Maritime Policy & Management

The flagship journal of international shipping and port research

ISSN: 0308-8839 (Print) 1464-5254 (Online) Journal homepage: https://www.tandfonline.com/loi/tmpm20

Mitigative and adaptive investments for natural

disasters and labor strikes in a seaport–dry port
inland logistics network

Zixiang Gong & Nan Liu

To cite this article: Zixiang Gong & Nan Liu (2019): Mitigative and adaptive investments for
natural disasters and labor strikes in a seaport–dry port inland logistics network, Maritime Policy &
Management, DOI: 10.1080/03088839.2019.1675195

To link to this article: https://doi.org/10.1080/03088839.2019.1675195

Published online: 15 Oct 2019.

Submit your article to this journal

Article views: 45

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=tmpm20
MARITIME POLICY & MANAGEMENT
https://doi.org/10.1080/03088839.2019.1675195

Mitigative and adaptive investments for natural disasters and

labor strikes in a seaport–dry port inland logistics network
Zixiang Gong1 and Nan Liu
Department of Data Science and Management Engineering, School of Management, Zijingang Campus, Zhejiang
University, Hangzhou, China

ABSTRACT KEYWORDS
Inland distribution is a crucial link between ocean shipping and terminal Inland logistics network; dry
consumers. However, such a distribution system is vulnerable to distur- port; mitigative and adaptive
bance caused by natural disasters and labor strikes. Pre-event invest- investment; natural disasters;
labor strikes
ments, specifically mitigative and adaptive investments, could alleviate
the adverse consequences of a disturbance to mitigate the possibility of
event occurrence and reduce damage after event occurrence, respec-
tively. This study investigates an inland distribution network that com-
prises one seaport and one major dry port. Moreover, this research
discusses the two existing management models in the network, namely,
centralized and decentralized systems. We analyze the optimal invest-
ment level when coping with natural disasters and labor strikes and
compare the results under different cases. Results show that decentraliza-
tion will increase the total expected cost, while cooperation is always
beneficial in terms of pre-event investment. However, an appropriate
cost-sharing plan is required to make the cooperation feasible for
a seaport and dry port.

1. Introduction
Maritime transport chains employ considerably large ships to deal with increasing demands and
facilitate low unit costs (Cullinane and Khanna 2000). With large vessels at the gateway ports,
hinterland operations must improve to completely achieve economies of scale. Despite heavy
investments, seaport congestion remains a severe problem that lowers the efficiency of the hinter-
land transport sector. Accordingly, dry ports have been introduced to relieve seaport congestion.
These ports function as buffers in the entire maritime supply chain. A dry port is also referred to as
intermodal freight center, inland port, inland clearance depot and container freight station
(Beresford and Dubey 1990; Roso, Woxenius, and Lumsden 2009). A dry port can be defined as
an inland intermodal terminal that is directly connected to seaports with high-capacity transport
corridors, where customers can directly leave or pickup their standardized units similar to a seaport
(Roso, Woxenius, and Lumsden 2009; Qiu, Lam, and Huang 2015). Typical examples of dry ports
include the Isaka dry port in Tanzania, Eskilstuna dry port in Sweden and Xi’an dry port in China
(Roso and Lumsden 2010; Beresford et al., 2012).
A seaport–dry port inland logistics network (SPDPILN) comprises one large-volume dry port and
multiple shippers located proximately to it (see Figure 1). This scenario is prevalent in practice. For
example, the Virginia Inland Port (VIP), which has a land area of 161 acres, serves 39 major

CONTACT Nan Liu [email protected] Department of Data Science and Management Engineering, School of Management,
Zhejiang University, Room 1005, Administration Building, Zijingang Campus, Hangzhou, China
1
Present address: Room1302, Administration Building, Zijingang Campus, Zhejiang University, Hangzhou 310058, China.
© 2019 Informa UK Limited, trading as Taylor & Francis Group
2 Z. GONG AND N. LIU

Figure 1. Structure of a seaport dry port logistics network.

companies located nearby. To reduce port congestion, a seaport will charge a storage fee if the
containers’ dwell time is longer than the free-time-limit. VIP functions as a seaport buffer. An
individual shipper could first deliver containers to VIP. Alternatively, containers may be directly
delivered to the seaport. Customer demand in the region is satisfied by a seaport and dry port (Qiu,
Lam, and Huang 2015). SPDPILN has two different management models. First, a dry port is initially
developed by a seaport to address port congestion. Thus, a seaport and dry port are run by the same
player. Guangzhou Seaport (GSP) attempted to expand its cargo handling process to follow the ‘One
Belt One Road’ policy proposed by the Chinese government. In 2015, GSP established the Kunming
dry port (KDP) to relieve congestion. Accordingly, GSP and KDP are managed by the Guangzhou
Port Authority. In the second model, a dry port is operated independently from the seaport. Actual
examples are Eskilstuna dry port in Sweden and Xi’an dry port in China.
However, SPDPILN shows vulnerability to various unexpected disturbances, which cause con-
siderable economic losses (Marufuzzaman et al. 2014). Lam and Su (2015) indicated that natural
disasters and labor strikes are the two main causes of disruption in SPDPILN. Thus, the current
study focuses on the two causes and discusses the mechanism of reducing the vulnerability in
SPDPILN. To avoid confusion, the investment to minimize the probability of a disaster occurrence
is called mitigative investment, whereas the investment to reduce the damage after a disaster
occurrence is termed adaptive investment.
Natural disaster is one of the major causes of disruption in SPDPILN. In March 2011,
a magnitude 9.0 earthquake hit Japan near the northeast coast of Honshu. This tremor was followed
by a series of tsunamis. After the disaster, nearly all ports on the northeast coast of Japan, including
Kashima, Sendai and Hachinoe, were shut down. Several other ports in the east, such as Hitachiwere
and Miyako, were damaged to different degrees. Approximately 1.3 million TEUs were affected by
the earthquake and subsequent tsunamis, thereby accounting for 7% of all the containers handled in
Japan in 2011 (Reuters 2011). Hurricane Sandy, which lasted from October 21 to 31 October 2012,
resulted in tremendous local economic loss and major damage to ports in the eastern US. Five ports
along the coasts, including the Port of New York and New Jersey, which is the third largest port in
the US, were severely affected by this disaster and closed for a certain period (Janić 2015). The
economic loss was estimated at US$1 billion. The likelihood of natural disaster occurrence is
 MARITIME POLICY & MANAGEMENT 3

