands in Milne Bay, PNG. Canberra: The
Australian National University; 2005.
[43] National Fisheries Authority. Socio-economic survey of small-scale sheries in
Milne Bay Province, Papua New Guinea. Kavieng: Papua New Guinea National
Fisheries Authority and Coastal Fisheries Management and Development
Project; 2006.
[44] Mitchell DK, Peters J, Cannon J, Holtz C, Kinch J, Seeto P. Sustainable use option
plan for the Milne Bay community-based coastal and marine conservation
program PNG/99/G41 [UNOPS contract for services report ref: C00-1076];
2001.
[45] National Research Institute. Papua New Guinea district and provincial proles.
Port Moresby: the National Research Institute; 2010.
[46] Skewes T, Kinch J, Polon P, Dennis D, Seeto P, Taranto T, et al. Distribution and
abundance of reef resources in Milne Bay Province, Papua New Guinea: giant
calms and other species. Cleveland: CSIRO Division of Marine Research Final
Report; 2003.
[47] Bohensky E, Butler JRA, Mitchell D. Scenarios as models for knowledge
integration: ecotourism futures in Milne Bay, Papua New Guinea. In: Anderssen RS, Braddock RD, Newham LTH, editors. Proceedings of the 18th World
IMACS Congress and MODSIM09 International Congress on modelling and

J.R.A. Butler et al. / Marine Policy 46 (2014) 113

[48]

[49]

[50]

[51]

[52]

[53]

[54]

[55]
[56]

[57]

[58]

[59]
[60]
[61]

[62]
[63]
[64]

[65]

[66]

[67]

simulation: interfacing modelling and simulation with mathematical and

computational sciences; 2009, p. 282632.
Department of Environment and Conservation. PNG marine program on coral
reefs, sheries and food security 20102015: national plan of action on the
Coral Triangle Initiative. Port Moresby: Papua New Guinea Department of
Environment and Conservation, Ofce of Climate Change and Carbon Trade,
and National Fisheries Authority; 2010.
Mitchell RK, Agle BR, Wood DJ. Towards a theory of stakeholder identication
and salience: dening the principle of who and what really counts. Acad
Manage Rev 1997;22(4):85386.
Folke C., Fabricius C. Communities, ecosystems and livelihoods. In: Millennium
Ecosystem Assessment: ecosystems and human well-being sub-global assessments. Washington DC: Island Press; 2005. p. 26176.
Christensen JH, Hewitson B, Busuioc A, Chen A, Gao X, Held I, et al. Regional
climate projections. In: Solomon S, Qin D, Manning M, Chen Z, Marquis M,
Averyt KB, Tignor M, Miller HL, editors. Climate Change 2007: the physical
science basis. Contribution of Working Group I to the Fourth Assessment
Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press; 2007.
Chamberlain M, Matear R, Feng M, Sun C. Dynamic downscaling of a climate
change projection of ocean dynamics and biogeochemistry for the Australian
region [Technical Report II]. Canberra: WAMSI Node 2 Project 2; 2008.
Leisz SJ. A spatial analysis on locations in Melanesia most vulnerable to
climate change. Honolulu: Department of Natural Sciences, Bishop Museum;
2009.
Church JA, White NJ, Hunter JR, McInnes KL, Mitchell WM, OFarrell SP, et al.
Sea level. In: Poloczanska ES, Hobday AJ, Richardson AJ, editors. A marine
climate change impacts and adaptation report card for Australia. NCCARF
Publication 05/09. Grifth University, Queensland, Australia: National Climate
Change Adaptation Research Facility; 2009.
Polhemus DA. Climate change in New Guinea and its potential effects on
freshwater ecosystems. Bishop Mus Tech Rep 2009;42(3).
Richards K, Timmermann A. Regional climate change projections for the
southwestern Pacic with a focus on Melanesia. Bishop Mus Tech Rep
2009;42(1).
International Panel on Climate Change. Climate change 2007 the physical
science basis. Contribution of Working Group I to the Fourth Assessment
Report of the IPCC. Cambridge: Cambridge University Press; 2007.
Organisation for Economic Cooperation and Development. OECD core set of
indicators for environmental performance reviews. Paris: OECD Environment
Monographs, vol. 83; 1993.
Bowen RE, Riley C. Socio-economic indicators and integrated coastal management. Ocean Coast Manage 2003;46:299312.
National Statistical Ofce. Preliminary gures: Papua New Guinea census 2011.
Port Moresby: PNG National Statistics Ofce; 2012.
Bourke RM. An overview of food security in PNG. In: Bourke RM, Allen MG,
Salisbury JG, editors. Food security for Papua New Guinea. Proceedings of the
Papua New Guinea food and nutrition 2000 conference. ACIAR proceedings
no. 99. Canberra: Australian Centre for International Agricultural Research;
2001, p. 514.
Hallegatte S. Strategies to adapt to an uncertain climate change. Global
Environ Change 2009;19:2407.
World Bank. Convenient solutions to an inconvenient truth: ecosystem-based
approaches to climate change. Washington DC: World Bank; 2009.
Andrade Prez A, Herrera Fernandez B, Cazzolla Gatti R. Building resilience to
climate change: ecosystem-based adaptation and lessons from the eld.
Gland: International Union for the Conservation of Nature; 2010.
Bayas JCL, Marohn C, Dercon G, Dewi S, Piepho HP, Joshi L, et al. Inuence of
coastal vegetation on the 2004 tsunami wave impact in west Aceh. Proc Natl
Acad Sci USA 2011:186127.
Ofce of Climate Change and Development. Interim action plan for climatecompatible development. Port Moresby: Papua New Guinea Ofce of Climate
Change and Development; 2010.
McAdoo BG, Ah-Leong JS, Bell L, Ifopo P, Ward J, Lovell E, et al. Coral reefs as
buffers during the 2009 South Pacic tsunami, Upolu Island, Samoa. Earth Sci
Rev 2011;107:14755.

