An integrated and quantitative vulnerability assessment for proactive hazard response and sustainability a case study on the chan may lang co gulf area, central vietnam

C A S E R E P O R TAn integrated and quantitative vulnerability assessment for proactive hazard response and sustainability: a case study on the Chan May-Lang Co Gulf area, Central Vietn

C A S E R E P O R T

An integrated and quantitative vulnerability assessment

for proactive hazard response and sustainability: a case study

on the Chan May-Lang Co Gulf area, Central Vietnam

Mai Trong Nhuan• Le Thi Thu Hien•

Nguyen Thi Hoang Ha•Nguyen Thi Hong Hue•

Tran Dang Quy

Received: 5 December 2012 / Accepted: 27 May 2013

Ó Springer Japan 2013

Abstract A natural factors-based approach was

devel-oped to examine proactive responses to hazards and

improving sustainability on the Chan May-Lang Co Gulf

area, Central Vietnam The approach was based on a

weight-of-evidence method within an integrated and

quantitative vulnerability assessment in which the spatial

relationship between a set of evidential factors (lithology,

distance to the coastline, altitude, slope, aspect, drainage,

wind speed during storms, and land use and cover) and a

set of hazard locations was combined with the prior

probability (total vulnerability) to obtain the posterior

probability of hazard occurrence The result showed that

44.3 % of the study area had high to very high total

vul-nerability, due to the high density of vulnerable objects and

frequency of severe damage from typhoons, floods,

land-slides, and erosion The result also demonstrated that the

contribution of natural factors was directly proportional to

total vulnerability in approximately 75 % of the study area,

indicating a high dependence of vulnerability on natural

factors In the remaining areas, low contributions were

found in the high and very high vulnerability areas

domi-nated by high anthropogenic activities In contrast, natural

factors were important contributors to total vulnerability in areas characterized by dense vegetation, consolidated rocks, and altitude greater than 300 m, reflecting high natural resilience The present study demonstrated that a proactive approach may provide appropriate measures to mitigate hazards and to increase the sustainability of the study area

Keywords Chan May-Lang Co gulf area Hazard  Proactive response Sustainability  Vulnerability assessment Weight of evidence

Introduction The Vietnam coastal zone plays an important role in socio-economic development, territorial sovereignty protection, and maintenance of biodiversity in Vietnam However, this region is vulnerable to natural hazards (e.g typhoons, floods, coastal erosion, salinity intrusion, and landslides) and anthropogenic impacts (e.g population growth, excessive aquaculture, and overfishing) These threats have the potential to limit sustainable development in the Viet-nam coastal zone, through severe and widespread damage

to human life and property as well as degradation of natural resources and the environment (Nhuan et al.2011a) Vulnerability and sustainability are two contrasting aspects of a system, in which local vulnerability can affect the system sustainability in a resilience framework (Eakin and Wehbe 2009) Vulnerability is one of the central ele-ments of dialogue in science, decision-making, and sus-tainability research (Turner et al 2003) Appropriate adaptive and preparedness planning, and mitigation mea-sures implemented at an appropriate time help to reduce vulnerability and the risk from potential hazards, thus

Handled by Soontak Lee, Yeungnam University, Korea.

M T Nhuan (&)  N T H Ha  T D Quy

Department of Geo-environment, VNU University of Science,

334 Nguyen Trai, Thanh Xuan, Hanoi, Vietnam

e-mail: nhuanmt@vnu.edu.vn

L T T Hien

Institute of Geography, Vietnam Academy of Science

and Technology, Hanoi, Vietnam

N T H Hue

VNU Sea and Islands Research Centre, Vietnam National

University, Hanoi, Vietnam

DOI 10.1007/s11625-013-0221-9

increasing the sustainability of a system (Winograd2007).

