Next Article in Journal
Development of a Proposal for a Program to Promote Positive Mental Health Literacy among Adolescents: A Focus Group Study
Previous Article in Journal
Quality of Life in Patients with Pancreatic Cancer—A Literature Review
Previous Article in Special Issue
Dual Carbon Goal-Based Quadrilateral Evolutionary Game: Study on the New Energy Vehicle Industry in China
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
IJERPH
Volume 20
Issue 6
10.3390/ijerph20064896
ijerph-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Identification of Emerging Roadkill Hotspots on Korean Expressways Using Space–Time Cubes

by
Minkyung Kim
and
Sangdon Lee
*
Department of Environmental Science and Engineering, Ewha Womans University, Seoul 03760, Republic of Korea
*
Author to whom correspondence should be addressed.
Int. J. Environ. Res. Public Health 2023, 20(6), 4896; https://doi.org/10.3390/ijerph20064896
Submission received: 15 January 2023 / Revised: 25 February 2023 / Accepted: 7 March 2023 / Published: 10 March 2023
(This article belongs to the Special Issue Sustainable Prosperity without Growth, or Damage to Nature)
Browse Figures
Review Reports Versions Notes

Abstract

:
Collisions with wild animals on high-speed expressways not only lead to roadkill but can also cause accidents that incur considerable human and economic costs. Based on roadkill data from 2004–2019 for four common wildlife species involved in collisions with vehicles on expressways in Korea (water deer, common raccoon dog, Korean hare, and wild boar), the present study conducted optimized hotspot analysis and identified spatiotemporal patterns using a space–time cube (STC) approach. Temporal and spatial differences in the roadkill data were observed between species. Water deer were the most common roadkill species of the four studied, with hotspots in the southern region of the capital area, in the Chungnam region, and in the western Chungbuk and Gangwon-do regions. However, the instances of water deer roadkill over time differed between each region. In addition, it was found that the number of cases of wild boar roadkill has increased recently. In particular, a number of new hotspot areas were observed centered on the metropolitan area Gyeonggi-do, which contains a high population and significant infrastructure. Overall, the emerging hotspot analysis based on STCs was able to determine cold spot and hotspot trends over time, allowing for a more intuitive understanding of spatiotemporal clustering patterns and associated changes than cumulative density-oriented hotspot analysis. As a result, it becomes easier to analyze the causes of roadkill and to establish reduction measures according to priority.

1. Introduction

The length of the road network in Korea has continuously increased in recent decades [1]; as of 2019, the total length of roads in Korea totaled 111,314 km, consisting of 4767 km of expressways (4.3%), 14,030 km of general national highways (12.6%), 4945 km of special city and metropolitan city roads (4.4%), 18,047 km of local roads (16.2%), 30,307 km of city roads (27.2%), 22,776 km of county roads (20.5%), and 16,442 km of district roads (14.8%), [2]. Expressways were first introduced to Korea with the completion of the Gyeongin Expressway in 1968 [3], providing high-speed transportation for vehicles and acting as the backbone of the transportation network connecting major cities. As of 2019, expressways accounted for 71.8% of the daily traffic volume [4]. In addition, from 2000, the total length of roads increased by 1.3% per year, while the expressways increased rapidly by 6.5% annually.
Road expansion is the major cause of land fragmentation [5], in which natural habitats are divided into smaller areas, where the total habitat area decreases. This can lead to a gradual decrease in wildlife populations due to inbreeding and a lack of natural resources [6], leading to a loss of species and deterioration in ecosystem functions [7]. In addition, wildlife–vehicle collisions (WVCs), several of which lead to roadkill, commonly occur when a wild animal attempts to cross a road within its natural range. Medium-sized and large mammals such as water deer (Hydropotes inermis) and wild boar (Sus scrofa), which require a relatively larger area compared to small mammals, are in danger of WVCs due to a lack of habitat [8].
WVCs not only pose a direct threat to wildlife species but can also lead to human injury and death and material damage; thus, their management is very important. Accordingly, the Korea Expressway Corporation (KEC), which manages Korea’s expressways, has recorded all cases of roadkill on expressways from 2004 to the present, including the species involved, spatial data such as the exact point and route, and temporal data such as the date and time. In addition, to establish comprehensive measures to prevent WVCs, the KEC conducted a detailed ecological survey in 2008 and research on facility evaluation and the development of alternatives [9,10]. As a result, the annual number of roadkill cases found on expressways in Korea has decreased, but, nevertheless, thousands of WVCs still occur annually, and there are differences in the roadkill incidence rates and hotspot areas between species [1].
In Korea, the first study on WVCs analyzed the characteristics of roadkill cases on a central highway [11], and various influencing factors such as species, year, season, and land use have subsequently been analyzed [12,13]. In addition, a roadkill hotspot study was conducted using spatial analysis [14], while roadkill clusters have also been investigated [15]. Recently, Kim et al. [1] analyzed the temporal and spatial trends of roadkill on highways across Korea over the past 15 years to determine priority roadkill mitigation measures. However, all of these studies targeted temporal or spatial data separately, and thus, could not accurately reflect spatiotemporal changes in roadkill patterns. Because WVCs have both spatial and temporal characteristics in terms of the location and the date of the accident, respectively, spatiotemporal analysis must be conducted [16].
The identification of WVC hotspots is crucial for traffic safety management, and an understanding of the trends in WVCs can help inform policy decisions designed to prevent these collisions and mitigate associated damage [17]. In this respect, the space–time cube (STC) model proposed by Hagerstrand [18] is a useful method of analysis that can consider the temporal dimension on a spatially two-dimensional plane [19] and has been -widely used in traffic safety, public health, and other fields [20,21,22]. Because spatiotemporal data can be aggregated while maintaining continuity using this method [23,24], it is useful for the spatiotemporal analysis of WVCs, making it possible to intuitively identify areas of low risk and provide information that can be used as the basis for establishing prevention and control strategies [24]. Nevertheless, few studies using hotspot analysis based on the STCs have been carried out to analyze the WVCs occurrence characteristics.
The present study identifies spatiotemporal roadkill patterns through STC-based analysis for four main wildlife species from 2004 to 2019 on expressways in Korea. In particular, it determines cold spots and hotspots for each species, with the aim to provide information that can be used to devise measures to reduce WVCs.

