Fish Diversity and Ichthyofauna of Areas Adjacent to the Demilitarized Zone in South Korea

: The Korean Demilitarized Zone (DMZ) was established in 1953 during the cessation of the Korean War, which divided the Korean Peninsula into North and South Korea. The DMZ is a representative biodiversity resource because it limits human activities. The current status of faunal diversity in adjacent areas of the DMZ was investigated in ﬁve regions at 91 sites from 2015 to 2019. A total of 19,562 individuals were collected and identiﬁed, including 81 species, 19 families, and 11 orders. Zacco koreanus was the most abundant ﬁsh species (relative abundance of 21.9%), while other key species were Zacco platypus (18.2%), Rhynchocypris oxycephalus (5.2%), Rhynchocypris steindachneri (4.9%), and Pungtungia herzi (3.7%). Notably, twenty-nige endemic and seven Red Data Book species were recorded, and three exotic species, Carassius cuvieri , Micropterus salmoides , and Lepomis macrochirus were found at ﬁve sites. Human activities and disturbances were found at 34.0% of all sampling sites, and the highest disturbance factors were river dredging and agricultural use (seven sites), sewage (ﬁve sites), and ﬁsheries and development (two sites). We recommend that the governments of South and North Korea collaboratively manage the DMZ to protect ecological diversity and maintain its status as a symbol of peace.


Introduction
The Korean Demilitarized Zone (DMZ) was established on 27 July 1953 as a result of signing the Armistice Agreement at the end of the Korean War [1]. The DMZ is situated 2 km from the Northern Limit Line (NLL), the Southern Limit Line (SLL), and the Military Demarcation Line (MDL) and covers an area of approximately 99,144 ha. The MDL (248 km) divides the Korean Peninsula in half, from the western end (Paju, Gyeonggi-do Province) to the eastern end (Goseong, Gangwon-do Province). The NLL and SLL have a cease-fire line, and the armies of the North and South are stationed on this line. There is a civilian control zone (CCZ) surrounding the MDL to protect military operations and facilities in the vicinity of the MDL and maintain security, and this area restricts the entry and exit of civilians. The civilian control line designates an additional buffer zone to the DMZ within a distance of 5 to 10 km from the SLL of the DMZ, with an area of approximately 231,405 ha [2].
The DMZ divides the Korean Peninsula into two areas and has had unique ecosystems under strict control over the last 70 years. The DMZ is a representative biodiversity record for the Korean Peninsula, revealing not only the world's rarest and endangered species but also some that have not yet been discovered [2]. In 1974, the Ministry of Culture and Public conducted the first academic survey of the DMZ and CCZ [3]. Subsequently, since the beginning of the 1990s, various reports on the environment and ecology of these zones began to be published at the national level, providing an inventory of the ecological status

Study Regions and Sites
Based on previous research data and available literature [2], this study investigated the current status of fish species inhabiting areas adjacent to the DMZ. Based on the climate characteristics and ecological features of the mountains, rivers, CCZ and DMZ, and their adjacent areas, the study area was divided into five regions with a total of 91 sampling sites: the eastern coastal zone (sites  in 2015, eastern mountainous zone (sites [25][26][27][28][29][30][31][32][33][34] in 2016, western plain zone (sites [35][36][37][38][39][40][41][42][43][44][45][46][47]  Additionally, streams in the eastern coastal zone were divided into Northeast Korean subdistricts (sites 1-15) flowing into the East Sea, and West Korean subdistricts (sites , which flow into the Bukhangang and West Sea centered on Taebaeksanmaek. The elevation at each survey site was recorded using an altimeter (SKT-102, Tlead, Seoul, Korea) for analyzing the relationship between species and elevation.