Appropriate adaptation and effective mitigation of hazard

effects requires a detailed knowledge of the vulnerability of

an area to potential hazards (Cutter et al.2000)

A number of vulnerability assessment methods have

been suggested for particular hazards, such as sea level rise

(Torresan et al.2008), storms (Bosom and Jimenez2011),

floods (FAO 2004; Snoussi et al 2008), erosion (Boruff

et al 2005), and landslides (Szlafsztein and Sterr 2007;

Uzielli et al 2008) Recently, the importance of a

multi-hazard approach to risk management has been emphasized

(Kappes et al.2011) However, few studies have presented

an integrated approach to multi-hazard assessment (Cutter

et al.2000; Kappes et al 2011; Kumar et al 2010;

Ma-hendra et al 2011; Nhuan et al 2009, 2011a, b; NOAA

1999; Pratt et al 2005)

Vulnerability has been assessed by qualitative,

semi-quantitative, and quantitative methods Quantitative

meth-ods involve statistical, geotechnical, and artificial neural

network methods that reduce subjectivity and are more

easily reproduced One quantitative method, a weight of

evidence model, uses evidence from previous events to

predict the probability of hazards occurring in the study

area The relative importance of each line of evidence is

estimated by a statistical method, based on the available

data (Mathew et al 2007) However, this model is used

primarily for vulnerability assessment of landsides, rather

than for multi-hazard environments (Barbieri and Cambuli

2009; Mathew et al.2007)

An object-related approach creates a clear separation

between the biophysical or natural dimension and the

socio-economic dimension when assessing vulnerability

(Adger 1999; Cutter et al 2000; Nhuan et al 2009,

2011a,b) Almost all studies using such an approach have

performed a vulnerability assessment subsequent to

haz-ard occurrence Although such studies provide some

useful results, their ability to assess the adaptability of a

system and the timeliness of the response to hazards is

limited

Natural factors such as geology, geography, hydrology

and meteorology are important components that influence

the vulnerability of a region (Birkmann2006; Furlan et al

2011; Marchand 2009; Nhuan et al 2009, 2011a, b)

Determining the contribution of natural factors to

vulner-ability by applying the weight-of-evidence method

pro-vides a reliable base for assessing and forecasting the

vulnerability of a region This proactive, prediction-based

approach is a fundamental requirement for outlining

appropriate strategies for community response to hazards

(Mimura2008), hazard adaptation, and hazard mitigation

The prospect of a proactive approach highlights the need to

conduct appropriate research on which this approach is

based

The objective of the present study was to propose a new approach for assessing and forecasting vulnerability based

on natural factors and evidence that can create proactive responses to hazards and thus enhance sustainability Subsequently, the approach developed was applied to determine the contribution of natural factors to the total vulnerability of the Chan May-Lang Co Gulf area, Central Vietnam Proposed measures for hazard mitigation and improvement of sustainability are also discussed

Study area The Chan May-Lang Co Gulf area is located in Central Vietnam (Fig.1) It is approximately 711 km2in area, and

is surrounded by 18 communes There are two lagoons (Cau Hai and An Cuu) and two gulfs (Chan May and Lang Co), which are the most popular and important wetlands of the Central Vietnam coastal zone In addition, the study area is a key economic zone in Central Vietnam, as it is on shipping routes to northern and southern Asia Land use in the study area is divided between scattered forest (48.8 %), anthropogenic construction (17.0 %), agriculture (12.1 %), aquaculture (17.4 %), and others (4.7 %) (PLPC2010) The major igneous rocks are biotite granite, two-mica granite, aplite, pegmatite, and granite (Nhuan and Tien

1993, 2011b) There are four main types of sedimentary materials: marine–river sediments (maQ2), lagoon sedi-ments (bmQ2), and two types of marine sedisedi-ments (mQ21–2 and mQ2) The sediments are composed primarily of sand, sand–mud, mud–sand, mud, mud–clay, and clay The geomorphology is typically characterized by erosion– denudation relief in the mountain area, and mixed depo-sitional relief of alluvium, deluvium, and proluvium in the coastal plain

The study area is located within a distinct monsoon climate zone, with a rainy season from August to January, and a dry season from February to July Annual average rainfall level is 2,800 mm and the annual average tem-perature is 25 °C The average annual wind speed and maximum wind speed are 1.5 and approximately 40 m/s, respectively The prevailing wind directions are northwest

in winter (14–34 %) and south–southwest in summer (10–17 %)

Analysis of historical data shows that typhoons, land-slides, floods, and erosion are the most frequently occurring hazards and cause the most severe damage (MONRE

