Available via license: CC BY 4.0
Content may be subject to copyright.
Article citation info:
Borucka A, Kozłowski E. Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19
and decarbonization requirements. Combustion Engines. 0000;XXX(X):xx-xx. https://doi.org/10.19206/CE-000000
COMBUSTION ENGINES, 0000;XXX(X) 3
Anna BORUCKA
Edward KOZŁOWSKI
Polish Scientific Society of Combustion Engines
Modeling the dynamics of changes in CO2 emissions from Polish road transport
in the context of COVID-19 and decarbonization requirements
ARTICLE INFO
Emissions from transport account for 20–25% of anthropogenic global carbon dioxide emissions [17, 37],
with more than 70% coming from road transport, making it an extremely important topic in the context of
decarbonization. The aim of the article is to analyze the trend of CO2 generated from road transport, taking into
account various sources, and also to examine how reduced mobility during the pandemic affected the emissions
at the time. For this purpose, a time series containing observations up to the pandemic outbreak and a time
series containing additional observations from the pandemic period were analyzed. For each time series, a trend
was determined and described by a polynomial and then verified to see if the pandemic phenomenon significant-
ly affects a parameter of the proposed model, using appropriate statistical tests.
Received: 13 May 2023
Revised: 8 July 2023
Accepted: 17 July 2023
Available online: 22 July 2023
Key words: transport decarbonization, CO2 emissions, road transport, COVID-19 pandemic, polynomial model
This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)
1. Introduction
Global warming, the main cause of which is the grow-
ing emission of carbon dioxide, causes drastic consequenc-
es for many ecosystems, bringing about irreversible chang-
es in them [27, 36, 38]. The use of petroleum fuels in
transport determines its significant share in greenhouse gas
emissions, which is why this sector faces the greatest de-
mands [5, 12, 23]. Meanwhile, the progressing globaliza-
tion, population growth, increasing demand for goods, as
well as the dynamic development of the tourism industry
and the number of people travelling do not make this task
easier, especially since it is estimated that by the 2050, the
EU economy will more than double [10].
Emissions from transport account for 20–25% of an-
thropogenic global carbon dioxide emissions [17, 37], of
which 71.7% are from road transport, but taking into ac-
count the production of cars and the construction of road
infrastructure, this number increases to 37% of all emis-
sions [7, 14, 16, 24]. Transport uses 30% of the world’
energy. Although only 7% of the population owns cars, this
translates into 40% of the world's petrol production.
Cars are considered the most polluting means of
transport and at the same time the most unsafe [3, 4, 18].
They are also the largest emitter of toxic chemical com-
pounds not subject to legal regulation, such as butadiene,
benzene and others [22, 25]. The area necessary to build
a road (30 to 40 m on average) is much larger than the re-
quirements for railway traction (10 to 14 m) [25]. What's
more, 30% of car journeys in the European Union do not
exceed 3 km, and 50% – 6 km [39], so they could be suc-
cessfully replaced by environmentally friendly natural
forms of transport, such as bicycles.
The share of Polish road transport in the total emissions
of the European Union is significant. Poland has been oc-
cupying leading positions for years [9]. Moreover, due to
the intensive increase in passenger and transport activity,
CO2 emissions are constantly growing, increasing in 2020
by almost 150% compared to 2000, while emissions
throughout the EU remain relatively constant [2].
The transport sector is therefore a challenge on the way
to achieving climate neutrality, related to the reduction of
greenhouse gases emissions, which is the result of the Eu-
ropean climate policy. Currently, the European Parliament
requires a 40% reduction in greenhouse gas emissions by
EU countries by the 2030, compared to the level of 2005
[29]. This is a recent change (March 2023). The previous
target was 30%.
Therefore, the article analyses the current dynamics of
changes in CO2 emissions from road transport, including
various types of transport means. Mathematical identification
of the examined time series and determination of the forecast
was aimed at relating the current level of CO2 emissions to
the requirements imposed by the EU in this regard.
In addition, the article features an analysis of the impact
of the COVID-19 pandemic on CO2 emissions. Global
movement restrictions, limited travel options and remote
work, as well as fear of contagion, have strongly influenced
changes in people's preferences in their choice of transport
and travel in general. More people staying indoors or using
safer forms of transport like cycling or walking may have
reduced these emissions, as confirmed by studies by
a number of authors [11, 13, 20, 32, 42]. On the other hand,
the fear of coming into contact with an infected person has
fostered a switch from public to private transport [1,6], and
a number of authors believe that the decrease in emissions
was too short-lived to have a relative effect [31, 33]. There-
fore, the article also features an attempt to answer the ques-
tion of how the COVID-19 pandemic in Poland affected the
trend in CO2 emissions, taking into account different modes
of transport.