13

[68] Hills T, Brooks A, Atherton J, Rao N, James R. Pacic Island biodiversity,

ecosystems and climate change adaptation: building on nature0 s resilience
[Report for the Secretariat for the Pacic Regional Environment Programme
and Conservation International]; 2011.
[69] Butler JRA, Wong G, Metcalfe D, Honzak M, Pert PL, Bruce C, et al. An analysis
of trade-offs between multiple ecosystem services and stakeholders linked to
land use and water quality management in the Great Barrier Reef, Australia.
Agric Ecosyst Environ 2013;180:17691.
[70] Daw T, Brown K, Rosendo S, Pomeroy R. Applying the ecosystem services
concept to poverty alleviation: the need to disaggregate human well-being
[Working Paper 30, DEV Working Paper Series]. University of East Anglia, The
School of International Development; 2011.
[71] Butler JRA. The challenge of knowledge integration in the adaptive comanagement of conicting ecosystem services provided by seals and salmon.
Anim Conserv 2011;14:599601.
[72] Butler JRA, Middlemas SJ, Graham IM, Harris RN. Perceptions and costs of seal
impacts on salmon and sea trout sheries in the Moray Firth, Scotland:
implications for the adaptive co-management of Special Areas of Conservation. Mar Policy 2011;35:31723.
[73] Butler JRA, Tawake L, Tawake A, Skewes T, McGrath V. Integrating traditional
ecological knowledge and sheries management in the Torres Strait, Australia:
the catalytic role of turtles and dugong as cultural keystone species. Ecol Soc
2012;17(4):34.
[74] Butler JRA, Gunn R, Berry H, Wagey GA, Hardesty BD, Wilcox C. A value chain
analysis of ghost nets in the Arafura Sea: identifying trans-boundary stakeholders, intervention points and livelihood trade-offs. J Environ Manage
2013;3:1425.
[75] Hein L, van Koppen K, de Groot RS, van Ierland EC. Spatial scales, stakeholders
and the valuation of ecosystem services. Ecol Econ 2006;57:20928.
[76] Hayes G. Mirab processes and development on small Pacic Islands: a case
study from the Southern Massim, Papua New Guinea. Pac Viewp
1993;34:15378.
[77] Wood AL, Butler JRA, Sheaves M, Wani J. Sport sheries: opportunities and
challenges for diversifying coastal livelihoods in the Pacic. Mar Policy
2013;42:30514.
[78] Fazey I, Pettorelli N, Kenter J, Wagatora D, Schuett D. Maladaptive trajectories
of change in Makira, Solomon Islands. Global Environ Change 2011;21:
127589.
[79] Allen B, Bourke RM. El Nio Southern Oscillation (ENSO) and food supply. In:
Bourke RM, Harwood T, editors. Food and agriculture in Papua New Guinea.
Canberra: ANU E Press, The Australian National University; 2009. p. 627.
[80] Oliver-Smith A. Sea level rise, local vulnerability and involuntary migration.
In: In Piguet E, Pcoud A, de Gauchteneire P, editors. Migration and climate
change. Paris, Cambridge: UNESCO and Cambridge University Press; 2011.
p. 16088.
[81] Butler JRA, Skewes TD, Wise RM, Bohensky EL, Peterson N, Bou N, et al. West
New Britain futures workshop report 1819 July 2012 [Report prepared for the
Australian and Papua New Guinea Government under the Coral Triangle
Initiative on Coral Reefs, Fisheries and Food Security]. Canberra: CSIRO Climate
Adaptation Flagship; 2012.
[82] Butler JRA, Skewes TD, Wise RM, Bohensky EL, Peterson N, Bou N, et al. Bali
Witu Local Level Government futures workshop report summary 2627
November 2012 [Report prepared for the Australian and Papua New Guinea
Government under the Coral Triangle Initiative on Coral Reefs, Fisheries and
Food Security]. Canberra: CSIRO Climate Adaptation Flagship; 2012.
[83] De Groot R. Function-analysis and valuation as a tool to assess land use
conicts in planning for sustainable, multi-functional landscapes. Landsc
Urban Plan 2006;75:17586.
[84] Butler JRA, Radford A, Riddington G, Laughton RL. Evaluating an ecosystem
service provided by Atlantic salmon, sea trout and other sh species in the
River Spey, Scotland: the economic impact of recreational rod sheries. Fish
Res 2009;96:25966.
[85] TEEB. The economics of ecosystems and biodiversity: the ecological and
economic foundations. London: Earthscan; 2010.
[86] Kenter JO, Hyde T, Christie M, Fazey I. The importance of deliberation in
valuing ecosystem services in developing countries evidence from the
Solomon Islands. Global Environ Change 2011:50521.