Passive Restoration Achieved through Natural Processes over 70 Years in the Korean DMZ

Abstract

The Korean Demilitarized Zone (DMZ) is an area where human disturbance has been strictly restrained for about seven decades since 1953. As a result, compared with rural and urban areas in Korea with similar ecological conditions, the landscape structure of the Korean DMZ today shows a big difference in that it has a riparian forest and lacks any of the artificial landscape elements, such as agricultural and residential areas. The vegetation maps made in the 1950s and those made in recent years of a model site that extends throughout the DMZ, the Civilian Control Zone (CCZ), and adjacent rural areas show that successional changes in the DMZ and CCZ areas are remarkable, while changes in the rural areas are not clear. Rice fields have been replaced by riparian forests, whereas young forests in areas previously subjected to excessive use and pine forests have been replaced by broad-leaved forests. Among the three landscape elements that changed, it was found that riparian zones, where natural disturbance is usually frequent, could mitigate the impacts of human disturbance and restore the original features of nature sooner than any other elements. The results of analyses on landscape change that focused on rice fields before the Korean War also showed similar results: most of these fields had turned into riparian vegetation. Stand ordination of riparian vegetation investigated in the CCZ, rural areas, and urban areas showed clear differences in species composition and diversity among regions. In this study, we confirmed the passive restoration of the Korean DMZ and CCZ through natural regeneration processes as a result of restricted human disturbance over a period of about seven decades.

1. Introduction

Korea is geographically characterized by abundant mountains and hills, which occupy about 65 percent of its territory. However, Korea is one of the countries with the highest population densities in the world [1]. Therefore, land use intensity in Korea is very high. In particular, land use in the alluvial plain land around rivers and hilly areas is excessive. The western part of the Korean peninsula is low and very gentle, with various types of plains, low hills, and basins developed along the rivers, while the eastern part is high and very steep, with no significant rivers or plains because high mountains are very close to the coast. Such topographical characteristics cause excessive land use in the western part and consequently lead to fragmentation of the green pathway, which runs from east to west [2,3].

In Asian countries, where rice is a dietary staple, most of the floodplains of rivers and/or streams were transformed into rice fields in the past, and large banks were constructed along waterways to prevent flooding damage. Therefore, the widths of most rivers and/or streams were greatly reduced. Further, many of those rice fields were not only again transformed into urbanized areas, including residential areas, but also meandering and complex channels were changed to straight and monotonous ones in urban areas [4,5,6,7]. In such continuing transformation processes, riparian communities have been greatly degenerated or destroyed by tree cutting, the introduction of exotic species, the diversion and channeling of water for agriculture, and the use of river beds and shores for cultivation or even as roads. Therefore, riparian vegetation is has greatly reduced and become rare in most Asian countries [5,6,7,8].

Riparian landscapes in Korea, including river ecosystems and surrounding environments, have hardly retained their original features owing to excessive use. To date, the riparian landscape has usually been managed in terms of use and disaster protection. However, its importance as a natural environment is being reevaluated [6,7].

Riparian ecosystems are spatially and temporally dynamic and are shaped by fluvial geomorphic processes; therefore, there are physical and biological links between terrestrial and aquatic environments and biotopes in which animals may seek refuge and food, which, in turn, enriches the soil with detritus [9,10,11]. Riparian ecosystems usually support higher species richness and densities of wildlife than other nearby ecosystems [12,13,14,15]. Riparian vegetation detains erosion materials, thus decreasing the amount of solids in suspension in the watercourses and improving the quality of the water [16,17,18,19,20,21]. Moreover, riparian vegetation slows down the flow of torrential rains and collects the material carried, thus reducing the effects of the rains downstream. Furthermore, the highly developed root systems of riparian vegetation reinforce the banks of streams [22,23,24].

All these advantages, together with the considerable enhancement of the landscape that this vegetation affords, justify considering this type of vegetation as being of primary importance [23]. Thus, the maintenance and/or restoration of riparian vegetation deserves to be given priority in land management projects [4,25,26].

The ecological imbalance caused by excessive land use is a fundamental cause of environmental problems, including climate change [27,28,29]. The damage and destruction of nature caused by human activities has harmed most other species and is causing adverse effects on humans themselves in various forms. Ecological restoration reverses such trends, providing opportunities for different living things to find new habitats, and providing humans with the possibility of escaping such effects [30,31]. Ecological restoration is a strategy to aid in the recovery of large areas of degraded ecosystems and/or landscapes [31,32,33]. Two general strategies of ecological restoration are recognized: (1) passive restoration, which eliminates the factors of disturbance and permits natural regeneration, and (2) active restoration, which eliminates the source of disturbance and implements strategies to accelerate recovery and overcome obstacles to that recovery [34,35,36]. Rates of natural regeneration through passive restoration are highly variable, and the process can be unsuccessful [37] since it depends on multiple factors, including the intensity and type of the previous land management, type of surrounding landscape, and resilience of the ecosystem [37,38]. Active restoration is where management techniques such as planting seeds or seedlings are implemented, and passive restoration is when no action is taken except to cease environmental stressors such as agriculture or grazing [34,35]. The probability that active restoration strategies are employed at a given site is highly dependent on the degree of environmental degradation [39,40]. Active restoration techniques are costly and tend to be used only in the most degraded ecosystems, where passive approaches would almost certainly fail [31,40].