Ichthyofauna Sampling and Geographic and Physical Surveys
The ichthyofauna sampling in each survey region was conducted one to three times (spring, summer, and autumn) each year from 2015 to 2019 in cooperation with the army: 24 sites in the eastern coastal zone from July to September 2015 (one to two times), 10 sites in the eastern mountainous zone from June to August 2016 (three times), 13 sites in the western plain zone from May to September 2017 (two times), 24 sites in the central mountainous zone from May to October 2018 (two to three times), and 20 sites in the western Imjin River estuary zone between May and September 2019 (two to three times) ( Figure 1; Table S1). The ichthyofauna sampling in each survey region was conducted one to three times 102 (spring, summer, and autumn) each year from 2015 to 2019 in cooperation with the army: 103 24 sites in the eastern coastal zone from July to September 2015 (one to two times), 10 sites 104 in the eastern mountainous zone from June to August 2016 (three times), 13 sites in the 105 western plain zone from May to September 2017 (two times), 24 sites in the central moun-106 tainous zone from May to October 2018 (two to three times), and 20 sites in the western 107 Imjin River estuary zone between May and September 2019 (two to three times) ( Figure 1; 108 Table S1). 109 The geographic and physical data surveys of river width (the distance between the 110 riverbanks), water width (the width of the flowing water), and water depth were meas-111 ured using a distance-measuring telescope (Yardage Pro Tour XL, BUSHNELL, Japan) 112 and a tape measure. In addition, river type (see Table S1 for definitions) and bottom struc-113 tures were assessed according to the methods proposed by Kani [22] and Cummins [23], 114 respectively. 115 Fish were collected using both a cast net (6-mm mesh; 12.5 m 2 = 2 m  2 m  3.14) 116 and a scoop-net (4-mm mesh; 1.35 m 2 = 1.5 m  0.9 m) depending on the water width and 117 depth for the investigation of various habitats and ichthyofauna. When the water width 118 was <10 m, the cast nets were cast five to 10 times and scoop-nets were used for 30 min at 119 water widths between 10 and 30 m; 11 to 20 casts were made and scoop-nets were used 120 for 30 to 60 min, and when the water width was >30 m, more than 20 casts were made, 121 and scoop-nets were used for over 60 min. These methods were in accordance with the 122 Guidelines of the National Natural Environment Survey (IV) [24]. 123 All survey habitat types, including riffles, runs, and pools, were investigated in an 124 upstream direction for a distance of 50 to 100 m. The reservoir and pond sites (39,42,44,125 72, 85, and 90), representative locations were selected as the place where the fish were in 126 various habitats and investigated to ensure that they were as reflective as possible. The 127 cast net survey was conducted on the inlet areas that enter the reservoir and pond, and 128 the kick net survey was performed only at the shallow edge on them. Although we could 129 The geographic and physical data surveys of river width (the distance between the riverbanks), water width (the width of the flowing water), and water depth were measured using a distance-measuring telescope (Yardage Pro Tour XL, BUSHNELL, Tokyo, Japan) and a tape measure. In addition, river type (see Table S1 for definitions) and bottom structures were assessed according to the methods proposed by Kani [22] and Cummins [23], respectively.
Fish were collected using both a cast net (6-mm mesh; 12.5 m 2 = 2 m × 2 m × 3.14) and a scoop-net (4-mm mesh; 1.35 m 2 = 1.5 m × 0.9 m) depending on the water width and depth for the investigation of various habitats and ichthyofauna. When the water width was <10 m, the cast nets were cast five to 10 times and scoop-nets were used for 30 min at water widths between 10 and 30 m; 11 to 20 casts were made and scoop-nets were used for 30 to 60 min, and when the water width was >30 m, more than 20 casts were made, and scoop-nets were used for over 60 min. These methods were in accordance with the Guidelines of the National Natural Environment Survey (IV) [24].
All survey habitat types, including riffles, runs, and pools, were investigated in an upstream direction for a distance of 50 to 100 m. The reservoir and pond sites (39,42,44,72,85, and 90), representative locations were selected as the place where the fish were in various habitats and investigated to ensure that they were as reflective as possible. The cast net survey was conducted on the inlet areas that enter the reservoir and pond, and the kick net survey was performed only at the shallow edge on them. Although we could not always survey two sites (35 and 37) because they were in hazardous areas with mines, we conducted at least one survey with the permission of the military units.
