A wetland is an area where the soil is either permanent-ly or intermittently submerged throughout the year, or for varying times of the year, including the growing season. As a result, hydrology determines soil formation and de-velopment (pedology), and in combination, the two vari-ables also affect the structure of the ecological communi-ties present in the wetlands. Because of the dual nature of a wetland, these variables provide conditions that sup-port both aquatic and terrestrial species. The prolonged presence of water creates conditions that favor the growth of specially adapted plants and animals and promotes the development of characteristic wetland soils (Cowardin et al. 1979, Mitsch and Gosselink 1993).

However, because of their characteristic traits, wet-lands create habitat islands in the surrounding landscape matrix, which may or may not be limited in size (areal extent). In typical mountainous areas, wetlands are more isolated and limited in areal extent (Forman 1995), and thus are described as azonal or extrazonal ecosystems rel-ative to the surrounding matrix (Spitzer and Danks 2006). In addition, in wetlands, the climate is cool, precipitation is relatively high, evaporation is often limited, and drain-age may be poor. Hence, for terrestrial species that prefer the dry habitats or are nonadapted to wet conditions, the aforementioned conditions have a negative effect on their hibernation and reproduction (Andersen 2006). In gen-eral, species inhabiting this type of habitats, which have limited resources, could choose special strategies such as temporal, spatial, and food resource partitioning to sur-vive and to maintain their population (Schoener 1974, Niemela 1993).

We investigated the carabid beetles in wetlands for explaining their survival strategies in wetlands that have restricted resources and limited habitat conditions. The carabid beetles show sufficient variation taxonomically and ecologically, and they are too abundant and sensi-tive to environmental changes to be considered a reli-able monitoring group. Specifically, carabid beetles have been observed to be highly sensitive to wetland types and wetland environmental characteristics (Do et al. 2007b). Their habitat distribution, annual and daily activity pat-terns, and diet are determined by abiotic factors (e.g., light, moisture, habitat structure) and biotic factors (e.g., competition, food supply) (Williams 1959, Thiele 1977, Loreau 1985, Kang et al. 2009, Do et al. 2011, Jung et al. 2011).

We tested the following predictions: 1) There is a sig-nificant difference in the assemblage of carabid beetles in the wetlands when compared with adjacent habitats (e.g., mountain), regarding species richness and abundance. 2) Carabid species that preferred the dry habitats limit their spatial distribution depending on the wet condition of the wetlands. 3) Carabid beetles inhabiting wetlands have different seasonal activities when compared with other habitats to survive and to maintain their population. This article also discusses the wetland environmental changes using the carabid assemblages as ecological indicators, for the establishment of wetland management regimes.

We studied the seasonal activity and spatial distribution patterns of carabid beetles at the Mujechi Wetlands (wet-land sites) of Ulsan Metropolitan City (Republic of Korea) and Mt. Jeongjok (mountain sites). The Mujechi Wetlands (35°27′51.13″ N, 129°08′38.94″ E) is a 1.5 ha montane peat bog at 520-530 m altitude on Mt. Jeongjok (700 m above sea level), which was designated as a National Wetland Conservation Area by the Republic of Korea in 1997. The wetlands originated from differential weathering and ero-sion, and it has existed since 1,785 ± 120 y BP (Choi 1998). The annual average temperature and precipitation were approximately 13.8℃ and 1,275 mm, respectively, as re-

corded at the Ulsan Weather Station, 19.4 km away from the Mujechi Wetlands. However, because the weather station is at an elevation of 35 m, it is likely that the ac-tual temperature of the wetlands is lower. Water flows in the wetlands from the south-southwest (SSW) to north-northeast (NNE) direction. The mean groundwater level was -34.7 m, which became shallower with increasing rainfall and deeper with less rainfall (Lee and Kim 2002).

An equal grid map (15 × 15 m quadrats) with simple vegetation, identifying the characteristic vegetation of the Mujechi Wetlands, was used to confirm the sampling sites and environmental variables (Kim and Kim 2003). The dominant wetland plant species are Molinia japonica, Miscanthus sinensis var. purpurascens, Pinus densiflora, and Alnus japonica, with the surrounding slopes being dominated by P. densiflora and Quercus serrata.

In the wetlands, M. japonica is most dominant cover-ing 85 quadrats (57%). Comparatively, mixed vegetation including M. sinensis var. purpurascens and P. densiflora occupied 35 quadrats (23.5%), whereas Q. serrata and A. japonica occur in 21 quadrats (14.1%) and 7 quadrats (4.7%), respectively. There is one quadrat, which consists of exposed soil located at the upper edge of the wetlands. The wetland is divided and isolated from the mountain slope by a forest road and ditch (Fig. 1).

