Ecological Complexity 32 (2017) 99–110

Contents lists available at ScienceDirect

Ecological Complexity
journal homepage: www.elsevier.com/locate/ecocom

Original Research Article

Effect of global warming on sea level rise: A modeling study

J.B. Shuklaa , Maitri Vermab,* , A.K. Misrab
a
International Internet University for Research in Science & Technology, India
b
Department of Mathematics, Institute of Science, Banaras Hindu University, Varanasi 221 005, India

A R T I C L E I N F O A B S T R A C T

Article history:
Received 28 April 2017 Global mean sea level has been rising in response to global warming since the past few decades and is
Received in revised form 24 October 2017 anticipated to potentially affect the coastal population. The main driver of global warming is the
Accepted 26 October 2017 enhanced concentration of the heat-trapping gas carbon dioxide in the atmosphere. In this paper, we
Available online 6 November 2017 propose a nonlinear mathematical model to study the effect of an increase in the anthropogenic carbon
dioxide emissions on sea level rise and its effect on the human population. The long-term behavior of the
Keywords: proposed system is analyzed using stability theory of differential equations. The model analysis shows
Mathematical model that an increase in the anthropogenic emission rate of carbon dioxide leads to increase in the equilibrium
Atmospheric carbon dioxide
levels of surface temperature and sea water level. Further, it is found that the increase in anthropogenic
Global warming
emission rate of carbon dioxide and melting rate of ice sheets lead to decrease in the equilibrium level of
Sea level rise
Stability human population as a result of crowding caused by the decrease in the total inhabitable land area due to
sea level rise. Numerical simulations are carried out to illustrate the effect of key parameters on the
dynamics of the system.
© 2017 Elsevier B.V. All rights reserved.

1. Introduction comparison to pre-industrial level (Lindsey, 2014). The rise in

Global warming is affecting all of us in many ways. Sea level rise
is one of the direst impacts of global warming which is affecting
the coastal population of many countries. Global warming is a
result of the increase in radiative forcing of climate system due to
the enhanced anthropogenic emissions of greenhouse gases,
primarily carbon dioxide (CO2). Human activities, between 1750
and 2011, have caused nearly 40% increase in the atmospheric
concentration of carbon dioxide. This upsurge in carbon dioxide
concentration has caused a radiative forcing of 1.82 watts per
square meter, which is nearly 64% of the total radiative forcing of all
the well-mixed greenhouse gases (Stocker et al., 2013). The
positive radiative forcing created by carbon dioxide and other
greenhouse gases has increased the Earth's energy budget, leading
to increase in the global surface temperature. The global mean
surface temperature has risen by nearly 0.84
C since pre-industrial
times and is expected to rise in future with the increase in
atmospheric concentration of CO2 and other greenhouse gases. A
doubling of carbon dioxide concentration is estimated to bring
nearly 1.5–4.5
C rise in global mean surface temperature in
* Corresponding author.
E-mail addresses: jbs@iitk.ac.in (J.B. Shukla), maitri.verma9@gmail.com
(M. Verma), akmisra@bhu.ac.in (A.K. Misra).
global mean temperature leads to melting of glaciers and ice caps,
resulting in sea level rise. It is found that the global mean sea level
has risen approximately 0.19 m between 1900 and 2010 with a rate
of nearly 1.7 millimeter per year (Church et al., 2013). The rate of
sea level rise is expected to increase in future with the further
increase in atmospheric temperature. The fifth assessment report
of IPCC (Intergovernmental Panel on Climate Change) projects that
the global mean sea level may rise by 0.52–0.98 m by the end of
21st century. The NOAA (National Oceanic and Atmospheric
Administration) projection of sea level rise is up to 2 m by 2100.
The ice sheets across the world hold sufficient water to raise the
sea level by several meters. Antarctica and Greenland ice sheets are
the two largest ice sheets containing nearly 99% of the freshwater
ice on the Earth. It is estimated that if the Greenland ice sheet melts
completely, sea level may rise about 6 m, while the melting of
Antarctic ice sheet may cause a sea level rise of about 60 m. It is
estimated that Antarctica ice sheet alone can contribute more than
a meter of sea level rise by the year 2100 and more than 15 m by the
year 2500 (DeConto and Pollard, 2016).
Sea level rise has potentially harmful consequences, such as
coastal flooding, erosion, saltwater contamination of fresh water
supplies, increase in salinity of agricultural soil, hurricane storm
surge, etc. Nearly 625 million people residing in low-elevation
coastal zones are at risk of exposure to these hazards of sea level
rise (Neumann et al., 2015). Hazards of sea level rise may force a

