Content uploaded by Sangdon Lee
Author content
All content in this area was uploaded by Sangdon Lee on Apr 24, 2024
Content may be subject to copyright.
Global Ecology and Conservation 51 (2024) e02949
Available online 10 April 2024
2351-9894/© 2024 Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/by-nc-nd/4.0/).
Short communication
The effect of COVID-19 movement restriction on Korean
expressway wildlife−vehicle collisions
Hyomin Park , Sangdon Lee
*
Department of Environmental Science & Engineering, College of Engineering, Ewha Womans University, Seoul 03760, South Korea
ARTICLE INFO
Keywords:
Wildlife management
Conservation strategy
Roadkill
Road ecology
Hotspots
ICT-Environmental Impact Assessment
ABSTRACT
The construction of roads obstructs animal movement, which directly results in fatal wild-
life–vehicle collisions (WVCs). Owing to the worldwide COVID-19 outbreak in 2020, movement
restrictions were imposed by the Korean government between February and March 2020, which
represented the beginning of the COVID-19 pandemic. During this period, this study examined the
effects of reduced trafc volume on WVCs. The study selected the Dangjin−Yeongdeok, Jungang,
and Jungbu Expressways as study sites since they have the highest occurrence of WVCs in Korea.
The sections of these expressways with the highest densities of WVCs were denoted as WVC
hotspots. We categorized the period from February to March 2020 as ‘strict pandemic’. In
addition, the period from February to March 2015–2019 was categorized as “pre-strict
pandemic”, and between April 2020 and December 2020 was classied as the “pandemic period”.
We analyzed the relationship between WVC and trafc volume for each period in the designated
hotspots. As a result, there were statistically signicant differences in WVCs and trafc volumes
across all routes for different periods. WVCs and trafc volume showed a strong negative cor-
relation during the strict pandemic period. WVC per trafc volume was positively correlated
between the strict pandemic period and the pandemic period, while it was negatively correlated
between pre-strict pandemic and strict pandemic periods. The deceleration in WVC continued
after all routes were out of strict pandemic. This study shows that WVC and trafc volume are
negatively correlated and that a strong, albeit temporary, trafc reduction can help reduce WVC.
Therefore, the overall quantitative conclusion of this study on the interaction can serve as a
valuable reference to develop practical strategies for preventing WVCs.
1. Introduction
The increasing number of new roadways has become a considerable threat to wildlife globally (Primack et al., 2021). With the
continuous expansion of towns and heavier trafc between cities, newly constructed roads have increasingly fragmented wildlife
habitats (Forman et al., 2003), eventually, isolating, and conning wildlife. Roadways obstruct animal movement, inuence species
diversity, and directly result in fatal wildlife–vehicle collisions (WVCs) (Forman et al., 2003; Laurance et al., 2009; Van der Ree et al.,
2015). Due to the use of wider lanes and adjacency to forests, expressways have higher trafc speeds and volumes than other road
types, thereby making them more prone to WVCs, and thus, a signicant threat to wildlife (Park et al., 2021). This problem is
exemplied in South Korea, where persistent industrial growth necessitates the construction of additional expressways, which
* Corresponding author.
E-mail address: lsd@ewha.ac.kr (S. Lee).
Contents lists available at ScienceDirect
Global Ecology and Conservation
journal homepage: www.elsevier.com/locate/gecco
https://doi.org/10.1016/j.gecco.2024.e02949
Received 30 October 2023; Received in revised form 8 April 2024; Accepted 10 April 2024
Global Ecology and Conservation 51 (2024) e02949
2
increases the threat to wildlife.
Wildlife–vehicle collisions (WVCs) are a signicant threat to many species, causing nancial loss, and posing a serious risk to driver
safety (Laube et al., 2023). Since WVCs have caused severe problems worldwide, researchers have actively investigated various
associated issues, including the socioeconomic costs of WVCs (Huijser et al., 2009) and mitigation measures (Lester, 2015). Several
Fig. 1. A map of study sites in South Korea showing sections of the surveyed expressways.
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
3
factors affect the incidence of WVCs, including spatial characteristics (Park et al., 2021), trafc volume changes (Bencin et al., 2019),
and seasonal migrations of wildlife (Main and Allen, 2002). Among these, trafc volume is a critical predictor of wildlife roadkill rates
(Taylor and Goldingay, 2010; Visintin et al., 2017; Pagany, 2020). The presence of roadways generally affects wildlife (Fahrig and
Rytwinski, 2009), and trafc volume is strongly correlated to the frequency and potential risks of WVCs (Litvaitis and Tash, 2008). In
South Korea, the escalating incidence of WVCs has also become a nationwide problem, thereby prompting further research. However,
the relationship between WVCs on expressways and trafc volume in the country requires further investigation.
Owing to the worldwide COVID-19 outbreak in 2020, movement restrictions were imposed by the Korean government between
February and March 2020 to curb the spread of the virus—the beginning of the COVID-19 pandemic. These restrictions signicantly
reduced human activities (Primack et al., 2021), thereby providing opportunities to observe the human inuence on WVCs (Driessen,
2021). Several studies have reported that changes in human activity during the pandemic altered the behavior of wild animals
(Manenti et al., 2020; Rutz et al., 2020). Thus, during this period, this study examined the effects of reduced trafc volume on WVCs.
Based on nationwide WVC data from 2015 to 2020, this study selected the Dangjin−Yeongdeok, Jungang, and Jungbu Expressways
as study sites since they have the highest occurrence of WVCs in Korea. The sections of these expressways with the highest densities of
WVCs were denoted as WVC hotspots, meaning that in these hotspots, the risk of a collision was statistically higher than in any other
section of the expressway (Seiler et al., 2016). The identication of the hotspot locations is important for the effective application of
mitigation measures (Bíl et al., 2016). Therefore, we analyzed the relationship between the WVCs and trafc volume in these hotspots.
To our knowledge, there has been no previous study conducted in Korea on the changes in the occurrence of WVCs during the
pandemic period. Therefore, this study seeks to establish important basic data necessary for the management plan of WVC occurrences
on expressways.