determined by the geophysical and climatic characteristics of a region (Sapir, Hargitt, and Hoyois
2004). Evidently, the natural characteristics of a particular region cannot be changed. Thus, only
adaptive investment can be adopted to reduce the severity of a disaster. The concrete form of an
investment may include promoting beach nourishments, improving groins and seawalls, building
storm-surge barriers and increasing the height of causeways. For example, the Netherlands built the
Delta Works to protect an area of land around the Rhine–Meuse–Scheldt delta. The Delta Works
are designed to deal with the risk associated with an increase in sea level and climate change (Xiao
et al. 2015).
Labor strikes also threaten SPDPILN. In 2012, a strike was staged at the ports of Los Angeles and
Long Beach. The strike initially started from a single terminal at the Port of Los Angeles but
eventually affected the other terminals in this area. Consequently, 10 of the 14 terminals at the two
ports were shut down. Thousands of containers were stacked at the storage area and the economic
loss was estimated at US$1 billion daily (China News 2012). Table 1 summarizes the number of
labor strikes in the transportation, storage and communication sectors from 2004 to 2016. In
general, labor strikes are staged in developed and developing countries. Argentina encounters
nearly 134 strikes annually. The average number of labor strikes in several countries is relatively
high (marked as red), with European countries accounting for nearly half of these actions, while the
rest are relatively minimal (i.e. approximately 2.85 times annually).
Source: International Labour Organization, 2017; (Chen, Lam, and Liu 2018)
The damage caused by a labor strike can be alleviated through effective pre-event investment.
Human factors and labor laws are often the root cause of labor strikes (Lam and Su 2015).
Therefore, mitigative investment in organization management, such as enhancing human resource
management and reforming labor laws, can mitigate the occurrence of such an event. The other is to

Table 1. Statistical summary of labor strikes in transportation, storage, and communication sectors in 2004–2016.
Average number of event Average number of event
Countries per year Countries per year
Asia Japan 16.10 Africa South Africa 11.00
India 12.13 Tunisia 8.00
Israel 4.85 Nigeria 3.33
Philippines 1.73 Central African 2.50
Republic
Hong Kong, China 1.43 Algeria 2.00
Sri Lanka 2.00 Oceania Australia 30.00
Occupied Palestinian 1.00 New Zealand 3.71
Territory
Europe Spain 99.69 America Argentina 133.64
Denmark 72.92 Canada 22.62
Italy 96.17 Brazil 54.50
Germany 60.14 Chile 13.88
UK 30.25 Peru 7.67
Finland 20.77 Mexico 4.30
Poland 18.80 US 2.20
Portugal 34.25 EI Salvador 2.50
The Netherlands 6.73 Uruguay 20.00
Cyprus 5.20 Jamaica 3.50
Sweden 2.91 Barbados 2.50
Ireland 2.00 Puerto Rico 2.00
Hungary 3.20 Costa Rica 1.00
Norway 3.00 Panama 1.00
Malta 1.86 Ecuador 1.00
Russian Federation 2.60 Dominican 1.00
Republic
Switzerland 1.29
Romania 2.67
Ukraine 1.75
Slovakia 2.50
4 Z. GONG AND N. LIU

make adaptive investment, such as arranging redundant resources, to control the adverse con-
sequences once labor strikes happen. In our interview with the director of the Port Management
Office, Transportation Commission of Ningbo, China, he mentioned that Ningbo port encountered
a labor strike in 2014 initiated by the container truck drivers outside the port area. The strike was
staged because of the complaints of truck drivers on a low freight rate. However, Ningbo port had
already established contingency mechanism with a redundant truck fleet and other resources. Thus,
this port responded immediately, maintained normal operations and mitigated the effects of the
strike.
The proposed model analyzes the investment strategy on disaster prevention against natural
disasters and labor strikes. Hence, the following questions have emerged in the context of deciding
the optimal investment level:

(1) Given the mitigative and adaptive investment, how do seaport and dry port operators decide
the optimal investment strategy to adapt to natural disasters and labor strike?
(2) What is the difference between a centralized system (i.e., seaport and dry port operated by
the same player or the damage of SPDPILN shared by the players) and decentralized system
(i.e., seaport and dry port are operated by different players) in terms of investment strategy
on disaster prevention?

The remainder of this paper is organized as follows. Section 2 presents a brief review of the related
literature. Section 3 mathematically models the issue. Section 4 analyzes the model and calculates
the optimal investment amount under different cases. Section 5 proposes several theorems by
comparing the equilibrium results under different cases. Section 6 concretizes the results of the
proposed model and offers management guidance. Lastly, Section 7 provides the conclusion of this
study.

2. Literature review
Several streams of literature are related to the present study. The first stream involves inland
transport development and dry port-related studies. Slack (1999) stated that the function of satellite
terminals it to ensure that major transport terminals can cope with traffic expansion without having
to undergo major site expansion. Hesse and Rodrigue (2004) considered transport as an integrated
demand where physical distribution and materials are interdependent. By defining physical dis-
tribution and materials management, they provided an analysis of the evolution of logistics.
Substantial attention has been provided to the inland transport sector within the entire maritime
logistics system. Notteboom and Rodrigue (2005) stated that current efficiency improvements in
maritime transportation are generally derived from inland distribution. Thus, inland accessibility
has become a cornerstone in port competitiveness (CEMT 2001). Several studies have addressed the
issue of hinterland access within the port competition context. Zhang (2009) examined the inter-
action between port competition and hinterland access. He explained that when ports compete in
quantities, an increase in corridor capacity will reduce the rival port’s output. Wan, Zhang, and
Yuen (2013) conducted an empirical study to investigate the impact of road congestion on the
competition among major container ports in the US. Their results indicated that delays on urban
roads and road congestion delays around the port have a significant impact on a port’s throughput.
Dry ports play an essential role in hinterland development and intermodal transportation system.
Several reviews have been conducted on the application of different techniques on intermodal
transport studies and the effects of dry port to different stakeholders (e.g., Roso and Lumsden 2010).
The second stream of literature is related to port supply chain disruption management. Previous
studies have focused on issues related to climate change (e.g., Yang et al. 2017; Ng et al. 2018) or
post-disaster reconstruction and optimal relief response under a certain budget (e.g., Chen,
Cullinane, and Liu 2017). Several studies have attempted to systematically quantify the risk in
 MARITIME POLICY & MANAGEMENT 5