Since the end of the Cold War in the 1990s, thousands of former military properties have been closed in various parts of the world due to international disarmament agreements and new security–political surroundings [41]. These abandoned military lands, like abandoned agricultural fields, can be a typical place for passive restoration [41,42]. The Korean Demilitarized Zone (DMZ) has been left relatively undisturbed for about seven decades since 1953, when the Korean War ended. The primary human interferences in this area have been occasional circumscribed fires deliberately set for military purposes, as well as some use of herbicides. To the south of the Korean DMZ, there exists an additional Civilian Control Zone (CCZ) of varying width, averaging 5.4 kilometers, which has also remained relatively undeveloped and only modestly disturbed (although some agriculture is permitted in this informal zone); moreover, a similar CCZ area is said to exist to the north. At present, the immediate surroundings of the Korean DMZ, including the DMZ itself, are heavily mined, and human disturbance has thereby been strictly restrained [43,44,45,46].

We hypothesized that the landscape damaged by human activities would be able to heal passively through the natural process if human interference was stopped. Furthermore, we assumed that passive restoration would proceed faster in dynamic ecological spaces, such as riparian zones. In order to test these hypotheses, we first analyzed landscape structure and change in the entire DMZ and the immediate surrounding area, the CCZ, and compared the differences between both zones as well as rural and urban areas with the same ecological conditions, except land use intensity, in the years 1979 and 2019. Secondly, we clarified the specificity of the DMZ and CCZ zones by comparing the results of our study with those of the rural and urban areas chosen as reference sites. Thirdly, we compared species composition and dynamics of riparian vegetation among several regions, including the DMZ and CCZ. Finally, we confirmed the passive restoration of the Korean DMZ and CCZ through natural regeneration processes as a result of restricted human disturbance for a period of about seven decades.

2. Materials and Methods

2.1. Study Area

The DMZ traverses a major river delta and old farmlands in the west and rugged mountains in the east (Figure 1). The DMZ ecosystems and landscapes represent a cross-section of the Korean peninsula, with central and western lowlands divided from the eastern highlands by the Taebaek Mountains and adjacent north–south mountain ranges. Geomorphologically, the western part is usually very gentle, with various types of plain, low hills, and basins developed along the rivers, while the eastern part is very steep and high, with no significant rivers or plains. Therefore, the intensity of land use is higher in the western part than in the eastern part. Biogeographically, this area is a transition zone where the southern and northern limits of the northern and southern plant species meet [47]. The topographical diversity and biogeographical position of this area provide a basis for high biodiversity.

The Korean DMZ contains six vulnerable, four endangered, and two critically endangered animal species [43,46,48,49]. Although a detailed survey of the ecology of the DMZ is not possible at present, the results obtained from the CCZ provide a glimpse of what may occur within the DMZ itself. The Korean DMZ is completely protected, clearly delimited, and demilitarized between two borders; these conditions allow the DMZ to function as a ready-made nature reserve [43,50].

To clarify the specificity of the land use regime of this area, three different areas, including the DMZ and CCZ, were designated. The survey area within the DMZ and CCZ was located throughout the Geonsol-ri, Bangsan-myun, and Yanggu-gun areas in Gangwon-do (Appendix A). Ri, myun, gun, and do are administrative units. Among them, gun and do correspond to county and province, respectively, in the English administrative system. The rural area, as a type of reference area, is located in Yanggu-eup and Yanggu-gun in Gangwon-do, and close to the DMZ in latitude. This area type is usually covered with forest, and land use is usually concentrated in the lowlands. Agricultural fields, residential areas, forests under artificial interference, and natural forests are distributed at a distance from the river. The urban area, another type of reference area, is located around Mt. Cheonggye, Yangjae-dong, and Seocho-gu in Seoul. Although the matrix of this area is also composed of a forest, it was classified as an urban area because the forest is surrounded by urban areas [51]. Urbanized areas, narrow belts of agricultural fields, forests under artificial interference, and natural forests are arranged in a concentric circle pattern in urban areas (Appendix A).

Model sites to clarify landscape change were designated in two areas. One site was designated in the watershed of the Suip stream, which runs from Mt. Geumgang, North Korea, through the DMZ and CCZ to Yanggu, South Korea. Other sites were designated in the abandoned rice fields where farming activities had been carried out before the Korean War.

2.2. Site Description

The Korean Demilitarized Zone (DMZ) was established in 1953 along the Military Demarcation Line by the Armistice Agreement between North Korea and the United Nations after the Korean War. This military buffer zone is a 4 km wide and 250 km long belt that traverses from east to west coast and divides the Korean peninsula into north and south (Figure 1). Civilian access has been prohibited; thus, the landscape of the DMZ has been protected from human disturbance, except for military patrols. To the south of the DMZ is the CCZ, a 1529 km² buffer zone with widths varying from 5 to 20 km. Civilian access to the CCZ has also been strictly controlled. For this reason, the CCZ has remained relatively undeveloped and only modestly disturbed (agriculture is permitted in limited areas); moreover, a similar CCZ is said to exist in the north [43]. Seven decades of protection from human intervention has allowed the DMZ/CCZ to recover its natural or close-to-natural landscapes. The recovered nature of the DMZ/CCZ is now a recipient of worldwide attention for its unusually high biological diversity [52]. Kim [43] labeled the DMZ/CCZ as “a unique nature preserve containing the last vestiges of Korea’s natural heritage”, and Westing [53] proposed to dedicate it to “a transfrontier (between North and South Koreas) reserve for peace and nature”.