For this purpose, time series of CO2 emissions from
transport, published by the European Environment Agency,
expressed in Gg [8], were analyzed. Time series with ob-
servations up to the outbreak of the pandemic and time
series with additional observations from the pandemic peri-
od were analyzed.
For each time series, a trend was determined and de-
scribed by a polynomial and then verified to see if the pan-
demic phenomenon significantly affects a parameter of the
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
4 COMBUSTION ENGINES, 0000;XXX(X)
proposed model, using appropriate statistical tests. A de-
tailed survey procedure is presented in Chapter 2. The lack
of confirmation of the significance of the impact of the
pandemic on the model coefficients means that the hazard
did not significantly affect CO2 emissions.
2. Materials and methods
2.1. Determining the trend in a time series
The trend in the time series is identified as
a polynomial [15, 41]
(1)
where is a sequence of independent random varia-
bles with a normal distribution and
. The
trend occurring in the series is identified using a degree
polynomial. The dependence (1) between the endogenous
variable and the predictors (transformations of the variable
) is linear. At the beginning we define objects:
therefore, the dependence (1) can be presented in linear
form [21,24]:
Using the Least Squares Method (LSM) [41] the values
of the model parameters can be estimated using the formula
(3)
The vector of residuals can be presented as
. Coefficient of determination [41] is deter-
mined as follows
(4)
where
and the estimator of the variance of
the residuals is equal
Values of variance of structural parameters [41] are de-
termined as follows
Thus, each of the structural parameters has a normal
distribution for . For each parameter
the significance of the influence of component on the
realizations of the series according to model (1) is tested. At
the significance level for each structural parame-
ter we create a null hypothesis
:
against an alternative hypothesis
: .
The test statistic
(5)
has – distribution with degrees of freedom
[15,34,40]. The test probability is equal:
where is – distribution function with de-
grees of freedom. If then for the parameter the
null hypothesis is rejected in favor of the alternative hy-
pothesis Therefore, the component significantly affects
the realization of the time series defined by formula (1).
The key in the analyzed mathematical equation is the
choice of the degree of the polynomial. To select the appro-
priate polynomial, the Ramsey RESET test (linearity test)
[21, 28] was used in the research. The degree of the poly-
nomial is chosen as the lowest natural number for which the
coefficient of determination is at a sufficiently high level
and there are no grounds to reject the null hypothesis for the
linearity test.
2.2. Study of the impact of COVID-19 on CO2 emissions
First, a time series of CO2 emissions is considered for
pre-pandemic data, i.e., data covering the years up to and
including 2019. Thus, for the series the model
(1) is identified and, taking into account the linearity test of
the models, the degree of the polynomial is deter-
mined. Using LSM the structural parameters of the model
and the variances of these structural parameters are deter-
mined, therefore for .
Next, the time series of CO2 emissions for data that in-
cludes the COVID-19 pandemic are analyzed, i.e., for data
including the year 2020. Thus, for the series the
model (1) for the degree of the polynomial is identi-
fied. Using LSM the structural parameters of the model and
the variances of these structural parameters are determined,
therefore for .
In the next step, for each indices, at signifi-
cance level the null hypothesis is created:
: (no impact of pandemic on parameter )
against an alternative hypothesis:
: (significant impact of pandemic on pa-
rameter )
The test statistic
(6)
has a normal distribution The test proba-
bility is equal:
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
COMBUSTION ENGINES, 0000;XXX(X) 5
where denotes the standard normal distribution
. If then the null hypothesis is rejected
in favor of .
If there is such index, for which , it is
considered that the estimator of parameter determined
from observations up to the pandemic and observations
with the onset of the pandemic are significantly different,
and therefore the pandemic had a significant impact on the
trend of CO2 emissions. If for each indices, the
condition is satisfied, then for each structural
parameter, the estimators obtained from observations up to
the pandemic and observations with the start of the pan-
demic are not significantly different from each other, there-
fore the pandemic did not have a significant impact on CO2
emissions.
3. Evaluation of the trend of CO2 emissions in road
transport
3.1. Road transport
First, CO2 emissions from road transport as a whole
were evaluated without breaking them down by mode of
transport. Figure 1 presents the time series under study
(black curve).
Fig. 1. CO2 Emission for Road Transportation [8]
Then, for the time series presented, a model built only for
the pre-pandemic period was proposed (blue line in Fig. 1).
The estimators of structural parameters, standard deviations,
t-statistic values and p-values are presented in Table 1.