port facility security assessment and enhance the resilience of the system. Chang and Nojima (2001)
developed post-disaster system performance measures and applied them to the transportation
system in Kobe, which was damaged by an earthquake in 1995. Wakeman (2013) investigated
Hurricane Sandy and attempted to uncover lessons for port resilience. He gathered information by
interviewing key stakeholders at New York Harbor and provided several recommendations to
improve the existing design codes for infrastructure. The prior theory relating to investment on
disruption probability mitigation in safety regulation and supply chain management has been
considerably studied (e.g., Bakshi and Kleindorfer 2009). However, studies on port disruption
from the predisaster investment perspective have been relatively rare. Xiao et al. (2015) proposed
an economic model for a marine disaster prevention mechanism between a port and its tenant.
They discussed the ideal time for investing and their model showed that regulatory intervention is
not consistently optimal when the decision-maker lacks sufficient information on disaster prob-
ability distribution. Wang and Zhang (2018) established a game theoretic model between two ports
competing for a common hinterland and studied the disaster adaptation investments issue. Their
model indicated that the expectation and variance of disaster occurrences have positive and
negative impact, respectively, on port adaption.
The third stream of literature is the governance issue in the hinterland development process.
Notteboom and Rodrigue (2005) identified several major governance issues that port authorities
and other stakeholders confront in the port regionalization phase. In particular, they indicated that
over-optimism and slow start are the main concerns. Privatization and corporatization schemes
have redefined the role of the public sector in the port industries (Goss 1990). Beresford et al. (2012)
studied the dry port development in China and determined that the regulatory framework is
complex. The central government has dealt with institutional matters, whereas municipal bodies
control the local decision-making process. Rodrigue et al. (2010) discussed the functions of inland
ports in Europe and North America and concluded that these ports are operated and planned by
multiple actors, including the public and private sectors. Several case studies have investigated the
spatial evolution and government policies related to dry port development within a certain country.
Padilha and Ng (2012) studied the evolutionary pattern of dry port configuration in Brazil. They
explained that institutional barriers have led to substantial deviations in the evolution of Brazilian
dry ports, thereby suppressing the development of efficient multimodal supply chains within Brazil.
Several studies have analyzed the development of dry ports in India (Ng and Cetin 2012). The result
shows that dry ports in this country are more cluster-oriented than those in advanced economies
owing to geographical diversifications.
The last stream of literature is network theory related to shipping network vulnerability. Hall
(2004) proposed an interesting theory that the estimated economic loss of a certain disruption
within the entire shipping network may be inflated. By mentioning the example of the 2002 West
Coast Port lockout, he stated that the possibility of substitution could reduce the impact of
disruption. Lee (2010) conducted an empirical study on the basis of a database of vessel movements
in the Northeast Asian liner network from 1996 to 2006. Their study revealed a strong relationship
between local port policies and the evolution of shipping network design. Ducruet and Lugo (2013)
adopted network analysis to bridge shipping and logistics studies and cities and urban research.
They emphasized that the current state of shipping and logistics research could be improved using
the literature related to cities and transport networks. Ducruet (2016) investigated the role of the
two major interoceanic canals (i.e., Suez and Panama) in global shipping flows and measured with
network topological properties and geographic coverage. The results show a decreasing importance
of canal shipping in the context of the increasing south–south trade exchanges.
Therefore, the locational and institutional problems related to dry port development have been
receiving increasing attention in the literature. However, only a few of these studies have analyzed
the vulnerability of the transportation network under several unconventional events, such as
natural disasters and labor strikes. Many studies have focused on risk quantity in the port sector
but only limited research has focused on pre-event investment in alleviating the adverse
6 Z. GONG AND N. LIU

consequence of a potential disturbance. Meanwhile, no research has investigated this issue within
SPDPLIN. Thus, our research intends to fill in the literature gap.

3. Model description
We consider the pre-event investment strategy in SPDPILN, in which a seaport operator (herein-
after ‘seaport’) and dry port operator (hereinafter ‘dry port’) can make a pre-event investment to
alleviate the damage of a potential disturbance caused by natural disasters or labor strikes. The
occurrence probability of such a disturbance in SPDPILN is denoted by a random variable x. The
expected damage received by a seaport and dry port are given by Ds x and Dd x, respectively, where
Ds and Dd are the two given parameters that reflect the maximum possible financial loss. The
precise values of Ds and Dd can be derived on the basis of the historic data (see Xiao et al. 2015;
Bakshi and Kleindorfer 2009).
For natural disasters, the occurrence probability x is represented with a uniform distribution in
the range of ½0; θ; ðθ < 1Þ. Given that pre-event investment will not reduce the occurrence prob-
ability of a natural disaster, θ is an exogenous parameter that reflects the geophysical and climatic
characteristics of a region. In addition, θ could be interpreted as the vulnerability of the network,
which is the estimated maximum disaster occurrence probability. A seaport and dry port can only
make adaptive investment (denoted by ys and yd ) to reduce the damage after a natural disaster
occurrence. Customer demand is satisfied by a seaport and dry port. A dry port functions as
a seaport buffer. Thus, the adaptive investment made by dry port yd will cause an indirect effect on
seaport. Similarly, ys also affects the dry port. Therefore, if a disaster occurs, then the adaptive
investments reduce damages to a seaport and dry port as follows:
αys þ βyd and αyd þ βys ;
where α and β are positive parameters that satisfy α > 1 and α > β > 0 and α > 1 indicates that the
damage reduction is higher than the investment. Otherwise, investing before a disaster would be
pointless because investment costs more than the possible damage reduction. Meanwhile, α > 0and
β > 0 imply that a seaport and dry port can benefit from adaptive investment from damage
reduction. In addition, they can benefit from their own investment and the other side’s investment.
Accordingly, α > β indicates that the damage reduction is higher than that emerging from the other
side’s investment (Xiao et al. 2015). The adaptive investments bring no additional benefit other than
reducing disaster damage to zero. Therefore, the actual damages received by a seaport and dry port
are as follows:
Maxf0; Ds x  ðαys þ βyd Þg and Maxf0; Dd x  ðαyd þ βys Þg:
Mitigative and adaptive investments are considered for labor strikes. On the one hand, a seaport
and dry port can make adaptive investment (arrange redundant resources) to control the adverse
consequence once a labor strike occurs. On the other hand, they can also make mitigative invest-
ment (i.e., enhancing human resource management, reforming labor laws), which is denoted by zs
and zd . The occurrence probability x follows a uniform distribution in the range of ½0; θs  and ½0; θd ,
where θs and θd are two endogenous variables dependent on zs and zd (i.e., θs ;θs ðzs Þ; θd ;θd ðzd Þ).
The objective of a seaport and dry port is to minimize the sum of their investments and expected
damage caused by natural disasters or labor strike. SPDPILN has two different management models,
namely, decentralized and centralized systems. In a decentralized system, a dry port is operated

Table 2. Four different cases in making pre-event investment decisions.

Decentralized system Centralized system
Natural disasters (Only make adaptive investment ys and yd ) Case 1 Case 2
Labor strikes (Make both mitigative and adaptive investment zs ; zd ; ys ; and yd .) Case 3 Case 4
 MARITIME POLICY & MANAGEMENT 7

independently from the seaport. Each port minimizes its own expected cost, denoted as Cs and Cd .
In a centralized system, a seaport and dry port are run by the same player. Therefore, their pre-event
investment decisions are coordinated to minimize the joint cost, denoted by C ¼ Cs þ Cd .
Accordingly, we develop the model with four cases (see Table 2).
The effect of coordination is analyzed by comparing the results in Case 1 (Case 3, respectively)
with the results in Case 2 (Case 4, respectively). We are optimistic that we can identify if
centralization is beneficial in terms of pre-event investment in alleviating the adverse consequence
of natural disasters and labor strikes. Sections 4 and 5 perform these analyses.