https://doi.org/10.1016/j.ecocom.2017.10.007
1476-945X/© 2017 Elsevier B.V. All rights reserved.
100 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110
portion of the coastal population to migrate from coastal zones to
inland zones, affecting the coastal population dynamics. Sea level
rise affects not only the dynamics of regional population, but also
the global population by affecting the net carrying capacity of
human population due to the reduction in inhabitable land area.
The elevated sea level has already begun to wipe out many coastal
areas. For instance, nearly half of Bhola Island in Bangladesh has
become permanently flooded due to sea level rise, leaving 500,000
climate refugees (Penna and Rivers, 2013). As most of the global
population is residing in coastal zones and many of the world's
major cities, such as New York, Los Angeles, California, Sydney, etc.,
are located in low-lying coastal zones, the future sea level rise is
likely to potentially influence the population dynamics.
From the above, it is pointed out that the global warming
caused by an increase in anthropogenic carbon dioxide emissions
may affect the sea water level, which in turn may affect the human
population levels. To address the problem of sea level rise, it is
crucial to comprehend the role of anthropogenic carbon dioxide
emissions and the associated temperature rise behind the sea level
rise. For this purpose, mathematical modeling is an effective tool.
The effect of climate change on sea level rise is usually evaluated by
using the process-based numerical models (Gregory et al., 2006;
Meehl et al., 2007) and semi-empirical models (Gornitz et al., 1982;
Rahmstorf, 2007; Vermeer and Rahmstorf, 2009; Grinsted et al.,
2010). These models are either numerical or statistical models and
use quantitative methods to assess the response of sea level to the
temperature change and rely completely on data. Apart from these
numerical and semi-empirical models, dynamical models that
utilize qualitative methods may also be used to gain an in-depth
understanding of the effect of temperature change on sea level rise.
In recent years, various differential equation models have been
proposed to address the dynamics of greenhouse gases and
pollutants in the atmosphere (Tennakone, 1990; Naresh and
Sundar, 2007; Naresh et al., 2007; Caetano et al., 2011; Sundar,
2013; Misra and Verma, 2013, 2014; Misra et al., 2015; Shukla et al.,
2015; Verma and Misra, 2016). In these studies, qualitative analysis
of the proposed models is carried out to assess the long-term effect
of various crucial factors on the dynamics of greenhouse gases and
pollutants. In the present study, we formulate a nonlinear
mathematical model comprising a set of differential equations
to study the effect of the enhanced concentration of carbon dioxide
and the associated temperature rise on the sea level rise, and the
effect of sea level rise on the human population. The qualitative
analysis of the proposed model is performed to assess the long-
term behavior of the system using stability theory of differential
equation.
2. Model formulation
Human activities increase the concentration of carbon dioxide
gas in the atmosphere. An increase in the concentration of carbon
dioxide gas causes an increase in the global mean temperature and
it causes melting of glaciers and ice caps, leading to sea level rise.
As the sea level rises, the land area submerged in water increases.
Due to the decrease in the inhabitable land area, carrying capacity
of human population declines. To model this scenario, we consider
six dynamical variables, namely, the human population ‘N(t)’, the
atmospheric concentration of carbon dioxide ‘C(t)’, the average
surface temperature ‘T(t)’, the mass of ice sheets ‘Ic(t)’, the mean
sea level ‘W(t)’ and the land area submerged in sea ‘L(t)’. It is
assumed that the carbon dioxide enters in the atmosphere due to
natural processes as well as due to human activities. The natural
source of carbon dioxide includes volcanoes, decomposition, fire,
ocean release, respiration, etc. The anthropogenic CO2 emission
consists mainly of the post-industrial CO2 emissions from fossil
fuel burning, industrial processes, deforestation and other human
related activities. The carbon dioxide is uptaken by the natural
sinks, such as forests, oceans, fossil fuels, carbonated rocks, etc.
Prior to the industrial revolution, the atmospheric concentration of
CO2 had been relatively stable as the natural source and sinks of
carbon dioxide were relatively balanced and the anthropogenic
emissions were insignificant relative to natural variability (IPCC,
2007). Let C0 be the equilibrium concentration of carbon dioxide in
the absence of anthropogenic carbon dioxide emissions (or pre-
industrial equilibrium concentration of carbon dioxide). Since the
industrial revolution, anthropogenic emission of carbon dioxide
has increased linearly with the increase in human population
(Newell and Marcus, 1987; Onozaki, 2009), so the emission rate of
carbon dioxide from anthropogenic sources is taken to be
proportional to human population ‘lN’. The emission of CO2 from
the natural sources is assumed to be a constant Q0. The uptake rate
of carbon dioxide by the natural sinks is assumed to be
proportional to the atmospheric concentration of CO2, i.e., ‘l0C’
(Nikol'skii, 2010). Let T0 be the average surface temperature in the
absence of anthropogenic carbon dioxide emissions (or the pre-
industrial average surface temperature). An increase in the
concentration of carbon dioxide leads to an increase in the
atmospheric temperature. It is observed that the increase in global
average temperature has linear correlation with the increase in CO2
concentration (IPCC, 2014). Hence, it is assumed that the
temperature increases at a rate proportional to the increase in
the concentration of CO2 due to human activities.
We have assumed that the ice sheets gain mass at a constant
rate q. The increase in temperature causes melting of ice sheets,
leading to sea level rise. In the study by Rahmstorf (2007), the rate
of sea level rise is taken to be proportional to the increase in
temperature. However, the rate of sea level rise also depends on
the melting of ice sheets (Meehl et al., 2007). Hence, we assume
that the ice sheets melt at a rate bIc(T  T0), where b is the melting
rate coefficient of ice sheets. The melting of ice sheets causes an
increase in the water level at a rate gbIc(T  T0), where g is a
proportionality constant which represents the conversion of ice
sheets into sea level rise. The surging sea level causes coastal
erosion and inundation, leading to an increase in the submerged
area. In some studies, it is found that the long-term shoreline
retreat rate is proportional to sea level rise (Leatherman et al.,
2000; Zhang et al., 2004). In view of this, it is assumed that the
submerged area increases at a rate proportional to the increase in
sea water level, i.e., f(W  W0), where f is the growth rate
coefficient of the submerged area and W0 is the sea water level
when the temperature is T0. As the submerged area increases, the
land inhabitable to human population decreases and this
negatively affects the growth of human population, via decreasing
the carrying capacity of the human population, at a rate r0N2(L
 L0), where r0 is the declination rate coefficient of human
population due to increase in the submerged area.
Under the above considerations, the dynamics of the problem is
governed by the following system of differential equations:
dIc
¼ q  a0 Ic  bIc ðT  T 0 Þ;
dt
dW
¼ gbIc ðT  T 0 Þ  g 0 ðW  W 0 Þ;
dt
dL
¼ fðW  W 0 Þ  f0 ðL  L0 Þ;
dt
 ð1Þ
dN N
¼ rN 1   r0 N2 ðL  L0 Þ;
dt K
dC
¼ Q 0 þ lN  l0 C;
dt
dT
¼ uðC  C0 Þ  u0 ðT  T 0 Þ;
dt
where C0 = Q0/l0, Ic(0) > 0, W(0)  W0, L(0)  L0, N(0)  0, C(0)  C0
and T(0)  T0.
 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110 101

Table 1
Description of model parameters and their unit.