The individual fish captured were identified and counted in situ with the naked eye and then immediately released; however, some ambiguous specimens were preserved in 10% formalin and later identified in the laboratory. All specimens were identified according to the process described by Kim and Park [25] and Kim et al. [26] using the classification system proposed by Nelson [27]. Scientific names and authorities followed those of the National Species List of Korea [28]. Although several of the species and family names have been revised (FishBase, 2022), we used the older system to maintain consistency and comparability with previous studies.

Data Analysis
Diversity [29], richness [30], evenness [31], and dominance [32] were calculated for each of the five regions. The hierarchical cluster analysis of species composition of the five regions was analyzed using the Bray-Curtis similarity index, performed with the PRIMER software (Primer-e 6, New Zealand Ins., Palmerston North, New Zealand). Statistical analyses were conducted using the SYSTAT software (Systat version 24.0, SPSS Inc., Chicago, IL, USA). To evaluate the relationship between geographical and physical factors and community composition of fish, principal component analysis (PCA) was performed using in the PRIMER software.
Pearson's correlation, linear regression, and quadratic analysis were used to test the relationships among the elevation, the total number of species, and the total number of endemic species at each survey site. Since the survey method varied depending on the elevation and river type, we limited the comparative survey sites to those where more than 100 individuals were collected. Statistical significance was considered when p was <0.05.

Human Activities and Disturbance
Human activities and disturbances were surveyed throughout the areas adjacent to the DMZ during the faunal sampling. We recorded human activities and use, both inside and outside the investigation sites, which had direct or indirect impacts on water quality, fish populations, and environment and classified them into four categories: (1) river dredging, (2) agricultural use, (3) sewage, and (4) fisheries and development.

Geographic and Physical Analysis
The river and water width, water depth, river type, and bottom structures in each of the survey sites are shown in Table S1. The study sites were mostly small-to-mid-sized streams with a water width of less than 50 m (range, 0.5-50 m) and water depth of 1.5 m (0.1-1.5 m, n = 84), except for six reservoir and pond sites (39,42,44,72,85, and 90), which had depths of 0.3-10 m. There were five large streams with a water width of more than 100 m, including two sites (80 and 83) in the western Imjin River estuary zone and three sites (48, 53, and 54) in the central mountainous zone. Although most sites (16 to 91) included streams that flowed into the Bukhangang, Hantangang, and Imjingang and were distributed in the West Korean subdistricts, a few sites (1 to 15) had independent streams flowing into the East Sea and these were distributed in the Northeast Korea subdistricts.
The site at the highest elevation was centered on Taebaeksanmaek; elevation decreased gradually in a westerly direction, whereas it dropped sharply in an easterly direction ( Figure 2). Of the five regions, the eastern mountainous zone had the highest mean ± SD elevation of 359.8 ± 60.3 m (range 280-469 m), followed by the central mountainous zone at 264.3 ± 56.5 m (197-381 m), the eastern coastal zone at 225.7 ± 181.5 m (10-556 m), the western plain zone at 164.0 ± 75.0 m (27-253 m), and the lowest mean elevation was in the western Imjin River estuary zone, 24.8 ± 22.4 m (2-82 m). The river types in the mountainous zone were Aa and Aa-Bb types, and they took the forms of Bb, Bb-Bc, and Bc types toward the Yellow Sea. At several sites (86-89 and 91) in the western Imjin River estuary zone, more than 99% of the bottom structure was composed of sand and mud (with particle sizes < 2 mm in diameter) owing to the influx of mud from estuaries. in the western Imjin River estuary zone, 24.8 ± 22.4 m (2-82 m). The river types in the mountainous zone were Aa and Aa-Bb types, and they took the forms of Bb, Bb-Bc, and Bc types toward the Yellow Sea. At several sites (86-89 and 91) in the western Imjin River estuary zone, more than 99% of the bottom structure was composed of sand and mud (with particle sizes < 2 mm in diameter) owing to the influx of mud from estuaries.