4. Solutions and analysis

4.1. Investment in alleviating the damage of natural disasters
This subsection discusses Cases 1 and 2, in which a seaport and dry port choose their adaptive
investment levels to alleviate the damage from natural disasters.

4.1.1. Case 1
In this case, a seaport and dry port decide their optimal investment levels independently by
attempting to minimize their expected costs. The decision problems are as follows:
θ
MinCs 1 ¼
0 maxf0; Ds x  ðαys þ βyd Þgdx þ ys (1)

θ
MinCd 1 ¼
0 maxf0; Dd x  ðαyd þ βys Þgdx þ yd (2);
where superscript ‘1’ indicates the case number. The optimal investment levels in the different cases
are calculated separately. For ease of notation, we define Ms ¼ αys þ βyd and Md ¼ αyd þ βys . In
simplifying (1) and (2), for i ¼ fs; dg, one has:
(

θ
θ
ðD x  M Þdx; M  D θ 1
ðD θ  Mi Þ2 ; Mi  Di θ

0 maxf0; Di x  Mi gdx ¼ Mi =Di i i i i ¼ 2Di i (3):
0; Mi > Di θ 0; Mi > Di θ

For a seaport and dry port, investing in more than the maximum possible financial loss Di θ is
pointless. Thus, the decision problem in Equation (3) can be simplified as follows:.
1
Minys Cs 1 ¼ ðDs θ  Ms Þ2 þ ys (4)
2Ds

s:t: Ms ¼ αys þ βyd  Ds θ

1
Minyd Cd 1 ¼ ðDd θ  Md Þ2 þ yd (5)
2Di

s:t: Md ¼ αyd þ βys  Dd θ

We can assume the existence of an interior solution by disregarding the constraints in Equations
(4) and (5) and deriving the solution adopting the first-order condition approach. Thereafter, we
determine the conditions to be true. The first-order conditions of a seaport and dry port are as follows:
(
1  ðDs θM
Ds
s Þα
¼0
(6)
1  ðDd θM
Dd
d Þα
¼0

By solving Equation (6), we derive the equilibrium investment levels as follows:

8 Z. GONG AND N. LIU

8
< ys 1 ¼ Ds ðα θαÞþD d ðβαβθÞ
2

αðα2 β2 Þ
(7):
: yd 1 ¼ Dd ðα θαÞþDs2ðβαβθÞ
2

αðα2 β Þ

From Equation (7), one has:

(
αys 1 þ βyd 1 ¼ ðαθ1
α
ÞDs
 Ds θ
(8)
αyd 1 þ βys 1 ¼ ðαθ1
α
ÞDd
 Dd θ

Equation (8) indicates that the constrains in Equations (4) and (5) are satisfied. Therefore, ys 1 and
yd 1 are the interior equilibrium investment for a seaport and dry port, respectively. The second-
order conditions have been verified to hold for all the optimal solutions obtained in the presented
model. The details are not reported for brevity. The expected cost incurred with a seaport (Cs 1 Þ and
dry port (Cd 1 Þ can be derived by converting Equation (7) into Equations (4) and (5).

4.1.2. Case 2
A centralized system is considered in this case, in which the investments made by a seaport and dry
port are coordinated. The aim is to minimize the expected total cost. Thus, the decision problem is
provided as follows:
θ θ
MinC2 ¼ Cs 2 þ Cd 2 ¼
0 maxf0; Ds x  Ms gdx þ
0 maxf0; Dd x  Md gdx þ ys þ yd (9)
Equation (3) indicates that the investment decision regions can be divided into four regions (see
Figure 2). Region ① indicates that Mi  θDi are satisfied for i ¼ fs; dg and region ④ indicates that
Mi > θDi for i ¼ fs; dg. The conditionMd > θDd ; Ms  θDs and Ms > θDs ; Md  θDd are reflected in
regions ② and ③.
8 1 2 2
>
> 2Ds ðDs θ  Ms Þ þ 2Dd ðDd θ  Md Þ þ yd þ ys ; ðyd ; ys Þ
1 ①
>
< 2
2Ds ðDs θ  Ms Þ þ yd þ ys ; ðyd ; ys Þ ②
1
Minys ;yd C2 ¼ (10)
> 2
> 2Dd ðDd θ  Md Þ þ yd þ ys ; ðyd ; ys Þ ③
1
>
:
ys þ yd ; ðyd ; ys Þ④

Figure 2. Investment decision regions in case 2.

 MARITIME POLICY & MANAGEMENT 9

Therefore, the objective function in Equation (9) can be rewritten as follows:

where the investment decision regions ①-④ are defined in Figure 2. By solving the constrained
optimization problems in four different regions with the first-order conditions @C
2
yi ¼ 0; i ¼ fs; dg,
we obtain the optimal solutions in Equation (10) as follows:
8
< ys 2 ¼ Ds αðαθþβθ1ÞþDd βð1αθβθ Þ
ðαβÞðαþβÞ2
(11)
: yd 2 ¼ Dd αðαθþβθ1ÞþDs βð21αθβθÞ
ðαβÞðαþβÞ

The expected total cost is as follows:

2ðDs þ Dd Þðαθ þ βθ  1Þ
C2 ¼ (12)
2ðα þ βÞ2

4.2. Investment in alleviating the damage of labor strikes

This subsection discusses Cases 3 and 4, where a seaport and dry port choose their mitigative and
adaptive investment levels to alleviate the damage from labor strikes. The occurrence probability
can be lowered by adopting mitigative investment. Thus, in Cases 3 and 4, the occurrence
probability is an endogenous variables (i.e., θs ;θs ðzs Þ; θd ;θd ðzd Þ).
To avoid confusion, we define φ as the default occurrence probability of a labor strike in
SPDPLIN, thereby indicating the occurrence probability without any mitigative investment. In
reality, φ depends on several exogenously determined factors, such as the geographical location of
SPDPLIN. For example, a strike is more likely to be staged in ports in developed countries than in
developing countries. We define these exogenous factors as ‘vulnerability’ and are represented as φ.
An SPDPLIN with high vulnerability means that the risk exposure of that to labor strike is high.

(i) The marginal occurrence probability of a labor strike is assumed to be strictly decreasing in
investment.
(ii) For the same investment level, the occurrence probability is high for SPDPLIN with a high
level of vulnerability: θi ðzÞ < θj ðzÞ; φi < φj ; "z 2 Rþ .