2008) Annually, there are 4–5 typhoons and tropical low-pressure storms, causing severe damage to property and loss of human lives (MONRE 2008) For example, Typhoon Tilda struck the Lang Co region on 22 September

1964 with wind speeds of 38 m/s and a storm surge of 1.7 m (MONRE2008) In addition, the Bach Ma mountain chain in the southwest of the study area affects the regional

rainfall regime, intensifying the occurrence of hazards For

example, in November 1999, severe rains caused a flood

and landslides which resulted in property damage in the

3,000 m2mountain area of the L Tien and L Vinh

com-munes (MONRE 2008) The flood and landslide hazards

threatened 50 households, and destroyed roads and

infrastructure

The study area is representative of the Central Vietnam

coastal zone, which is characterised by a contrast between

flat lagoons and river plains, and adjacent mountains

ran-ges The Central Vietnam coastal zone is experiencing

rapid economic development while facing increasing

nat-ural hazards Therefore, an integrated quantitative

vulner-ability assessment for proactive responses to hazards is

crucial to the continued development of the study area and

the Central Vietnam coastal zone

Methodology

Proactive approach

Previous vulnerability assessments and reduction measures

have used two major approaches: (1) post-event or (2)

pre-event A number of vulnerability assessments have

focus-sed on the former (e.g Adger1999; Snoussi et al 2008;

Uzielli et al 2008) This approach, shown in Fig.2, is

largely considered a passive response, as damage from the

event has already occurred In contrast, a proactive

approach would provide more effective and active

responses prior to any event occurring (Fig.2)

Analysis of natural factors in a region can provide evi-dence for the probability of a hazard occurring (Birkmann

2006; Furlan et al.2011; Marchand2009) Natural factors that may generate, intensify, or mitigate natural hazards include the geology, geography, hydrology, oceanography, meteorology, and land cover in the region For example, landslides can often be attributed to the local geology, geo-morphology, land cover, and drainage (Mathew et al.2007) Similarly, mean tidal range, coastal slope, rate of relative sea level rise, shoreline erosion or accretion rates, and mean wave height, are key indicators of erosion vulnerability (Boruff et al.2005) In addition, Furlan et al (2011) revealed that the geomorphology, geology, pedology, and vegetation are important criteria in assessing natural vulnerability Vulnerability assessments of multi-hazards based on all natural factors are extremely complex Therefore, evidence and damage of major hazards in the study area needs to be assessed and weighted, thus enabling less important factors

to be disregarded A deficiency of reliable data and infor-mation, which is a problem in developing countries such as Vietnam, can also restrict the assessment of all natural factors In addition, some attribute parameters (e.g., rain-fall, geodynamic features) are also disregarded due to the paucity of spatial differentiation in the small study area Therefore, in this study, several major parameters have been selected to assess the natural component of hazard vulnerability (Table 1) This selection was based on evi-dence from field surveys, existing data, further data anal-ysis, and spatial differentiation of parameters

As shown in Fig.2, the contribution of natural param-eters to total vulnerability was calculated using the Fig 1 Map showing the study area

weight-of-evidence method The relative contribution of

natural parameters is assumed to be constant and is used to

estimate the total vulnerability of an area when natural

parameters change

Total vulnerability assessment

Vulnerability is considered as the potential for loss or

damage to objects and systems from hazards (Cutter1996;

Cutter et al.2000; Mitchell1989; Nhuan and Tien2011b)

The vulnerability of natural and social systems has been

assessed using three components: danger level of hazards,

density of vulnerable objects, and resilience (Nhuan et al

2009,2011a) It is noteworthy that the level of probability

caused by a hazard depends on both the danger of the

hazard and the resilience of the system For example, given

a particular probability hazard, a region of low resilience

will experience more damage than a region of high resil-ience Damage caused by an event is considered to be a practical and reliable method for weighting evidence in vulnerability assessments

In this study, the total vulnerability of the Chan May-Lang Co Gulf area was evaluated using the following components: proportion of people evacuated per year, total economic losses, and density of vulnerable objects Each component was then divided into five levels based on the damage caused by hazards (for evacuations and economic losses) or the level of vulnerability (for density of vulner-able objects; Tvulner-able2) The number of people evacuated and the economic losses for the period 2004–2010 were determined from existing data and from field surveys conducted in 2010 Vulnerable objects included humans, natural resources, economic assets (agriculture, aquacul-ture, and tourism), and infrastructure (construction, roads,