Fish diversity and fauna
In total, 19562 individual organisms were collected between 2015 and 2019. Examination of these specimens revealed 81 species belonging to 19 families and 11 orders, and their distributions in the five regions are shown in Table 1 and Table S2. The family Cyprinidae was the most dominant with 43 species identified, representing approximately 46.3% of all identified species. Zacco koreanus (relative abundance (RA), 21.9%) was the most abundant fish species overall, while other key species were Zacco platypus (18.2%), Rhynchocypris oxycephalus (5.2%), Rhynchocypris steindachneri (4.9%), and Pungtungia herzi (3.7%).

Fish Diversity and Fauna
In total, 19,562 individual organisms were collected between 2015 and 2019. Examination of these specimens revealed 81 species belonging to 19 families and 11 orders, and their distributions in the five regions are shown in Tables 1 and S2. The family Cyprinidae was the most dominant with 43 species identified, representing approximately 46.3% of all identified species. Zacco koreanus (relative abundance (RA), 21.9%) was the most abundant fish species overall, while other key species were Zacco platypus (18.2%), Rhynchocypris oxycephalus (5.2%), Rhynchocypris steindachneri (4.9%), and Pungtungia herzi (3.7%). In the eastern coastal zone, 2544 individuals belonging to 38 species and 13 families were collected, and the dominant species was Zacco koreanus (appearance, 13 of 24 sites; RA, 43.6%). In the eastern mountainous zone, 19 species and eight families (from 1516 individuals) were sampled, and Zacco koreanus (8 of 10 sites; 38.7%) was the dominant species. In the central mountainous zone, 43 species and 12 families were classified from 7744 individuals investigated, and Zacco koreanus (20 of 24 sites; 30.3%) was again the most dominant species. In the western plain zone, 46 species and 10 families were found among 3405 individuals collected, and Zacco koreanus (10 of 13 sites; 21.4%) was the dominant species. Of the 4353 individuals collected in the western Imjin River estuary zone, 58 species and 14 families were classified, and Zacco platypus (12 of 20 sites; 28.6%) was the dominant species.
Community indices of each survey site in the five regions are shown in Table 2. The diversity and richness index was higher in the western Imjin River estuary zone than in the other regions. The evenness index was highest in the western plain zone and the dominance index was highest in the eastern mountainous zone. Hierarchical clustering showed that the five regions tended to cluster into two groups according to species composition ( Figure 3). The first group included the WP, WI, and CM zones and the second group included the EC and EM zones.
In the eastern coastal zone, 2544 individuals belonging to 38 species and 13 families 202 were collected, and the dominant species was Zacco koreanus (appearance, 13 of 24 sites; 203 RA, 43.6%). In the eastern mountainous zone, 19 species and eight families (from 1516 204 individuals) were sampled, and Zacco koreanus (8 of 10 sites; 38.7%) was the dominant 205 species. In the central mountainous zone, 43 species and 12 families were classified from 206 7744 individuals investigated, and Zacco koreanus (20 of 24 sites; 30.3%) was again the most 207 dominant species. In the western plain zone, 46 species and 10 families were found among 208 3405 individuals collected, and Zacco koreanus (10 of 13 sites; 21.4%) was the dominant 209 species. Of the 4353 individuals collected in the western Imjin River estuary zone, 58 spe-210 cies and 14 families were classified, and Zacco platypus (12 of 20 sites; 28.6%) was the dom-211 inant species. 212 Community indices of each survey site in the five regions are shown in Table 2. The 213 diversity and richness index was higher in the western Imjin River estuary zone than in 214 the other regions. The evenness index was highest in the western plain zone and the dom-215 inance index was highest in the eastern mountainous zone. 216 Hierarchical clustering showed that the five regions tended to cluster into two groups 219 according to species composition ( Figure 3). The first group included the WP, WI, and CM 220 zones and the second group included the EC and EM zones.  Table 1. 226 The PCA was used to identify possible relationships between eight factors (river 227 width, water width, water depth, river type, elevation, human activity, survey regions, 228 and districts) and a number of species at the 89 sites except for two sites (80 and 83) due 229 to appear to be extremely wide of river and water widths (Figure 4). Eigenvalues (lambda) 230 were 0.301, 0.214, 0.165, and 0.134 for the first to fourth axes, respectively. The number of 231 species was significantly correlated with elevation, average river and water widths, and 232 water depth and poorly correlated with human activity and disturbance, survey regions, 233 districts, and river type. 234 Figure 3. Dendrogram produced from a hierarchical cluster analysis of species composition in the five regions of the adjacent areas of the DMZ. Study region codes are the same as those defined in Table 1.