The flora and fauna of the DMZ/CCZ have been studied by numerous authors in Korea (e.g., [45,46,52,54,55,56,57,58,59,60]). However, these studies were largely limited to a description or catalog of the species found in the local habitats of the DMZ/CCZ. Very little, if anything, has been studied on the extent and processes of natural recovery in the landscape or ecological communities of the DMZ/CCZ. The aims of our study were (1) to examine the extent of recovery in the natural landscapes of the DMZ/CCZ from the 1950s to 2000s, (2) to investigate the process of riparian forest development on abandoned rice paddies as an example of vegetation recovery, and (3) to address the enhanced biological diversity after such recovery by comparing the DMZ/CCZ with two reference sites near a rural village and an urban center.

The eastern part of the DMZ/CCZ is best described as rugged mountains with very steep slopes and high peaks. The central and western landscape is characterized by low hills and delta plains with gentle slopes located along river basins. Because of its fertile soil and easy access to water, the western plains were cultivated heavily for rice and other agricultural products before the Korean War (1950–1953). The annual mean temperature and precipitation are 10.2 °C, 10.2 °C, and 10.9 °C and 1252.9 mm, 1385.7 mm, and 1395.9 mm, respectively, in the eastern, central, and western parts of the DMZ [61]. Brown forest soil weathered from granite dominates the forest soil, but alluvial soil occupies wide areas in the lowlands, especially in the western part with lower elevation. Soil types are composed of Entisols, Inceptisols, Alfisols, and Ultisols [46].

Protection from human disturbance for more than seven decades has allowed an unusually high degree of biological diversity to exist in the DMZ and CCZ. Together, the DMZ and CCZ cover less than 3% of the total land area of South Korea [46]; however, the area supports more than 1854 plants, 2189 terrestrial insects, 351 benthic invertebrates, 136 fishes, 34 amphibians and reptiles, 266 birds, and 43 mammals. These species constitute approximately 41.9%, 13.3%, 23.0%, 10.8%, 65.4%, 51.0%, and 47.8%, respectively, of the total taxa for each group found in South Korea. Among them are six vulnerable, four endangered, and two critically endangered animal species, following the IUCN category [43,45,52,60]. At present, human access to the DMZ is strictly limited due to military security and countless buried land mines [44]; thus, a detailed survey of the ecology of the DMZ is not feasible. However, the results obtained from the CCZ can provide a good estimate of what may occur within the DMZ itself [43].

2.3. Methods

2.3.1. Landscape Analysis

Using Landsat^TM images and aerial photos, followed by field checks, we made a GIS map at a scale of 1:25,000 with ArcView (ESRI, Redlands, CA, USA) by overlaying the identified elements on a topographical map for each area. Small patches (<625 m²) were excluded from the map to solve the problem of uncertainty due to their size and their shape on the 1:25,000 map [62]. The cover areas of the elements were calculated and compared for the study areas.

To clarify the characteristics of the DMZ from the landscape perspective, land use patterns of the DMZ and CCZ were compared with those of the whole country. In addition, the topographic map based on aerial photos taken in the 1950s and the Landsat^TM images of the entire DMZ and CCZ taken in 1979 and 2019 were compared to examine temporal changes in their landscapes. Further, we compared the landscapes of the DMZ and the CCZ with two reference sites. The first reference site was a rural village and its vicinity outside of but adjacent to the CCZ in Yanggu-gun (127°53′~128°00′ E, 38°07′~38°11′ N), and the other site was an urbanized area (127°00′~127°06′ E, 37°23′~37°28′ N) in southern Seoul. Modest and extreme levels of human disturbance were assumed for the rural and the urban sites, respectively.

The natural recovery process at the landscape level was investigated by performing temporal (the 1950s and 2019) and spatial (from the DMZ through the CCZ to one model site) comparisons of land use patterns in the Suip stream watershed in Yanggu-gun (gun: county or district), Gangwon Province. Another analysis of the landscape change focused on the rice paddy before the Korean War, which shows a clear boundary. The boundary of the rice fields was identified by referring to a topographic map made by the National Institute of Geography of Korea. Changes in landscape elements identified as rice paddy before the Korean War in the DMZ and its surrounding areas, including the CCZ, were investigated by analyzing satellite images (Landsat TM) taken in 2019.

The degree of landscape fragmentation among DMZ and CCZ and two reference areas, a rural area and an urban area, were compared based on landscape indices obtained using Fragstats ver. 4.2 (University of Massachusetts, Amherst, MA, USA). The landscape shape index (LSI), fractal dimension (FD), and Euclidean nearest-neighbor distance (ENN) applied for landscape analysis were calculated using the following formulas:

$$\mathrm{LSI}=\frac{0.25E}{\sqrt{A}}$$

E = total length (m) of the edge in the landscape; A = total landscape area (m²).

$$\mathrm{Frac}=\frac{2\ln \left(0.25p_{ij}\right)}{\ln a_{ij}}$$

p_ij = perimeter (m) of patch ij; a_ij = area (m²) of patch ij; 0.25 is an adjustment coefficient.