(iii) Initially, SPDPLIN with a high level of vulnerability is significantly responsive to incre-
  @θ ðzÞ
mental investment: @θ@zi ðzÞ <  @zj  ; φi < φj .
z¼0 z¼0

An exponential functional form is adopted in the current study to depict the relationship
between mitigative investment and occurrence probability of a strike:

θi ¼ φexpðγzi Þ (13)

where θi ; zi ; i ¼ fs; dg are the occurrence probability of a labor strike and mitigative investment in
a seaport and dry port, respectively. This model is extensively adopted in academic research. Bakshi
and Kleindorfer (2009) discussed risk management in a supply chain using this equation. Major
(2002) applied this model to study the probability of a terrorist attack. Although the background of
various studies may vary with different research topics, the intention to quantify the relationship
between investment level and probability of occurrence in small probability events is similar (Table
1 indicates that a labor strike can be considered a small probability event in a certain SPDPLIN).
Moreover, γ is a positive parameter that represents the marginal investment efficiency and is
affected by such factors as local educational and economic development levels. The three assump-
tions can be verified via Equation (13), while θs and θd are equal to the default occurrence
probability φ when zs ¼ zd ¼ 0.
10 Z. GONG AND N. LIU

4.2.1. Case 3
In this case, a seaport and dry port decide their optimal mitigative and adaptive investment level by
independently attempting to minimize their respective expected costs. The decision problems are as
follows:
θs
MinCs 1 ¼
0 maxf0; Ds x  ðαys þ βyd Þgdx þ ys þ zs (14)

θd
MinCd 1 ¼
0 maxf0; Dd x  ðαyd þ βys Þgdx þ yd þ zd : (15)
With Equation (3), for i ¼ fs; dg, the problem in Equations (14) and (15) can be simplified as
follows (Given that it makes no sense for a seaport and dry port to invest more than maximum
possible financial loss Di θi ; i ¼ fs; dg):.
1
Minys; zs Cs 3 ¼ ðDs θs  Ms Þ2 þ ys þ zs (16)
2Ds

s:t: Ms ¼ αys þ βyd  Ds θs

1
Minyd ;zd Cd 3 ¼ ðDd θ  Md Þ2 þ yd þ zd (17)
2Di

s:t: Md ¼ αyd þ βys  Dd θd

By disregarding the constraint in Equations (16) and (17) and adopting the same method in
obtaining Equation (6), the optimal mitigative and adaptive investment levels can be derived as
follows (Detailed calculation process is also provided):
   
1 α 1 α
zs ¼  ln
3
; zd ¼  ln
3
(18)
γ Ds γφ γ Dd γφ

Dd β  D s α α Ds β  Dd α α
ys 3 ¼  2 
2 þ ; yd 3 ¼  2 
2 þ (19)
α α β γðα þ βÞ α α β γðα þ βÞ

4.3. Proof of Equation (18) and (19)

By converting Equations (13) into (16) and 17, we have:.
   
1 1 #s 1 1 #d
Cs 3 ¼ ðDs θs  Ms Þ2 þ ys  ln ; Cd 3 ¼ ðDd θd  Md Þ2 þ yd  ln
2Ds γ φ 2Dd γ φ
Using the envelope theorem yields the first-order conditions, we derive:.
1 1
Ds θs  αys   βyd   ¼ 0; Dd θd  αyd   βys   ¼0
γθs γθd
For a given value of θs and θd , Equation (6) indicates:.
ðDs θs  Ms Þα ðDd θd  Md Þα
1 ¼ 0; 1  ¼0
Ds Dd
By combining them, Equations (18) and (19) are obtained. □
 MARITIME POLICY & MANAGEMENT 11

4.3.1. Case 4
In a centralized system, a seaport and dry port make coordinated investment decisions to minimize
the expected total costs. The decision problem is given as follows:

θs θd
MinC4 ¼ Cs 4 þ Cd 4 ¼
0 maxf0; Ds x  Ms gdx þ
0 maxf0; Dd x  Md gdx þ ys þ yd þ zs þ zd
(20)

A similar decision region as that in Case 2 can be obtained in this case (see Figure 3). The meanings
of four different regions are similar to those in Figure 2.
The decision problem is converted into Equation (21):

8 1 2 2
> 2D ðDs θs  Ms Þ þ 2Dd ðDd θd  Md Þ þ yd þ ys þ zd þ zs ; ðyd ; ys ; zd ; zs Þ ①
1
>
>
< s 2
②
2Ds ðDs θs  Ms Þ þ yd þ ys þ zd þ zs ; ðyd ; ys ; zd ; zs Þ
1
Minys ;yd ;zs ;zd C4 ¼ (21)
>
> 1 2
ðDd θd  Md Þ þ yd þ ys þ zd þ zs ; ðyd ; ys ; zd ; zs Þ ③
>
: 2D d
ys þ yd þ zs þ zd ; ðyd ; ys ; zd ; zs Þ ④

Thereafter, we analyze the optimal decision problem in Equation (21) separately from regions ①
to ④. The optimal solution is obtained as follows:

Dd β  Ds α 1 Ds β  D d α 1
ys 4 ¼ þ ; yd 4 ¼ þ (22)
ðα  βÞðα þ βÞ2 γ ðα  βÞðα þ βÞ2 γ

   
1 αþβ 1 αþβ
zs 4 ¼  ln ; zd 4 ¼  ln (23)
γ γDs φ γ γDd φ

Figure 3. Investment decision regions in case 4.

12 Z. GONG AND N. LIU

4.4. Proof of Equations (22) and (23)

Meanwhile, verifying that no interior optimal solution in decision regions 173174175is easy. Given
the decision region 172, the decision problem is simplified as follows by converting Equation (13)
into Equation (21):
   
1 2 1 2 1 #s 1 #d
Minys ;yd ;zs ;zd C ¼
4
ðDs θ s  M s Þ þ ðDd θd  Md Þ þ yd þ ys  ln  ln
2Ds 2Dd γ φ γ φ
The envelope theorem yields the first-order conditions as follows:.
1
Ds θs  αys   βyd   ¼0
γθs

1
Dd θd  αyd   βys   ¼0
γθd
From Equation (10), under given θs and θd , one has:.
ðDs θs  αys  βyd Þα ðDd θd  αyd  βys Þβ
1  ¼0
Ds Dd

ðDs θs  αys  βyd Þβ ðDd θd  αyd  βys Þα

1  ¼0
Ds Dd
Thereafter, we derive the optimal solution in Equations (22) and (23) combined with Equation
(13). We convert the result into the constrains as follows:.
αþβ
αys 4 þ βyd 4  θs 4 Ds ¼  <0
γDs

αþβ
αyd 4 þ βys 4  θd 4 Dd ¼  <0
γDd
Therefore, Equations (22) and (23) are the interior optimal mitigative and adaptive investments
for a seaport and dry port. □

5. Comparison of investment levels under different cases

This section further analyzes the information in Sections 3 and 4. The following theorems are
proposed by comparing the results under different cases.