2. Materials and methods
2.1. Study sites
To choose the appropriate areas for this study, we analyzed the annual numbers of WVCs per kilometer of expressway in Korea
between 2015 and 2019. Expressways under 100 km in length were excluded owing to suggestions that the WVC rate for short
roadways is often overestimated (Lee et al., 2014). Based on this analysis, Dangjin−Yeongdeok Expressway (DYE), Jungang
Expressway (JAE), and Jungbu Expressway (JBE) were included in the study (Fig. 1, Appendix 1). All expressways selected in this
study have maximum and minimum speed limits of 110 km/h and 50 km/h, respectively.
The DYE (route number 30) has a total length of 278.6 km and connects the eastern and western regions of South Korea, thereby
traversing the central inland regions of Chungcheongnam-do, Chungcheongbuk-do, and Gyeongsangbuk-do. Its rst section opened in
November 2007, with subsequent additional sections being completed in stages thereafter.
Next, the JAE (route number 55) passes through Gyeongsangbuk-do, Chungcheongbuk-do, and Gangwon-do, cutting across the
eastern inland area of the country, north to south. These regions are enclosed by the Taebaek and Sobaek Mountain Ranges. The JAE is
288.9 km long, while the rst section of the expressway opened to trafc in December 1994.
The JBE (route number 35) connects Tongyeong-si, Gyeongsangnam-do, Hanam-si, and Gyeonggi-do. Initially, it was 117.2 km
long between Hanam-si, Gyeonggi-do, Nami-myeon, Cheongwon-gun, and Chungcheongbuk-do. The Tongyeong–Daejeon Expressway
was built and joined to the original JBE; therefore, the current total length of the expressway is now 332.5 km, and it has been renamed
the Tongyeong–Daejeon Jungbu Expressway. Notably, it was rst opened in December 1987.
2.2. Data collection
The analysis was conducted based on the nationwide data for trafc volume and WVCs provided by the Korea Expressway Cor-
poration (KEC) (http://data.ex.co.kr) and spanned six years from January 2015 to December 2020. WVC investigation methods are
investigated in accordance with the ‘Guidelines for Roadkill Investigation and Management’ set by the government as administrative
rules. WVC data are collected by the Safety Patrol of the KEC, which records the number of carcasses discovered along with other
related information during their expressway patrol, which occurs three times a day. Carcasses found are photographed by safety
patrols and sent to the ‘Animal Roadkill Information System’, and the National Institute of Ecology uses the photos to identify the
animal species. The recorded items are as follows: the date; the jurisdictional authority; the route name; the distance from the
expressway starting point; the direction; the species of the animal killed in the WVC; the longitude and latitude of the WVC location.
Following the analysis, the carcass was removed from the site to prevent the formation of duplicate records. WVC investigations were
conducted in the same way as before, even when government movement restrictions due to COVID-19 were in place.
Toll collection system (TCS) data from the KEC was used to measure the trafc volume. TCS data contain detailed information
regarding the vehicles entering and leaving the expressways via the toll booths. Such data facilitates the easy and precise identication
of expressway usage patterns.
2.3. Data analysis
2.3.1. WVC hotspot modelling
Kernel density estimations (KDEs) were used to identify specic WVC hotspots in each expressway (Eq. 1). KDE calculates the
spatial density based on the distribution of individual points within a specic search radius (bandwidth) and generates kernel function
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
4
K. In the estimation process, the larger the bandwidth, the greater the hotspot area (Da Silva et al., 2022). Thus, the spatial relevance
decreases as the distance between the center and an individual point increases.
f(x,y) = 1
nh2=∑
n
i=1
K(di
h)(1)
where f (x, y) is the density estimation at (x, y), n is the number of WVCs, h is the bandwidth, di is the location (x, y), and i is the
distance between occurrence points (Eq.1). A Gaussian kernel function is used to determine K (Eq. 2):
K(u) = 1
2
π
√e−1
2u2(2)
where u is the ratio of the distance at location (x, y) within the bandwidth.
To identify specic WVC hotspots in the expressways, the WVC density was estimated using KDEs based on the WVC data from
2015 to 2019. The WVC data used for the KDEs was based on the longitude and latitude of the WVC occurrence points to obtain
cumulative data for each WVC occurrence point. KDE was then used to estimate the WVC density for the entire route, which was then
categorized into ve sections (I, II, III, IV, and V) with equal ranges for WVC density using the equal interval method: Section I with a
WVC density ranging from 100% to 80%; Section II from 80% to 60%; Section III from 60% to 40%; Section IV from 40% to 20%;
Section V from 20% to 0%. Section I covers the areas with extremely high WVC occurrence densities (top 20%). By contrast, Section V
covers the areas with extremely low or zero WVC occurrence density (bottom 20%) (Park et al., 2021). Therefore, we classied Section
I as the highest WVC occurrence, Section II as high WVC occurrence, Section III as medium, Section IV as low, and Section V as very
low.
We identied the Section I regions with the highest WVC density in DYE, JAE, and JBE. We extracted the roads passing between the
toll booths located within the Section I as hotspot areas. This is to measure the trafc volume of hotspots accurately. Subsequently, the
trafc volume and WVCs within the hotspot zone between 2015 and 2020 were analyzed by period (before, during and after
restrictions).
2.3.2. Statistical Analysis
Owing to the severe risk of COVID-19, governments implemented strict restrictions during the initial stage of the pandemic to
contain the spread of the virus (Rilett et al., 2021). A wide range of studies has been conducted worldwide that have focused on the
early stages of the pandemic. Similarly, in Korea, strict regulations were implemented, particularly between February and March 2020.
Thus, this study focused on the initial period of the pandemic in Korea. We categorized the period from February to March 2020, when
strong measures were implemented to contain COVID-19 infections, as a ‘strict pandemic’. In addition, the period from February to
March between 2015 and 2019 was categorized as pre-strict pandemic, and between April 2020 and December 2020, when strong
measures were lifted, was categorized as the pandemic period.
We conducted an analysis of ANOVA by a period (strict pandemic, pre-pandemic, and pandemic) on WVC data from all routes to
determine whether there were statistically signicant differences in WVC incidence due to the COVID-19 pandemic. Paired t-test were
also used to identify periods with statistically signicant differences in WVC. Additionally, a paired t-test was conducted for each
period of each route using WVC data. This derived the period when differences in WVC occurrence for each route were observed. In
addition, we derived the trafc volume for each route to determine whether there is a statistically signicant difference in the change
in trafc volume due to the COVID-19 pandemic by period (DYE, JAE, JBE). Then, we derived trafc volumes by period (strict
pandemic, pre-pandemic, pandemic) and conducted an ANOVA analysis.