ENN = h_ij

h_ij = distance (m) from patch ij to the nearest-neighboring patch of the same type (class).

2.3.2. Vegetation Analysis

The natural recovery process was also discussed by comparing the species composition and diversity of riparian vegetation in the CCZ and rural and urban areas. In addition, the structure and dynamics of the riparian vegetation established in the CCZ were analyzed. A vegetation survey within the DMZ was not possible; thus, it was replaced by the results of the CCZ survey.

In the field investigation, the occurrence and dominance of all plant species were recorded in 58 random plots in each area [63,64]. Plot sizes were 1 m × 1 m in the riparian plains dominated by herbaceous vegetation immediately adjacent to a stream channel, 2 m × 2 m in the shrublands, and 10 m × 10 m in the forests distant from the stream channel [6,7]. The nomenclature followed Lee [65] and KPNI [66]. Dominance was estimated with an ordinal class scale (from 1 for <1% to 5 for >75%) of Braun Blanquet [63], converted to the median value of the percent cover range in each class, and subjected to detrended correspondence analysis (DCA) for ordination [67]. A rank–abundance curve was constructed following Magurran [68] and Kent and Cocker [69], species diversity (H’) was calculated following Shannon and Weaver [70], and richness was determined by the number of species that occurred in each area. In addition, the stem diameters of major species were measured and sorted into two diameter classes (≤10 cm and >10 cm) for the four major types of riparian vegetation (Salix-, Robinia-, Acer-, and Alnus-dominated) near the Suip stream in the DMZ/CCZ site.

3. Results

3.1. Characteristics of Landscape Structure in the DMZ and CCZ Compared with the Whole Country

As a result of comparing the landscape structure of the DMZ and CCZ with that of Korea as a whole, the biggest difference was found in the ratio of forests to built-in areas (Figure 2). In Korea as a whole, the proportion of forests was slightly over 60%, but in the DMZ, the proportion was over 80%, and in the CCZ, the proportion was over 70%. The proportion of built-in areas was close to 6% in Korea as a whole, but the proportion was as low as 1% in the DMZ and CCZ. The ratio of agricultural fields was similar to the whole of Korea and the CCZ, at a little over 20%, but in the DMZ, the ratio was as low as 5%. On the other hand, the ratio of grasslands was as low as 2% in the whole of Korea and the CCZ, but more than 5% in the DMZ, which was higher than in both regions.

3.2. Landscape Change in the DMZ and CCZ

Appendix A reveals the recovery of the natural landscape in the DMZ/CCZ during the past seven decades. In the DMZ, the cover of mature forest (colored dark green in Appendix A) increased significantly, from 55.2% to 75.4%, while bare ground (gray) and rice paddies decreased from 25.1% and 18.4% to 1.8% and 2.6%, respectively. Similar changes occurred in the CCZ as well. The cover of mature forests (dark green) increased from 58.2% to 72.0%, while bare ground (gray) decreased from 24.9% to 4.0%. On the other hand, in CCZ, where human access is partially permitted, rice paddies remain without big changes, unlike in the DMZ (Appendix A and Figure 3).

As a result of the analysis of the landscape transfer from the 1950s to 2019 in the DMZ, forests, grasslands, paddy fields, and waters were usually transferred to forests or young forests, and bare lands were converted to rice fields and forests (

Table 1. Landscape transfer matrix in the DMZ and CCZ from 1950s to 2019 (ha). The number in parentheses indicates the ratio to the total area.

	Landscape Element		Forest	Young Forest	Shrubland	Grassland	Paddy Field	Bare Ground	Wetland	Waters	Built-in
1950s–2019	DMZ	Forest	31,543.1 (44.3)	4226.5 (5.9)	496.4 (0.7)	1206.2 (1.7)	550.4 (0.8)	1008.2 (1.4)	8.3 (0.0)	336.6 (0.5)	62.9 (0.1)
		Grassland	11,084.6 (15.6)	4093.8 (5.7)	333.4 (0.5)	1396.2 (2.0)	156.0 (0.2)	399.8 (0.6)	23.4 (0.0)	262.4 (0.4)	27.9 (0.0)
		Paddy field	5994.5 (8.4)	2793.5 (3.9)	56.2 (0.1)	1021.0 (1.4)	2318.6 (3.3)	342.9 (0.5)	80.0 (0.1)	223.4 (0.3)	264 (0.4)
		Bare ground	16.2 (0.0)	17.5 (0.0)	-	-	53.9 (0.1)	11.6 (0.0)	-	2.1 (0.0)	-
		Waters	443.5 (0.6)	105.1 (0.1)	1.5 (0.0)	63.0 (0.1)	72.5 (0.1)	34.6 (0.0)	8.1 (0.0)	84.5 (0.1)	6.2 (0.0)
	CCZ	Forest	50,353.3 (46.8)	2408.7 (2.2)	10.8 (0.0)	68.5 (0.1)	7444.1 (6.9)	1985.6 (1.8)	39.6 (0.0)	349.4 (0.3)	-
		Grassland	18,706.8 (17.4)	289.7 (0.3)	-	18.8 (0.0)	6772.2 (6.3)	289.8 (0.3)	26.4 (0.0)	554.2 (0.5)	-
		Paddy field	5249.0 (4.9)	689.9 (0.6)	1.0 (0.0)	119.2 (0.1)	8983.2 (8.4)	232.7 (0.2)	191.7 (0.2)	526.2 (0.5)	75.6 (0.1)
		Wetland	74.3 (0.1)	4.4 (0.0)	-	30 (0.0)	174.2 (0.2)	9.4 (0.0)	139.9 (0.1)	203.2 (0.2)	-
		Waters	391.7 (0.4)	8.4 (0.0)	-	36.1 (0.0)	520.5 (0.5)	36.6 (0.0)	132.7 (0.1)	360.6 (0.3)	-