Fig 2 Formal and proactive

approaches in vulnerability

assessment

Table 1 Parameters used to

assess natural dimension

vulnerability

Natural factors Parameters Hazard/resilience

to hazards

Calculated methods

Geology Lithology Landslide, erosion Classification of rock types based on

consolidated levels Geography Distance to coastline Typhoon, erosion Calculation for each cell the Euclidean

distance to the closest coastline

Slope Landslide, flood 3D analyst: interpolation of slope

landslide

3D analyst: interpolation of aspect Hydrology Drainage Flood, landslide Drainage density (km/km 2 ): length of the

stream channels per calculated unit area Meteorology Wind speed during

storms

Typhoon, flood, landslide

Classification of wind speed levels corresponding to the wind in the storm Land use and

cover

Land use and cover Flood, landslide,

erosion

Classification of land cover and land use patterns

and houses) Analysis and calculation of total vulnerability,

as well as the contribution of natural factors, were

per-formed using ArcGIS 10

Weight of evidence

The weight of evidence was based on a log-linear Bayesian

model using the prior and posterior probabilities (Jeffreys

1998) The method has been used for mineral potential

mapping (Agterberg et al 1990; Bonham-Carter et al

1989) and landslide hazard mapping (Barbieri and Cambuli

2009; Hien2010; Mathew et al.2007) This approach uses

the prior probability of an occurred hazard to find the

posterior probability based on the relative contribution of

the subject by evidence Prior and posterior probabilities of

a hazard (S), given the presence or absence of any binary

pattern (Bior Bi), are calculated using Eqs.1 and2:

PPrior¼ P Sf g ¼NpixðHazardÞ

and,

P SjBf ig ¼P Sf \ Big

P Bf gi ¼

NpixfS \ Big

where Npix (Hazard) and Npix (Total) are the number of

pixels affected by the hazard and the total number of pixels

in the study area, respectively

Positive and negative weights (wþi and w

i ) are devel-oped from these conditional probabilities as defined by

Eqs.3and4:

wþi ¼ logeP Bf ijSg

and,

wi ¼ logeP BijS

The difference between the positive and negative weights

is termed the contrast (Cw) for each parameter class and is

calculated to reflect the spatial combination between the

evidence of vulnerability and the occurrence of the hazard, as

shown in Eq.5(Barbieri and Cambuli2009):

In addition, Cw/S(Cw), where S(Cw) is the standard deviation, provides an indication of the reliability of the relationship calculated between the hazard parameters A higher Cw/S(Cw) value reflects a closer relationship between the hazard and the parameters used in the calculation (Barbieri and Cambuli2009)

In the present study, the spatial relationship between a set of evidential themes and a set of hazard locations is combined with the prior probability (total vulnerability) to derive the posterior probability of hazard occurrence This enables the contribution of natural factors to total vulner-ability to be calculated

Results and discussion Total vulnerability assessment

A total of 755 billion Vietnamese Dong (US $36 million) was lost in the period from 2004 to 2010 as a result of natural hazards in the study area (Table3; PLPC 2009) The damage from the hazards was scattered throughout the study area The highest economic losses occurred in several communes in the northwest of the Chan May-Lang Co Gulf area (L Bon, L Son, and L Dien communes) However, more than 90 % of the populations in the L Tri, L Tien, and

L Co communes were affected by the hazards (Table3) More than 20 % of the populations of the L Binh, L Vinh,

V Hien, and V Hai communes were evacuated each year (Table3)

Total vulnerability is shown in Fig.3 The vulnerability level is classified into 5 levels: very high (4–5), high (3–4), medium (2–3), low (1–2), and very low (0–1) These classes account for 10.0, 34.3, 12.8, 23.8, and 19.0 % of the Chan May-Lang Co Gulf, respectively The result showed that approximately 44.3 % of the study area has high to very high vulnerability levels, encompassing the coastal and the northwestern communes of the Chan May-Lang Co Gulf region (Fig.3) These regions have a high density of vulnerable objects and frequently suffer severe damage from typhoons (L Vinh and V Hai communes), floods

