Modeling Dynamics of the Carbon Market: A
System Dynamics Approach on the CO2
Emissions and its Connections to the Oil Market
Mohammadhussein Rafieisakhaei!, Babak Barazandch?,
Amirbahador Moosavi?, Masoud Fekri*, and Kaveh Bastani?
* Dept. of Electrical and Computer Engineering, Texas A&M University, USA
? Dept. of Industrial & Systems Engineering, Virginia Tech, USA
* IFP School, France
4 Dept. of Industrial Engineering, Tehran University, Iran
* mrafieis@tamu.edu
Abstract. Global warming poses a real threat to the sustainable de-
velopment. The temperature data shows that unless the greenhouse gas
are the global can rise critically by
the end of the century. The Emission Trading System (ETS) has been
introduced to control the emissions in the participating countries by
providing economic incentives to the industries and manufacturers to
shift towards cleaner energy resources and technologies. In this paper,
we study the main variables and causes behind the emissions, and inves-
tigate the mechanisms on the carbon market, particularly the European
Union ETS. We also provide an updated oil market model to include the
recent data and developments in the global oil market. We connect the
two models to each other so that the effects of the oil price on the climate
change can be investigated. Next, we train our model with the historic
data and simulate it to show the capability of the proposed model in
predicting the trends of the historic data. Finally, we provide simulation
results to support our model.
1 Introduction
Climate change is real. During the past century the average temperature
of our planet has increased significantly. According to the temperature data,
the global average temperature has increased from -0.375 Celsius degrees in
1850 to an estimate value of 0.894 Celsius degrees in 2016 [i]. The land-surface
air temperature indicates even a higher change from -0.496 degrees to 1.345
degrees between 1850 and 2016 [2]. Moreover, a similar data set shows that
the northern hemisphere’s temperature has increased more than that of the
southern hemisphere. Fig. [I] shows the average global land-ocean temperature
index between 1880 and 2012. The overall message is that human activities
leading to greenhouse gas (GHG) emissions should be controlled heavily.
An initial global attempt towards recognizing the climate change was in 1997,
which led to the establishment of the Kyoto Protocol (KP) [3]. Through the the
jus
Global Land-Ocean Temperature Index
os
Annual Mean (Celsius Degrees)
1988,
1994
2000
2006
212
Dates (Year)
Fig. 1. Average global land-ocean temperature index between 1880 and 2012.
EU-28 Total GHG Emissions Excluding LULUCF
7,000,000.00
6,000,000.00
k=
4,000,000.00
(Thousand Tonnes)
3,000,000.00
2,000,000.00
1,000,000.00
1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012
Dates (Year)
Fig. 2. Historic total EU GHG emissions, excluding Land Use, Land Use Change, and
Forestry (LULUCF) and including international aviation.
Kyoto mechanisms, the KP enabled the participating 37 industrialized countries
to set goals in reducing their carbon emissions by 5% between 2008 and 2012
compared to the Kyoto base year of 1990 [/]. One of the mechanisms was the in-
troduction of an Emission Trading System (ETS). A successful implementation
of an ETS is in the EU zone, which we refer to as the EU ETS. Another mecha-
nism, which also provides financial incentives for industries, is providing means
of investment in low-carbon projects in the developing countries, and transfer-
[If]. These methods, have all been
stem. Figure 2] shows the overall EU
ring cleaner technologies to those countri
adapted and integrated in the EU ETS sy
emissions excluding the Land Use, Land Use Change, and Forestry (LULUCF)
and including international aviation.
In this paper, we provide a model on the main paramete
play a role in an emission trading system. Particularly, we consider the European
implementation of this approach. We utilize the system dynamics methodology
in our modeling and provide our model based on the variables extracted
through a literature study on the main sour
build a casual model between the derived fact
s and variables that
of carbon emissions. Then, we
and parameters. For our sim-
ulation, we build a stock and flow model in which the stock variables and their
ors
rate variables are related through differential equations. The overall model is
reduced to a 4
and is solved numerically using methods such as Runge-Kutta method [10}{13).