). Meanwhile, more diverse changes appeared in the CCZ. Forests usually remained as forests or were converted to rice fields. Grassland was converted to forests or rice fields, and paddy fields remained as rice fields or were converted to forests. Wetlands and waters were usually converted to rice fields, but they remained themselves or were converted to forests (Table 1).

3.3. Landscape Structure of the DMZ and CCZ Compared with Two Reference Sites

The landscape element types identified by analyzing satellite images of the DMZ and CCZ, as well as a rural area and an urban area with different land use intensity levels, are composed of secondary forest, introduced plantation, agricultural field, riparian forest, and others (Figure 4 and

Table 2. The configurations of landscape elements identified from landscape ecological maps of the study areas: the DMZ and CCZ and two reference areas, a rural area and an urban area.

Landscape Element Type	DMZ and CCZ		Rural Area		Urban Area
	(ha)	%	(ha)	%	(ha)	%
Secondary forest
Quercus mongolica Fisch. ex Ledeb.	1219.0	76.6	3832.9	25.2	1684.3	37.9
Betula costata Trautv.	141.8	8.9	-	-	-	-
Quercus aliena Blume	-		-	-	1.2	0
Quercus variabilis Blume	9.2	0.6	-	-	441.9	37.6
Quercus acutissima Carruth.			-	-	412.7	9.2
Pinus densiflora Siebold & Zucc.	4.4	0.3	2805.1	18.5	131.3	2.9
Mixed forest	87.4	5.5	4782.7	31.5	-	-
Subtotal	1461.8	91.9	11,420.7	75.2	2671.3	59.6
Introduced plantation
Pinus rigida Mill.	-	-	89.9	0.6	100.4	2.2
Pinus koraiensis Siebold & Zucc.	-	-	121.6	0.8	16.5	0.4
Larix kaempferi (Lamb.) Carrière	-	-	381.4	2.5	69.8	1.6
Populus × tomentiglandulosa T.B.Lee	-	-	-	-	22.7	0.5
Castanea crenata Siebold & Zucc.	-	-	-	-	169.3	3.8
Robinia pseudoacacia L.	-	-	-	-	439.6	9.8
Subtotal	0.0	0.0	595.9	3.9	818.3	18.3
Agricultural field
Cropland	-	-	2578.2	17.0	527.4	11.8
Orchard	-	-	-	-	-	-
Abandoned field	16.5	1.0	-	-	-	-
Subtotal	16.5	1.0	2578.2	17.0	527.4	11.8
Riparian forest
Riparian forest	112.2	7.1	-	-	-	-
Subtotal	112.2	7.1	-	-	-	-
Others
Stream and reservoir	-	-	601.6	4.0	-	-
Graveyard	-	-	-	-	11.3	0.3
Urbanized area	-	-	-	-	449.4	10.0
Subtotal	0.0	0.0	601.6	4.0	460.7	10.3
Total	1590.5	100.0	15,193.4	100.0	4477.7	100.0

). The DMZ and CCZ have riparian forests, whereas they lack introduced plantations and agricultural fields, except for abandoned fields, and other element types, and thereby show a difference from rural and urban areas. Such results of the landscape structure show that the DMZ and CCZ are hardly fragmented and retain connectivity among natural landscape elements.

The landscape shape index was the largest in the rural area, followed by the urban area and the DMZ and CCZ, while the fractal dimension and Euclidean nearest-neighbor distance indices showed a reverse order (Table 2). These results imply that the landscape of the DMZ and CCZ is less fragmented compared with that of the urban and rural areas; the shape of the patch is not simple, but complex, and connectivity between patches is high [71,72]. Consequently, the DMZ and CCZ showed a significant difference in connectivity compared with rural and urban areas (Appendix A and

Table 3. A comparison of landscape shape indices of the study areas: the DMZ and CCZ and two reference areas, a rural area and an urban area.

Districts	Landscape Shape Index	Fractal Dimension	Euclidean Nearest-Neighbor Distance
DMZ and CCZ	7.46	1.15	424.42
Rural	17.44	1.10	178.34
Urban	15.69	1.11	250.03

).