Theorem 1. If the occurrence of a natural disaster in a certain SPDPILN is high, then

a decentralized system will reduce the incentive for a seaport and dry port to invest in alleviating
the damage caused by natural disasters.
Proof:
Simplifying ys 1 and yd 1 as:

ys 1 ¼ ðDs αD d βÞðαθ1Þ

αðα2 β2 Þ
, yd 1 ¼ ðDd αDs βÞðαθ1Þ
αðα2 β2 Þ
.

From Equations (7) and (11):

ys 2  ys 1 ¼ αððDs αDd βÞβ

αþβÞðα2 β2 Þ
, yd 2  yd 1 ¼ αððDd αDs βÞβ
αþβÞðα2 β2 Þ
.
 MARITIME POLICY & MANAGEMENT 13

Given that ys 1  0; yd 1  0. If θ  α1 , Ds α  Dd β  0; Dd α  Ds β  0, then ys 2  ys 1 and

yd  yd 1 . □
2

The management understanding behind this theorem is logical and intuitive. If a natural disaster
is likely to happen in a certain SPDILN, then a centralized system would motivate additional
investments. Operators decide to cooperate when their objective is to minimize the total expected
costs. Therefore, they will consider the strategy of other ports when deciding on the investment
level. This measure results in adaptive investments of a seaport and dry port.

Theorem 2. For labor strikes, decentralization will increase the mitigative investment amount.
However, the adaptive investment amount is not related to the management model.

Proof:
From Equations (18) and (23), zs 3 > zs 4 and zd 3 > zd 4 can be verified easily. However,
Equations (19) and (22) indicate as follows:

ys 4  ys 3 ¼ βðDγα
s αγDd βγþα α βÞ
, yd 4  yd 3 ¼ βðDγα
d αγDs βγþα α βÞ
3 2 3 2

ðαþβÞðα2 β2 Þ ðαþβÞðα2 β2 Þ

The sign of the preceding equations cannot be determined. □

In a centralized system, a seaport and dry port share the cost. From a seaport’s perspective,
reducing its own labor strike occurrence does not guarantee a total cost reduction because such
a loss incurred by a labor strike in a dry port could also cause damage. Therefore, the incentive to
make mitigative investment is reduced compared with that in a decentralized system. This scenario
is similar for the dry port. For adaptive investment, the result depends on the value of β. When β is
sufficiently low, additional adaptive investments will be placed in a centralized system. In
a decentralized system, substantially adaptive investment merely reduces its own damage if β is
sufficiently low. However, cost is shared in a centralized system. Therefore, a seaport and dry port
will benefit from their respective investments and the other side’s investments. Moreover, they have
additional incentives to make adaptive investments.

Theorem 3. A centralized system is more beneficial than a decentralized system in terms of

investing in the alleviation of the adverse consequences of natural disasters and labor strikes in
SPDPLIN. Coordination is consistently effective in reducing the total expected cost
(i.e., C2 < Cs 1 þ Cd 1 , C4 < Cs 3 þ Cd 3 ).

Proof:
Cs 1 and Cd 1 can be obtained by converting Equation (7) into Equations (4) and (5). With
Equation (13):.

ðDs þ Dd Þβ2
Cs 1 þ Cd 1  C2 ¼ >0
2ðα þ βÞ2 α2
Cs 3 þ Cd 3 and C4 can be obtained with Equations (16)–(19) and (21)—(23), respectively.
Thus:

Δ ¼ Cs 3 þ Cd 3 C4  
1 αþβ  4 
¼ 2 2 4 ln α þ 2α3 β þ α2 β2 þ ðDs þ Dd Þγβ2  4α2 βðα þ βÞ
2γðα þ βÞ α α
 4 
Let Γ ¼ ln αþβ
α α þ 2α3 β þ α2 β2  α2 βðα þ βÞ; t ¼ αβ > 1. Thus, Γ ¼ lnðt þ 1Þðt 4 þ t 3 þ t2 Þ
@Γ
t3  t2 , @t > 0 for all t > 1. Γ is monotonic increasing. Thus, we have Γ > 0 for all t > 1. With ΔC > Γ,
Δ > 0. □
14 Z. GONG AND N. LIU

In a decentralized system, a seaport and dry port are managed by different operators. Theorem 3
indicates that cooperation could decrease the total expected cost between these two operators when
it comes to investing in alleviating the adverse consequence of natural disasters and labor strikes.
The concrete form of cooperation could be jointly establishing a committee by a seaport and dry
port to make the pre-event investment decision in the SPDPILN. Another form could be the
government mandating the investment amounts of a seaport and dry port. To make the cooperation
mechanism feasible, the individual rationality should be satisfied. Mathematically, let C2 ¼ Cs 2 þ
Cd 2 and C4 ¼ Cs 4 þ Cd 4 ; in which the conditions Cs 2 < Cs 1 ; Cd 2 < Cd 1 and
Cs 4 < Cs 3 ; Cd 4 < Cd 3 need to hold. Therefore, cooperation requires an appropriate cost-sharing
plan. Section 6 provides additional managerial insights.

6. Managerial and policy implications

This section focuses on concretizing the analysis in previous sections and offers guidance for
a seaport and dry port in terms of pre-event investment to alleviate the adverse consequences of
natural disasters and labor strikes in SPDPLIN. The result of the optimal investment level could be
interpreted as determining an appropriate budget for a potential disaster.
If operators in SPDPILN plan to introduce a project to reduce the damage of a potential labor
strike, then the presented model could be adopted via the following procedure.

(1) Estimate the value of the parameters of the model with historic data. A seaport and dry port
could evaluate the expected damage from a labor strike (e.g., collating the data from the past
20 years). Calculate the number of the strikes and average damage caused by a labor strike in
the SPDPILN.
(2) Propose a project including the concrete form of investment, such as enhancing human
resource management, reforming labor laws to lower the event occurrence, and arranging
redundant resources, to reduce the damage once a labor strike happens. Thereafter, estimate
the return of different forms of investment.
(3) If SPDPILN is managed in a centralized system, then the manager should consider the
potential disaster of a seaport and dry port and coordinately decide the optimal investment
amount. For a decentralized system, a seaport and dry port should seek cooperation to
reduce the total expected cost.
(4) Thereafter, a seaport and dry port could propose a budget to alleviate the damage from
a labor strike by adopting the presented model. The budget includes the optimal investment
amount that should be spent on each practice (i.e., how much money should be spent on
enhancing human resource management and how much money should be spent on prepar-
ing redundant resources).