Out of a total of 149 quadrats in the wetlands, pitfall traps (plastic cups, 7 cm diameter) were installed in 64 of them (43.0%). Installing pitfall traps in wetland-grass sites is very inconvenient because of the high water level and the protected status of the M. japonica community. In the mountainous area, 20 pitfall traps were installed at three sites every 3 m. The sites investigated in this study are designated as Wetland Conservation Areas and, hence, the use of baits and preservation liquids is prohibited for the protection of the species inhabiting the wetlands. As much as possible, the traps were installed in the center of the site in homogenous stands of vegetation at each site. The trapping period covered most of the growing season (from January to December 2005), and the traps were emptied after 48 h.

The carabid species richness was compared between treatments using individual-based rarefaction curves. This technique is based on a random resampling of the pool of captured individuals, and it is used to estimate expected richness at lower sample sizes (Magurran 1988, Gotelli and Colwell 2001). The distribution of the domi-nant carabid species and large-sized species that inhabit the wetlands were marked on an equal grid map, which designated where the individuals from each species were caught.

The difference in species composition and seasonal ac-tivities of carabid beetles between the wetlands and the mountain were tested for statistical significance using a paired t-test on data transformed using ln (x + 1). The spatial distribution of some species in the wetlands was calculated using a paired t-test and one-way analysis of variance (ANOVA) on the abundance and appearance of frequency data of the carabid beetles. All statistical analy-ses were performed using Minitab 16 software (Minitab Inc., State College, PA, USA).

A total of 1,733 individual beetles representing 30 ca-rabid beetle species (including brachinid species) were collected from the studied sites (Table 1). The wetland sites had 16 species (53.33% of total species richness), and mountain sites had 26 species (86.66% of total species richness), respectively. The rarefied species richness was observed to be higher for the mountain sites when com-pared with the wetland sites (Fig. 2). Moreover, carabid abundances between the wetland and the mountain sites

were significantly different (paired t-test value = -2.75, P = 0.008), although 12 species (70.58% of total species rich-ness in the wetlands) occurred in both areas. Some spe-cies such as Amara chalcites, Demetrias marginicollis, and Cymindis vaporariorum immaculatus were found exclu-sively in the wetland sites.

Both the wetland and mountain sites were dominated by Carabus sternbergi sternbergi (29.25% and 20.51%, re-spectively). Twenty-six species accounted for less than 5% of the caught (less than approximately 50 individuals) (cf. Table 1). In the wetland sites, Calosoma inquisitor cya-nescens, A. chalcites, and C. jankowskii jankowskii were relatively abundant species, but their proportional abun-dances were much lesser in the wetlands when compared with that in the mountain sites. Of the abundant species on the mountain sites, C. sternbergi sternbergi and C. jankowskii jankowskii accounted for 32.89%, whereas the other species recorded on the mountain sites comprised less than 5% of the total individuals.

In this study, we investigated Calosoma inquisitor cyanescens, C. sternbergi sternbergi, and C. jankowskii jankowskii, which were relatively abundant and were the largest sized species on both the wetland and the moun-tain sites.

C. inquisitor cyanescens showed a single-peak cycle in both the habitats. Seasonal activity patterns of this spe-cies showed no statistical differences between the areas (paired t-test value = 2.84, P = 0.708) (Fig. 3a). However, in the wetland sites, adult C. inquisitor cyanescens was active from June to October, and the maximum activity period of the species was very short. However, in the mountain sites, this species was active from May to September, which indicated that this species occurred at an earlier time when compared with the wetland sites.

The main activity period of C. sternbergi sternbergi ex-tended from March to November, with two peaks in April and August (Fig. 3b). The maximum activity period of the species was more clearly distinguished in the wetland sites than on the mountain sites. Their seasonal changes in both areas were significantly different (paired t-test value = -3.14, P =0.009).

C. jankowskii jankowskii showed significantly differ-ent activity patterns in the two areas (paired t-test value = -3.27, P = 0.007) (Fig. 3c). In the wetland sites, activity of this species began in May, but peaked in June, and the abundance during May-June was followed by a decline in August. However, on the mountain sites, C. jankowskii jankowskii had two peaks in May and September. The mountain sites-resident individuals showed a first period of activity in March-August that was always followed by a phase of inactivity, which was in turn followed by a period of activity again in September when it reached its maxi-mum.

In the wetland sites, the three studied species had significantly different seasonal patterns (F = 3.73, P = 0.035) (cf. Fig. 3d). They were observed to have different maximum activity periods, with C. sternbergi sternbergi dominating, whereas both C. inquisitor cyanescens and C. jankowskii jankowskii had a comparatively lower popula-tion. As a result, maximum activity period of C. jankowskii jankowskii was delayed, and thus, the peak activity period did not interfere with that of C. sternbergi sternbergi.