3.4. Landscape Change in the Suip Stream Watershed

The landscape elements of the Suip stream watershed obtained by analyzing a map drawn up in the 1950s are composed of broad-leaved forest, pine forest, young forest, upper field, rice field, sand bar, and stream (Figure 5). Among these landscape elements, young forest occupied the largest area, and the areal sizes were larger in the order of broad-leaved forest, pine forest, rice field, river, and upper field. The vegetation map obtained by analyzing the Landsat satellite image taken in 2019 showed a simpler pattern than that taken in the 1950s. For example, pine forest, young forest, and upper field in the upper northern part, which is attributed to the DMZ, usually changed to broad-leaved forest, and rice field usually changed to riparian forest. However, rice paddies in the middle and lower parts, which corresponded to the CCZ and rural area, respectively, have partially and mostly remained.

3.5. Landscape Change in Abandoned Rice Paddies

Landscape change in abandoned rice paddies was reviewed by analyzing satellite images of how rice paddies, which were identified as rice paddies before the Korean War, changed during the 70 years after the war. Landscape elements that changed from the landscape element identified as paddy fields before the Korean War were classified into seven types: forest, young forest, shrubland, wetland, paddy field, aquatic area, and bare ground (

Table 4. A configuration of landscape elements identified as paddy fields before the Korean War in the DMZ and CCZ as of 2019.

Landscape Element	DMZ		CCZ
	Area (ha)	Percentage (%)	Area (ha)	Percentage (%)
Forest	3361.5	60.0	4041.9	30.3
Young forest	865.9	15.5	494.1	3.7
Shrubland	550.5	9.8	791.4	5.9
Grassland	95.1	1.7	167.7	1.3
Paddy field	344.6	6.2	6402.2	47.9
Bare ground	191.2	3.4	1000.4	7.5
Aquatic area	189.7	3.4	456.0	3.4
Total	5598.0	100.0	13,153.7	100.0

). In the DMZ, among the landscape elements that changed, the forest occupied the largest area, and the areal size followed in the order of young forest, shrubland, paddy field, aquatic area, and bare ground, with wetland occupying the smallest area. In the CCZ, the paddy field occupied the largest area, and the areal size followed in the order of forest, bare ground, shrubland, young forest, aquatic area, and wetland (Table 4). Compared with the landscape structure between both zones, the forest and the young forest occupied a higher proportion in the DMZ, whereas the rice paddy and bare ground occupied a higher proportion in the CCZ (Table 4).

3.6. A Comparison of Riparian Vegetation among Areas with Different Land Use Intensity

The results of stand ordination based on riparian vegetation collected in areas with different land use regimes showed that stands were arranged in the order of the DMZ and CCZ, the urban area, and the rural area on Axis I (Figure 6). On Axis II, the stands of the DMZ and CCZ were dispersed widely, whereas those of the rural and urban areas were concentrated in the middle and lower parts of the axis, respectively. This result implies that the species composition of riparian vegetation in the DMZ and CCZ is different from that of the rural and urban areas.

Species diversity was compared using the species rank–dominance curve. The results showed that the highest riparian vegetation was established in the DMZ and CCZ and the lowest was in the urban area (Figure 7).

3.7. Ordination of Riparian Vegetation Established in the DMZ and CCZ

The results of the ordination of stands investigated in the DMZ and CCZ were arranged in the order of herbaceous plants and shrub-dominated stands, alder-dominated stands, willow-dominated stands–maple-dominated stands, and black locust–maple-dominated stands on Axis I (Figure 8). On Axis II, the forb-dominated and grass-dominated stands were divided into upper and lower parts, respectively, and the other stands were distributed in the center.

Black locust stands would be dispersed from plantations introduced in mountain feet or riverbanks before the Korean War [73]. In fact, black locust plantations are frequently found at mountain feet or on riverbanks in many regions in Korea [74,75,76]. Many young trees of maple appear in a sub-tree layer of these black locust stands, and the succession of these stands into maple stands is thereby persuasively predicted. Maple stands have not developed into a mature forest yet, as the height of the trees composing the over-story is 6–7 m, although the species composition shows typical features of a floodplain. Willow stands are usually established on sites with a concave micro-topography.

3.8. Vegetation Dynamics

The vegetation dynamics of four major vegetation types, willow, alder, maple, and black locust stands, were analyzed (Figure 9). In the willow community, willow (Salix pierotii) and maple (Acer tartaricum subsp. ginella) trees are mixed; willow trees tend to dominate in the larger-diameter classes, while maple trees dominate in the lower-diameter classes. The diameter class distribution pattern of maple trees shows a reverse J-shaped distribution, whereas the diameter class distribution pattern of willow trees is close to the normal distribution type. The alder community shows a typical reverse J-shaped distribution pattern. Shrubby willow (Salix gracilistyla Miq.) appears together in the lower layer of the community, but its occupancy is very low. The maple community shows a normal distribution pattern. In the black locust–maple community, black locust and maple trees are mixed, with the black locust population showing a normal distribution pattern, whereas the maple population shows a reverse J-shaped one.