Cooperation in a decentralized system requires an appropriate cost-sharing plan. If a seaport or

dry port suffers more loss with cooperation than that in a decentralized system, then they will have
no incentive to join the cooperation mechanism. Mathematically, we consider labor strike as an
example. Let C4 ¼ Cs 4 þ Cd 4 ; the condition Cs 4 < Cs 3 ; Cd 4 < Cd 3 needs to hold. A ‘cost
sharing’ problem can be derived through numerous methods with the help of cooperative game
theory (dual method, Shapley value). We provide a solution that uses one of these methods, namely,
an equal-distributed cost-sharing plan. ΔC ¼ Cd 3 þ Cs 3  C4 , the cost that should be assumed by
a seaport and dry port are
Cs 4 ¼ Cs 3  12 ΔC, and Cd 4 ¼ Cd 3  12 ΔC. Accordingly, the individual rationality is satisfied.
However, the factor that affects the decision of operators in a real-life scenario may be compli-
cated because the decision-making process is not ideal. For example, the operators of a seaport
 MARITIME POLICY & MANAGEMENT 15

and dry port may not accurately estimate the related parameters. If a seaport does not fully
understand the condition of a dry port and proposes a cost-sharing plan based on incorrect
information, then the former may not reduce its total cost by adopting a cooperation strategy.
Subsequently, the possibility of inefficient cooperation mechanism will be high. Therefore, an
information-sharing platform within SPDPILN is crucial when deciding the optimal investment
amount.

7. Conclusion
This study investigated SPDPILN in investing to alleviate the damage of natural disasters and labor
strikes. Such a system is within the entire maritime supply chain and consists of one major dry port
and one seaport. Two types of investment are considered: mitigatve (lower the probability of the
event occurrence) and adaptive (reduce the damage after disaster occurrence) investments. For
natural disasters, the operators can only make adaptive investments because of the impossibility of
changing the occurrence probability of a natural disaster. Mitigative and adaptive investments are
considered for labor strikes. The analytical solutions derived from the presented model provide us
with several important conclusions.
We provide the optimal level of investment under different cases. In Section 4, we calculate the
optimal adaptive investment levels for Cases 1 and 2 and optimal mitigative and adaptive invest-
ments in Cases 3 and 4. The results provide guidance for operators in seaports and dry ports when
deciding the amount that they should spend in coping with the potential damage caused by natural
disasters and labor strikes.
Three theorems are proposed to compare two different management models (i.e., decentralized
and centralized systems). The results indicate that a centralized system is more beneficial than
a decentralized system in terms of investing in alleviating the adverse consequences of natural
disasters and labor strikes in SPDPLIN. Coordination is consistently effective in reducing the total
expected cost. If the occurrence of a natural disaster in a certain SPDPILN is high, then
a decentralized system will reduce the incentive for a seaport and dry port to make pre-event
investment. For labor strikes, decentralization will increase the mitigative investment amount.
However, the adaptive investment amount is not related to the management model.
Several avenues for future research are available. First, the logistics network in the presented
model comprises one major dry port and one seaport. Port competition is a relevant topic and could
affect the investment strategies of the involved actors. Meanwhile, port competition increases the
complexity of the logistics network because of the numerous ports in a particular region. The
decision process of different ports may affect one another. Such a framework, referring to competi-
tion between ‘transportation chains’, is worthy of further investigation (Zhang et al. 2007). Second,
we assume that seaports and dry ports make decisions simultaneously. A seaport operator is more
dominant compared with a dry port operator. Consequently, the seaport operator can make
a decision first and the dry port player has to follow. Other methods, such as the Stackelberg
model, could be applied to answer the present research question. Another important future avenue
is related to the economic loss of a disruption. Accordingly, an appropriate investment level is
affected by the damage caused by disruption. Therefore, an accurate estimation method to the
potential damage caused by disruption could increase the validity of the result.

Disclosure statement
No potential conflict of interest was reported by the authors.
16 Z. GONG AND N. LIU

Funding
This work was supported by the National Natural Science Foundation of China (Grant No. 71471162), NSFC-
Zhejiang Joint Fund for the Integration of Industrialization and Informatization (Grant No. U150920049), and
Zhejiang University Social Science Special Fund for Risk Identification and Security System of Port Maritime
Logistics under the ‘One Belt One Road’ (Grant No. 102000-541903/111).