Table 1. Structural parameters, standard errors, values of t-statistic and p-
values for the trend to the pre-pandemic period
Estimate
Std. error
t value
p-value
7115.71935
5166.94560
1.37716
0.18172
22985.07637
6162.91969
3.72958
0.00110
–12041.9597
2495.18762
–4.82607
0.00007
2769.26980
479.62213
5.77386
0.00001
–327.12113
50.15569
–6.52211
< 1e–5
21.43286
3.01362
7.11200
< 1e–5
–0.78550
0.10381
–7.56667
< 1e–5
0.01505
0.00190
7.90711
<1e–5
–0.00012
0.00001
–8.15249
<1e–5
The coefficient of determination is equal to 0.9907, and
the standard deviation of the residuals is equal to 1524.28.
At the significance level , the hypothesis was
rejected in favor of for the structural parameters of
, therefore the predictors of these
parameters significantly affect the polynomial trend in the
time series. The value of the statistic for the Ramsey
RESET test is 1.6778, while the p-value is 0.2496.
In the next step, a model was built for data covering the
period of the pandemic (red line in Fig. 1). Table 2 presents
the estimators of structural parameters, standard deviations,
values of t-statistics and p-values for the entire period.
Table 2. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the entire period
Estimate
Std. error
t value
p-value
12173.08505
5969.19873
2.03932
0.05258
15486.44602
6946.95856
2.22924
0.03542
–8500.03179
2739.16084
–3.10315
0.00485
2003.84146
512.08133
3.91313
0.00066
–239.32353
52.03960
–4.59887
0.00012
15.74571
3.03706
5.18452
0.00003
–0.57695
0.10158
–5.67979
0.00001
0.01102
0.00181
6.09446
< 1e–5
–0.00009
0.00001
–6.43854
< 1e–5
The coefficient of determination is equally high at
0.9873. The standard deviation of the residuals is equal to
1836.77. At the significance level , the hypothesis
was rejected in favor of for the parameters
. Thus, predictors at these param-
eters significantly affect the polynomial trend in the time
series.
For the polynomials constructed above, their parameters
were compared. Table 3 presents the calculated values of
statistics (6) and p-values for the test of differences of
structural parameters.
Table 3. T-statistic values and p-value for the tests of differences in struc-
tural parameters
T
p-value
–3.10942
0.00187
3.91841
0.00009
–4.63753
< 1e–5
5.29107
< 1e–5
–5.88978
< 1e–5
6.44278
< 1e–5
–6.95724
< 1e–5
7.43869
< 1e–5
–7.89137
< 1e–5
At the significance level , the hypothesis
was rejected in favor of for all structural parameters.
Thus, the parameters are significantly different for the poly-
nomial trend determined for the CO2 emission series up to
the time of the pandemic outbreak and the polynomial trend
determined for the CO2 emission series containing the first
year of the pandemic. Thus, the COVID-19 pandemic sig-
nificantly affected the CO2 emissions trend analyzed for
road transport in Poland, as a whole.
In the next stage of the study, an analogous analysis was
made, however, taking into account the different modes of
transport. Passenger cars, light duty trucks, heavy duty
trucks and motorcycles were studied.
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
6 COMBUSTION ENGINES, 0000;XXX(X)
3.2. Passenger cars
Passenger cars make up the majority of the vehicle mar-
ket in Poland (more than 60% of the market). The latest
available data shows that at the end of 2022, 26.675 million
passenger cars were registered in the database of the Cen-
tral Register of Vehicles and Drivers (CEPiK), 577,000
more than a year earlier [35]. In Fig. 2 CO2 emissions
(black line) are also on an upward trend.
Fig. 2. CO2 emissions for cars [8]
According to the adopted algorithm, the model was first
built for the pre-pandemic period only (blue line in Fig. 2).
Its structural parameters, standard deviations, t-statistic
values and p-values are presented in Table 4.
Table 4. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the pre-pandemic period
Estimate
Std. error
t value
p-value
–902.69310
2570.25660
–0.35121
0.72863
12031.29715
3065.69611
3.92449
0.00068
–5793.93596
1241.21153
–4.66797
0.00011
1326.84643
238.58427
5.56133
0.00001
–156.72823
24.94956
–6.28180
< 1e–5
10.23508
1.49910
6.82748
< 1e–5
–0.37295
0.05164
–7.22225
< 1e–5
0.00710
0.00095
7.49445
< 1e–5
–0.00005
0.00001
–7.66937
< 1e–5
The coefficient of determination equals 0.993, and the
standard deviation of the residuals equals 758.24. At the
significance level , the hypothesis was rejected
in favor of for the structural parameters of
, therefore the predictors of these
parameters significantly affect the polynomial trend in the
time series. The value of the statistic for the Ramsey RE-
SET test is 0.9118, while the p-value is 0.5895.