4. Discussion

4.1. Ecological Characteristics of the DMZ and CCZ Based on Landscape Structure

The landscape of the DMZ maintains high ecological connectivity, with more than 90% natural landscape elements (Figure 2). The fractal dimension and Euclidean nearest-neighbor distance indices were the largest in the DMZ and CCZ, followed by the urban area and rural area; however, the landscape shape index tended to show a reverse order (Table 2). These results imply that the landscape of the DMZ is less fragmented compared with that of the urban and rural areas; the shape of the patches is not simple, but complex, and connectivity between the patches is high [71,72]. These results are believed to have resulted from the strictly limited human access in the area for about 70 years following the Korean War [77]. In particular, the recovery of the riparian forest that has disappeared in most areas of Korea contributed to the connectivity of natural landscape elements. This riparian forest can function in corridors of the landscape mosaic scale and thereby connect major landscape features [14,78]. Landscape elements, such as agricultural fields and urbanized areas in rural areas and urban areas, respectively, sever natural landscapes and thereby lead to fragmentation of the landscape. Fragmentation is usually the greatest threat to biodiversity [79,80,81]. Taking into consideration that maintaining a large connected intact landscape is the only solution to the fragmentation problem, the DMZ and CCZ could be nature reserves by themselves [47].

4.2. Evidence of Passive Restoration

One of the most important considerations in ecological restoration is the pace of change from a degraded site toward a more desirable community/ecosystem [82]. Natural succession is known to function in a passive form of restoration [8,83,84,85]. The first evidence of passive restoration can be found in landscape changes in the Suip stream watershed (Figure 5). Landscape changes in the Suip stream watershed progressed differently between the DMZ and CCZ and nearby rural areas. In the DMZ and CCZ, the young forest was changed to a broad-leaved forest as a result of succession due to the restriction of human access since the Korean War. The rice field was changed to a riparian forest, which was also a result of the limited disturbance in the area since the war. Lee et al. [83] clarified that succession in the abandoned paddy fields progressed to alder-dominated stands through a serial seral stage of annual grassland, perennial grassland, and willow-dominated stands.

Further evidence of passive restoration can be found in landscape change in the abandoned rice fields, which were rice paddies before the Korean War. Rice paddies were replaced by wetlands, shrublands, or young forests and, furthermore, by forests through the successional process over about 70 years after the Korean War. In fact, Lee et al. [83] clarified the successional process by reporting that an abandoned rice paddy was replaced by a riparian forest dominated by alder after about 50 years of abandonment. In addition, Lim et al. [8] also clarified that an abandoned rice paddy succeeded to a riparian forest, and the successional process resembled the spatial distribution pattern of riparian vegetation, which is under frequent flooding disturbance. Based on these viewpoints, this temporal change or land use pattern in the DMZ could be regarded as a successional change or passive restoration. Compared with the results for the CCZ, where human disturbance is partially permitted, such an evaluation is more persuasive. In the CCZ, rice paddies occupied 47.9%, whereas forests, young forests, shrublands, and wetlands, which correspond to riparian vegetation, occupied a lower proportion of 41.2%. These results suggest that rice cultivation continued, although it was limited to the CCZ. In addition, higher connectivity of DMZ and CCZ landscapes compared with other areas (Table 2), higher percentage of natural landscape elements in the DMZ compared with the CCZ, where human access is partly allowed (Figure 3), and similarity of vegetation established in the CCZ reach of the Suip Stream to vegetation established in the late stages in succession of the abandoned rice fields (Figure 6, [8,83]) could be used as pieces of evidence that the DMZ, which has extremely limited human interference, has recovered from past disturbed conditions through natural processes.

4.3. The DMZ and CCZ as Hot Spots of Biodiversity in Korea

The DMZ is known as a thesaurus of biodiversity like any other abandoned military lands [41,42]. Biological resources in the DMZ are abundant, with plant, insect, fish, amphibian, reptile, bird, and mammal species attributed to 1220, 1215, 83, 12, 16, 200, and 52 taxa, respectively [45,46,60]. In terms of the percentages of the total taxa that appear in Korea, they occupy 38.6%, 10.9%, 61.5%, 80.0%, 55.2%, and 68.4%, respectively [45,46,60]. Considering that the areal percentage of this zone corresponds to only 0.44% of the total area of Korea, such occupancy percentages are indeed surprising [47].

The most alarming immediate global threat to life on Earth is that of biological impoverishment caused by an accelerating downward spiral of biological diversity [86]. This spiral, in turn, is driven by accelerating rates of loss of species and their habitats [87,88].

The threats to nature and its life support systems are even graver because the human population is already destroying a significant part of all terrestrial photosynthetic production. This destruction includes crop wastes, forest burning and clearing, desertification, and agricultural and urban–industrial land conversion—both of which adversely affect much more of our life support systems through air, soil, and water pollution. We can assume that more species will become more susceptible to extinction, and their prospects for survival will be diminished further because of the degradation and fragmentation of their habitats in natural and human-modified landscapes. These threats are caused by the great variety of intentional and unintentional effects of modern agricultural and urban–industrial, recreational, and other land uses [86].

Habitat fragmentation is generally considered to have two components: (1) reduction in the total amount of a habitat type, or perhaps of all natural habitat, in a landscape, and (2) apportionment of the remaining habitat into smaller, more isolated patches [89,90,91].