2. Materials and Methods

This study analyzed roadkill patterns on expressways managed by the KEC in Korea (Figure 1). The KEC employees in charge of roadkills record data, detailed information including dates, times, longitude and latitude, the number and type of carcasses discovered during daily expressway patrols in three shifts on a daily basis regularly. Thereafter, they remove the animal carcasses from the road to ensure that the same carcass is not recorded twice [25].
Overall, 36,863 cases of roadkill were recorded by the KEC from 2004 to 2019. The species most frequently involved was water deer with 28,045 cases, followed by the common raccoon dog (Nyctereutes procyonoides) with 5246 cases, the Korean hare (Lepus coreanus) with 1511 cases, and wild boar with 749 cases, though wild boar were only recorded from 2011 [1]. Overall, approximately 96.44% of the 36,863 animals killed on expressways in Korea were from these four species; thus, they were targeted for subsequent hotspot analysis. We analyzed both the analysis of the optimized hotspot and the emerging hotspot of the space-time cube in ArcGIS Pro.
First, optimized hotspot analysis was conducted to determine the areas with a high roadkill density. In this way, it was possible to analyze hotspots and cold spots based on the number of roadkill occurrences and to identify regions with statistically significant patterns. For optimized hotspot analysis, the individual roadkill cases within a specific management boundary or grid cell were combined and analyzed. For this hotspot analysis, the grid cell size was set at 4 × 4 km, and the Getis–Ord Gi* statistic was employed [26] for 90%, 95%, and 99% confidence intervals (Equation (1)).
G i = ∑ j = 1 n w i , j x j − X ¯ ∑ j = 1 n w i , j S n ∑ j = 1 n w i , j 2 − ∑ j = 1 n w i , j 2 n − 1
where xj is the attribute value for feature j, wi,j is the weighted value of the space between feature i and j, and n is the total number of features.
The Gi* statistic is expressed as a z-score indicating statistically significant hotspots and cold spots, with values of >2.58, 1.96–2.58, and 1.65–1.96 representing hotspots or cold spots with confidence levels of 99%, 95%, and 90%, respectively.
Furthermore, emerging hotspot analysis was conducted to examine the temporal changes in roadkill occurrences over the 16-year study period. For this, it was first necessary to create a three-dimensional STC at a specific location based on the temporal trend in the roadkill data (Figure 2). The time series of the z-scores for that location was analyzed using the Mann–Kendall statistic. This analysis returns the clustering trend z-score, p-value, and binning category for each location. The time axis for the STCs was set to 1 year in order to examine the annual roadkill trends over the 16-year period. However, because the STC approach requires at least 10 individual time periods, the time axis for wild boar employed 6-month time intervals. The identified hotspots and cold spots were classified into eight types each based on their following temporal patterns: new, consecutive, intensifying, persistent, diminishing, sporadic, oscillating, and historical (Table 1).