Table 2 Classification of

vulnerability criteria on the

Chan May-Lang Co Gulf area

Proportion of people evacuated per year

Value Total economic losses

(million VND/person)

Value Density of

vulnerable objects

Value

(L Vinh and L Tien communes), landslides (L Tien, L Son,

X Loc, and L Vinh communes), and erosion (L Vinh and V

Hai communes) Conversely, the regions with very low and

low vulnerability levels corresponded to areas that have a

medium density of vulnerable objects, but have suffered

little damage from natural hazards

Contribution of natural factors to total vulnerability The weight of evidence is shown in Tables4, 5, 6,

7, 8, 9, 10, 11 The weight of evidence was calculated for various parameter classes used in the study (Table1)

Table 3 Population, affected

and evacuated people, and

economic loss on the

Chan-Lang Co Gulf area due to

natural hazards in the period

from 2004 to 2010

Source: PLPC ( 2009 )

a

1 US dollar is equal to 20,850

VND (2012)

No Commune Population

(people)

Population density (people/km 2 )

Proportion of people affected per year (%)

Proportion of people evacuated per year (%)

Economic losses (million VND) a

Fig 3 Map of the total

vulnerability in the period from

2004 to 2010 on the Chan

May-Lang Co Gulf area

Among the lithological classes, the

shale–sandstone–con-glomerate has the highest w?

and Cw values (Table4), indicating that landslides and erosion could occur, resulting

in high vulnerability In contrast, the consolidated rocks

composed of biotite granite and binary granite have the

lowest w?

and Cwvalues

Distance to coastline

Previous observations indicate that areas close to the

coast-line experience more frequent and more intense coastal

erosion This parameter contributes significantly to vulnera-bility in areas located 0–0.7 km from the coastline (Table5) This result is in accordance with the high frequency of typhoons and erosion and high proportion of people evacu-ated in the L Vinh and V Hai communes (Table3)

Altitude The w?

and Cwvalues showed a positive correlation with vulnerability at 0–50 m altitude (accounting for approxi-mately 34 % of the study area; Table 6) It is noteworthy that the majority of the population and infrastructure are distributed within this altitude range Therefore,

Table 4 Weights and contrast

(square km)

w ?

w

-Contrast (Cw)

Cw/S(Cw)

a,am: loam/sandy/pebble-gravel 1 121 0.6104 - 0.1416 0.7520 40.7750 Biotite granite/binary granite 3 303 - 0.6266 0.3762 - 1.0027 - 62.3451 Gabbro/olivine gabbro/

gabbronorite

m,bm,vm: sand/calcareous sand/

coral/peat

m,m(v): sand/calcareous sand/

coral

Sandstone/siltstone/shale/

limestone

Shale/sandstone/conglomerate 14 38 1.7432 - 0.1123 1.8556 55.9522

Table 5 Weights and contrast

values for the distance to

coastline

Distance to coastline (km) Class Area (square km) w ?

w

-Contrast (C w ) C w /S(C w )

Table 6 Weights and contrast

values for the altitude Altitude (m) Class Area (square km) w

?

w

-Contrast (C w ) C w /S(C w )

Table 7 Weights and contrast

values for the slope Slope (°) Class Area (square km) w

?

w

-Contrast (C w ) C w /S(C w )

vulnerability is heightened due to the high density of

vul-nerable objects

Slope

Slopes of 0°–6° were found to be a significant contributor

to landslides and other hazards (Table7) This contradicts

the results reported by Mathew et al (2007) that slopes

under 30° were insignificant in terms of hazard

vulnerability The difference between the two studies is attributable to the high population density in areas of slope gentler than 6° in the present study area

Aspect The w?

and Cwvalues are high in regions with southerly, southeasterly, westerly, southwesterly, and northwesterly aspects (Table8) This pattern indicates that the prevailing

Table 8 Weights and contrast

values for the aspect Aspect (degree according

to the north direction)

Class Area (square km) w ?

w

-Contrast (Cw) Cw/S(Cw)

Table 9 Weights and contrast

values for the drainage Drainage (km/km

2 ) Class Area (square km) w ?

w

-Contrast (C w ) C w /S(C w )