The mathematical relations between the main variables are derived using regres:
em of ordinary differential equations with variable
on the available historic data which are used to train the model. For
the cannot be derived, we use look-up
tables. A main loop of the oil market is also provided, and the relations between
the oil market variables and the carbon market are explained. Our simulations
support our results.
sion anal
‘elation:
t of variables where analytic
2 A System Dynamics Model
In this section, we elaborate the main causal loops, and identify the type of
the variables and relations between them. The model co: of two main parts.
The first part of the model provides the governing relations in the carbon market,
whereas the second part systematically models the global oil price which is the
resource in the world. We will explain the carbon market model
scond part. Lastly, we explain
how these two parts connect to each other and interact.
General modeling strategies: Our model is a system dynamics based
model. Like any other system dynamics model, first, the main variables and
parameters are extracted and the casual relations among them are established.
Then, the mathematical relations between each two variable that are connected
through a es represent the variables
whose values have a momentum with respect to time accumulating values over
time, and are generally the outcome of an integral whose relations are modeled
through a first order ordinary differential equation that is ly solved.
The change rates of the stock variable is the variable that controls the stock
main ener;
first, and then elaborate the oil model in the
usal line are developed. The stock variabl
numeric:
variable over the course of time. The other variables which lack the charac
of ion, and take instantaneous values are modeled as flow variab
Main loop in carbon market: The main loop consists of three main variables.
The Emission Allowances Price (EAP), Emission Allowances Demand (EAD)
and the Emission Allowances Supply (EAS) [I7]{19]. The supply for emissi
allowances is considered to be constant over a year. The only way it can change
is through the regulatory policies of the ETS. For i , in the third period of
the EU ETS implementation the supply for allowances is reduced by 1.74% each
Regulatory
Policies of ETS
Transport Sector
Emission
‘Allowances Demand aeaee
: Other Eneray
Emission
Emissi Allowances Price
Allowances Supply Fugitive Emissions —
from Fuels Manufacturing
Industries and
Construction
Industrical
External Variables on Processes
Electricity Supply Electricity
Demand Aptedture
Energy Prices : we
Economic Growth Solvent and Other
External Variables on Product Use
Electricity Demand Temperatures
Fig. 3. Carbon market main loop.
The External Variables on
. The overall causal loop is shown in Fig.
Demand summarize the variables that model the demand side. Similar to any
other ec system, the price of the EU Allowances (EUAs) emerges from
In our model, the price and
the balance between the supply and demand forc
demand of allowances are stock variables [20}{22].
Carbon allowances demand loop: On the demand side, the main factor is the
is mainly due to the fact that the power sector,
which is mainly affected by the electricity consumption, is the biggest CO2 emit-
electricity consumption. This
. Therefore, it is
ity demand
and supply. Generally, the electricity demand changes with the economic activ-
ter, and thus it has the largest demand for the allowances
important to analyze the main factors that determine the electric
ity, temperature and the amount of daylight (the latter two depend on the time
of the year). On the other hand, it is indeed the electricity supply that generates
the emissi mainly the ¢ demand that drives the
he demand through
Based on these, we
ons. Even though it is
electricity supply, the elec
its external imposing facto’
uich as the fuel pric
have modeled the demand for allowances to be affected by the main four factors
weather conditions, and regulatory policies
igated that the deviations from the
predicted (or expected) values of cach of these variables are the main reasons for
big changes in the EUA prices.
cold in winter or extreme high temperatures in summer time (which are above
the average or expected value of the variables) increase the electricity demand,
thus, increasing the fuel consi ion and sub ly i ing the demand
for carbon allowan
of the economic growth, fuel prices
of the ETS. It has been previously inv
For instance, a harsh weather such as extreme
due to the predicted increase in the emi
Oil market main loop: The main factors behind the oil market are the Global
Oil Supply and Global Oil Demand which determine the Global Oil Price. In the
oil market, other than the aggregated oil demand and supply, there are other
factors that determine the oil price, most of which we refer to as the expectational
or anticipated behavioral parameters [I0|31]. These factors are generally formed
(5
instance, in the occurrence of a conflict or riot in an oil produci
oil prices usually react to the events immediately [8]. In such a s
it is expected that the (short-run or long-run depending on the intensity of the
events) oil supply of that country will face volatility, adding to the volatility of
the whole market [32], Although in many situations, the actual supply does not
change as much as expected (or at least as much as the reaction of the oil price),
the predicted trend reflects its influence on the price much earlier than the real
data. Therefore, prediction of the market trends in the oil market
determining the future price. The oil market part of the model utilizes previous
work, [32] where we have updated the modeling approach and incorporated the
latest market data and events ecurred in 2015. The main loop is shown in
Fig.
is crucial in
Expected Global Expected Gloabl
Oil Demand Oil Supply
+
i ,
Expected Trends of ‘ Expected Trends
Demand _ 7 + of Supply
SK _ oil Price
Fig. 4. Oil market main loop.