3. Results

3.1. Optimized Hotspot Analysis by Animal Species

Hotspot analysis was conducted on 36,863 cases of roadkill that occurred over 16 years on Korean expressways, identifying those regions with high or low roadkill numbers. In Figure 3, Figure 4, Figure 5 and Figure 6, the darker the red (hotspot) or blue (cold spot) region, the less likely that the clustering was the result of chance. White represents statistically insignificant areas (i.e., no clustering patterns).
Figure 3a presents the optimized hotspot results for water deer roadkill, which accounted for 76.08% of the 36,863 roadkill cases. Significant hotspots were observed in the southern part of the capital area, around the Chungnam region, in the western part of Chungbuk, and in the western region of Gangwon province. In contrast, significant cold spots were found in the western part of the capital area, Jeonnam province, South Gyeongbuk province, and the Gyeongnam area. Raccoon dog showed the optimized hotspot in Gangwon, Jeonnam, and Jeonbuk areas (Figure 4a), and the Korean hare showed the optimized hotspot only in the Gangwon area (Figure 5a). In addition, since 2011, 778 wild boar roadkill have accounted for 2.03% of the total roadkill numbers, with hotspots in the eastern region of Chungbuk, and parts of Chungnam, Jeonbuk, and Gyeongbuk (Figure 6a).

3.2. Spatiotemporal Analysis of Roadkill by Species

As a result of the spatiotemporal analysis of roadkill accidents by species on expressways from 2004 to 2019, a total of 185 hotspot regions were identified for water deer, consisting of 49 consecutive, 41 persistent, 1 Diminishing, 9 sporadic, 78 oscillating, and 7 historical hotspots. There was a total of 500 cold spots for water deer, consisting of 4 new, 36 consecutive, 24 intensifying, 20 persistent, 20 diminishing, 339 sporadic, and 57 oscillating cold spots (Table 2 and Figure 3b).
Looking at Gangwon province and the capital area, cumulative road-kill hotspots for 16 years are widely distributed, as shown on the left of Figure 7. However, the result of classifying into subdivided types according to the time trend is shown on the right of Figure 7, and the types can be divided into persistent or intensifying or consecutive hotspot sections and so on. Persistent hotspots (those that are present without an increase or decrease over at least 90% of the individual time periods) were most common on the Yeongdong Expressway, which connects the southeast of the capital area to the western area of Gangwon. This expressway was also associated with consecutive hotspots, which are defined as areas with continuously high numbers of WVCs in recent years. Therefore, it is necessary to prioritize reduction measures for such persistent hotspots or intensifying hotspots, which are hotspot areas that are gradually strengthened according to the trend of time, rather than other areas. In addition, a part of the capital area and the entire Chungnam area contained oscillating hotspots. Compared with the optimized hotspots (Figure 3a), the spatiotemporal analysis revealed that the roadkill frequency varied over time, even for the hotspots in the central region of Korea. Based on these results, it will be possible to establish priorities and response strategies to reduce the incidence of WVCs.
For raccoon dogs, hotspot analysis revealed no significant hotspot areas (Figure 4b and Table 2), while there were 625 cells representing cold spots, made up of 10 new, 44 consecutive, 65 sporadic, and 506 oscillating cold spots. Some parts of the capital area and some regions of Gyeongbuk and Gyeongnam were associated with consecutive cold spots. However, most recent cold spots were of an oscillating cold spot type, i.e., were a hotspot in the past. For example, Jeonbuk and Jeonnam province were classified as hotspots in the optimized hotspot analysis but were classified as oscillating cold spots based on recent occurrences (Figure 4).
There were 1513 cases of Korean hare roadkill, but as shown in Figure 5b, neither hotspots nor cold spots were identified using spatiotemporal analysis (Table 2), meaning that there were no statistically significant clusters in the study range during the study period. However, with the optimized hotspot analysis, hotspots were identified in the vicinity of Gangwon province.
Two types of statistically significant hotspot were found for wild boar using the 6-month intervals (Figure 6), with 49 new hotspots and 25 sporadic hotspots identified (Table 2). Significant new hotspots appeared within the capital area, some areas in Chungnam, and some areas in Gyeongbuk, where sporadic hotspots were also found. Thus, this spatiotemporal analysis can be used to identify priority areas for reducing the occurrence of wild boar roadkill, even in regions that were not classified as hotspots using optimized hotspot analysis.