Parameter Description Units

q Rate of formation of ice sheets teraton per year
a0 Natural depletion rate coefficient of ice sheets per year
b Melting rate coefficient of ice sheets per
C per year


T0 Average surface temperature in the absence of anthropogenic CO2 emissions C
g Conversion factor of sheet melt into sea level rise meter per teraton
g0 Natural depletion rate coefficient of sea level per year
W0 Mean sea level when the average surface temperature is T0 meter
f Growth rate coefficient of the submerged area due to sea level rise Mm2 per meter per year
f0 Declination rate coefficient of the submerged area per year
L0 Submerged land area when sea level is W0 Mm2
r Intrinsic growth rate of human population per year
K Carrying capacity of human population million persons
r0 Declination rate of carrying capacity of human population due to submerging of land area in sea per Mm2 per million persons per year
Q0 Natural emission rate of carbon dioxide ppm per year
l Anthropogenic emission rate coefficient of CO2 ppm per million persons per year
l0 Depletion rate coefficient of carbon dioxide per year
C0 Equilibrium concentration of carbon dioxide in the absence of anthropogenic CO2 emissions ppm
u Growth rate coefficient of the average surface temperature

C per ppm per year
u0 Depletion rate coefficient of the average surface temperature per year

 
The description of parameters of model system (1) and their (i) E q ; W ; L ; 0; Q 0 ; T ,
0 a0 0 0 l0 0
unit are given in Table 1. In Table 1, Mm2 stands for square mega- (ii) E*(I *, W*, L*, N*, C*, T*).
c
meter and ppm stands for parts per million.
The dynamics of the coastal population is assumed to be
The existence of equilibrium E0 is obvious. The components of
governed by the following equation
interior equilibrium E* can be obtained by solving the following set
dNc of algebraic equations:
¼ sN  s0 Nc  s1 Nc ðL  L0 Þ ð2Þ
dt
q  a0 Ic  bIc ðT  T 0 Þ ¼ 0; ð3Þ
where 0  Nc(0)  N(0). In the above equation, the constants s and
s0 are the growth rate and natural death rate coefficients of the
coastal population. The constant s1 is the migration rate coefficient gbIc ðT  T 0 Þ  g 0 ðW  W 0 Þ ¼ 0; ð4Þ
of the coastal population due to sea level rise. Further, it is assumed
that the coastal population increases at a rate proportional to the
total population. Once the dynamics of the system (1) is fðW  W 0 Þ  f0 ðL  L0 Þ ¼ 0; ð5Þ
determined, dynamics of the coastal population can be obtained
using Eq. (2). We assume that s  s0, so that the coastal population
cannot exceed the total population.

N
r 1  r0 NðL  L0 Þ ¼ 0; ð6Þ
Lemma 1. The set K
 
q
G ¼ ðIc ; W; L; N; C; TÞ 2 R6þ : 0  Ic  ; W 0  W  W m ; L0  L  Lm ; 0  N  K; C 0  C  Cm ; T 0  T  T m ;
a0

where W m ¼ W 0 þ gbqðT m T 0 Þ
g a0 , Lm ¼ L0 þ ff ðW m  W 0 Þ, C m ¼ C 0 þ llK0
0 0

and T m ¼ T 0 þ uullK , is positively invariant and for all Q 0 þ lN  l0 C ¼ 0; ð7Þ

0 0

ðIc ð0Þ; Wð0Þ; Lð0Þ; Nð0Þ; Cð0Þ; Tð0ÞÞ 2 R6þ satisfying

the initial condi-
tions of the model system (1), (Ic(t), W(t), L(t), N(t), C(t), T(t)) ! G as
t ! +1. The set G is called region of attraction for the model system uðC  C 0 Þ  u0 ðT  T 0 Þ ¼ 0: ð8Þ
(1). From (7), we have
For the proof of the above lemma see Appendix A.
Q0 lN lN
C¼ þ ¼ C0 þ : ð9Þ
3. Model analysis l0 l0 l0
From (8) and (9), we have
A qualitative analysis of the nonlinear system (1) is carried out
using the stability theory of differential equations. The stability ulN
T ¼ T0 þ : ð10Þ
theory addresses the stability properties of the solutions of a u0 l0
dynamical system. In the following, we examine the stability of the
From (3) and (10), we have
stationary solutions (i.e., the equilibrium points) of the model
system (1). q
Ic ¼ : ð11Þ
a0 þ ubul
0 l0
N
3.1. Equilibria
From (4), (10) and (11), we have
The model system (1) exhibits the following two non-negative gbqulN
equilibrium points: W ¼ W0 þ ð12Þ
g 0 fu0 l0 a0 þ bulNg:
102 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110

 dI  dN
Using Eq. (12) in (5), we have Remark 2. It is found that ddTl0 < 0, dlc0 > 0, dW
dl 0
< 0, dl0c > 0 and
dN
> 0. This shows that the equilibrium levels of surface
fgbqulN dl 0
L ¼ L0 þ ð13Þ
f0 g 0 fu0 l0 a0 þ bulNg: temperature and sea water level decrease and the equilibrium
levels of ice sheet mass, coastal population and total human
Using Eq. (13) in (6), we get the following quadratic equation in N: population increase as the depletion rate coefficient of carbon


 dioxide increases.
bul r r0 fg q 2 ra0 rbul
FðNÞ ¼ þ N þ  N  ra0 ¼ 0: ð14Þ
u0 l0 K f0 g 0 K u0 l0   dN
Remark 3. It is found that dL
db
< 0, dN
db
< 0 and dbc < 0, showing
From (14), we find that F(0) < 0 and F(K) > 0. Thus, there exists a that the equilibrium level of land area submerged in water
unique positive root of Eq. (14), say N*, in the interval (0, K), which increases whereas the equilibrium levels of total human
is given as population and coastal population decline with an increase in
r ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
 the melting rate coefficient of ice sheets.
2
raK0 þ rubul
0 l0
þ ra0
K þ rubul
0 l0
þ 4rau00 bul r0 fg q
l0 f g

 
0 0
N ¼ : ð15Þ
2bul
u 0 l0
r
K þ rf0 fg
g
q 3.2. Stability of equilibria
0 0

Using the above value of N* in Eqs. (9), (10), (11), (12) and (13), we The local stability of the equilibrium E0 is determined by
get the positive values of C*, T*, Ic , W* and L*, respectively. observing the sign of eigenvalues of the Jacobian matrix J0 of the
The equilibrium level of coastal population is given by model system (1) evaluated at the equilibrium E0. The eigenvalues
of J0 are found to be a0, g 0, f0, r, l0 and u0. Since one of the
sN
Nc ¼ : ð16Þ eigenvalues of J0 is r, which is positive, the equilibrium E0 has
s0 þ s1 ðL  L0 Þ locally unstable manifold in N-direction. Thus, the system will
never approach to the equilibrium E0.
 dI   dN
We determine the local as well as global stability behavior of
Remark 1. It is found that dT
dl
> 0, dlc < 0, dW
dl
> 0, dL
dl
> 0, dlc < 0 the equilibrium E* by using the Lyapunov's stability theory. The
dN
and dl < 0. Thus, the model predicts that the equilibrium levels results obtained from stability analysis of E* are stated in the form
of surface temperature, sea water level and land area submerged of following theorems.
in water increase, while the equilibrium levels of ice sheet mass,
Theorem 1. The equilibrium E* is locally asymptotically stable
coastal population and total human population decrease as the
provided that the following inequality is satisfied:
anthropogenic emission rate coefficient of carbon dioxide
( )
2
increases. b2 N2 Ic 2ðT   T 0 Þ2 3  2 1 rl0 u0 f0 g 0
max ; I 2N < : ð17Þ
a20 4c 36 Kr0 lfgub