Effects of the oil market on the carbon emissions: Finally, we connect the
two parts of the model by connecting the economic growths from two parts and
relating the energy prices of the carbon market to the oil prices. In the carbon
market, we assume that the energy prices is a portfolio of the oil price, natural
gas, and coal which are the main carbon emitting resources. However, we provide
the dynamics of oil price from the oil market part of the model, and assume a
linear change of the natural gas and coal prices based on 2015 price data.
3 Simulation Results
In this section, we provide our simulation results to support our model. First,
we provide an analysis of the predicted changes on the average global temper-
ature. Then we provide our Vensim [83] simulation results. In the simulation
Carbon Prices
SERS Rae eRe 88 8 eee
SHAR AKASHRHeARAR SG aA Sa
Seed egseaaagaeas a4 44
Dates (Daily Data)
Fig. 5. Historic carbon prices for EUAs.
scenario, we train our model to reproduce the trends for the historic changes of
the carbon prices. We use the EUA price data from [34] beginning from the 2013
year. A historic data of BUA prices is depicted in Fig
Climate change: As shown in figures [7] and } we have performed first,
second and third order analyses of the temperature changes, respectively. Using
the analysis of these figures, it is conjectured that by the end of the 21st century,
the mean temperature of the earth can increase to 1.0494, 2.7952, and 3.7518
degrees Celsius as a result of first, second and third order analyses, respectively.
This is compliant with the data of the Intergovernmental Panel on Climate
Change (IPCC) which has reported that the projected increase of temperature
in the current century can be anywhere between 0.3 degrees to 4.8 degrees
‘These analysis confirms’ that the global warming can be a threatening event
able development, unless the GHG emissions are controlled more
for the susta
seriously.
Historic data fit: In this scenario, we train our model and its variables accord-
ing to the historic data of the oil and carbon market. Particularly, we consider
the WTI oil price data taken from economic growth data from
supply data from
between Ist of Feb. 2013 and Ist of Feb. 2016. As it is shown in the figure, the
EUA price starts from 5.8 Euros, which we have chosen to be the starting value
in our simulations, as well. The prices drop during April 2014 period and have a
trend of growth until the end of 2015, and start to fall sharply in the beginning
of 2016. As shown in the Fig. J] our results follow the historic data’s trend.
Global Land-Ocean Temperature Index
¥=0,0068x - 0.4466
R= 0.7436
Annual Mean (Celsius Degrees)
Fig. 6. A first order analysis on the global temperature data, with prediction of the
temperature for the rest of the 21st century.
Global Land-Ocean Temperature Index
y= BE-05x7 - 0,004x - 0.1968 at
R
Annual Mean (Celsius Degrees)
Fig. 7. A second order analysis on the global temperature data, with prediction of the
temperature for the rest of the 21st century.
4 Conclusion and Future Work
In this paper, we provided a model of the carbon market, and connected
it to the oil market through a system dynamics approach. We investigated the
main factors and variables involved in the carbon market, particularly, the major
emission resources. We also provided a stock and flow model on the cap-and-trade
market mechanisms that provides an incentive to the industries to move forward
towards the least costly solutions regarding their emissions. By connecting the
oil market model to the carbon market, we could model the effects of oil price on
Global Land-Ocean Temperature Index
# -0,0008x -0.2346 :
0.8684
35 y= 3E-O7x" + 2-0
Re
Annual Mean (Celsius Degrees)
Fig. 8. A third order analysis on the global temperature data, with prediction of the
temperature for the rest of the 21st century.