References
Bakshi, N., and P. Kleindorfer. 2009. “Co-opetition and Investment for Supply-chain Resilience.” Production and
Operations Management 18 (6): 583–603. doi:10.1111/poms.2009.18.issue-6.
Beresford, A., and R. Dubey. 1990. “Handbook on the Management and Operation of Dry Ports, 1-2. UNCTAD.”
RDP/LDC 7.
Beresford, A., S. Pettit, Q. Xu, and S. Williams. 2012. “A Study of Dry Port Development in China.” Maritime
Economics & Logistics 14 (1): 73–98. doi:10.1057/mel.2011.17.
CEMT. 2001. “Land Access to Seaports.” Round Table 113,European Conference of Ministers of Transport. Paris:
OECD.
Chang, S. E., and N. Nojima. 2001. “Measuring Post-disaster Transportation System Performance: The 1995 Kobe
Earthquake in Comparative Perspective.” Transportation Research Part A Policy & Practice 35 (6): 475–494.
doi:10.1016/S0965-8564(00)00003-3.
Chen, H., J. S. L. Lam, and N. Liu. 2018. “Strategic Investment in Enhancing Port–Hinterland Container
Transportation Network Resilience: A Network Game Theory Approach.” Transportation Research Part
B Methodological 111: 83–112. doi:10.1016/j.trb.2018.03.004.
Chen, H., K. Cullinane, and N. Liu. 2017. “Developing a Model for Measuring the Resilience of a Port-hinterland
Container Transportation Network.” Transportation Research Part E: Logistics and Transportation Review 97:
282–301. doi:10.1016/j.tre.2016.10.008.
China News. 2012. “A Massive Strike Happened in Ports along the West Coast in American.” http://www.chinanews.
com/gj/2012/12-03/4376328.shtml doi:10.1094/PDIS-11-11-0999-PDN.
Cullinane, K., and M. Khanna. 2000. “Economies of Scale in Large Containerships: Optimal Size and Geographical
Implications.” Journal of Transport Geography 8 (3): 181–195. doi:10.1016/S0966-6923(00)00010-7.
Ducruet, C. 2016. “The Polarization of Global Container Flows by Interoceanic Canals: Geographic Coverage and
Network Vulnerability.” Maritime Policy & Management 43 (2): 242–260. doi:10.1080/03088839.2015.1022612.
Ducruet, C., and I. Lugo. 2013. “Cities and Transport Networks in Shipping and Logistics Research.” Asian Journal of
Shipping & Logistics 29 (2): 145–166. doi:10.1016/j.ajsl.2013.08.002.
Goss, R. 1990. “Economic Policies and Seaports: Are Port Authorities Necessary?” Maritime Policy & Management 17
(4): 257–271. doi:10.1080/03088839000000032.
Hall, P. V. 2004. ““We’d Have to Sink the Ships”: Impact Studies and the 2002 West Coast Port Lockout.” Economic
Development Quarterly 18 (4): 354–367. doi:10.1177/0891242404269500.
Hesse, M., and J. P. Rodrigue. 2004. “The Transport Geography of Logistics and Freight Distribution.” Journal of
Transport Geography 12 (3): 0–184. doi:10.1016/j.jtrangeo.2003.12.004.
International Labour Organization (ILO), 2017. “Ilostat –Ilo Database Of Labour Statistics.” http://www.ilo.org/
global/statistics-and-databases/lang–en/index.htm
Janić, M. 2015. “Modelling the Resilience, Friability and Costs of an Air Transport Network Affected by a Large-scale
Disruptive Event.” Transportation Research Part A: Policy and Practice 71: 1–16.
Lam, J. S. L., and S. Su. 2015. “Disruption Risks and Mitigation Strategies: An Analysis of Asian Ports.” Maritime
Policy & Management 42 (5): 415–435. doi:10.1080/03088839.2015.1016560.
Lee, S. W. 2010. “Centrality and Vulnerability in Liner Shipping Networks: Revisiting the Northeast Asian Port
Hierarchy.” Maritime Policy & Management 37 (1): 17–36. doi:10.1080/03088830903461175.
Major, J. A. 2002. “Advanced Techniques for Modeling Terrorism Risk.” Journal of Risk Finance 4 (1): 15–24.
doi:10.1108/eb022950.
Marufuzzaman, M., S. D. Eksioglu, X. Li, and J. Wang. 2014. “Analyzing the Impact of Intermodal-related Risk to the
Design and Management of Biofuel Supply Chain.” Transportation Research Part E Logistics & Transportation
Review 69 (69): 122–145. doi:10.1016/j.tre.2014.06.008.
Ng, A. K. Y., and I. B. Cetin. 2012. “Locational Characteristics of Dry Ports in Developing Economies: Some Lessons
from Northern India.” Regional Studies 46 (6): 757–773. doi:10.1080/00343404.2010.532117.
Ng, A. K. Y., T. Wang, Z. Yang, K. X. Li, and C. Jiang. 2018. “How Is Business Adapting to Climate Change Impacts
Appropriately? Insight from the Commercial Port Sector.” Journal of Business Ethics 150: 1–19. doi:10.1007/
s10551-016-3179-6.
Notteboom, T., and J. P. Rodrigue. 2005. “Port Regionalization: Towards a New Phase in Port Development.”
Maritime Policy & Management 32 (3): 297–313. doi:10.1080/03088830500139885.
 MARITIME POLICY & MANAGEMENT 17

Padilha, F., and A. K. Y. Ng. 2012. “The Spatial Evolution of Dry Ports in Developing Economies: The Brazilian
Experience.” Maritime Economics & Logistics 14 (1): 99–121. doi:10.1057/mel.2011.18.
Qiu, X., J. S. L. Lam, and G. Q. Huang. 2015. “A Bilevel Storage Pricing Model for Outbound Containers in A Dry
Port System.” Transportation Research Part E Logistics & Transportation Review 73 (73): 65–83. doi:10.1016/j.
tre.2014.10.009.
Reuters. 2011. “Impacts of the Great Earthquake in Japan on Energy, Bulk Commodity, and Ports.” http://cn.reuters.
com/article/2011/03/24/idCNCHINA-4029920110324?pageNumber=1>
Rodrigue, J. P., J. Debrie, A. Fremont, and E. Gouvernal. 2010. “Functions and Actors of Inland Ports: European and
North American Dynamics.” Journal of Transport Geography 18 (4): 519–529. doi:10.1016/j.jtrangeo.2010.03.008.
Roso, V., J. Woxenius, and K. Lumsden. 2009. “The Dry Port Concept: Connecting Container Seaports with the
Hinterland.” Journal of Transport Geography 17 (5): 338–345. doi:10.1016/j.jtrangeo.2008.10.008.
Roso, V., and K. Lumsden. 2010. A Review of Dry Ports. Maritime Economics & Logistics 12 (2): 196–213. doi:10.1057/
mel.2010.5.
Sapir, D. G., D. Hargitt, and P. Hoyois. 2004. Thirty Years of Natural Disasters 1974-2003: The Numbers. Presses univ.
de Louvain.
Slack, B. 1999. “Satellite Terminals: A Local Solution to Hub Congestion?” Journal of Transport Geography 7 (4):
0–246. doi:10.1016/S0966-6923(99)00016-2.
Wakeman, T. H. 2013. Lessons from Hurricane Sandy for Port Resilience, 15-28. University Transportation Research
Center.
Wan, Y., A. Zhang, and A. C. L. Yuen. 2013. “Urban Road Congestion, Capacity Expansion and Port Competition:
Empirical Analysis of US Container Ports.” Maritime Policy & Management 40 (5): 417–438. doi:10.1080/
03088839.2013.797615.
Wang, K., and A. Zhang. 2018. “Climate Change, Natural Disasters and Adaptation Investments: Inter- and
Intra-port Competition and Cooperation.” Transportation Research Part B: Methodological 117: 158–189.
doi:10.1016/j.trb.2018.08.003.
Xiao, Y. B., X. Fu, A. K. Y. Ng, and A. Zhang. 2015. “Port Investments on Coastal and Marine Disasters Prevention:
Economic Modeling and Implications.” Transportation Research Part B 78: 202–221. doi:10.1016/j.
trb.2015.04.009.
Yang, Z., A. K. Y. Ng, T. W. Lee, T. Wang, Z. Qu, V. S. Rodrigues, S. Pettit, I. Harris, D. Zhang, and Y. Y. Lau. 2017.
“Risk and Cost Evaluation of Port Adaptation Measures to Climate Change Impacts.” In Transportation Research
Part D Transport & Environment 61: 444-458.
Zhang, A. 2009. “The Impact of Hinterland Access Conditions on Rivalry between Ports.” In: Port Competition and
Hinterland Connections, 129–160. Paris: International Transport Forum, OECD.
Zhang, A., C. Lang, Y. V. Hui, and L. Leung. 2007. “Intermodal Alliance and Rivalry of Transport Chains: The Air
Cargo Market.” Transportation Research Part E: Logistics and Transportation Review 43 (3): 0–246. doi:10.1016/j.
tre.2006.10.003.