Fig. 1. Comparison between model fit and actual data of atmospheric concentration of carbon dioxide, human population, average surface temperature and mean sea level.
 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110
For the proof of the above theorem see Appendix B.
Theorem 2. The equilibrium E* is globally asymptotically stable in
G provided that the following inequality is satisfied:
( ) !2
b2 q2 ðT   T 0 Þ2 3 2 1 r a 0 l0 u 0 f 0 g 0
max ; q < : ð18Þ
a20 4 36 K 2 r0 lfgub
For the proof of the above theorem see Appendix C.
4. Numerical simulation
To check the feasibility of analysis and to depict the effect of key
parameters on the dynamics of the system (1), we have conducted
numerical simulation using MATLAB 7.5.0. The snow accumulation
rates for the Antarctica and Greenland ice sheets are 1850 gigatons
per year and 500 gigatons per year, respectively (Steffen et al.,
2008). Thus, the rate of formation of ice sheets q is taken to be
2.35  103 teratons (or million gigatons) per year. We have taken
r = 0.032 per year and l = 0.000576 ppm per million persons per
year from Verma and Misra (2016). The carrying capacity of human
population K is taken to be 10,000 million which is according to the
United Nations Population Division's medium-variant scenario
(UNPD, 2001). The atmospheric life-time of carbon dioxide is 30–
95 years (Jacobson, 2005), thus the value of l0 lies in the range
0.0105–0.0333 per year. We take C0 = 280 ppm, the pre-industrial
CO2 concentration (IPCC, 2001). To decide the value of other
parameters of the model system (1), we use the annual time series
data of world population, carbon dioxide concentration, global
average temperature and mean sea level. The annual data of
atmospheric carbon dioxide for the period 1961–2011 is taken
from NOAA-ESRL data set (NOAA, 2016), while that of world
population for the period 1961–2011 is obtained from UNPD
(2012). We have taken the annual data of the global average surface
temperature for the period of 1961–2013 from NOAA (2017b). The
annual time series data of the mean sea level observed at a
particular station Reykjavik, Iceland between the years 1961–2011
is taken from NOAA (2017a). We assume the value of W0 = 6.8 m
which is 0.26 m below the 2001 level. The pre-industrial
temperature is taken to be T0 = 13.5
C. We calibrate the model
system for different values of a0, b, g , g 0, f, f0, L0, u, u 0, r0 and l0,
and find the best fit for a0 = 0.78  104 per year, b = 105 per
C per
Fig. 2. The global stability of E* inside the region of attraction G in Ic  W  L and N  C  T spaces.
103
year, g = 20 m per teraton, g 0 = 0.008 per year, f = 0.03 Mm2 per
meter per year, f0 = 0.01 per year, L0 = 2 Mm2, u = 0.003
C per ppm
per year, u0 = 0.3 per year, r0 = 0.8  107 per Mm2 per million
persons per year and l0 = 0.02 per year. The actual data and the
model projections are illustrated in Fig. 1. In the absence of annual
time series data of the land area submerged in water and mass of
ice sheets, the above parameter values provide a very rough
estimate. However, we try to be as realistic as possible while
selecting the parameter values. The parameter values used in
simulation for the model system (1) and (2) are listed as
q ¼ 2:35  103 ; a0 ¼ 0:78  104 ; b ¼ 105 ; g 0 ¼ 0:008; g ¼ 20; f ¼ 0:03;
f0 ¼ 0:01; W 0 ¼ 6:8; r ¼ 0:032; K ¼ 10; 000; L0 ¼ 2; r0 ¼ 0:8  107 ;
l0 ¼ 0:04; C0 ¼ 280; Q 0 ¼ 5:6; l ¼ 0:000576; u ¼ 0:003; u0 ¼ 0:3; T 0 ¼ 13:5;
s ¼ 0:03; s0 ¼ 0:04; s1 ¼ 0:04:
ð19Þ
For the above set of parameter values, the interior equilibrium of
system (1) is obtained as E*(22.6078, 8.2665, 6.3994, 9009.1269,
539.4628, 16.0946). The eigenvalues of the Jacobian matrix for the
model system (1) calculated at the equilibrium E* are 0.300000,
0.030918, 0.0079799 + 0.003883i, 0.0079799  0.003883i,
0.000102, and 0.023124. All the eigenvalues are either negative
or having negative real part, showing that the equilibrium E* is
locally asymptotically stable. The global stability condition (18) is
also satisfied for the set of parameter values (19), showing that the
equilibrium E* is globally asymptotically stable in G.
In Fig. 2, we have plotted the solution trajectories in Ic  W  L
and N  C  T spaces for different initial starts inside the region of
attraction G. From this figure, it can be observed that all the
solution trajectories starting inside the region of attraction finally
approach to ðIc ; W  ; L Þ and (N*, C*, T*), respectively. This shows the
global stability of E* in the region of attraction G in Ic  W  L and
N  C  T spaces. Fig. 3 is plotted to show the effect of an increase in
the anthropogenic emission rate coefficient of carbon dioxide (i.e.
l) on the equilibrium levels of average surface temperature, mean
sea level, land area submerged in water and human population.
From this figure, it can be observed that an increase in the
anthropogenic emission rate coefficient of carbon dioxide leads to
increase in the equilibrium levels of temperature, sea water and
the land area submerged in water, while the equilibrium level of
the human population decreases due to increase in the
104 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110

Fig. 3. Variation of surface temperature, sea level, land area submerged in sea water and human population with respect to time for different values of the anthropogenic
emission rate of carbon dioxide (l).

anthropogenic emission rate coefficient of carbon dioxide. This dioxide increases, which leads to increase in the temperature.
happens because as the anthropogenic emission rate of carbon The increase in temperature causes melting of ice sheets leading to
dioxide increases, the concentration of atmospheric carbon an increase in sea level and consequently the land area submerged

Fig. 4. Variation of surface temperature and sea level with respect to time for different values of the intrinsic growth of human population (r).
 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110 105

Fig. 5. Variation of surface temperature and sea level with respect to time for different values of the melting rate coefficient of ice sheets (b).