Meanwhile, we analyzed the correlation between WVC and trafc volume during the strict pandemic using Pearson’s correlation
coefcient. To examine the relationship between the changes in trafc volume and WVCs during the COVID-19 pandemic, we obtained
monthly data on the trafc volume and WVCs for the study routes for each period (strict pandemic, pre-strict pandemic, and
pandemic). To analyze the changes in WVCs per trafc volume, the percentage of WVCs per trafc volume was derived using monthly
data on trafc volume and WVCs. Then, a paired t-test was used to examine the difference in the percentage change of the percentage of
WVCs per trafc volume during each period. The correlation between WVC and trafc volume was analyzed for each period using the
Pearson correlation coefcient. All statistical analyses were performed using R (ver. 4.1.2).
3. Results and discussion
3.1. WVC hotspots
The roads passing between the toll booths located within the Section I derived using the KDE method were identied as hotspots.
The WVC hotspots for each study site were identied as follows: DYE: Gongju–Namsejong IC; JAE: Hoengseong–Namwonju IC; JBE:
Gonjiam-Seoicheon IC (Fig. 2).
WVCs per kilometer have been decreasing steadily since 2015 throughout the total length of the DYE and JAE, whereas they have
increased since 2017 for the JBE. This trend was the same in the hotspots for each expressway. Despite this trend, the WVC/km of the
hotspots was observed to be higher than that of the WVC/km for the total length of the expressways. The percentage of WVCs that
occurred at each hotspot relative to the entire expressway has also grown (Table 1).
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
5
Therefore, the WVCs in the hotspots have remained high, so we analyzed the WVC characteristics of each hotspot. We analyzed the
WVCs on the DYE, JAE, and JBE from 2015 to 2019 and our ndings revealed that water deer (Hydropotes inermis) were the most
frequently involved in collisions (Table 2), while April, May, and June were the months that exhibited the highest number of WVCs.
This suggests that WVCs exhibit seasonality, which is associated with the behavioral characteristics of water deer. According to Garrah
et al., (2015), WVCs are strongly seasonal, and this seasonality is extensively species-specic. The Korean water deer have a high
population density and are widely distributed throughout the country. During their breeding season in spring, they tend to expand
their home range (Park and Lee, 2013). Thus, during this time of year, young deer are more likely to be involved in vehicle accidents
compared to other seasons, thereby contributing to the high rate of WVCs in spring.
3.2. The effect of the COVID-19 movement restriction on WVCs
To examine the impact on WVCs by the pandemic, we analyzed the relative change in the monthly number of WVCs during the
pandemic period versus that during the pre-pandemic period, focusing specically on the period from February to March 2020, when
the Korean government implemented movement restrictions. The WVCs recorded on all studied routes were analyzed by dividing them
into pre-COVID (2015–2019), strict pandemic (February–March 2020), and pandemic (April–December 2020). Thereafter, it was
conrmed that during the strict pandemic period, the number of WVCs increased compared to pre-COVID, and during the pandemic,
WVCs decreased compared to pre-COVID (Table 3). An ANOVA test was conducted to evaluate the differences in WVCs by period for all
Fig. 2. Categorization of the Wildlife density into ve sections (I, II, III, IV, V) for the (1) Dangjin−Yeongdeok Expressway (DYE), (2) Jungang
Expressway (JAE), and (3) Jungbu Expressway (JBE), using kernel density estimations (KDEs). Sections with a high WVC density are marked in red
(I), while those with a low WVC density are marked in green (V). (I: 100–80%, II: 80–60%, III: 60–40%, IV: 40–20%, and V: 20–0%).
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
6
study routes. As a result, the difference in the number of WVCs according to the period (pre-COVID, strict pandemic, and pandemic)
was statistically signicant (df=2, F value=4.772, p =0.012). Paired t-test were used to identify periods with statistically signicant
differences in WVC. There was a statistically signicant difference in the WVCs between the strict pandemic (February–March 2020)
and the pandemic period (April–December 2020) (t=–2.988, df=5, p =0.031).
After the strict pandemic, the WVCs continued to decrease on all routes in this study compared to the previous year (Fig. 3). A study
from northwest Spain demonstrated results similar to our study, whereby an increase in WVCs was observed during the early months of
2020 before becoming reduced following the enforced COVID-19 lockdowns. Moreover, this trend persisted following the end of the
COVID-19 lockdown. According to that study, the number of WVCs was 30.22% lower in 2020 than in 2019, with a considerable
decline of 66.77% observed during the strictest lockdown period (García-Martínez-de-Alb´
eniz et al., 2022). This study also revealed a
signicant reduction in WVCs on the DYE, JAE, and JBE during 2020 (Jan–Dec) compared to 2019 (Jan–Dec), with declines of 50.09%,
42.55%, and 57.78%, respectively.
Meanwhile, the WVCs recorded for studied routes were analyzed by dividing them into pre-COVID (2015–2019), strict pandemic
(February–March 2020), and pandemic (April–December 2020). On DYE and JAE, the number of WVCs increased during the strict
pandemic (February to March 2020) compared to the pre-pandemic period. Specically, the DYE exhibited a 66.67% and 25% in-
crease, and the JAE exhibited a 100% and 42.86% increase. However, during this period, the number of WVCs on the JBE signicantly
decreased to 0. The number of WVCs showed a decreasing trend during the pandemic period (April to December 2020), compared to
the pre-COVID period, except for specic months, which were October for the DYE, November for the JAE, and August and December
for the JBE (Fig. 3).
Therefore, the results from the paired t-test to compare WVCs between the pre-strict pandemic and strict pandemic periods for each
study route revealed a statistically signicant difference in WVCs for the DYE (t=2.360, df=10, p=0.040) and JAE (t=2.562, df=10,
p=0.028). Contrastingly, for the JBE, the difference in WVCs during the same period was not statistically signicant (t=1.854, df=10,
p=0.093).
Numerous studies have investigated the potential impact of reduced human activities during various periods of the COVID-19
pandemic on worldwide incidences of WVCs. In this study, WVCs increased during the strict pandemic period on the DYE and JAE.