The PCA was used to identify possible relationships between eight factors (river width, water width, water depth, river type, elevation, human activity, survey regions, and districts) and a number of species at the 89 sites except for two sites (80 and 83) due to appear to be extremely wide of river and water widths (Figure 4). Eigenvalues (lambda) were 0.301, 0.214, 0.165, and 0.134 for the first to fourth axes, respectively. The number of species was significantly correlated with elevation, average river and water widths, and water depth and poorly correlated with human activity and disturbance, survey regions, districts, and river type.
In the West Korea subdistrict, 73 species and 17 families (17,603 individuals) were found, of which 55 species and 15 families were only found in these subdistricts, e.g., Lethenteron reissneri, Cyprinus carpio, Carassius cuvieri, Rhodeus ocellatus, and Rhodeus uyekii. Whereas in the Northeast Korea subdistricts, 21 species and 11 families (1959 individuals) were identified, of which eight species and seven families were unique to this subdistrict: Squalidus multimaculatus, Rhodeus semotilus, Orthrias toni, Iksookimia pacifica, Plecoglossus altivelis, Oncorhynchus masou, Pungitius kaibarae, and Cottus hangiongensis. There was a negative correlation between elevation and the total number of species; 261 however, there was a positive correlation between elevation and the total number of en-262 demic species (linear regression analysis, total species, R 2 =0.04, p = 0.110; endemic species, 263 R 2 =0.05, p=0.077). High species richness was confirmed at an elevation of approximately 264 200 m for all species and for endemic species ( Figure 5). There was a negative correlation between elevation and the total number of species; however, there was a positive correlation between elevation and the total number of endemic species (linear regression analysis, total species, R 2 = 0.04, p = 0.110; endemic species, R 2 = 0.05, p = 0.077). High species richness was confirmed at an elevation of approximately 200 m for all species and for endemic species ( Figure 5).

Endemism
Of the 81 species identified in all regions, eight were endangered in Korea (three families; RA, 9.8%: Lethenteron reissneri, Acheilognathus signifer, Pseudopungtungia tenuicorpa, Gobiobotia macrocephala, Gobiobotia brevibarba, Rhodeus semotilus, Brachymystax lenok tsinlingensis, and Cottus hangiongensis). One species, Hemibarbus mylodon, is a natural monument (1 family; 1.2%), and 29 species were endemic to Korea (eight families; 35.8%). The number of endemic species in the five regions was highest in the western plain zone (20 species) and the central mountainous zone (20 species) and lowest in the eastern mountainous zone (11 species ilies; RA, 9.8%: Lethenteron reissneri, Acheilognathus signifer, Pseudopungtungia tenuicorpa, 268 Gobiobotia macrocephala, Gobiobotia brevibarba, Rhodeus semotilus, Brachymystax lenok 269 tsinlingensis, and Cottus hangiongensis). One species, Hemibarbus mylodon, is a natural mon-270 ument (1 family; 1.2%), and 29 species were endemic to Korea (eight families; 35.8%). The 271 number of endemic species in the five regions was highest in the western plain zone (20 272 species) and the central mountainous zone (20 species) and lowest in the eastern moun-273 tainous zone (11 species 290 Human activities and disturbances were consistently observed throughout the sur-291 vey sites and evidence of these was found at 31 of 91 sites (RA, 34.0%). The highest dis-292 turbance factors were identified at 17 sites; river dredging and agricultural use, sewage, 293 and fisheries and development were recorded at seven, five, and two sites, respectively. 294 In particular, river dredging and agricultural use directly sieve or disrupt river sediments, 295 causing large amounts of mud, sand, and gravel to become suspended, thereby reducing 296 water quality and destroying faunal habitats. In the present study, the number of species 297 was poorly affected by human activity and disturbance, presumably, because some sur-298 vey sites will reflect the habitat features facing the prevailing currents owing to discontin-299 ued monitoring (Figure 4).