Consistently, in a further step, a model was built for da-
ta covering the time of the pandemic (red line in Fig. 2).
Table 5 presents structural parameters, standard deviations
of parameters, values of t-statistics and p-values for the
entire period.
The coefficient of determination is equal to 0.9921, and
the standard deviation of the residuals is 825.34. At the
significance level , the hypothesis was rejected
in favor of for the
parameters. Thus, also in this case, the predictors at these
parameters significantly affect the polynomial trend in the
time series.
Table 5. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the entire period
Estimate
Std. Error
t value
p-value
801.28414
2682.23759
0.29874
0.76771
9504.78508
3121.59039
3.04485
0.00558
–4600.55486
1230.83189
–3.73776
0.00102
1068.95081
230.10187
4.64555
0.00010
–127.14660
23.38380
–5.43738
0.00001
8.31891
1.36469
6.09581
< 1e–5
–0.30269
0.04564
–6.63143
< 1e–5
0.00574
0.00081
7.06237
< 1e–5
–0.00004
0.00001
–7.40639
< 1e–5
For polynomials constructed for CO2 emitted from
passenger vehicles, their parameters were similarly
compared. Table 6 presents the estimated values of
statistics (6) and p-values for the test of differences of
structural parameters.
Table 6. T-statistic values and p-value for the tests of differences of struc-
tural parameters
T
p-value
–2.22406
0.02614
2.79922
0.00512
–3.30850
0.00094
3.76942
0.00016
–4.18986
0.00003
4.57649
< 1e–5
–4.93458
< 1e–5
5.26820
< 1e–5
–5.58048
< 1e–5
At the significance level , the hypothesis
was rejected in favor of for all structural parameters.
The parameters are significantly different for the polynomi-
al trend determined for the CO2 emission series up to the
pandemic and the polynomial trend determined for the CO2
emission series containing the first year of the pandemic.
Thus, in the case of passenger cars, the COVID-19 pandem-
ic has significantly affected the trend in CO2 emissions.
3.3. Light and heavy duty trucks
Trucks, which are the second largest group of vehicles
registered in Poland [19, 26, 35] were then examined. As of
the end of 2021, the number of registered trucks (including
goods and passenger carrying vehicles) will reach 3.6 mil-
lion [30], 3.0% more than a year ago. CO2 emissions from
light duty tracks (black line in Fig. 3) and heavy duty trucks
(black line in Fig. 4) were analyzed.
Fig. 3. CO2 emissions for light duty truck [8]
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
COMBUSTION ENGINES, 0000;XXX(X) 7
Fig. 4. CO2 emissions for heavy duty trucks [8]
The models built for the pre-pandemic period (blue line
in Fig. 3 and 4), their structural parameters, standard devia-
tions of the parameters, values of the t-statistic and p-values
are presented in Table 7.
Table 7. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the pre-pandemic period
Estimate
Std. error
t value
p-value
Light duty truck
7347.38426
585.58715
12.54704
< 1e–5
–1103.45576
237.77329
–4.64079
0.00008
136.30530
28.67462
4.75352
0.00006
–5.78733
1.29616
–4.46498
0.00013
0.08198
0.01950
4.20497
0.00026
Heavy duty truck
5669.98087
2480.91621
2.28544
0.03183
4898.82097
2959.13457
1.65549
0.11141
–3354.68039
1198.06785
–2.80008
0.01017
850.59185
230.29124
3.69355
0.00120
–106.96717
24.08233
–4.44173
0.00019
7.34736
1.44699
5.07767
0.00004
–0.27943
0.04984
–5.60608
0.00001
0.00552
0.00091
6.03490
< 1e–5
–0.00004
0.00001
–6.37530
< 1e–5
For CO2 emissions from light duty trucks, the coeffi-
cient of determination is equal to 0.7679 and the standard
deviation of the residuals is 542.63. At the significance
level , it was found that the structural parameters
are significantly different from zero (Table
7), thus, the predictors of these parameters significantly
affect the polynomial trend in the time series. The value of
the statistic for the Ramsey RESET test is 2.1032, while the
p-value is 0.0877.
At the significance level for heavy duty
trucks, it was found that parameters
are significantly different from zero (Tab. 7), thus
predictors at these parameters significantly affect the trend
of CO2 emissions. The value of the statistic for the Ramsey
test is 1.4937, while the p-value is 0.305. The coefficient of
determination is higher and equal to 0.9856. The standard
deviation of the residuals, meanwhile, is 731.88.
The models built for the entire study period (red line in
Fig. 3 and 4), their structural parameters, standard deviations,
t-statistic values and p-values are presented in Table 8.