Corrective action to solve problems related to fragmented landscapes includes the maintenance or restoration of large, intact core areas that span large portions of regional landscapes [92]. The nature of the DMZ and CCZ recovered after a severe disturbance during the Korean War is a product of passive restoration achieved through natural processes. In this respect, it can be concluded that this region solved fragmentation by itself. The results demonstrate the real fact of the surprisingly high and significant biodiversity in this region. Indeed, this region comprises less than 1% of the total area in terms of spatial extent. However, this region houses from ca. 11% to 80% of all species in South Korea, depending on the taxon; moreover, the region contains several vulnerable, endangered, and critically endangered species [43]. In this respect, this region could occupy a position as a hot spot of biodiversity in Korea [44].

4.4. Ecological Value of the Restored DMZ

Firstly, the DMZ can be regarded as a model for restoration, in particular, of degraded riparian ecosystems. In Oriental countries, where rice is a dietary staple, most of the floodplains of rivers and/or streams were transformed into rice fields in the past, and many of those rice fields were more recently transformed into urban areas, including residential areas. In such continuing transformation processes, riverside communities have been greatly degenerated or destroyed by tree cutting, the introduction of exotic species, the diversion and channeling of water for agriculture, and the use of riverbeds and shores for cultivation or even as roads. Therefore, we need a reference river with an integrated structure and function in order to restore such degraded urban rivers; the riparian ecosystem revived in the DMZ and CCZ can provide satisfactory information as such a model [93].

Secondly, this area can play a role in wetland deficiency in Korea. As mentioned earlier, rice is a staple of the Korean diet; therefore, most wetlands, including the floodplains of rivers and/or streams, have been transformed into rice fields. However, most rice fields in the DMZ and CCZ, which have been left for about seven decades, have recovered almost all of their original features through natural processes. The greater diversity in wildlife and natural habitats (Figure 1 and Table 2, Table 3 and Table 4) of the DMZ and CCZ compared with the other regions in Korea can be regarded as evidence of ecosystem recovery.

Finally, a long conservation belt amounting to 250 km that the DMZ and CCZ form can function as an ecological corridor connecting the eastern mountainous zone and the western plain zone in Korea. In terms of topography, Korea is low and gentle in the western part, whereas it is high and steep in the eastern part. Therefore, the western part, which is composed of various types of plains, low hills, and basins, has been dominated by diverse development pressures, while the eastern part has been relatively well conserved due to its steep conditions that are difficult to exploit. Accordingly, the ecological corridor linking the south and the north is relatively well connected, but the link between the east and the west hardly maintains connectivity. The DMZ and CCZ belts, which have escaped human disturbance for a long period and thus recovered the entire landscape, connect the east and west of the Korean peninsula and could play a role as an ecological corridor.

Since the end of the Cold War in the 1990s, international disarmament agreements and a new security political environment have led to the abandonment of thousands of former military facilities in various parts of the world [41]. Abandoned military lands can be positioned as typical places to recover valuable natural resources through passive restoration, as seen in the Korean DMZ [94]. Military lands, in particular, are becoming important ecological reserves because they cover large areas and are often protected from intensive agricultural and urban development. In this respect, there is an urgent need for international attention and efforts to preserve abandoned military areas.

5. Conclusions

The Korean DMZ, where artificial interference has been strictly controlled for about 70 years, has a significant proportion of natural landscape elements compared to other regions in Korea. As a result of comparing landscape structures with rural and urban areas with similar ecological conditions, the Korean DMZ is different from other regions in that it maintains the connectivity of natural landscape elements by restoring riparian forests that have disappeared from most regions in Korea through passive restoration. As a result of comparing the vegetation of the 1950s with the recently made vegetation map of the model site leading to the DMZ–CCZ–nearby rural area, successional changes in vegetation and land use types caused by artificial interference were evident in the DMZ and CCZ, and in particular, the change of rice paddies to riparian forests was noticeable. Most of the land used as rice paddies before the Korean War was also changed to riparian vegetation, and these results showed the same trend as the results of a study on the succession of abandoned rice fields.

A comparison of riparian vegetation in the CCZ, rural, and urban areas showed distinct differences in species composition among the areas. In particular, the riparian vegetation of the CCZ reflected the spatial difference in the frequency and intensity of flooding, showing the typical diversity of riparian vegetation by establishing all grasslands, shrublands, and tree-dominated forests. These results are interpreted as the result of passive restoration that occurred through natural processes in the CCZ, where artificial interference has been controlled for a long time. This restoration of nature has led to the high ecological connectivity of the region, and it could provide a background for the region to have high biodiversity.

However, there are still many restrictions on field surveys in this area because of the many mines that are buried. Thus, it should be noted that the results obtained in this study are not satisfactory. Despite the restrictions on conducting a detailed survey, this area is known to have very high biodiversity. Therefore, if a systematic and detailed investigation is conducted, the information is expected to increase significantly. In this respect, we would like to propose a systematic investigation into the biodiversity of the region through the cooperation of relevant authorities in the future. Furthermore, we would like to recommend designating and preserving this area, which functions as a refuge and ecological pathway for various animals and plants, as a natural heritage conservation area, and as a park of peace between South and North Korea.