Human Activities and Disturbance
Human activities and disturbances were consistently observed throughout the survey sites and evidence of these was found at 31 of 91 sites (RA, 34.0%). The highest disturbance factors were identified at 17 sites; river dredging and agricultural use, sewage, and fisheries and development were recorded at seven, five, and two sites, respectively. In particular, river dredging and agricultural use directly sieve or disrupt river sediments, causing large amounts of mud, sand, and gravel to become suspended, thereby reducing water quality and destroying faunal habitats. In the present study, the number of species was poorly affected by human activity and disturbance, presumably, because some survey sites will reflect the habitat features facing the prevailing currents owing to discontinued monitoring ( Figure 4).

Geographic Characteristics and Ichthyofauna
The ichthyofauna survey in the DMZ areas was conducted by various government agencies such as the Ministry of Environment, the Korea Forest Service, and the Korea Cultural Heritage Administration until the late 2000s. Subsequently, the National Institute of Ecology has conducted this survey as a commissioned task of the Ministry of Environment since 2014 [2]. The sampling regions, based on the available literature, were broadly divided into the Eastern, Central, and Western regions and the geographic and physical characteristics of the study sites were similar to those in the present study: water width of fewer than 30 m and water depth of 1.5 m (Table S1). The eastern mountainous zone, centered on Taebaeksanmaek and consisting of a high mountain range, exhibited the narrowest water width and the highest elevation. The water width increased and elevation decreased as sites were sampled in the central mountainous zone, western plain zone, and western Imjin River estuary zone [2,33].
In the present study, the highest number of species was found in the western Imjin River estuary zone (58 species) because of the inclusion of brackish water and a combination of primary and secondary freshwater fish. The western plain zone had the second highest species diversity (46 species), followed by the central mountainous zone (43 species), which included the midstream and upstream regions of Imjingang and had relatively diverse habitats. The West and Northeast Korea subdistricts were included in the eastern coastal zone (38 species) to reflect the variety of fish species occurring in this zone due to an independent stream flowing into the East Sea. The lowest number of species was recorded in the eastern mountainous zone (19 species) and this was attributed to the absence of large rivers and relatively simple habitats [33,34].
Based on 21 previously published reports, a comprehensive list of ichthyofauna identified in the adjacent areas of the DMZ included 136 species belonging to 33 families and 14 orders (a total of 91,964 individuals) (Table S2). Zacco platypus (RA, 14.9%) was the most abundant species, while other dominant species were Zacco koreanus (14.4%), Rhynchocypris steindachneri (9.5%), Hypomesus nipponensis (3.9%), Rhynchocypris oxycephalus (3.5%), and Rhinogobius brunneus (2.9%). In these reports, the scientific names were updated using the most recent taxonomic reviews and publications and eliminated the fish listed such as ambiguous, misidentified, or diadromous species and reported approximately 109 species (Table S2). Only one species, Squaliobarbus curriculus, was newly added.
The most recent ichthyofauna investigation of the DMZ areas was conducted by Mintongseon monitoring from 2012 to 2014 [13][14][15][16]. In this monitoring, 33,384 individuals belonging to 85 species and 18 families were collected, and four more species (13,822 individuals) were identified than in the present study. However, 13 species that were confirmed in the Mintongseon monitoring were not found in this study, and nine species were newly added by the present study (Table S2). Although the length of the study periods did not differ much, there were 22 different species between the two surveys. In addition, for these study results to be reliable enough to base future conservation and human-use management decisions on, it is estimated that three surveys must be conducted continuously at the same site throughout the spring, summer, and autumn [35,36].