Table 8. Structural parameters, standard errors, values of t-statistic and
p-value for the trend for the entire period
Estimate
Std. error
t value
p-value
Light duty truck
7166.73855
585.40978
12.24226
< 1e–5
–993.26539
230.94782
–4.30082
0.00019
119.85381
27.04951
4.43091
0.00013
–4.92894
1.18699
–4.15246
0.00028
0.06762
0.01733
3.90227
0.00055
Heavy duty truck
8345.93187
2968.73528
2.81128
0.00967
931.14915
3455.01664
0.26951
0.78984
–1480.57710
1362.30067
–1.08682
0.28791
445.58876
254.67973
1.74960
0.09296
–60.51175
25.88150
–2.33803
0.02805
4.33818
1.51046
2.87209
0.00839
–0.16909
0.05052
–3.34697
0.00269
0.00338
0.00090
3.76141
0.00096
–0.00003
0.00001
–4.11824
0.00039
For CO2 emissions from light duty trucks, the coeffi-
cient of determination for the entire study period is 0.7621
and the standard deviation of the residuals is 554.42. At the
significance level , thus all parameters of the
model are significantly different from zero. For the trend
model of CO2 emissions generated by heavy duty trucks,
the coefficient of determination is 0.9792, and the standard
deviation of the residuals is 913.5. At the significance level
, it was found that the structural parameters
significant are different from zero, thus
the predictors at these parameters significantly affect the
trend occurring in the time series.
Comparing the two models with each other for each of
the modes of transport analyzed, it can be seen again that
the COVID-19 pandemic also affected the trend of CO2
emissions from trucks. Table 9 presents the values of statis-
tic (6) and p-values for the test of differences in the struc-
tural parameters. For light duty trucks at the significance
level , the hypothesis was rejected in favor of
for the structural parameters , , , while for heavy
duty trucks the parameters , , , , , , , ,
are significantly different for the polynomial trend deter-
mined for the CO2 emission series up to the pandemic and
the polynomial trend determined for the CO2 emission
series containing the first year of the pandemic.
Table 9. T-statistic values and p-value for the tests of differences of struc-
tural parameters
T
p-value
T
p-value
Light duty truck
Heavy duty truck
1.14388
0.25267
–3.35853
0.00078
–1.74252
0.08142
4.23413
0.00002
2.18704
0.02874
–5.01350
< 1e–5
–2.55875
0.01051
5.72280
< 1e–5
2.88339
0.00393
–6.37357
< 1e–5
6.97556
< 1e–5
–7.53647
< 1e–5
8.06220
< 1e–5
–8.55730
< 1e–5
3.4. Motorcycles
The analyses conducted ended with a study of motorcy-
cles. Despite their growing popularity — 23,910 new motor-
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
8 COMBUSTION ENGINES, 0000;XXX(X)
cycles were registered in 2022, 10.8% more than in the pre-
vious year [30, 35]. It is clear that CO2 emissions have been
on a downward trend over the years (black line Fig. 5).
Fig. 5. CO2 Emission for Motorcycles [8]
The estimators of parameters of the model constructed
for the pre-pandemic period (blue line in Fig. 5), standard
deviations of parameters, t-statistics and p-values are pre-
sented in Table 10.
Table 10. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the pre-pandemic period
Estimate
Std. error
t value
p-value
452.59779
17.62924
25.67313
< 1e–5
–0.94515
0.16588
–5.69780
< 1e–5
0.02087
0.00535
3.90082
0.00052
The coefficient of determination is 0.8293 and the
standard deviation of the residuals is 48.66. At the signifi-
cance level , the hypothesis was rejected in
favor of for the structural parameters , , , thus the
predictors of these parameters significantly affect the poly-
nomial trend in the time series. The value of the statistic for
the Ramsey RESET test is 1.3867, while the p-value is
0.2638.
The estimator of model parameters for the entire study
period (red line in Fig. 5), standard deviations of parame-
ters, t-statistics and p-values are presented in Table 11.
Table 11. Structural parameters, standard errors, values of t-statistic and
p-value for the trend to the entire period
Estimate
Std. error
t value
p-value
451.31832
17.10369
26.38718
< 1e–5
–0.91973
0.15152
–6.07003
< 1e–5
0.01992
0.00474
4.20027
0.00022
The coefficient of determination is 0.8368, and the
standard deviation of the residuals is 47.97. At the signifi-
cance level , the hypothesis was rejected in
favor of for the structural parameters , , , thus the
predictors at these parameters significantly affect the poly-
nomial trend in the time series.