Table 10 Weights and contrast

values for the wind speed during

storms

Wind speed during storms (m/s)

Class Area (square km) w ?

w

-Contrast (C w ) C w /S(C w )

Table 11 Weights and contrast

values for the land use and

cover

Land use and cover Class Area

(square km)

w ?

w

-Contrast (Cw) Cw/S(Cw)

Spare forest and afforestation 2 117 - 0.8012 0.1385 - 0.9397 - 46.0253

Agriculture/aquaculture/road 4 113 1.3047 - 0.2326 1.5374 74.0988

wind direction (northwest in winter and south–southwest in

summer) has a major influence on hazard vulnerability

Drainage

Drainage significantly influences slope stability by

con-trolling toe erosion and the saturation of slope material

(Gokceoglu and Aksoy 1996; Mathew et al 2007) The

efficiency of the river system also controls the extent of

flooding The intensity of hazards increased in areas where

drainage density ranged from 3.6 to 10.0 (Table9),

resulting in increased vulnerability This is due to the

distribution of these areas within regions of complex

topography The distribution of high drainage density in a

relatively flat area is considered to minimize the occurrence

of flash floods in that area

Wind

Wind speed during storms contributes significantly to

hazard intensity The maximum wind speed in storms

occurred most frequently in classes 1–3 The highest w?

and Cw values correlated to winds of 26.0 to 26.6 m/s

(Table10), showing that high wind speeds result in high

vulnerability This is due to substantial storm damage in

areas of high population density and low altitude, without

adjacent mountains acting as wind barriers

Land use and cover

Vegetation plays a crucial role in slope stability and the

regulation of surface flow In the absence of other factors,

areas of dense vegetation should be less susceptible to

landslides and erosion than bare areas The present results showed that the agriculture, aquaculture, roads, and human settlement had the highest contrast values (Table 11), reflecting high vulnerability associated with weakly cohe-sive materials (Mathew et al 2007) This result was sup-ported by the evidence of landslides observed in the northern L Tien, L Son, and L Vinh communes In addi-tion, these land uses were also classified as vulnerable objects, consequently enhancing their vulnerability The contribution of natural factors to total vulnerability is shown in Fig 4in which the negative and positive values indicate the low and high contribution The result showed that the contribution of natural factors was directly propor-tional to total vulnerability in approximately 75 % of the study area (Figs.3,4) This pattern reflected the fact that vulnerability is highly dependent on natural factors The result also indicated that social resilience was so low that it contributed little to resisting natural hazards in the study area Social resilience remains low as a result of an outdated forecasting system for hazards, low community awareness of hazards, and low income In contrast, social resilience is an important contributor to total vulnerability in developed countries (Boruff et al 2005; Cutter 1996; Harvey and Woodroffe2008; NOAA1999) In the high and very high vulnerability areas, two contrast trends of the contribution of the natural factors to total vulnerability were found The first trend showed a high contribution in the L Son, L Binh, southern L Tien, and southern L Tri communes (Figs.3,4) Natural factors were dominant in regions characterized by dense vegetation, consolidated rocks, and altitude greater than 300 m (Fig.4) This demonstrates the role of natural factors in enhancing natural resilience In contrast, natural factors contributed little to total vulnerability in the regions Fig 4 Contribution of natural

factors to total vulnerability on

the Chan May-Lang Co Gulf

area

dominated by high anthropogenic activities such as the

northern V Hai, L Vinh, X Loc, northern L Tien, and northern

L Tri communes (Fig.4)

The present study clearly demonstrates that natural factors

influence the resilience of both natural and socio-economic

systems Mangrove and terrestrial forests, mountainous areas,

consolidated rocks, and distance from the coast increase

nat-ural resilience Low elevation, unconsolidated rocks, high

wind speed, and natural hazards decrease natural resilience

The location of vulnerable socio-economic objects in these

areas of low natural resilience results in low socio-economic

resilience Based on this, appropriate measures for proactive

responses to hazards can be proposed to reduce this risk of

disaster, and increase the sustainability of the study area The

results of the present study indicate that proposed measures

should aim to increase social resilience Three groups of

solutions can be implemented to achieve this, as follows:

1 Natural vulnerability assessment and forecasting-based

planning such as sustainable resource use (Adger et al

2005); implementation of sustainable livelihood

solu-tions (e.g the Satoyama–Satoumi model, sustainable

economic development models, diverse agriculture,

eco-tourism, and community frameworks); locating

evacuation channels, technical infrastructure, and

social infrastructure in areas of low natural

vulnera-bility; installation of early warning systems in

high-vulnerability areas; and ensuring that vulnerable

communities have access to emergency health

ser-vices, safe havens, and evacuation channels

2 Management strategies, such as creating and

imple-menting proactive policies for responses to natural

hazards; and enhancing sustainability, adaptive

man-agement of wetlands, integrated community-based

coastal zone management (Nunn and Mimura 2007)

and integrated mountainous area management

3 Hazard mitigation plans, policies, and measures based

on the results of the present study such as installation

of updated early warning systems, policies for

proac-tive mitigation of hazards, afforestation and

reforesta-tion of mangrove areas, construcreforesta-tion of coastal

protection structures, and maintenance of the natural

sediment balance (Winchester et al.2007) In addition,

community awareness and education campaigns,

reg-ular training, and guidance materials should be

imple-mented with reference to natural hazards, disasters,

and factors contributing to vulnerability

Conclusions

Eight natural parameters (lithology, distance to coastline,

altitude, slope, aspect, drainage, storm wind speed, and

land use and cover) were used to evaluate the influence of natural factors on total vulnerability The contribution of natural factors was directly proportional to total vulnera-bility in approximately 75 % of the study area This result indicated that the vulnerability was highly dependent on natural factors In contrast, low contribution was found in the high and very high vulnerability areas dominated by high anthropogenic activities

The results of this study highlight the need for increas-ing resilience and sustainability of natural and socio-eco-nomic systems by implementing management practices, sustainable resource use planning, and proactive hazard mitigation measures Future research should focus on forecasting and verifying vulnerability based on natural and socio-economic factors Using a proactive approach to hazard response will help to increase the resilience and sustainability of important ecosystems such as coastal waters, marine ecosystems, and mangrove and terrestrial forests

Acknowledgments This research was supported by the Vietnam’s National Foundation for Science and Technology Development (NAFOSTED) (No 105.09.82.09) The authors gratefully acknowl-edge the People’s Committee of Phu Loc District, Thua Thien Hue Province (Vietnam), the VAST Institute of Marine Resources and Environment for their help with data collection.

References Adger WN (1999) Social vulnerability to climate change and extremes in coastal Vietnam World Dev 27(2):249–269 Adger WN, Arnell NW, Tompkins EL (2005) Successful adaptation

to climate change across scales Global Environ Chang 15:77–86 Agterberg FP, Bonham-Carter GF, Wright DF (1990) Statistical pattern integration for mineral exploration: In: Gaal G, Merriam DF (eds) Computer applications in resource estimation: predictions and assessment for metals and petroleum, Pergamon, Oxford, pp 1–21 Barbieri G, Cambuli P (2009) The weight of evidence statistical method in landslide susceptibility mapping of the Rio Pardu Valley (Sardinia, Italy) 18th World IMACS/MODSIM Con-gress, Cairns, Australia, 13–17 July

Birkmann J (2006) Measuring vulnerability to natural hazards: towards disaster resilient societies United Nations University Press, Tokyo

Bonham-Carter GF, Agterberg FP, Wright DF (1989) Weights of evidence modeling: a new approach to mapping mineral potential In: Agterberg FP, Bonham-Carter GF (eds) Statistical applications in the earth sciences, Canadian Government Pub-lishing Centre, pp 171–183

Boruff BJ, Emrich C, Cutter SL (2005) Erosion hazard vulnerability

of US coastal counties J Coastal Res 21(5):932–942 Bosom E, Jimenez JA (2011) Probabilistic coastal vulnerability assessment to storms at regional scale—application to Catalan Beaches (NW Meditrrranean) Nat Hazard Earth Sys 11:475–484 Cutter SL (1996) Vulnerability to environmental hazards Prog Hum Geog 20:529–539

Cutter SL, Mitchell JT, Scott MS (2000) Revealing the vulnerability

of people and places: a case study of Georgetown County, South Carolina Ann Assoc Am Geogr 90(4):713–737