Because ichthyofauna has limited distribution patterns due to the interaction of various ecological and historical conditions related to geographical events, the study of species distribution patterns has become an important topic for the study of biogeography [19]. The Taebaeksanmaek is a barrier that has separated the water systems in the east and west districts since the beginning of the Third Era (Pliocene, the epoch in the geologic time scale that extends from 5.333 million to 2.58 million years ago), so that differences in freshwater fish fauna can be found in each region [34]. The geographic barrier of the Baekdudaegan mountain ranges in the Korean Peninsula is divided into West, South, and Northeast Korean subdistricts and has different fish fauna. However, the appearance of many new species not previously recorded in these water systems in recent years has progressed due to artificial phenomena, such as releasing fish, rather than natural processes. Among the species that did not appear before the 1960s-70s in the Northeast Korea subdistricts, 11 species, Pungtungia herzi, Pseudorasbora parva, Squalidus multimaculatus, Misgurnus mizolepis, Silurus asotus, Odontobutis interrupta, Zacco koreanus, Zacco platypus, Silurus microdorsalis, Liobagrus andersoni, and Coreoperca herzi, were identified in this study and were presumed to have appeared since the 1980s-90s [37]. The artificial introduction of native species into a water system in which they did not previously occur can also pose a risk because these species are effectively exotic species; therefore, fish management and efforts to elucidate the source of introduction of domestic exotics are required at the government level [21,35,38].

Endemism
This survey showed that the areas adjacent to the DMZ play an important role in freshwater fish diversity in South Korea, especially for endemic species. In the Mintongseon monitoring a higher number of Korean endemic species were recorded (RA, 40.7%, [13][14][15][16]) than in the present study (35.8%), the report of the DMZ areas (30.3%, [33]), and the Korean national parks survey (33.3%). Additionally, the Mintongseon monitoring revealed a higher level of endemism than the proportion of endemic species in the entire Korean Peninsula (23.6%, [39]). These results demonstrate the conservation value of the DMZ areas because the level of endemism identified in several survey sites was higher than the overall Korean endemic frequency of 30%.
There were 11 and 16 endangered species (I rank: an endangered species and II rank: a species that may become extinct in the near future if current threats are not removed, respectively) in Korea, and four, 13, and seven threatened species (Critically endangered, Endangered, Vulnerable, respectively) in the Red Data Book of Korea, designated by the Ministry of Environment [40]. When comparing the present study with Mintongseon monitoring, two endangered species (Lethenteron japonica and Pungitius sinensis) were not found in the present study; this may be because they have restricted distributions that were not sampled in the eastern coastal zone [13][14][15][16]. Although five endangered species (Lethenteron japonica, Acheilognathus signifer, Gobiobotia marcocephala, Gobiobotia brevibarba, and Gobiobotia naktongensis) were previously identified in the adjacent areas of sites 78 and 79 in the western Imjin River estuary zone, only two of these (Gobiobotia marcocephala and Gobiobotia brevibarba) were identified in the current study. It is presumed that severe habitat disturbance due to river dredging for restoration of flood damage in 2018 resulted in the loss of the other three endangered species. Considering that a variety of endangered and endemic species were found in the main regions, the DMZ areas in Korea reflect biogeographical patterns and the diversity of distinct fish communities, which can be valuable from a conservation and preservation perspective.
The proportion of endemic species was highest at elevations of approximately 200 m, and the richness was similar to previous surveys of the DMZ areas and the Korean national parks survey [33,39]. Additionally, the frequency of endemic species associated with the total number of species was highest in mountain valleys and headwater streams. Three exotic species (Carassius cuvieri, Micropterus salmoides, and Lepomis macrochirus) were identified at five sites, where four sites were reservoirs or streams just below the reservoir, and one site was the estuary of Imjingang. Site 44 of the Togyo reservoir was one area where Micropterus salmoides and Lepomis macrochirus were introduced in the 1970s for aquaculture as a protein supply [41,42]. However, it is presumed that these exotic species escaped from the aquafarm to other nearby rivers and will have negative effects on the native fish communities in Korea [43][44][45]. This is an important finding and highlights the need for management to ensure that exotic species do not enter the eastern mountainous or eastern coastal zones, which currently do not have active native fish conservation practices in place.