Comparing the parameters of the models for each period
this time led to different conclusions. At the significance
level , there is no basis to reject the hypothesis
for all structural parameters. Thus, the parameters are not
significantly different for the polynomial trend determined
for the CO2 emission series up to the pandemic and the
polynomial trend determined for the CO2 emission series
containing the first year of the pandemic. The trend pa-
rameters of the period up to the pandemic and the period
with the onset of the pandemic are not significantly differ-
ent, thus there was no significant effect on the trend. Table
12 presents the values of statistic (6) and p-values for the
structural parameter difference test.
Table 12. T-statistic values and p-value for the tests of differences of
structural parameters
T
p-value
0.28281
0.77732
–0.61391
0.53927
0.72418
0.46895
This result of the study of CO2 emissions from motorcy-
cles is probably due to the fact that traveling on a motorcy-
cle was not associated with an increased risk of danger, as
all single-track vehicles were a safe means of transport
from the point of view of virus infection. This is why the
result is so different from other modes of transport.
4. Conclusion
The primary objective of the study was to assess wheth-
er the COVID-19 pandemic significantly affected CO2
emissions from road transport. A general analysis was made
first, without distinguishing between different modes of
transport, and then passenger cars, light and heavy-duty
trucks and motorcycles were examined. It turned out that
only in the case of motorcycles was the impact of the pan-
demic not significant. Thus, global mobility restrictions and
probably the fear of becoming infected have influenced
public behavior. Despite the woeful pandemic period,
meaningful conclusions can also be drawn. The study
shows that changes are possible regarding CO2 emissions
from transport, but they require comprehensive, systemic
solutions.
The obtained results clearly prove that it is possible to
implement mechanisms for controlling society on a global
scale and to achieve the desired results in terms of harmful
emissions. The solutions implemented in this area require
the use of appropriate analysis methods that will allow for
a reliable assessment of whether the implemented changes
bring the expected results and whether the obtained effect is
statistically significant. The method proposed in the article
can be successfully used for such analyses. The algorithm
proposed in the methodological chapter is universal and can
also be applied to factors other than the COVID-19 pan-
demic. This is an additional advantage of the article.
In early 2023, the European Parliament approved new
targets to reduce CO2 emissions produced by new passenger
and goods carrying vehicles by 100 percent by 2035 com-
pared to 2021. With regard to the results presented, these
assumptions seem difficult to implement. Over the entire
period studied, for every mode of transport except motorcy-
cles, the research presented shows an upward trend. Only
the COVID-19 pandemic caused small declines but only for
passenger cars and heavy duty trucks. Thus, changes in this
area are necessary. Given the short time left for Poland and
the European Union as a whole to achieve climate neutrali-
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
COMBUSTION ENGINES, 0000;XXX(X) 9
ty, it is necessary to introduce the principles of sustainable
development and take a holistic view of the environmental
impact of individual modes of transport.
In the article, the study was conducted at a general level,
observing global trends and considering whether they are
possible to change. This provides the basis for more de-
tailed analyzes and the search for relationships between
specific solutions and CO2 emissions.
In addition, due to the fact that the publication of emis-
sions data is delayed, so the authors have not had the oppor-
tunity to make a study based on complete historical data, it
is necessary to continue research and analyze trends. How-
ever, this will be the direction of further research planned in
this area.
Bibliography
[1] Advani M, Sharma N, Dhyani R. Mobility change in Delhi
due to COVID and its’ immediate and long term impact on
demand with intervened non motorized transport friendly in-
frastructural policies. Transport Policy. 2021;111:28-37.
https://doi.org/10.1016/j.tranpol.2021.07.008
[2] Antosiewicz M, Gorzałczyński A, Rabiega W. The role of
public transport in Poland in the pursuit of climate neutrality
(in Polish). 2023.
https://managerplus.pl/rola-transportu-publicznego-w-
polsce-w-dazeniu-do-neutralnosci-klimatycznej-54351 (ac-
cessed on 30.04.2023).
[3] Borucka A, Kozłowski E, Oleszczuk P, Świderski A. Predic-
tive analysis of the impact of the time of day on road acci-
dents in Poland. Open Engineering. 2021;11(1):142-150.
https://doi.org/10.1515/eng-2021-0017
[4] Borucka A, Pyza D. Influence of meteorological conditions
on road accidents. A model indexed by: for observations
with excess zeros. Eksploat Niezawodn. 2021;23(3):586-
592. https://doi.org/10.17531/ein.2021.3.20
[5] de Blas I, Mediavilla M, Capellán-Pérez I, Duce C. The
limits of transport decarbonization under the current growth
paradigm. Energy Strateg Rev. 2020;32:100543.
https://doi.org/10.1016/j.esr.2020.100543
[6] Eisenmann C, Nobis C, Kolarova V, Lenz B, Winkler C.