Fig. 6. Variation of coastal population with respect to time for different values of anthropogenic emission rate coefficient of carbon dioxide (l), melting rate coefficient of ice-
sheets (b), growth rate coefficient of submerged area due to sea level rise (f) and migration rate coefficient of the coastal population (s1).
106 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110

Fig. 7. Effect of change in the anthropogenic emission rate coefficient of carbon dioxide (l) and depletion rate coefficient of carbon dioxide (l0) on the equilibrium levels of
surface temperature and sea water level.

Fig. 8. Semi-relative sensitivity solutions for the state variables with respect to the parameters l (the anthropogenic emission rate coefficient of carbon dioxide), b (the
melting rate coefficient of ice-sheets), l0 (the depletion rate coefficient of carbon dioxide) and f (the growth rate coefficient of submerged land area due to sea level rise).
 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110
in water increases. The effect of an increase in the intrinsic growth
rate of human population on the atmospheric temperature and sea
level is shown in Fig. 4. From this figure, we can observe that as the
growth rate of human population increases, the equilibrium levels
of surface temperature and sea water increase.
Fig. 5 shows that as the melting rate coefficient of ice-sheets (i.e.
b) increases, the equilibrium level of land area submerged in water
increases while that of human population decreases. Fig. 6 is
plotted to show the effect of changes in parameters l, b, f and s1
on the equilibrium level of coastal population. This figure shows
that the coastal population settles to a lower level as the values of
parameters l, b, f and s1 increase. The sea level rise may be
controlled by cutting down the carbon dioxide emissions as well as
by increasing the depletion rate of carbon dioxide. To show the
effects of changes in the anthropogenic emission rate and
depletion rate of carbon dioxide on the equilibrium levels of
atmospheric temperature and sea water, we have plotted Fig. 7.
This figure shows that the atmospheric temperature and sea water
level approach low values for low emission rates and high
depletion rates of carbon dioxide, while for high emission rates
and low depletion rates of carbon dioxide, the temperature and sea
water levels approach higher values.
To assess that how sensitive the model solutions are to change
in some of the key parameters l, b, l0 and f, a basic differential
sensitivity analysis of the equations in the model system (1) is
performed with respect to these parameters following Bortz and
Nelson (2004). The sensitivity function of a state variable X with
respect to a parameter w is given by X w ðt; wÞ ¼ @Xðt;wÞ
@w . We have
calculated the semi-relative sensitivity solutions (i.e., w X w ðt; wÞ);
those represent the change in state of the system on doubling of a
parameter value. These solutions are plotted in Fig. 8. From this
figure we can see that the doubling of parameter b has significant
influence over the mass of ice sheets Ic(t), sea level W(t), land area
submerged in water L(t) and human population N(t), while the
change in parameters l and l0 significantly influence all the
variables. The change in parameter f has large influence over L(t)
and N(t), while its influence over other variables is negligible. From
the second plot of Fig. 8, we can see that doubling of parameter b
causes an increase of 0.7014 m in the sea level while the doubling
of the parameter l leads to a rise of 0.5359 m in the sea level over a
period of 100 years. The doubling of parameter l0 causes a
decrease of 0.3808 m in the sea level over a period of 100 years.
From the fourth plot of Fig. 8, we can see that the doubling of
parameter f causes a decrease of 174.8 million in the human
population over a period of 100 years.
5. Conclusion
The enhanced concentration of carbon dioxide gas in the
atmosphere due to human activities has contributed to the
increase in global temperature and consequently to the sea level
rise. Sea level rise, in turn, has adversely affected the human
population. In this paper, we have proposed a nonlinear
mathematical model to study the response of sea level to the
temperature rise caused by the anthropogenic carbon dioxide
emissions and its effect on the human population. In the modeling
process, six dynamical variables are considered, namely, the
human population, the atmospheric concentration of carbon
dioxide, the average surface temperature, the mass of ice sheets,
the mean sea level and the land area submerged in water. A
qualitative analysis of the proposed model is carried out. It is found
that the model system has two non-negative equilibria; a
boundary equilibrium and an interior equilibrium. The boundary
equilibrium is always unstable. The conditions for the local and
global stability of the interior equilibrium have been derived by
107
using Lyapunov's direct method. The model analysis suggests that
an increase in the anthropogenic emission rate of carbon dioxide
leads to an increase in the equilibrium level of temperature and
consequently the increase in the equilibrium sea level and the land
area submerged in water. The equilibrium levels of ice sheet mass
and human population are found to decline due to an increase in
the anthropogenic emission rate of carbon dioxide. This suggests
that declination in the anthropogenic carbon dioxide emission rate
may slow down the melting of glacier and ice caps and the sea level
rise across the globe. An increase in the intrinsic growth rate of
human population is also found to increase the equilibrium surface
temperature and the equilibrium sea level. It has been found that
an increase in the depletion rate of atmospheric carbon dioxide
leads to increase in the equilibrium mass of ice sheets and decrease
in the equilibrium sea level. This suggests that apart from cutting
down the anthropogenic carbon dioxide emissions, increasing the
sinks of carbon dioxide via plantation, carbon capture and storage
technologies, etc., may also aid in slowing down the sea level rise.
The increase in melting rate coefficient of ice sheets is found to
increase the submerged land area (decrease the inhabitable land
area) leading to decrease in human population due to crowding.
The effect of changes in the key parameters on the coastal
population is also investigated. Sensitivity analysis is carried out to
show the influence of changes in some of the key parameters over
the dynamics of the system. Overall, the present work provides a
simple framework to assess the effect of temperature rise due to
anthropogenic carbon dioxide emissions on the sea level rise and
the effect of sea level rise on the human population.
Acknowledgements
Authors are thankful to the handling editor and the anonymous
reviewers for their useful suggestions, which improved the quality
of the paper. The second author is thankful to National Board of
Higher Mathematics, Department of Atomic Energy, Government
of India for providing financial support in form of the post-doctoral
fellowship (No: 2/40(7)/2015/R & D-II/4951).
Appendix A. Proof of Lemma 1
We prove the lemma 1 following Freedman and So (1985). First
of all, we show that the set G is positively invariant, i.e. for all (Ic(0),
W(0), L(0), N(0), C(0), T(0)) 2 G, (Ic(t), W(t), L(t), N(t), C(t), T(0)) 2 G
for all t  0.
From the fourth equation of the model system (1), we have