Real Data vs. Simulation
—fetoats —smuaton Resuts
0
°
°
re
ds
gs
g4
53
2
1
ee
ee ee ee
ARR AAR AARRARRA RA RAR RRR
Dates (Year)
Fig. 9. Emission allowance price simulation result.
the carbon emissions. Our simulations showed that our trained model is capable
of providing a historic fit on the data. Our future works, will extend the model
to include sub-models of the electricity market, as well as a better modeling of
the other energy markets such as natural gas and coal. Moreover, we will provide
more analysis on the sensitivity of the model of the changes of different factors
to study the effects of policies on those parameters.
References
[1] Metoffice.gov.uk (2016)
[2] Climate Change: Vital Signs of the Planet, http://climate.nasa.gov/vital-
signs/global-temperature/ (2015)
[3] Shrestha, R.M., Timilsina, G.R.: The i ity criterion for identifying clean
development mechanism projects under the kyoto protocol. Energy Policy 30(1)
(2002) 73-79
[4] Wara, M.: Is the global carbon market working? Nature 445(7128) (2007) 595-596
[5] Lele, S.M.: Sustainable development: a critical review. World development 19(6)
(1991) 607-621
[6] Jacobson, M.Z., Delucchi, M.A.: Providing all global energy with wind, water,
srials. Energy Policy 39(3) (2011) 1154-1169
Chan, W.V., Moon, L., Roeder, T., Macal, C., Rosetti, M.: A system
ics model on the reasons of car price shocks after economic sanctions
[8] RafieiSakhaei, M., Jabbari, M.: Modeling the impacts of middle east and north
africa unrest on the global oil price. In: Proceedings of the international system
dynamics conference. St. Gallen, Switzerland. (2012)
[9] Barazandeh, B., Rafieisakhaei, M.: A system dynamics model on the reasons of
car price shocks after economic sanctions. In: 2015 Winter Simulation Conference
(WSC), IEEE (2015) 3220-3221
[10] Rafieisakhaei, M., Barazandeh, B., Bolurs
namics of expectations on global oil price.
of the System Dynamics Society. (2015)
[Il] Azadeh, A., Fekri, M., Asadzadeh, S., Barazandeh, B., Barrios, B.: A unique math-
ematical model for maintenance strategies to improve energy flows of the electrical
power sector. Energy Exploration & Exploitation (2016) 0144598715623665
[12] Butcher, J.C.: The numerical analysis of ordinary differential equations: Runge-
Kutta and general linear methods. Wiley-Interscience (1987)
[13] Barazandeh, B., RafieiSakhaei, M., Moosavi, A., Bastani, K.: Effect of localization
on the car market under intense sanctions; em dynamics approach. In: The
34rd International Conference of the System Dynamics Society. (2016)
[14] Sterman, J.D.: Business dynamics: systems thinking and modeling for a complex
world, Volume 19. Irwin/McGraw-Hill Boston (2000)
[15] Rafieisakhaei, M., Barazandeh, B.: The effects of oil market events on carbon
emissions: A 2016 case study. In: SPE Health, Safety, Security, Environment, &
Social Responsibility Conference North America, Society of Petroleum Engineers
(2017 (accepted))
[16] Rafieisakhaei, M., Barazandeh, B.: The efficacy of marketbased emission control
systems: A system dynamics approach. In: SPE Health, Safety, Security, Environ-
ment, & Social Responsibility Conference North America, Society of Petroleum
Engineers (2017 (accepted))
[17] Grubb, M., Neuhoff, K.: Allocation and iti in the eu
trading scheme: policy overview. Climate Policy 6(1) (2006) 7-30
[18] Ellerman, A.D., Convery, F.J., De Perthuis, C.: Pricing carbon: the European
Union emissions trading scheme. Cambridge University Press (2010)
[19] Commission, B., et al.: Sustainable development in the european union. 2009
monitoring report of the EU sustainable development strategy, Luxembourg: Of-
fice for Official Publications of the European Communities (2009)
a
z, M., Assadzadeh, M.: Modeling dy-