Transport mode use during the COVID-19 lockdown period
in Germany: The car became more important, public
transport lost ground. Transp Policy. 2021;103:60-67.
https://doi.org/10.1016/j.tranpol.2021.01.012
[7] European Environment Agency, Air quality in Europe 2022.
https://www.eea.europa.eu/publications/air-quality-in-
europe-2022
[8] European Environment Agency. National emissions reported
to the UNFCCC and to the EU Greenhouse Gas Monitoring
Mechanism.
https://www.eea.europa.eu/en/datahub/datahubitem-
view/3b7fe76c-524a-439a-bfd2-a6e4046302a2
[9] Global Carbon Atlas. CO2 Emissions |Global Carbon Atlas.
Globalcarbonatlas.org. 2020.
http://www.globalcarbonatlas.org/en/CO2-emissions
[10] Going climate-neutral by 2050. A strategic long-term vision
for a prosperous, modern, competitive and climate-neutral
EU economy. 2050 long-term strategy (europa.eu)
[11] Gössling S, Humpe A, Fichert F, Creutzig F. COVID-19 and
pathways to low-carbon air transport until 2050. Environ
Res Lett. 2021;16(3):034063.
https://doi.org/10.1088/1748-9326/abe90b
[12] Graba M, Bieniek A, Prażnowski K, Hennek K, Mamala J,
Burdzik R et al. Analysis of energy efficiency and dynamics
during car acceleration. Eksploat Niezawodn. 2023;25(1):
17. https://doi.org/10.17531/ein.2023.1.17
[13] Griffiths S, Furszyfer Del Rio D, Sovacool B. Policy mixes
to achieve sustainable mobility after the COVID-19 crisis.
Renew Sust Energ Rev. 2021;143:110919.
https://doi.org/10.1016/j.rser.2021.110919
[14] Groll W, Sowiński B, Konowrocki R. Study of transitional
phenomena in rail vehicle dynamics in relation to the relia-
bility and operational state of the continuous welded rail
track in terms of rail joints. Eksploat Niezawodn. 2023;
25(1):7. https://doi.org/10.17531/ein.2023.1.7
[15] Hamilton JD. Time series analysis, Princeton University
Press. 1994. https://doi.org/10.1515/9780691218632
[16] IEA. Net Zero by 2050 – Analysis. IEA. 2021.
https://www.iea.org/reports/net-zero-by-2050
[17] International Energy Agency. CO2 Emissions in 2022.
https://iea.blob.core.windows.net/assets/3c8fa115-35c4-
4474-b237-1b00424c8844/CO2Emissionsin2022.pdf
[18] Kozłowski E, Borucka A, Świderski A, Skoczyński P. Clas-
sification trees in the assessment of the road–railway acci-
dents mortality. Energies. 2021;14(12):3462.
https://doi.org/10.3390/en14123462
[19] Kukulski J, Lewczuk K, Góra I, Wasiak M. Methodological
aspects of risk mapping in multimode transport systems. Ek-
sploat Niezawodn. 2023;25(1):19.
https://doi.org/10.17531/ein.2023.1.19
[20] Le Quéré C, Jackson RB, Jones MW, Smith AJP, Abernethy
S, Andrew RM et al. Temporary reduction in daily global
CO2 emissions during the COVID-19 forced confinement.
Nat Clim Change. 2020;10(10):1-7.
https://www.nature.com/articles/s41558-020-0797-x
[21] Lee TH, White H, Granger CWJ. Testing for neglected
nonlinearity in time series models. J Econometrics. 1993;
56(3):269-290.
https://doi.org/10.1016/0304-4076(93)90122-L
[22] Li B, Wang J, Wang J, Zhang L, Zhang Q. A comprehensive
study on emission of volatile organic compounds for light
duty gasoline passenger vehicles in China: Illustration of
impact factors and renewal emissions of major compounds.
Environ Res. 2021;193:110461.
https://doi.org/10.1016/j.envres.2020.110461
[23] Mączak J, Więcławski K, Szczurowski K. New approach of
model based detection of early stages of fuel injector fail-
ures. Eksploat Niezawodn. 2023;25(1):6.
https://doi.org/10.17531/ein.2023.1.6
[24] Melnik R, Koziak S, Dižo J, Kuźmierowski T, Piotrowska
E. Feasibility study of a rail vehicle damper fault detection
by artificial neural networks. Eksploat Niezawodn. 2023;
25(1):5. https://doi.org/10.17531/ein.2023.1.5
[25] National Cluster of Innovative Enterprises, Poland 3.0 – A
response to modern transport needs (in Polish: Ogólnopolski
Klaster Innowacyjnych Przedsiębiorstw, Polska 3.0 – Od-
powiedź na współczesne potrzeby transportu). 2016.
http://klasterip.pl/polska-3-0-odpowiedz-na-wspolczesne-
potrzeby-transportu-jedyny-spojny-projekt-
infrastrukturalny-w-polsce/
[26] Oszczypała M, Ziółkowski J, Małachowski J. Semi-Markov
approach for reliability modelling of light utility vehicles.
Eksploat Niezawodn. 2023;1:25(2).
https://doi.org/10.17531/ein/161859
Modeling the dynamics of changes in CO2 emissions from Polish road transport in the context of COVID-19…
10 COMBUSTION ENGINES, 0000;XXX(X)
[27] Papalexiou SM, Montanari A. Global and regional increase
of precipitation extremes under global warming. Water Re-
sour Res. 2019;55(6):4901-4914.
https://doi.org/10.1029/2018WR024067
[28] Ramsey JB. Tests for specification errors in classical linear
least-squares regression analysis. J Roy Stat Soc B Met.
1969;31(2):350-371. https://www.jstor.org/stable/2984219
[29] Regulation of the European Parliament and of the Council
amending Regulation (EU) 2018/842 on binding annual
greenhouse gas emission reductions by Member States from
2021 to 2030 contributing to climate action to meet com-
mitments under the Paris Agreement.
https://eur-lex.europa.eu/legal-
content/PL/TXT/?uri=CELEX:52021PC0555
[30] Road transport in Poland in the years 2020 and 2021 (in
Polish), Główny Urząd Statystyczny Warszawa, Szczecin
2023.
[31] Rojas C, Muñiz I, Quintana M, Simon F, Castillo B, de la
Fuente H et al. Short run “rebound effect” of COVID on the
transport carbon footprint. Cities. 2022;131:104039.
https://www.sciencedirect.com/science/article/pii/S0264275
122004784.
[32] Rojas-Rueda D, Morales-Zamora E. Built environment,
transport, and COVID-19: a review. Current Environmental
Health Reports. 20215;8(2):138-145.
https://doi.org/10.1007/s40572-021-00307-7
[33] Schulte-Fischedick M, Shan Y, Hubacek K. Implications of
COVID-19 lockdowns on surface passenger mobility and re-
lated CO2 emission changes in Europe. Appl Energ. 2021;
300:117396.
https://doi.org/10.1016/j.apenergy.2021.117396
[34] Shumway RH, Stoffer DS. Time series analysis and its
applications, Springer International Publishing. 2017.
[35] Statistics | Information Portal of the Central Register of
Vehicles and Drivers (in Polish). www.cepik.gov.pl.
http://www.cepik.gov.pl/statystyki
[36] Thompson JR, Gosling SN, Zaherpour J, Laizé CLR. In-
creasing risk of ecological change to major rivers of the
world with global warming. earth’s future. Earth’s Future.
2021;9(11):1-20. https://doi.org/10.1029/2021EF002048
[37] United Nations. Fact Sheet Climate Change.
https://www.un.org/sites/un2.un.org/files/media_gstc/FACT
_SHEET_Climate_Change.pdf
[38] Wang J, Guan Y, Wu L, Guan X, Cai W, Huang J et al.
Changing lengths of the four seasons by global warming.
Geophys Res Lett. 2021;22:48(6).
https://doi.org/10.1029/2020GL091753
[39] Weiss M, Paffumi E, Clairott M, Drossinos Y, Vlachos T,
Bonnel P et al. Including cold-start emissions in the Real-
Driving Emissions (RDE) test procedure. Publications Of-
fice of the European Union. 2017.
https://publications.jrc.ec.europa.eu/repository/bitstream/JR
C105595/kjna28472enn.pdf
[40] Woodward WA, Gray HL, Elliott AC. Applied time series
analysis with R, CRC Press. 2017.
https://doi.org/10.1201/9781315161143
[41] Wooldridge JM. Introductory econometrics: a modern ap-
proach, South Western Educ Pub. 2019.
[42] Zhang R, Zhang J. Long-term pathways to deep decarboni-
zation of the transport sector in the post-COVID world.
Transp Policy. 2021;110:28-36.
https://doi.org/10.1016/j.tranpol.2021.05.018
Anna Borucka, DSc., DEng. – Faculty of Security,
Logistics and Management, Military University of
Technology, Poland.
e-mail: anna.borucka@wat.edu.pl
Edward Kozłowski, DSc., DEng. – Faculty of Man-
agement, Lublin University of Technology, Poland.
e-mail: e.kozlovski@pollub.pl