4. Discussion

In this study, based on 16 years of roadkill data on high-speed expressways in Korea, we analyzed the spatiotemporal patterns of roadkill accidents using STCs and optimized hotspot/cold spot analysis for four major species. There were differences in the temporal and spatial occurrence of roadkill between these species. Water deer, the species with the highest number of roadkill cases, had hotspots in the southern part of the capital area, around the Chungnam region, in the western part of Chungbuk, and in the western part of Gangwon province. However, the pattern of occurrence over time differed for each region. In addition, all four species had hotspots in areas near the Jungang Expressway and Yeongdong Expressway in Gangwon-do. For raccoon dogs, hotspots were found in the Jeonbuk region, northern Jeonnam, and the western part of Gangwon-do; however, spatiotemporal analysis revealed that most of these were gradually decreasing or oscillating between hotspots and cold spots, while the Korean hare did not have any statistically significant hotspots in the spatiotemporal analysis. In contrast, there was an increase in the number of wild boar roadkill cases, with new hotspots found in several areas (the capital area and some areas in Chungnam and Gyeongbuk Province). In particular, a number of new hotspot areas were centered in the capital area (Gyeonggi-do), where the population and infrastructure are concentrated. The present study found that there were differences in the location of roadkill hotspots between animal species on expressways in Korea over the 16-year study period, which is likely to be related to the density of their populations and their habitat range [1]. Our results, which classify types in detail, help to identify the causes and identify priority areas for mitigation measures. In addition, hotspots were found in the vicinity of Jungang Expressway and Yeongdong Expressway, which are in the western part of Gangwon-do. Therefore, our results suggest that this area should be managed as a roadkill hotspot area regardless of species.
Given the severity of accidents involving WVCs, decision-makers and safety managers need to better understand the characteristics of WVCs and identify high-risk locations to mitigate the effects of these crashes, as well as appropriate responses to action should be prioritized. However, understanding the occurrence of roadkill is very difficult because it usually changes in both spatial and temporal dimensions. Moreover, many studies separately analyzed and discussed the spatial distribution and the temporal trend so far [1,27]. Previous studies have analyzed roadkill distributions to determine spatial hotspots, but these studies have not considered changes over time [28]. Therefore, in this study, we employed STCs that aggregate spatiotemporal information for the same data set to simultaneously and comprehensively analyze spatiotemporal information. This represents a new method for spatiotemporal analysis, meets the requirements of retention time and spatial continuity, and has the advantage of visualizing the analysis results [24,29]. In recent years, the emerging hot spot analysis based on STCs has gradually been applied to different scientific fields [30,31,32]. The results of the emerging hotspot analysis based on STCs offer important information on the locations and times of WVCs, allowing decision-makers to observe spatiotemporal clustering patterns. Thus, it is easier to grasp the causes of WVCs and establish measures according to priority to reduce them.
Recently, studies on the reliability and accuracy of roadkill data collected to implement appropriate roadkill mitigation measures are being conducted [33,34,35]. The potential biases of roadkill data could be due to (1) carcass removal, for example, by scavenging animals or people, (2) detectability, as small road-killed animals are less likely to be observed than are large animals and (3) injury; that is, animals are not immediately killed by the collision but are injured and die some distance, out of sight of the road [36,37]. This could lead to an underreporting of the actual number of deaths. In Korea, efforts are being made to reduce such under-reporting and redundant reporting through a system that collects data on a regular basis (3 shifts, 24 h). In addition, in this paper, in order to supplement this point, the analysis was conducted focusing on large animals. However, since the reduction of the size of the data is a very important issue in roadkill analysis, it is judged that additional research to compensate for this needs to be supplemented.

5. Conclusions

Korea’s highway management institutions, the KEC, and the Ministry of Environment, aim to reduce roadkill using reduction measures such as ecological passages, warning signs, and guided fences. It would be best to establish roadkill reduction measures on all roads, but if that is not possible, it would be desirable to apply measures based on the priority of the areas in need (e.g., persistent hotspots or intensifying hotspots, etc.). This study thus conducted nationwide expressway hotspot analysis by animal species over 16 years; in particular, it introduced spatiotemporal analysis for the first time in Korea. This study characterized the distribution of eight types of hotspots and eight types of cold spots based on their trends over time and identified high- and low-risk areas for roadkill using spatiotemporal trend analysis. Using this information, it is possible to establish specific and priority measures for these hotspot/cold spot types in order to reduce WVCs. In particular, it is suggested that strategies are needed to prevent collisions with water deer in persistent and consecutive hotspot areas, while accidents involving wild boar need to be addressed in new hotspots, mainly in the capital area and parts of Chungnam province and Gyeongbuk province. Therefore, the results of this study are expected to contribute greatly to the establishment of roadkill reduction measures at the national level for individual animal species.