The three species of carabid beetles were present in 22 quadrats (30.99% of investigated quadrats, 12.86% of to-tal quadrats) in the wetlands (Fig. 4). They occurred in 14 quadrats of Q. serrata, 7 quadrats of M. sinensis var. pur-purascens (with P. densiflora), and 1 quadrat of exposed soil. The 6 quadrats had two species, whereas the other quadrats recorded just one species (Fig. 4a). The associa-tion between each vegetation type (number of quadrats) and each recorded carabid species showed no significant difference (F = 1.36, P = 0.325) (Fig. 5a), but each vegeta-tion type displayed significantly different abundances of each carabid beetle (F = 6.626, P = 0.0303) (Fig. 5b).

C. inquisitor cyanescens was distributed in 10 quadrats (5 grids and 6 quadrats of M. sinensis var. purpurascens and Q. serrata, respectively) (Fig. 4b). C. sternbergi stern-bergi was distributed over the largest area (16 quadrats) of the wetlands (Fig. 4c). This species was recorded in Q. serrata (11 quadrats) and M. sinensis var. purpurascens (4 quadrats) communities, as well as in one exposed soil site (1 quadrat). Both C. inquisitor cyanescens and C. sternbergi sternbergi showed a linear distribution along the wetland edge, adjoining the forest road. C. jankowskii jankowskii showed numerical limitations for only 2 quadrats (Fig. 4d).

We showed that the ecological characteristics of wet-lands affect carabid assemblages. Our results indicate the following: 1) The wetlands have a lower species richness and abundance of carabid beetles when compared with the mountain sites where there are relatively lower num-ber of limiting and disturbance factors. 2) The spatial-temporal distribution of carabid beetles inhabiting wet-land sites was clearly different from those inhabiting the mountain sites. These results show that carabid beetles inhabiting wetlands, where adverse habitat conditions are prevalent, have special strategies (e.g., spatial-tempo-ral partitioning) for maintaining their population.

The carabid species richness and abundance can be determined by the soil and vegetation community struc-ture in wetlands (Brose 2003), especially since soil water contents are a major factor affecting the suitability of a habitat for carabid beetles (Luff et al. 1989, Frambs 1990, Neve 1994). High soil water content is a negative influ-ence during the immature stages of the carabid beetles (Thiele 1977, den Boer 1981). Do et al. (2007a) reported that seasonal flooding of wetlands significantly decreased the species richness and abundance of carabid beetles, although some species adapted and preferred the wet soil condition and were able to recover their population after flooding. Therefore, Mujechi Wetlands, where high soil water contents are present, have lower species richness and abundance of carabid beetles than the mountain sites.

In this article, the seasonal activities of the three cara-bid species (C. inquisitor cyanescens, C. sternbergi stern-bergi, and C. jankowskii jankowskii) were distinguished between the wetland and mountain sites. In the early stages of the season, these activities increased quickly, indicating that the carabid beetles in the wetlands could be selecting a strategy to quickly produce a population within a short time during the productive season in the wetland areas, resource availability being the principal factor affecting productivity. Furthermore, species utiliz-ing similar ecological niches and seasonal activity pat-terns (spring or autumn breeders) often exhibit divisions in their activity time to reduce competition (Grum 1986, Sota and Ishikawa 2004).

Although the three carabid species were widely distrib-uted in the Mujechi Wetlands, they showed greater con-centration in the drier vegetation areas. It is proposed that the wet habitat acted as an obstacle to the movement of large-sized species, and hence carabid beetles moved lin-early along the relatively dry edges. It was demonstrated by Riecken and Raths (1996), who studied movement pat-terns using the telemetry method (radio tag), that cara-bid beetles inhabiting wetlands moved linearly because moist sites disturb the movement and diffusion of carabid beetles and drier vegetation types are usually distributed linearly in the wetland edges.

Mujechi Wetlands are undergoing a period of water loss with the soil becoming drier and the wetland thus under-going terrestrialization. Kim et al. (2005) described the distribution and age of alders in the Mujechi Wetlands. The alders extended from the inside to the outside edge of the wetlands and were associated with terrestrialization of the wetlands. However, it is suggested that the distri-bution of the carabid beetles progresses from the outside edge to inside of the wetland on the basis of the desic-cation of the wetland. If terrestrialization of the wetland progresses further, the distribution of the carabid beetles in the wetland can be extended throughout the wetland. Furthermore, when the wetland dries out, the carabid composition in the wetland and the adjacent mountain shows increasing similarity. It will result in a distribution-al range extension, although the carabid species compo-sition of the Mujechi Wetlands was significantly different from the adjacent mountain sites. For improving the wet-land habitat quality, the soil and organic matter migrat-ing from the forest road should be prevented, as it pro-motes the invasion of drier vegetation into the wetlands. In addition, continuous ecological surveys are needed for confirming ecological changes of the wetlands that are as-sociated with anthropogenic and ecological successions.