Current Status and Conservation
The DMZ areas include ecosystems that are not easily accessible to people owing to military installations and access control measures; therefore, many people assume that these areas are well-conserved in their natural, undisturbed conditions [21]. However, in the present study, proximal forests and habitats were only found in the eastern mountainous zone and in the upper streams of Namgang in the eastern coastal zone. We found that most of the regions were affected by development and illegal cultivation, whereby wetlands had been converted into rice fields in the western plain zone and the western Imjin River estuary zone, and military installations in the form of military camps, were discharging raw sewage into the streams in the central mountainous zone. Sung et al. [38] classified land damage in the DMZ areas into six categories: alteration of initial landform, loss of surface layer, land pollution, alteration of soil chemical properties, a decline of vegetation, and invasion of foreign species. Battisti et al. [46] reported the disturbance ecology of anthropogenic threats and divided natural disturbance into seven events: Flooding and other events connected to river dynamics, fires, extreme meteorological events, water fluctuation, geological events, biotic agents (pathogens), and biotic agents (animal disturbance). These diverse disturbances could potentially cause alterations to water quality, hydrologic regimes, and fish habitats.
Frequent wildfires have also affected the DMZ stream ecosystems in terms of visibility acquisition resulting from intentional military activities after political corporation [47]. In addition, wildfires sometimes occur naturally and are difficult to extinguish owing to the restricted access to the DMZ areas; therefore, many areas are burnt. Some researchers argue that the DMZ areas have no conservation value because they are disturbed ecosystems due to human interference from wildfires, noise, and so on [48,49]. Fires can influence water quality, water chemistry, discharge regime, stream stability, coarse woody debris, and native freshwater fish and lead to the introduction of exotic plants [50,51]. However, the DMZ areas are a good habitat for herbivorous wildlife because of the lack of large predators and the increasing prey for wild animals. Although vegetation is in the early or mid-term stages of secondary succession because of frequent wildfires, it can provide a wider variety of food for mammals and birds. In addition, wetlands are not used as agricultural land but are preserved, providing valuable habitats for many fish, amphibians, and reptiles [21]. Though there was secondary succession due to wildfires, the DMZ maintains a high degree of biodiversity that would allow it to return to its natural forest condition if wildfires were banned.
Many researchers have reported that the Baekdudaegan mountain range is an ecological axis in the north-south direction of the Korean Peninsula, and the DMZ is the east-west axis. This is because the DMZ has been less artificially affected for a long period of time and provides an important biodiversity corridor between East and West on the Korean Peninsula [33,52,53]. Therefore, the DMZ is considered to represent areas of national and international importance that are worth protecting owing to their ecological diversity and value [17]. However, negotiations on the conservation of the DMZ are slow and difficult because South and North Korea are still in a state of the ceasefire. Since cooperation with North Korea is necessary for the conservation of the DMZ, UNESCO, an international organization, can play an important role. UNESCO has designated two international nature reserves: one is a Biosphere Reserve designated by the Man and the Biosphere program and the other is a World Heritage Site under the World Heritage Act. It will be of great help to the conservation and utilization of the DMZ ecosystems if the zones are designated as UNESCO international protected areas.
It is necessary to establish a conservation strategy based on regional characteristics and to use available methods for evaluating the extinction risk of endangered fish species that include the distributional data and life traits of the species identified in the current study and previous surveys [21,33]. To achieve this, we recommend that the governments of South and North Korea should co-operatively manage the area to protect ecological diversity, minimize military installations, including prescribed fires and development, and improve relations between the two countries. In this study, investigations of the NLL and SLL from the MDL and its surrounding area were rarely conducted due to military issues. Therefore, a future survey near the MDL area is needed and is expected to facilitate the identification of more diverse fish species.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/d14121011/s1. Table S1: Geographic and physical data of each survey site in the area adjacent to the DMZ, South Korea. Table S2. A list of fish species was recorded and collected in the areas adjacent to the DMZ, South Korea. References [54][55][56][57][58][59][60][61] are cited in the supplementary materials.