Author Contributions

Methodology, M.K.; Writing—original draft, M.K.; Project administration, S.L.; Supervision, S.L. All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by the Korea Environment Industry and Technology Institute (KEITI), funded by the Korea Ministry of Environment (MOE) (2020002990006; 2022003640003). This research was supported by the National Research Foundation funded by the Korean government (NRF-2021R1A2C1011213; 2020R1I1A1A01058327). The authors also wish to thank the Korea Expressway Corporation for providing data.

Institutional Review Board Statement

Not applicable.

Informed Consent Statement

Not applicable.

Data Availability Statement

Not applicable.

Conflicts of Interest

The authors declare no conflict of interest.

References

  1. Kim, M.; Park, H.; Lee, S. Analysis of roadkill on the Korean expressways from 2004 to 2019. Int. J. Environ. Res. Public Health 2021, 18, 10252. [Google Scholar] [CrossRef]
  2. Ministry of Land, Infrastructure and Transport. Road Status Report. 2020. Available online: http://www.index.go.kr/potal/main/EachDtlPageDetail.do?idx_cd=1206 (accessed on 11 August 2020).
  3. Lee, G.; Tak, J.H.; Pak, S.I. Spatial and temporal patterns on wildlife road-kills on highway in Korea. J. Vet. Clin. 2014, 31, 282–287. (In Korean) [Google Scholar] [CrossRef]
  4. Ministry of Land, Infrastructure and Transport. Road Traffic Volume Statistical Yearbook. 2020. Available online: http://www.index.go.kr/potal/main/EachDtlPageDetail.do?idx_cd=1212#quick_02 (accessed on 11 August 2020).
  5. Forman, R.T.T.; Alexander, L.E. Roads and their major ecological effects. Annu. Rev. Ecol. Syst. 1998, 29, 207–231. [Google Scholar] [CrossRef] [Green Version]
  6. Sher, A.A.; Primack, R.B. An Introduction to Conservation Biology, 2nd ed.; Oxford University Press: Oxford, UK, 2020. [Google Scholar]
  7. Trombulak, S.C.; Frissell, C.A. Review of ecological effects of roads on terrestrial and aquatic communities. Conserv. Biol. 2000, 14, 18–30. [Google Scholar] [CrossRef] [Green Version]
  8. Caceres, N.C. Biological characteristics influence mammal road kill in an Atlantic Forest–Cerrado interface in south-western Brazil. Ital. J. Zool. 2011, 78, 379–389. [Google Scholar] [CrossRef]
  9. Korea Expressway Corporation. 2008 Highways Precision Ecological Survey and Comprehensive Measures for the Prevention of Traffic Accidents in Korea; Korea Expressway Corporation: Gimcheon-si, Republic of Korea, 2008. (In Korean) [Google Scholar]
  10. Korea Expressway Corporation. Research on the Effects of Existing Prevention Facilities for Roadkill and the Alternative; Korea Expressway Corporation: Gimcheon-si, Republic of Korea, 2014. (In Korean) [Google Scholar]
  11. Lee, S.D.; Cho, H.S.; Kim, J.G. A study of wildlife roadkill in Joongang highway. J. Environ. Impact Assess. 2004, 13, 21–31. (In Korean) [Google Scholar]
  12. Choi, T.Y.; Park, C.H. The effects of land use on the frequency of mammal roadkills in Korea. J. Korean Inst. Landsc. Archit. 2006, 34, 52–58. (In Korean) [Google Scholar]
  13. Min, J.H.; Han, G.S. A study on the characteristics of road-kills in the Odaesan national park. Korean J. Environ. Ecol. 2010, 24, 46–53. (In Korean) [Google Scholar]
  14. Seok, S.; Lee, J. A study on the correlation between road-kill hotspot and habitat patches. J. Environ. Impact Assess. 2015, 24, 23–243. (In Korean) [Google Scholar] [CrossRef]
  15. Song, E.; Seo, H.; Kim, K.; Woo, D.; Park, T.; Choi, T. Analysis of Roadkill Hotspot According to the Spatial Clustering Methods. J.Environ. Impact Assess. 2019, 28, 580–591. (In Korean) [Google Scholar]
  16. Son, S.; Kang, Y. Spatio-Temporal Pattern of Traffic Accident of Female Drivers in Seoul; The Korean Cartographic Association: Seoul, Republic of Korea, 2017; pp. 89–98. (In Korean) [Google Scholar]
  17. Cheng, Z.; Zu, Z.; Lu, J. Traffic crash evolution characteristic analysis and spatiotemporal hotspot identification of urban road intersections. Sustainability 2019, 11, 160. [Google Scholar] [CrossRef] [Green Version]
  18. Hägerstrand, T. What about people in regional science? Reg. Sci Assoc. 1970, 24, 6–21. [Google Scholar] [CrossRef]