dN N
 rN 1  :
dt K
Comparing the above differential inequality with differential
equation


dN N
¼ rN 1  ;
dt K
and using a standard comparison theorem (Hale, 1969), we get N
(t)  K for all t  0.
From the fifth equation of the model system (1), we get
dC
 Q 0 þ lK  l0 C:
dt
This implies that


Q 0 þ lK Q 0 þ lK
C ðtÞ    C ð0Þ el0 t
l0 l0
Q 0 þ lK
 ¼ Cm ; 8 t  0:
l0
108 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110

Now, from the sixth equation of the model system (1), we get Proceeding same as above, from the first three equations of the
model system (1), we get
dT
 uðC m  C 0 Þ þ u0 T 0  u0 T: q
dt lim supIc ðtÞ ¼ ;
t!þ1 a0
This implies that gbqðT m  T 0 Þ

 lim supWðtÞ ¼ W0 þ
¼ Wm;
ulK ulK t!þ1 g 0 a0
TðtÞ  T 0 þ  T0 þ  Tð0Þ eu0 t fðW m  W 0 Þ
u 0 l0 u 0 l0 and lim supLðtÞ ¼ L0 þ :
ulK t!þ1 f0
 T0 þ ¼ T m ; 8 t  0:
u 0 l0
Proceeding same as above, from the first three equations of the Appendix B. Proof of Theorem 1
model system (1), it can be found that
In the following, we obtain the characterization of the local
q gbqðT m  T 0 Þ stability of the interior equilibrium E* using the Lyapunov's direct
Ic ðtÞ  ; WðtÞ  W 0 þ ¼ W m and LðtÞ
a0 g 0 a0 method (La Salle and Lefschetz, 1961). We first linearize the model
fðW m  W 0 Þ system (1) about E* by using the following transformations:
 L0 þ ; 8 t  0:
f0 Ic ¼ Ic þ ic , W ¼ W  þ w, L = L* + l, N = N* + n, C = C* + c and T = T* + p,
Now we will show that for all satisfying the initial conditions of the where ic, w, l, N, c and p are small perturbations about E*.
model system (1)ðIc ð0Þ; Wð0Þ; Lð0Þ; Nð0Þ; Cð0Þ; Tð0ÞÞ 2 R6þ , (Ic(t), W Now, we consider the following positive definite function:
(t), L(t), N(t), C(t), T(t)) ! G as t ! +1. 1 2 1 1 2 1 n2 1 1
From the fourth equation of the model system (1), we have V¼ i þ m1 w2 þ m2 l þ m3  þ m4 c2 þ m5 p2 ; ðB:1Þ
2 c 2 2 2 N 2 2


dN N where m1, m2, m3, m4 and m5 are positive constants to be chosen
 rN 1  :
dt K suitably. Differentiating V with respect to t along the solutions of
the linearized system of (1), we get
Noting the solution of

 dV
¼ a0 ic  bðT   T 0 Þic  m1 g 0 w2  m2 f0 l 
2 2 2 m3 r 2
n  m3 r0 ðL  L0 Þn2
dN N dt K
¼ rN 1  ; Nð0Þ  0;
dt K m4 l0 c2  m5 u 0 p2  bIc pic þ m1 gbðT   T 0 Þic w þ m1 gbIc pw

and using a standard comparison theorem (Hale, 1969), we have þm2 flw  m3 r0 N nl þ m4 lnc þ m5 upc:
ðB:2Þ
K
NðtÞ    :
1þ K
 1 ert dV/dt will be negative definite provided the following conditions
Nð0Þ
hold:
Let e > 0 be given. Then there exists a t1 > 0 such that
3b Ic 2
2
< m5 ; ðB:3Þ
NðtÞ  K þ e; 8 t  t1 : 2a0 u0
Hence
lim supNðtÞ  K: 2a0 g 0
t!þ1 m1 < ; ðB:4Þ
3g 2 b ðT   T 0 Þ2
2

From the fifth equation of the model system (1), we get

dC
 Q þ lðK þ eÞ  l0 C; 8 t  t1 :
9g 2 b Ic 2
2
dt
m1 < m5 ; ðB:5Þ
As before, there exist t2  t1 such that 4u 0 g 0

Q 0 þ lðK þ eÞ
CðtÞ  þ e ¼ Ce ; 8 t  t2 :
l0 2f0 g 0
m2 < m1 ; ðB:6Þ
3f
2
Hence
Q 0 þ lK
lim supCðtÞ  ¼ Cm :
t!þ1 l0 r f0
m3 < m2 ; ðB:7Þ
Similarly, from the last equation of the model system (1), we find Kr20 N2
that there exist t3  t2 such that

uðC e  C 0 Þ
TðtÞ  T 0 þ þ e ¼ Te; 8 t  t3 : rl0
u0 m4 < m3 ; ðB:8Þ
K l2
Hence

uðCm  C 0 Þ
lim supTðtÞ  T 0 þ ¼ Tm: 2 u 0 l0
t!þ1 u0 m5 < m4 : ðB:9Þ
3u
2
 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110 109

We choose We choose
a0 g 0 f0 g 0 m1 rf0 m2 a0 g 0 f0 g 0 k1 rf k2 rl0 k3
m1 ¼ ; m2 ¼ ; m3 ¼ ; and m4 k1 ¼ ; k2 ¼ ; k3 ¼ 0 3 ; and k4 ¼ :
2g 2 b ðT   T 0 Þ2 2f 2Kr20 N2 2g 2 b ðT   T 0 Þ2 2f 2K l2
2 2 2 2
2r20 K
r l0 m 3
¼ : For this selection of k1, k2, k3 and k4, the inequalities (C.4), (C.6),
2K l2
(C.7) and (C.8) are satisfied. Now, from the rest of inequalities, we
For this selection of m1, m2, m3 and m4, the inequalities (B.4), (B.6), can select k5 > 0 provided the condition (18) holds.
(B.7) and (B.8) are satisfied. Now from the rest of inequalities, we
can select m5 > 0 provided the condition (17) holds. References