In Scotland, WVCs increased during the COVID-19 lockdown (Bíl et al., 2021), and in some states in the US, WVCs increased during the
pandemic (Abraham and Mumma, 2021). This suggests that wildlife may have increased their use of roads and areas near roads during
the pandemic in response to reduced trafc (Rutz et al., 2020).
While in this study, WVCs decreased during the strict pandemic period on the JBE. A study conducted in several US states reported a
decline in WVCs of 34% during the pandemic period, while some states showed a signicant relationship between the implementation
of COVID-19 restrictions and the change in WVCs (Shilling et al., 2021). Additionally, there was a 48% decrease in WVCs in Tasmania,
Australia, during the pandemic, with statistically signicant changes observed between WVCs during the pre-pandemic and pandemic
periods. Furthermore, a study conducted across 11 European countries revealed a 40% decrease in WVCs during the pandemic period
in Spain, Israel, Estonia, and the Czech Republic compared to previous periods (Driessen, 2021).
In this study, we found that WVCs differed by the duration of travel restrictions due to COVID-19, and the differences in WVCs by
duration were statistically signicant for each study route. Therefore, we can conrm that travel restrictions due to COVID-19 affect
the occurrence of WVCs.
3.3. Relationship between trafc volume and WVCs
To analyze the change in trafc volume due to the COVID-19 restriction, we compared the change in trafc volume in 2020 to the
average trafc volume from 2015 to 2019. We found that compared to pre-COVID-19, trafc volume on DYE and JAE decreased during
the COVID-19 restriction, while trafc volume on JBE increased. The decrease in trafc volume on JAE has been maintained even after
Table 1
Wildlife–vehicle collision (WVC) analysis results in each study site and hotspot: WVC occurs per 1 km on the entire route, WVC occurs per 1 km in
each hotspot, and percentage of WVCs occurring at hotspots relative to total routes.
Year Entire expressway (WVC/km) Hotspot (WVC/km) WVC rate in hotspots (%)
Dangjin−Yeongdeok Expressway
(DYE)
2015 2.83 4.24 17.96
2016 1.92 4.1 25.63
2017 1.06 2.9 21.38
2018 0.79 3.03 29.82
2019 0.85 2.72 24.9
Jungang Expressway
(JAE)
2015 0.35 4.74 10.86
2016 0.22 2.78 12.33
2017 0.18 1.76 15.38
2018 0.16 2.75 19.3
2019 0.16 2.33 17.54
Jungbu Expressway
(JBE)
2015 0.1 1.95 22.12
2016 0.08 1.67 17.52
2017 0.08 1.56 17.69
2018 0.09 1.37 19.9
2019 0.14 2.19 20.41
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
7
Table 2
The status of wildlife involved in wildlife–vehicle collisions (WVCs) in each hotspot zone.
Species DYE JAE JBE
2015 2016 2017 2018 2019 2015 2016 2017 2018 2019 2015 2016 2017 2018 2019
Water deer
(Hydropotes inermis)
91 88 59 66 57 92 55 41 40 40 33 24 27 30 40
Raccoon dog
(Nyctereutes procyonoides)
3 2 2 1 2 8 5 3 5 5 1 1 1
Wild boar
(Sus scrofa)
1 1 1 2 6 1 1 4
eopard cat
(Prionailurus bengalensis)
1 1
Asian badger
(Meles leucurus)
1 2 2 1 1 1
Korean hare
(Lepus coreanus)
2
Others 1
Total 95 92 65 68 61 101 63 53 45 47 34 26 27 30 45
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
8
the restrictions were lifted (Fig. 4).
Therefore, we analyzed the trafc volume of each route during the COVID-19 restrictions by dividing it into pre-strict pandemic
(February –March 2015–2019), strict pandemic (February-March 2020), and pandemic (April-December 2020) (Fig. 5). The results
show that trafc volume on DYE and JAE decreased signicantly during the strict pandemic compared to pre-strict pandemic trafc
volume. In contrast, after the strict pandemic, trafc volume increased signicantly. For JBE, trafc volume during the strict pandemic
tended to be higher than the pre-strict pandemic, and trafc volume continued to increase during the pandemic.
Table 3
Monthly WVC average across the entire study route. The gray blocks represent the period when strict restrictions began in Korea during the COVID-
19 pandemic (February–March 2020).
2015–2019 2020
month average variance average variance
Jan 1.867 1.310 2.333 0.943
Feb 0.800 0.980 1.333 0.943
Mar 1.067 1.236 1.333 0.943
Apr 4.667 2.936 1.667 1.247
May 17.533 9.142 6.000 0.816
Jun 13.667 5.861 4.000 0.816
Jul 5.533 3.964 2.000 0.816
Aug 2.467 2.306 1.667 1.247
Sep 1.867 1.454 0.667 0.471
Oct 2.000 1.461 1.333 0.471
Nov 2.667 1.776 1.000 1.414
Dec 2.667 2.119 1.333 1.247
Fig. 3. Monthly percentage change in wildlife–vehicle collisions (WVCs) from the pre-pandemic period to the pandemic period on the Dang-
jin−Yeongdeok Expressway (DYE), Jungang Expressway (JAE), and Jungbu Expressway (JBE). The area within the black dotted lines refers to the
period from February to April 2020, when strict COVID-19 restrictions were implemented to prevent the spread of the virus.
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
9
In the case of JAE, trafc volume has declined during and after the strict pandemic. This is due to the continuous dispersion of
vehicles due to the opening of the Jungbu Naeryuk Expressway near JAE. For this reason, it can be seen that the JAE route shows a
decreasing trafc volume trend every year (Appendix 2). On the other hand, it can be seen that the JBE has seen a decrease in the peak
value of trafc volume during the strict pandemic but an increase in the median and minimum values. The JBE is an expressway
connecting the capital region, and the hotspot section is located close to Seoul, which tends to show a signicant increase in trafc
volume every year (Appendix 2). In addition, the JBE is an expressway with high freight trafc volume, with more than 30% of the
trafc volume being trucks (https://stat.molit.go.kr/). While human movement has decreased during the COVID-19 lockdown, freight
transportation has expanded globally. JBE is also an expressway with a high proportion of freight vehicles, which may explain the
increase in trafc volume on JBE during this period.
Therefore, we used ANOVA analysis to determine if there was a statistically signicant difference in the trafc volume of each route
by period. As a result, we found that the difference in trafc volume by period was statistically signicant for DYE, JAE, and JBE routes
(DYE: df=2, F value=9.679, p=0.001**, JAE: df=2, F value=10.696, p=0.001**, JBE: df=2, F value=16.183, p=0.000***). There-
fore, the study found that COVID-19 restrictions impacted trafc volume.
To analyze the impact of trafc volume changes on WVCs due to movement restrictions during the COVID-19 pandemic, this study
analyzed the Pearson correlation between WVCs and trafc volume for all routes during the strict pandemic. The results showed a
strong negative correlation between WVCs and trafc volume during the strict pandemic (Pearson correlation coefcient =-0.873).
To analyze the inuence of trafc volume changes due to the COVID-19 pandemic on WVCs, our study calculated the WVCs by
trafc volume for each period over the entire route. This analysis found that the mean of the pre-strict pandemic was 0.329 (SD =
0.642), the strict pandemic was 0.541 (SD =0.675), and the pandemic was 0.499 (SD =0.784). Thus, the strict pandemic period
exhibited the highest mean as well as a greater range of change in WVC occurrences per trafc (Fig. 6). This is because trafc volume
and WVCs are strongly negatively correlated, with the decrease in trafc volume during the strict pandemic period leading to an
increase in WVCs.
Fig. 4. Monthly percentage changes in trafc volume from the pre-pandemic period to the pandemic period on the Dangjin−Yeongdeok Expressway
(DYE), Jungang Expressway (JAE), and Jungbu Expressway (JBE). The area within the black dotted lines refers to the period from February to April
2020, when strict COVID restrictions were implemented to curb the spread of the virus.
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
10
Fig. 5. Trafc volume from 2015 to 2020 on each route, broken down by pre-strict COVID (February-March 2015–2019), strict pandemic
(February-March 2020), and pandemic (April-December 2020) (Left is Dangjin−Yeongdeok Expressway (DYE), center is Jungang Expressway (JAE)
and right is Jungbu Expressway (JBE).
Fig. 6. Wildlife–vehicle collision (WVC) occurrence data per trafc volume by period for all study routes. (pandemic: April 2020 to December 2020,
Pre-strict pandemic: February to March between 2015 and 2019, strict pandemic: February to March 2020).
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
11
Therefore, a paired t-test was employed to examine the statistical signicance of the mean difference in WVC occurrence per trafc
volume between periods. There was a statistically signicant difference in the incidence of WVCs per trafc volume between the strict
pandemic and the pandemic periods (t=–4.2939, df=5, p =0.008). There is a positive correlation between the strict and pandemic
periods and a negative correlation between the pre-strict pandemic and the strict pandemic (Table 4).
During the COVID-19 pandemic, social distancing measures and associated restrictions considerably decreased human activities
worldwide. Several researchers predicted that the reduced trafc volume would lead to fewer WVCs (Primack et al., 2021; Coelho
et al., 2008; D’Amico et al., 2015; Lin, 2016). A study on trafc volume and WVCs on the Anuran highway in Ontario, Canada, found a
negative relationship between the two variables. Saeki and Macdonald (2004) also discovered collisions in Japan and trafc volume
whereby WVCs occurred more frequently when trafc volume decreased.
During the pandemic period, certain sections of the expressways exhibited a higher incidence of WVCs than in any other period.
This can be attributed to wildlife behavioral characteristics, including breeding, foraging, dispersal patterns, and their preferred routes
for crossing in search of better habitats (Dean et al., 2019; Saint-Andrieux et al., 2020). In this study, the behavioral characteristics of
water deer are among the most critical factors inuencing WVCs on the DYE, JAE, and JBE routes. These collisions mostly occurred in
the spring, while winter and spring are the known breeding seasons for water deer. During this period, the water deer mainly used
wetlands and farmland; however, they can also expand their home range (Park and Lee, 2013). Thus, the lockdown period coincided
with the water deer’s breeding season, thereby increasing the risk of a vehicle collision. Some wildlife restrains from roads when the
trafc volume increases (van Langevelde and Jaarsma, 2005). According to Seiler (2005), as wildlife tends to intrude on less crowded
roadways, WVCs would occur more frequently in sections where the trafc volume is moderate compared to high-trafc sections. For
example, WVCs involving moose (Alces alces) occurred most frequently in areas with moderate trafc volumes. Similarly, Rendall
(Rendall et al., 2021) discovered that the WVC rate was highest on roadways with moderate trafc volume.
WVC increased on sections with reduced trafc during strict COVID, and the reduction in WVC was maintained after strict COVID.
Other studies also demonstrated similar results. In a study of an expressway in northwestern Spain, the WVC occurrence rate decreased
in number after the lockdown (García-Martínez-de-Alb´
eniz et al., 2022). The sharp decrease in trafc volume, similar to that during
the strict pandemic, can temporarily increase the occurrence of WVCs. However, it can also have a lasting impact on reducing such
incidences. A signicant temporary reduction in trafc volume can help to reduce WVCs.
The causes for WVC are extensive and their interplay is complex. Previous studies have suggested that changes in trafc volume
might impact the behavior of certain wildlife species, particularly large mammals, which might complicate the relationship between
trafc volume and WVCs (Seiler and Helldin, 2006; Jaarsma and Willems, 2002; Alexander et al., 2005). The frequency of wildlife
crossings is determined by the animal’s behavioral patterns (K¨
ammerle et al., 2017), and the surrounding environment often inuences
behavioral patterns. The occurrence of WVCs follows a very specic spatiotemporal pattern (Park et al., 2021; García-Martí-
nez-de-Alb´
eniz et al., 2022), so the characteristics of the road and the distribution of trafc must also be considered (García-Martí-
nez-de-Alb´
eniz et al., 2022). Thus, it is essential to understand the ecological characteristics of wildlife to comprehend their
relationship with human activities and, consequently, the causes of WVCs.
4. Conclusion
Newly constructed expressways increase the risk of WVCs, leading to signicant socioeconomic consequences. Thus, the rela-
tionship between trafc volume and WVCs has been widely studied. Human interactions with nature were fundamentally changed by
the COVID-19 lockdowns, introduced globally to mitigate the risks associated with the disease (Bates et al., 2021). During the
pandemic, containment measures restricted human activities, thereby offering opportunities to quantitatively assess the human in-
uence on WVCs.
During the strict COVID period, when movement restrictions were in place due to the COVID-19 outbreak, we found a negative
correlation between trafc and WVC. After strict COVID, WVC maintained a downward trend. Therefore, this study is of great sig-
nicance in quantitatively revealing the relationship between WVC and trafc volume on Korean expressways due to movement
restrictions caused by the COVID-19 outbreak. It provides a basis for establishing alternatives to WVC occurrence on expressways.
It is important to note that this study primarily focused on the impact of trafc volume, among several factors contributing to
wildlife–vehicle collisions, and only utilized data from the initial phase of the pandemic. Therefore, there may be certain limitations to
the generalizability of the study’s ndings. Subsequently, further studies that examine the relationship between trafc restrictions and
WVCs for a longer period of time, from the beginning to the end of the pandemic, are needed to provide ample information on WVC
mitigation (García-Martínez-de-Alb´
eniz et al., 2022). Additional research on various causes of WVCs, such as land cover, landscape
structure, speed, and road characteristics, can also help to expand the understanding of the human inuence on WVCs on expressways.
WVCs are the result of an interaction between humans and wild animals, and thus, it is essential to comprehend the relationship and
Table 4
Correlation of ratio (WVC/trafc volume) analysis results for each period using Pearson correlation (pandemic: April 2020 to December 2020;
pre-strict pandemic: February to March between 2015 and 2019; strict pandemic: February to March 2020). **P<.01.
Pre-strict pandemic Strict pandemic Pandemic
Pre-strict pandemic 1.000 - -
Strict pandemic -0.543 1.000 -
Pandemic 0.246 0.740** 1.000
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
12
relevance of the interactions between humans and wildlife. In this sense, it is noteworthy that this study drew a quantitative conclusion
on the interaction between humans and animals. Moreover, it can serve as a reference to establish feasible strategies for WVC
prevention.
CRediT authorship contribution statement
Hyomin Park: Methodology, Data curation. Sangdon Lee: Supervision, Investigation, Funding acquisition, Conceptualization.
Declaration of Competing Interest
We conrm that there is no conict among authors.
Data availability
Data will be made available on request.
Acknowledgements
The nancial support was from KRF-2021R1A2C1011213, MOE-2020002990006 and 2021003360002.
Appendix 1
Wildlife–vehicle collision (WVC) occurrence on the expressways in Korea between 2015 and 2019: the three routes with the highest
WVC occurrence were selected as study sites.
Route no. Route km WVC
30 Dangjin–Yeongdeok Expressway 278.6 1665
55 Jungang Expressway 288.8 1560
35 Jungbu Expressway 332.5 1364
1 Gyeongbu Expressway 416.05 911
50 Yeongdong Expressway 234.4 739
15 Seohaean Expressway 340.8 713
45 Jungbunaeryuk Expressway 301.7 404
151 Seocheon–Gongju Expressway 61.4 370
25 Honam Expressway 194.2 317
65 Donghae/Ulsan–Pohang Expressway 175.86 244
20 Ikan–Pohang Expressway 130.3 233
251 Honam Branch Expressway 54 226
27 Suncheon Wanju Expressway 117.8 209
12 Muan–Gwangju /Gwangju Daegu Expressway 223.2 188
10 Namhae Expressway 273.1 160
40 Pyeongtaek Jecheon Expressway 109.4 130
37 Second Jungbu Expressway 31.1 119
100 Capital Region First Ring Expressway 91.7 88
102 Namhae 1st Branch Expressway 17.9 57
300 Daejeon Southern Belt Expressway 13.3 39
60 Seoul–Yangyang Expressway 88.8 36
253 Gochang-Damyang Expressway 42.5 29
104 Namhae 2nd Branch Expressway 20.6 19
451 Jungbunaeryuk Expressway 30 18
16 Ulsan Expressway 14.3 15
600 Busan Ring Expressway 48.8 6
110 Second Gyeongin Expressway 26.7 5
551 Jungang Branch Expressway 8.2 2
Appendix 2
Average monthly trafc volume from 2015 to 2020 by study route (vehicles/day). The gray blocks represent the period when strict
restrictions began in Korea during the COVID-19 pandemic (February–March 2020).
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
13
DYE JAE JBE
2015 2016 2017 2018 2019 2020 2015 2016 2017 2018 2019 2020 2015 2016 2017 2018 2019 2020
Jan 131.74 148.84 158.90 125.10 133.55 213.29 1065.68 1082.32 1126.58 1050.97 1047.10 1017.42 4259.29 4426.77 4379.35 5296.06 6049.35 6047.03
Feb 151.36 157.10 166.43 171.07 157.93 140.21 1159.57 1175.59 1139.21 1120.71 1051.07 930.41 4327.36 4521.24 4729.43 5578.50 5923.86 5538.00
Mar 152.06 162.84 158.32 157.94 151.55 144.84 1218.32 1234.65 1267.68 1229.48 1145.81 885.23 4601.87 4986.26 5152.90 5802.84 6159.61 5704.39
Apr 172.20 184.07 181.07 167.13 184.53 163.67 1302.20 1309.60 1387.60 1308.27 1239.47 1048.67 4848.00 5356.53 5374.80 6246.20 6277.07 6330.00
May 167.94 190.32 168.32 170.58 173.81 188.77 1366.00 1424.90 1433.29 1331.29 1318.39 1112.58 5123.03 5215.55 5526.45 6318.84 6528.13 6385.10
Jun 158.27 185.00 178.60 173.67 172.20 187.80 1312.00 1426.60 1435.47 1350.73 1274.73 1277.33 4883.60 5054.33 5484.27 6235.33 6099.27 6533.53
Jul 169.29 168.71 156.65 164.97 126.84 176.19 1362.58 1413.10 1403.35 1374.26 1253.35 1249.55 5008.32 5038.06 5358.32 6006.39 6099.61 6304.13
Aug 178.97 171.74 153.10 155.55 123.42 178.32 1458.84 1569.35 1430.26 1365.81 1300.32 1106.06 5182.32 5383.23 5618.71 6032.71 6356.32 6018.90
Sep 207.60 199.00 184.73 241.47 163.20 193.67 1439.93 1495.80 1460.67 1322.33 1251.07 1123.47 5565.80 5519.73 6318.93 6804.20 6449.87 6818.60
Oct 196.39 178.65 233.94 178.45 182.90 225.03 1460.90 1587.10 1433.03 1366.84 1344.90 1195.35 5547.42 5313.23 6551.42 6726.52 6415.61 6837.42
Nov 174.53 176.40 169.73 164.47 215.93 186.60 1348.07 1450.93 1344.80 1334.20 1295.53 1169.80 5091.40 5915.60 6797.33 6688.67 6438.07 6731.60
Dec 164.26 151.94 135.94 129.48 174.77 148.84 1259.81 1285.03 1204.00 1142.84 1165.61 1011.61 4826.26 4632.52 6160.77 6011.29 6059.29 6281.94
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
14
References
Abraham, J.O., Mumma, M.A., 2021. Elevated wildlife-vehicle collision rates during the COVID-19 pandemic. Sci. Rep. 11 (1), 20391 https://doi.org/10.1038/
s41598-021-99233-9.
Alexander, S.M., Waters, N.M., Paquet, P.C., 2005. Trafc volume and highway permeability for a mammalian community in the Canadian Rocky Mountains. Can.
Geogr. 49 (4), 321–331. https://doi.org/10.1111/j.0008-3658.2005.00099.x.
Bates, A.E., Primack, R.B., Biggar, B.S., et al., 2021. Global COVID-19 lockdown highlights humans as both threats and custodians of the environment. Biol. Conserv.
263, 109175 https://doi.org/10.1016/j.biocon.2021.109175.
Bencin, H.L., Prange, S., Rose, C., Popescu, V.D., 2019. Roadkill and space use data predict vehicle-strike hotspots and mortality rates in a recovering bobcat (Lynx
rufus) population. Sci. Rep. 9 (1), 15391 https://doi.org/10.1038/s41598-019-50931-5.
Bíl, M., Andr´
aˇ
sik, R., Svoboda, T., Sedoník, J., 2016. The KDE+software: a tool for effective identication and ranking of animal–vehicle collision hotspots along
networks. Land. Ecol. 31 (2), 231–237. https://doi.org/10.1007/s10980-015-0265-6.
Bíl, M., Andr´
aˇ
sik, R., Cícha, V., Arnon, A., Kruuse, M., Langbein, J., N´
ahlik, A., Niemi, M., Pokorny, B., Colino-Rabanal, V.J., Rolandsen, C.M., Seiler, A., 2021. COVID-
19 related travel restrictions prevented numerous wildlife deaths on roads: a comparative analysis of results from 11 countries. Biol. Conserv. 256, 109076
https://doi.org/10.1016/j.biocon.2021.109076.
Coelho, I.P., Kindel, A., Coelho, A.V.P., 2008. Roadkills of vertebrate species on two highways through the Atlantic Forest Biosphere Reserve, southern Brazil. Eur. J.
Wildl. Res. 54, 689–699. https://doi.org/10.1007/s10344-008-0197-4.
D’Amico, M., Rom´
an, J., De los Reyes, L., Revilla, E., 2015. Vertebrate road-kill patterns in Mediterranean habitats: who, when and where. Biol. Conserv. 191,
234–242. https://doi.org/10.1016/j.biocon.2015.06.010.
Da Silva, A.C.F.B., De Menezes, J.F.S., Santos, L., G.R.O, 2022. Roadkill risk for capybaras in an urban environment. Landsc. Urban Plan 222, 104398. https://doi.org/
10.1016/j.landurbplan.2022.104398.
Dean, W.R.J., Seymour, C.L., Joseph, G.S., Foord, S.H., 2019. A review of the impacts of roads on wildlife in semi-arid regions. Diversity 11 (5), 81. https://doi.org/
10.3390/d11050081.
Driessen, M.M., 2021. COVID-19 restrictions provide a brief respite from the wildlife roadkill toll. Biol. Conserv. 256, 109012 https://doi.org/10.1016/j.
biocon.2021.109012.
Fahrig, L., Rytwinski, T., 2009. Effects of roads on animal abundance: an empirical review and synthesis. Ecol. Soc. 14, 21.
Forman, R.T., Sperling, D., Bissonette, J.A., Clevenger, A.P., Cutshall, C.D., Dale, V.H., Fahrig, L., France, R., Goldman, C.R., Heanue, K., Jones, J.A., Swanson, F.J.,
Turrentine, T., Winter, T.C., 2003. Road ecology. Science and solutions. Island Press, Washington.
García-Martínez-de-Alb´
eniz, ´
I., Ruiz-de-Villa, J.A., Rodriguez-Hernandez, J., 2022. Impact of COVID-19 lockdown on wildlife-vehicle collisions in NW of Spain.
Sustainability 14 (8), 4849. https://doi.org/10.3390/su14084849.
Garrah, E., Danby, R.K., Eberhardt, E., Cunnington, G.M., Mitchell, S., 2015. Hot spots and hot times: wildlife road mortality in a regional conservation corridor.
Environ. Manag 56 (4), 874–889. https://doi.org/10.1007/s00267-015-0566-1.
Huijser, M.P., Dufeld, J.W., Clevenger, A.P., Ament, R.J., McGowen, P.T., 2009. Cost-benet analyses of mitigation measures aimed at reducing collisions with large
ungulates in the United States and Canada: a decision support tool. Ecol. Soc. 14 (2), 15.
Jaarsma, C.F., Willems, G.P., 2002. A Reducing habitat fragmentation by minor rural roads through trafc calming. Landsc. Urban Plan. 58, 125–135. https://doi.org/
10.1016/S0169-2046(01)00215-8.
K¨
ammerle, J.L., Brieger, F., Kr¨
oschel, M., Hagen, R., Storch, I., Suchant, R., 2017. Temporal patterns in road crossing behaviour in roe deer (Capreolus capreolus) at
sites with wildlife warning reectors. PLoS One 12 (9), e0184761. https://doi.org/10.1371/journal.pone.0184761.
Laube, P., Ratnaweera, N., Wr´
obel, A., Kaelin, I., Stephani, A., Reier-Baechtiger, M., Graf, R.F., Suter, S., 2023. Analysing and predicting wildlife–vehicle collision
hotspots for the Swiss road network. Land. Ecol. 38, 1765–1783. https://doi.org/10.1007/s10980-023-01655-5.
Laurance, W.F., Goosem, M., Laurance, S., 2009. Impacts of roads and linear clearings on tropical forests. Trends Ecol. Evol. 24, 659–669. https://doi.org/10.1016/j.
tree.2009.06.009.
Lee, G., Tak, J., Pak, S., 2014. Spatial and temporal patterns on wildlife road-kills on highway in Korea. J. Vet. Clin. 31, 282–287.
Lester, D., 2015. Effective wildlife roadkill mitigation. J. Trafc Transp. Eng. 3, 42–51. https://doi.org/10.17265/2328-2142/2015.01.005.
Lin, S.C., 2016. Landscape and trafc factors affecting animal road mortality. J. Environ. Eng. Landsc. 24, 10–20. https://doi.org/10.3846/16486897.2015.1098652.
Litvaitis, J., Tash, J., 2008. An approach toward understanding wildlife-vehicle collisions. Environ. Manag. 42, 688–697. https://doi.org/10.1007/s00267-008-9108-
4.
Main, M.B., Allen, G.M., 2002. Landscape and seasonal inuences on roadkill of wildlife in southwest Florida. Fla Sci. 65, 149–158.
Manenti, R., Mori, E., Di Canio, V., Mercurio, S., Picone, M., Caf, M., Brambilla, M., Ficetola, G.F., Rubolini, D., 2020. The good, the bad and the ugly of COVID-19
lockdown effects on wildlife conservation: insights from the rst European locked down country. Biol. Conserv. 249, 108728 https://doi.org/10.1016/j.
biocon.2020.108728.
Pagany, R., 2020. Wildlife-vehicle collisions - inuencing factors, data collection and research methods. Biol. Conserv. 251, 108758 https://doi.org/10.1016/j.
biocon.2020.108758.
Park, H., Lee, S., 2013. Habitat use pattern of Korean waterdeer based on the land coverage map. J. Wetl. Res. 15 (4), 567–572. https://doi.org/10.17663/
JWR.2013.15.4.567.
Park, H., Kim, M., Lee, S., 2021. Spatial characteristics of wildlife-vehicle collisions of water deer in Korea Expressway. Sustainability 13 (24), 13523. https://doi.org/
10.3390/su132413523.
Primack, R.B., Bates, A.E., Duarte, C.M., 2021. The conservation and ecological impacts of the COVID-19 pandemic. Biol. Conserv. 260, 109204 https://doi.org/
10.1016/j.biocon.2021.109204.
Rendall, A.R., Webb, V., Sutherland, D.R., White, J.G., Renwick, L., Cooke, R., 2021. Where wildlife and trafc collide: roadkill rates change through time in a
wildlife-tourism hotspot. Glob. Ecol. Conserv. 27, e01530 https://doi.org/10.1016/j.gecco.2021.e01530.
Rilett, L.R., Tufuor, E., Murphy, S., 2021. Arterial roadway travel time reliability and the COVID-19 pandemic. J. Transp. Eng. A Syst. 147 (7), 04021034 https://doi.
org/10.1061/JTEPBS.0000559.
Rutz, C., Loretto, M.C., Bates, A.E., Davidson, S.C., Duarte, C.M., Jetz, W., Johnson, M., Kato, A., Kays, R., Mueller, T., Primack, R.B., Ropert-Coudert, Y., Tucker, M.A.,
Wikelski, M., Cagnacci, F., 2020. COVID-19 lockdown allows researchers to quantify the effects of human activity on wildlife. Nat. Ecol. Evol. 4, 1156–1159.
https://doi.org/10.1038/s41559-020-1237-z.
Saeki, M., Macdonald, D.W., 2004. The effects of trafc on the raccoon dog (Nyctereutes procyonoides viverrinus) and other mammals in Japan. Biol. Conserv. 118,
559–571. https://doi.org/10.1016/j.biocon.2003.10.004.
Saint-Andrieux, C., Calenge, C., Bonenfant, C., 2020. Comparison of environmental, biological and anthropogenic causes of wildlife–vehicle collisions among three
large herbivore species. Popul. Ecol. 62, 64–79. https://doi.org/10.1002/1438-390X.12029.
Seiler, A., 2005. Predicting locations of moose-vehicle collisions in Sweden. J. Appl. Ecol. 42, 371–382. https://doi.org/10.1111/j.1365-2664.2005.01013.x.
Seiler, A., Helldin, J.O., 2006. Mortality in wildlife due to transportation. The ecology of transportation: Managing mobility for the environment. Springer
Netherlands, Dordrecht.
Seiler, A., Olsson, M., Rosell, C., Van Der Grift, E.A., 2016. SAFEROAD safe roads for wildlife and people: cost-benet analyses for wildlife and trafc safety.
SAFEROAD Tech. Rep. 4, 40.
Shilling, F., Nguyen, T., Saleh, M., Khant Kyaw, M., Tapia, K., Trujillo, G., Bejarano, M., Waetjen, D., Peterson, J., Kalisz, G., Sejour, R., Croston, S., Ham, E., 2021.
A reprieve from US wildlife mortality on roads during the COVID-19 pandemic. Biol. Conserv. 256, 109013 https://doi.org/10.1016/j.biocon.2021.109013.
Taylor, B.D., Goldingay, R.L., 2010. Roads and wildlife: impacts, mitigation and implications for wildlife management in Australia. Wildl. Res. 37 (4), 320–331.
https://doi.org/10.1071/WR09171.
H. Park and S. Lee
Global Ecology and Conservation 51 (2024) e02949
15
Van der Ree, R., Smith, D.J., Grilo, C., 2015. The ecological effects of linear infrastructure and trafc: challenges and opportunities of rapid global growth. Handb.
Road. Ecol. 1–9.
van Langevelde, F., Jaarsma, C.F., 2005. Using trafc ow theory to model trafc mortality in mammals. Land. Ecol. 19, 895–907. https://doi.org/10.1007/s10980-
005-0464-7.
Visintin, C., Van Der Ree, R., McCarthy, M.A., 2017. Consistent patterns of vehicle collision risk for six mammal species. J. Environ. Manag. 201, 397–406. https://doi.
org/10.1016/j.jenvman.2017.05.071.
H. Park and S. Lee