  19. Ha, J.; Kim, S.; Lee, S. Analysis of spatio-temporal characteristics of small business sales by the spread of COVID-19 in Seoul, Korea: Using Space-Time Cube Model. J. Korea Plan. Assoc. 2021, 56, 218–234. (In Korean) [Google Scholar] [CrossRef]
  20. Huang, J.X.; Wang, J.F.; Li, Z.J.; Wang, Y.; Lai, S.J.; Yang, W.Z. Visualized exploratory spatiotemporal analysis of hand-foot-mouth disease in Southern China. PLoS ONE 2015, 10, e0143411. [Google Scholar] [CrossRef] [PubMed]
  21. Kang, Y.; Cho, N.; Son, S. Spatiotemporal characteristics of elderly population’s traffic accidents in Seoul using space-time cube and space-time kernel density estimation. PLoS ONE 2018, 13, e0196845. [Google Scholar] [CrossRef] [Green Version]
  22. De Cos Guerra, O.; Castillo Salcines, V.; Cantarero Prieto, D. Are spatial patterns of COVID-19 changing? Spatiotemporal analysis over four waves in the region of Cantabria, Spain. Trans. GIS 2022, 26, 1981–2003. [Google Scholar] [CrossRef]
  23. Gatalsky, P.; Andrienko, N.; Andrienko, G. Interactive analysis of event data using space-time cube. In Proceedings of the Eighth International Conference on Information Visualisation, 2004. IV 2004, London, UK, 16 July 2004; pp. 145–152. [Google Scholar] [CrossRef] [Green Version]
  24. Mo, C.; Tan, D.; Mai, T.; Bei, C.; Qin, J.; Pang, W.; Zhang, Z. An analysis of spatiotemporal pattern for COIVD-19 in China based on space-time cube. J. Med. Virol. 2020, 92, 1587–1595. [Google Scholar] [CrossRef] [PubMed] [Green Version]
  25. Park, H.; Kim, M.; Lee, S. Spatial Characteristics of Wildlife-Vehicle Collisions of Water Deer in Korea Expressway. Sustainability 2021, 13, 13523. [Google Scholar] [CrossRef]
  26. Getis, A.; Ord, J.K. The analysis of spatial association by use of distance statistics. Geogr. Anal. 1992, 24, 189–206. [Google Scholar] [CrossRef]
  27. Canova, L.; Balestrieri, A. Long-term monitoring by roadkill counts of mammal populations living in intensively cultivated landscapes. Biodivers. Conserv. 2019, 28, 97–113. [Google Scholar] [CrossRef]
  28. Silveira Miranda, J.E.; de Melo, F.R.; Keichi Umetsu, R. Are roadkill hotspots in the Cerrado equal among groups of vertebrates? Environ. Manag. 2020, 65, 565–573. [Google Scholar] [CrossRef]
  29. Bach, B.; Dragicevic, P.; Archambault, D.; Hurter, C.; Carpendale, S. A Review of Temporal Data Visualizations Based on Space-Time Cube Operations. In Proceedings of the Eurographics Conference on Visualization, Swansea, UK, 9–13 June 2014. [Google Scholar]
  30. Betty, E.L.; Bollard, B.; Murphy, S.; Ogle, M.; Hendriks, H.; Orams, M.B.; Stockin, K.A. Using emerging hot spot analysis of stranding records to inform conservation management of a data-poor cetacean species. Biodivers. Conserv. 2020, 29, 643–665. [Google Scholar] [CrossRef]
  31. Chambers, S.N. The spatiotemporal forming of a state of exception: Repurposing hot-spot analysis to map bare-life in Southern Arizona’s borderlands. GeoJournal 2020, 85, 1373–1384. [Google Scholar] [CrossRef]
  32. Xu, B.; Qi, B.; Ji, K.; Liu, Z.; Deng, L.; Jiang, L. Emerging hot spot analysis and the spatial–temporal trends of NDVI in the Jing River Basin of China. Environ. Earth Sci. 2022, 81, 55. [Google Scholar] [CrossRef]
  33. Yang, X.; Zou, Y.; Wu, L.; Zhong, X.; Wang, Y.; Ijaz, M.; Peng, Y. Comparative Analysis of the Reported Animal-Vehicle Collisions Data and Carcass Removal Data for Hotspot Identification. J. Adv. Transp. 2019, 2019, 3521793. [Google Scholar] [CrossRef]
  34. Zou, Y.; Zhong, X.; Tang, J.; Ye, X.; Wu, L.; Ijaz, M.; Wang, Y. A Copula-Based Approach for Accommodating the Underreporting Effect in Wildlife—Vehicle Crash Analysis. Sustainability 2019, 11, 418. [Google Scholar] [CrossRef] [Green Version]
  35. Ahmed, I.U.; Ahmed, M.M. Investigating the Safety Effectiveness of Wildlife–Vehicle Crash Countermeasures using a Bayesian Approach with a Comparison between Carcass Removal Data and Traditional Crash Data. Transp. Res. Rec. 2022, 2676, 475–489. [Google Scholar] [CrossRef]
  36. Winton, S.A.; Taylor, R.; Bishop, C.A.; Larsen, K.W. Estimating actual versus detected road mortality rates for a northern viper. Glob. Ecol. Conserv. 2018, 16, e00476. [Google Scholar] [CrossRef]