Appendix C. Proof of Theorem 2 Bortz, D.M., Nelson, P.W., 2004. Sensitivity analysis of a nonlinear lumped
parameter model of HIV infection dynamics. Bull. Math. Biol. 66, 1009–1026.
Caetano, M.A.L., Gherardi, D.F.M., Yoneyama, T., 2011. An optimized policy for the
In the following, we derive sufficient conditions for the global reduction of CO2 emission in the Brazilian Legal Amazon. Ecol. Model. 222,
stability of the interior equilibrium E* in G by using the Lyapunov's 2835–2840.
Church, J.A., Clark, P.U., Cazenave, A., Gregory, J.M., Jevrejeva, Levermann, S.A.,
direct method. Consider the following positive definite function: Merrifield, M.A., Milne, G.A., Nerem, R.S., Nunn, P.D., Payne, A.J., Pfeffer, W.T.,
Stammer, D., Unnikrishnan, A.S., 2013. Sea level change. In: Stocker, T.F., Qin, D.,
1 1 1
U ¼ ðIc  Ic Þ2 þ k1 ðW  W  Þ2 þ k2 ðL  L Þ2 Plattner, G.-K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex, V.,
2 2 2 Midgley, P.M. (Eds.), Climate Change 2013: The Physical Science Basis,

 Contribution of Working Group I to the Fifth Assessment Report of the
N 1
þk3 N  N  N ln  þ k4 ðC  C  Þ2 Intergovernmental Panel on Climate Change. Cambridge University Press,
N 2 Cambridge, United Kingdom and New York, NY, USA.
1 DeConto, R.M., Pollard, D., 2016. Contribution of Antarctica to past and future sea-
þ k5 ðT  T  Þ2 ; ðC:1Þ level rise. Nature 531, 591–597.
2
Freedman, H.I., So, J.W.H., 1985. Global stability and persistence of simple food
where k1, k2, k3, k4 and k5 are positive constants to be chosen chains. Math. Biosci. 76, 69–86.
suitably. Differentiating U with respect to t along the solutions of Gornitz, V., Lebedeff, S., Hansen, J., 1982. Global sea-level trend in the past century.
Science 215, 1611–1614.
the model system (1), we get Grinsted, A., Moore, J.C., Jevrejeva, S., 2010. Reconstructing sea level from paleo and
projected temperatures 200 to 2100 AD. Clim. Dyn. 34 (4), 461–472.
dU
¼ a0 ðIc  Ic Þ2  bðT   T 0 ÞðIc  Ic Þ2  k1 g 0 ðW  W  Þ2  k2 f0 ðL  L Þ2 Gregory, J.M., Lowe, J.A., Tett, S.F.B., 2006. Simulated global-mean sea level changes
dt
over the last half-millennium. J. Clim. 19 (18), 4576–4591.
k3 r
 ðN  N Þ2  k3 r0 ðL  L0 ÞðN  N Þ2  k4 l0 ðC  C  Þ2  k5 u0 ðT  T  Þ2 Hale, J.K., 1969. Ordinary Differential Equations. Wiley-Inscience, New York.
K IPCC, 2001. The carbon cycle and atmospheric carbon dioxide. In: Houghton, J.T.,
bIc ðT  T  ÞðIc  Ic Þ þ k1 gbðT   T 0 ÞðIc  Ic ÞðW  W  Þ Ding, Y., Griggs, D.J., Noguer, M., Van der Linden, P.J., Dai, X., Maskell, K., Johnson,
þk1 gbIc ðT  T  ÞðW  W  Þ þ k2 fðL  L ÞðW  W  Þ  k3 r0 NðN  N ÞðL  L Þ C.A. (Eds.), Climate Change 2001: The Scientific Basis, Contribution of Working
Group I to the Third Assessment Report of the Intergovernmental Panel on
þk4 lðN  N ÞðC  C  Þ þ k5 uðT  T  ÞðC  C  Þ: Climate Change. Cambridge University Press, Cambridge.
ðC:2Þ IPCC, 2007. The carbon cycle and the climate system. In: Solomon, S., Qin, D.,
Manning, M., Chen, Z., Marquis, M., Averyt, K.B., Tignor, M., Miller, H.L. (Eds.),
dU/dt will be negative definite inside G provided the following Climate Change 2007: The Physical Science Basis, Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on
inequalities are satisfied:
Climate Change. Cambridge University Press, Cambridge.
IPCC, 2014. In: Core Writing Team, Pachauri, R.K., Meyer, L.A. (Eds.), Climate Change
3b q2
2
2014: Synthesis Report, Contribution of Working Groups I, II and III to the Fifth
< k5 ; ðC:3Þ
2a30 u0 Assessment Report of the Intergovernmental Panel on Climate Change. IPCC,
Geneva, Switzerland.
Jacobson, M.Z., 2005. Correction to “Control of fossil-fuel particulate black carbon
and organic matter, possibly the most effective method of slowing global
2a0 g 0 warming”. J. Geophys. Res. 110 D14105 (1–5).
k1 < ; ðC:4Þ La Salle, J., Lefschetz, S., 1961. Stability by Lyapunov's Direct Method with
3g 2 b ðT   T 0 Þ2
2
Applications. Academic Press Inc., New York.
Leatherman, S.P., Zhang, K., Douglas, B.C., 2000. Sea level rise shown to drive coastal
erosion. EOS Trans. 81 (6), 55–57.
Lindsey, R., 2014. How much will Earth warm if carbon dioxide doubles pre-
9k1 g 2 b q2
2
industrial levels? National Oceanic and Atmospheric Administration, USA. .
< k5 ; ðC:5Þ (accessed 10.01.2017) https://www.climate.gov/news-features/climate-qa/
4u0 g 0 a20 how-much-will-earth-warm-if-carbon-dioxide-doubles-pre-industrial-levels.
Meehl, G.A., et al., 2007. Global climate projections. In: Solomon, S. (Ed.), Chapter 10,
Climate Change 2007: The Physical Science Basis. Contribution of Working
Group I to the Fourth Assessment Report of the Intergovernmental Panel on
2f0 g 0 k1 Climate Change. Cambridge University Press, Cambridge, U.K.
k2 < ; ðC:6Þ
3f
2 Misra, A.K., Verma, M., 2013. A mathematical model to study the dynamics of carbon
dioxide gas in the atmosphere. Appl. Math. Comput. 219, 8595–8609.
Misra, A.K., Verma, M., 2014. Modeling the impact of mitigation options on methane
abatement from rice fields. Mitig. Adapt. Strateg. Glob. Change 19, 927–945.
rf0 k2 Misra, A.K., Verma, M., Venturino, E., 2015. Modeling the control of atmospheric
k3 < ; ðC:7Þ carbon dioxide through reforestation: effect of time delay. Model. Earth Syst.
r20 K 3 Environ. 1 24 (17 pages).
Naresh, R., Sundar, S., 2007. A nonlinear dynamical model to study the removal of
gaseous and particulate pollutants in a rain system. Nonlinear Anal.: Model.
Control 12, 227–243.
rl0 k3 Naresh, R., Sundar, S., Shukla, J.B., 2007. Modeling the removal of gaseous pollutants
k4 < ; ðC:8Þ
Kl and particulate matters from the atmosphere of a city. Nonlinear Anal.: Real
2
World Appl. 8, 337–344.
Neumann, B., Vafeidis, A.T., Zimmermann, J., Nicholls, R.J., 2015. Future coastal
population growth and exposure to sea-level rise and coastal flooding – a global
2u 0 l0 k4 assessment. PLoS ONE 10, e0118571.
k5 < : ðC:9Þ Newell, N.D., Marcus, L., 1987. Carbon dioxide and people. Palaios 2, 101–103.
3u
2
Nikol'skii, M.S., 2010. A controlled model of carbon circulation between the
atmosphere and the ocean. Comput. Math. Model. 21, 414–424.
110 J.B. Shukla et al. / Ecological Complexity 32 (2017) 99–110
NOAA, 2016. Trends in Atmospheric Carbon DioxideNational Oceanic and
Atmospheric Administration, USA. . (accessed 06.03.2017) www.esrl.noaa.gov/
gmd/ccgg/trends/.
NOAA, 2017a. Tide and Currents, Mean Sea Level Trends 010-001 Reykjavik,
IcelandNational Oceanic and Atmospheric Administration, USA. . (accessed
20.07.2017) https://tidesandcurrents.noaa.gov/sltrends/
sltrends_global_station.htm?stnid=010-001#.
NOAA, 2017b. Annual global land and ocean temperature anomalies. . (accessed
20.07.2017) https://www.ncdc.noaa.gov/cag/time-series/global/globe/
land_ocean/ytd/12/1880-2017.csv.
Onozaki, K., 2009. Population is a critical factor for global carbon dioxide increase. J.
Health Sci. 55, 125–127.
Penna, A.N., Rivers, J.S., 2013. Natural Disasters in a Global Environment. Wiley-
Blackwell, Chichester.
Rahmstorf, S., 2007. A semi-empirical approach to projecting future sea-level rise.
Science 315, 368–370.
Shukla, J.B., Chauhan, M.S., Sundar, S., Naresh, R., 2015. Removal of carbon dioxide
from the atmosphere to reduce global warming: a modeling study. Int. J. Global
Warm. 7, 270–292.
Steffen, K., Clark, P.U., Cogley, J.G., Holland, D., Marshall, S., Rignot, E., Thomas, R.,
2008. Rapid changes in glaciers and ice sheets and their impacts on sea level.
Abrupt Climate Change. Synthesis and Assessment Product 3.4. U.S. Geological
Survey, Reston, VA.
Stocker, T.F., Qin, D., Plattner, G.K., Alexander, L.V., Allen, S.K., Bindoff, N.L., Breon, F.
M., Church, J.A., Cubasch, U., Emori, S., Forster, P., Friedlingstein, P., Gillett, N.,
Gregory, J.M., Hartmann, D.L., Jansen, E., Kirtman, B., Knutti, R., Krishna Kumar,
K., Lemke, P., Marotzke, J., Masson-Delmotte, V., Meehl, G.A., Mokhov, I.I., Piao,
S., Ramaswamy, V., Randall, D., Rhein, M., Rojas, M., Sabine, C., Shindell, D.,
Talley, L.D., Vaughan, D.G., Xie, S.P., 2013. Technical summary. In: Stocker, T.F.,
Qin, D., Plattner, G.K., Tignor, M., Allen, S.K., Boschung, J., Nauels, A., Xia, Y., Bex,
V., Midgley, P.M. (Eds.), Climate Change 2013: The Physical Science Basis,
Contribution of Working Group I to the Fifth Assessment Report of the
Intergovernmental Panel on Climate Change.. Cambridge University Press,
Cambridge.
Sundar, S., 2013. An ecological type nonlinear model for the removal of carbon
dioxide from the atmosphere by introducing liquid species. Comput. Ecol.
Softw. 3, 33–34.
Tennakone, K., 1990. Stability of the biomass-carbon dioxide equilibrium in the
atmosphere: mathematical model. Appl. Math. Comput. 35, 125–130.
UNPD, 2001. World Population Monitoring 2001: Population, Environment and
Development, A report of United Nations Population DivisionDepartment of
Economics and Social Affairs. . (accessed 20.07.2017) http://www.un.org/esa/
population/publications/wpm/wpm2001.pdf.
UNPD, 2012. United Nations Population Division, World Population Prospects: The
2012 Revision, Total Population Both Sexes. . (accessed 20.07.2017) http://esa.
un.org/wpp/Excel-Data/population.htm.
Verma, M., Misra, A.K., 2016. Optimal control of anthropogenic carbon dioxide
emissions through technological options: a modeling study. Comput. Appl.
Math. doi:http://dx.doi.org/10.1007/s40314-016-0364-2.
Vermeer, M., Rahmstorf, S., 2009. Global sea level linked to global temperature. Proc.
Natl. Acad. Sci. U.S.A. 106 (51), 21527–21532.
Zhang, K.Q., Douglas, B.C., Leatherman, S.P., 2004. Global warming and coastal
erosion. Clim. Change 64, 41–58.