In: The 33rd International Conference
[20] RafieiSakhaei, M., Barazandeh, B., Moosavi, A., Fekri, M., Bastani, K.: Supply
and demand dynamics of the oil market: A system dynamics approach. In: The
34rd International Conference of the System Dynamics Society. (2016)
[21] Rafieisakhaei, M., Barazandeh, B.: Modeling dynamics of a Market-Based emis-
sion control system: Efficacy analysis. In: 2016 IEEE Conference on Technologies
for Sustainability (SusTech) (SusTech 2016), Phoenix, USA (October 2016)
[22] Barazandeh, B., Rafieisakhaei, M.: Effect of localization on the sustainable de-
velopment in iran’s car industry. In: 2016 IEEE Conference on Technologies for
Sustainability (SusTech) (SusTech 2016), Phoenix, USA (October 2016)
[23] Peters, G.P., Marland, G., Le Quéré, C., Boden, T., Canadell, J.G., Raupach,
MLR.: Rapid growth in co2 emissions after the 2008-2009 global financial crisis
Nature Climate Change 2(1) (2012) 2-4
Chiodi, A., Gargiulo, M., Rogan, F., Deane, J., Lavigne, D., Rout, U.K., Gal-
lachdir, B.P.O.: Modelling the impacts of challenging 2050 european climate mit-
igation targets on irelands energy system. Energy Policy 58 (2013) 169-189
[25] Bel, G., Joseph, S.: Emission abatement: Untangling the impacts of the eu ets
and the economic crisis. Energy Economics 49 (2015) 531-539
[26] Klassen, R.D., Whybark, D.C.: The impact of environmental technologies on
P ce, Academy of M: journal 42(6) (1999) 599
py
615
Daskalakis, G., Psychoyios, D., Markellos, R.N.: Modeling co 2 emission allowance
prices and derivatives: evidence from the european trading scheme. Journal of
Banking & Finance 33(7) (2009) 1230-1241
[28] Afra, S., Nasr-El-Din, H., Soci, D., Cui, Z., et al.: A novel viscosity reduction
plant-based diluent for heavy and extra-heavy oil. In: SPE Improved Oil Recovery
Conference, Society of Petroleum Engineers (2016)
[29] Schobeiri, M.T., Ghoreyshi, $.M.: The ultrahigh efficiency gas turbine engine with
stator internal ion, Journal of Engineering for Gas Turbines and Power
138(2) (2016) 021506
[30] Ghoreyshi, M., Saidi, M.S., Navabi, M.A., Firoozabadi, B.D., Shabanian, R.: Nu-
merical investigation of antegrade flow effects on flow pulsations in fontan op-
eration. International Journal of Biomedical Engineering and Technology 10(3)
(2012) 221-238
[31] Rafieisakhaei, M., Barazandeh, B., Tarrahi, M.: stem dynamics approach on
oil market modeling with statistical data analysis. In: SPE Middle East Oil &
Gas Show and C Society of Petroleum E s (2017 (accepted))
[32] Rafieisakhaei, M., Barazandeh, B., Tarrahi, M.: Analysis of supply and de-
mand dynamics to predict oil market trends: A case study of 2015 price data,
In: SPE/IAEE Hydrocarbon Economics and Eval ium, Society of
Petroleum Engineers (2016)
[33] Eberlein, R.L., Peterson, D.W.: Understanding models with vensim, European
journal of operational research 59(1) (1992) 216-219
[34] : Ice eua futures (c) - data from quandl. Quandl.com (2016)
[35] Stocker, T.F.: Climate change 2013: the physical science basis: Working Group
I contribution to the Fifth report of the Inter Panel on
Climate Change. Cambridge University Press (2014)
Company, D. Spot oil price: West texas intermediate (discontinued se-
ries) (oilprice). https: //research. stlouisfed. org/fred2/series/OILPRICE/|
downloaddatal (2015) [Online; accessed 13-July-2015].
[37] Databank.worldbank.org: The world bank databank — explore . create . share
(2015)
27)
[36
[38] Eia.gov: International energy statistics - eia. |nttp: //www.eia.gov/cfapps/
{ipdbproject /IEDIndex3. cfm?tid=50&pid=53kaid=1| (2015)
[39] TEA: Tea - oil market report public. Tea.org (2015)
[40] : Greenhouse gas emissions (source: Eea). Appsso.eurostat.ec.europa.eu (2016)
[41] : Statistics of the basque country. En.eustat.eus (2016)