  37. Lee, T.S.; Rondeau, K.; Schaufele, R.; Clevenger, A.P.; Duke, D. Developing a correction factor to apply to animal–vehicle collision data for improved road mitigation measures. Wildl. Res. 2021, 48, 501–510. [Google Scholar] [CrossRef]
Ijerph 20 04896 g001 550
Figure 1. Map of the expressways and the eight administrative provinces in Korea. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Figure 1. Map of the expressways and the eight administrative provinces in Korea. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Ijerph 20 04896 g001
Ijerph 20 04896 g002 550
Figure 2. Structure of space–time cube (http://pro.arcgis.com). (a) Space–time bins in 3D; (b) Generated bins 2D for emerging hotspot analysis.
Figure 2. Structure of space–time cube (http://pro.arcgis.com). (a) Space–time bins in 3D; (b) Generated bins 2D for emerging hotspot analysis.
Ijerph 20 04896 g002
Ijerph 20 04896 g003 550
Figure 3. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for water deer (Hydropotes inermis) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Figure 3. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for water deer (Hydropotes inermis) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Ijerph 20 04896 g003
Ijerph 20 04896 g004 550
Figure 4. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for raccoon dog (Nyctereutes procyonoides) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Figure 4. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for raccoon dog (Nyctereutes procyonoides) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Ijerph 20 04896 g004
Ijerph 20 04896 g005 550
Figure 5. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for Korean hare (Lepus coreanus) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Figure 5. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for Korean hare (Lepus coreanus) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Ijerph 20 04896 g005
Ijerph 20 04896 g006 550
Figure 6. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for wild boar (Sus scrofa) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Figure 6. Optimized hotspot analysis (a) and emerging hotspot analysis (b) for wild boar (Sus scrofa) roadkill on Korean expressways between 2004 and 2019. CP: Capital area, GW: Gangwon province, CN: Chungnam province, CB: Chungbuk province, JB: Jeonbuk province, GB: Gyeongbuk province, JN: Jeonnam province, GN: Gyeongnam province, JJ: Jeju-do.
Ijerph 20 04896 g006
Ijerph 20 04896 g007 550
Figure 7. Optimized hotspot and Emerging hotspot of water deer roadkill in Gangwon Province and Capital area. CP: Capital area, GW: Gangwon province.
Figure 7. Optimized hotspot and Emerging hotspot of water deer roadkill in Gangwon Province and Capital area. CP: Capital area, GW: Gangwon province.
Ijerph 20 04896 g007
Table
Table 1. Definition of the types of hotspot and cold spot identified in the spatiotemporal analysis (http://pro.arcgis.com).
Table 1. Definition of the types of hotspot and cold spot identified in the spatiotemporal analysis (http://pro.arcgis.com).
Ijerph 20 04896 i001No Pattern DetectedDoes Not Fall into Any of the Patterns Described Below.
HotspotsIjerph 20 04896 i002New HotspotA location that is a statistically significant hotspot in the final time period and has not been a statistically significant hotspot before.
Ijerph 20 04896 i003Consecutive HotspotA location with a single uninterrupted run of statistically significant hotspots in the final time periods. The location has not been a statistically significant hotspot prior to the final hotspot run and fewer than 90% of all time periods are statistically significant hotspots.
Ijerph 20 04896 i004Intensifying HotspotA location that has been a statistically significant hotspot for 90% of the time periods, including the final period. In addition, the intensity of the counts for each time period is increasing overall, and this increase is statistically significant.
Ijerph 20 04896 i005Persistent HotspotA location that has been a statistically significant hotspot for 90% of the time periods with no discernible increase or decrease in intensity over time.
Ijerph 20 04896 i006Diminishing HotspotA location that has been a statistically significant hotspot for 90% of the time periods, including the final time period. In addition, the intensity in each time period is decreasing overall, and this decrease is statistically significant.
Ijerph 20 04896 i007Sporadic HotspotA location that is an on-again, off-again hotspot. Less than 90% of the time periods were statistically significant hotspots and none were statistically significant cold spots.
Ijerph 20 04896 i008Oscillating HotspotA statistically significant hotspot for the final period that has a history of also being a statistically significant cold spot during a prior time period. Less than 90% of the time periods are statistically significant hotspots.
Ijerph 20 04896 i009Historical HotspotThe most recent time period is not a hot spot, but at least 90% of the periods are statistically significant hotspots.
Cold spotIjerph 20 04896 i010New Cold SpotA location that is a statistically significant cold spot for the final time period and has not been a statistically significant cold spot before.
Ijerph 20 04896 i011Consecutive Cold SpotA location with a single uninterrupted run of statistically significant cold spots in the final time periods. The location has not been a statistically significant cold spot prior to the final cold spot run and less than 90% of all time periods are statistically significant cold spots.
Ijerph 20 04896 i012Intensifying Cold SpotA location that has been a statistically significant cold spot for 90% of the time periods, including the final period. In addition, the intensity of the cold spots is increasing overall and this increase is statistically significant.
Ijerph 20 04896 i013Persistent Cold SpotA location that has been a statistically significant cold spot for 90% of the time periods with no discernible increase or decrease in the intensity of the count over time.
Ijerph 20 04896 i014Diminishing Cold SpotA location that has been a statistically significant cold spot for 90% of the time periods, including the final period. In addition, the intensity of the cold spots in each period is decreasing overall and this decrease is statistically significant.
Ijerph 20 04896 i015Sporadic Cold SpotA location that is an on-again, off-again cold spot. Less than 90% of the time periods are statistically significant cold spots and none are statistically significant hotspots.
Ijerph 20 04896 i016Oscillating Cold SpotA statistically significant cold spot for the period that has a history of also being a statistically significant hotspot during a prior time period. Less than 90% of the time periods are statistically significant cold spots.
Ijerph 20 04896 i017Historical Cold SpotThe most recent time period is not cold, but at least 90% of the time periods are statistically significant cold spots.
Table
Table 2. Number of cells (4 × 4 km) by type derived by emerging hotspot analysis.
Table 2. Number of cells (4 × 4 km) by type derived by emerging hotspot analysis.
SpeciesWater DeerRaccoon DogKorean HareWild Boar
TypeHotColdHotColdHotColdHotCold
New0401000490
Consecutive49360440000
Intensifying024000000
Persistent4120000000
Diminishing120000000
Sporadic933906500250
Oscillating785705060000
Historical70000000
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kim, M.; Lee, S. Identification of Emerging Roadkill Hotspots on Korean Expressways Using Space–Time Cubes. Int. J. Environ. Res. Public Health 2023, 20, 4896. https://doi.org/10.3390/ijerph20064896

AMA Style

Kim M, Lee S. Identification of Emerging Roadkill Hotspots on Korean Expressways Using Space–Time Cubes. International Journal of Environmental Research and Public Health. 2023; 20(6):4896. https://doi.org/10.3390/ijerph20064896

Chicago/Turabian Style

Kim, Minkyung, and Sangdon Lee. 2023. "Identification of Emerging Roadkill Hotspots on Korean Expressways Using Space–Time Cubes" International Journal of Environmental Research and Public Health 20, no. 6: 4896. https://doi.org/10.3390/ijerph20064896

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop