1. INTRODUCTION

Rivers exhibit variations in flow velocity, water depth, physicochemical properties, and substrate composition along their upper, middle, and lower reaches. Despite these changes, continuity is maintained across sections, forming heterogeneous abiotic environments in each reach. This physical heterogeneity creates specialized habitats for diverse freshwater organisms and is a key factor enhancing the functional diversity of river ecosystems (Son and Song 1998). Although freshwater ecosystems account for less than 1% of the Earth’s surface, they support over 10% of global biodiversity. Freshwater fish, in particular, act as higher-level consumers in river ecosystems by preying on insects and invertebrates, thereby maintaining energy flow (Ji et al. 2020). They are also highly sensitive to changes in water quality and habitat conditions, functioning as indicator species for assessing the health of freshwater ecosystems (Park et al. 2017).

Bongseonsa Stream, located in Sohul-eup, Pocheonsi, Gyeonggi Province, is approximately 6.4 km long. Most of its midstream section is included in the protected area of the Korea National Arboretum, where natural forests help maintain a relatively intact stream ecosystem. The Korea National Arboretum was designated as the fourth UNESCO Biosphere Reserve in Korea in 2010 and continues to sustain high biodiversity (Lim et al. 2024). In particular, habitat restoration and management of physical and water quality environments have been implemented in Bongseonsa Stream within the arboretum.

The Korean chub (Zacco koreanus), the focal species of this study, is an endemic freshwater cyprinid fish found in Korea that inhabits clean streams in mid- to upperreach rivers. It thrives only in pristine water environments (Hur and Seo 2011). As a higher consumer feeding on aquatic insects and invertebrates, it is widely used as a bioindicator for evaluating the ecological health of regional river systems due to its sensitivity to water quality and habitat changes (Kim et al. 2005). However, industrialization and urbanization have led to water pollution and artificial modifications of river structures, resulting in rapid population declines of Z. koreanus in many rivers and in some areas, local extinction is presumed to have occurred (Choi and Byeon 2009).

Historical records indicate the presence of Z. koreanus in Bongseonsa Stream, based on surveys conducted by the Forestry Experiment Station of the Government- General of Korea in 1932 (The Government- General of Korea 1932). However, subsequent studies did not confirm its occurrence, reporting the possibility of local extinction (Choi and Byeon 2009). To restore the species, the Korea National Arboretum translocated 300 individuals (males and females) from the Jojong and Sudong Streams, which have similar habitats, to the Bongseonsa Stream in 2008, after which monitoring studies were conducted (Byeon 2011), and comparative studies between the reintroduced population in Bongseonsa Stream and source populations in the Jojong and Sudong Streams were performed in 2016 (Wang et al. 2017a). Both studies confirmed that the reintroduced population had maintained a stable population size above a certain threshold.

Most fish restoration studies to date have focused primarily on endangered species, with monitoring efforts emphasizing presence/absence, density, and distribution following reintroduction (Ko et al. 2012;Ko et al. 2019;Han et al. 2020). However, relatively few studies have assessed long-term population growth, reproductive capacity, and other indicators of population viability.

Therefore, this study aimed to (i) evaluate the current population status and habitat adaptation of reintroduced Z. koreanus in Bongseonsa Stream through longterm monitoring, and (ii) assess reproductive ability by comparing the reintroduced population with the source populations in the Jojong and Sudong Streams.

2. MATERIALS AND METHODS

2.1. Survey period and sites

Surveys of the reintroduced Z. koreanus population in Bongseonsa Stream (BS) were conducted three times: June 29-30, 2015; July 2, 2021; and June 24, 2025. In addition, surveys comparing reproductive ability among streams were conducted as follows: Bongseonsa Stream (June 27, 2016; July 2, 2021; June 24, 2025), Jojong Stream (JJ: June 29, 2016; July 7, 2021; June 25, 2025), and Sudong Stream (SD: June 29, 2016; July 7, 2021; June 26, 2025). Survey sites and GPS coordinates are as follows (Fig. 1).

- Survey sites for length-weight measurements of Z. koreanus

- Survey sites for reproductive ability GSI (Gonadosomatic index) comparisons

2.2. Survey methods

Sampling of Z. koreanus populations were conducted at each survey site using cast nets (7×7 mm) and hand nets (4×4 mm), and sampling was performed 12 times with cast nets and for 30 minutes with hand nets. Both juveniles and adults were collected in this survey. For individuals used in the length-weight analysis, total length (TL) was measured on site with a scale ruler to the nearest 0.1 mm, and body weight was measured with an analytical balance to the nearest 0.1 g. For individuals used in reproductive ability comparisons, fieldcollected Z. koreanus that were more than one year old (>90 mm) (Im et al. 2010) were used. These individuals were randomly sampled and transported to the laboratory, where gonad weights were measured.

2.3. Population analysis of Z. koreanus

To compare and analyze the growth conditions of Z. koreanus populations, the growth rate and condition factor (K) were analyzed using the length-weight relationship. These indices are used to assess fish growth status and reproductive ability, and provide information on water quality, feeding efficiency, and habitat quality (Anderson and Gutreuter 1983;Ney 1993).

The length-weight relationship was determined following Anderson and Gutreuter (1983), W=aTL^b (W= weight, TL=total length, a, b=parameters). The condition factor (K) was determined according to Anderson and Neumann (1996), K=W/TL³ (W=weight, TL=total length).

Length frequency distribution was also analyzed to assess the age structure and growth stages of populations indirectly. Collected individuals were grouped into length classes, and frequencies were calculated. This method is used to evaluate growth patterns, size diversity, recruitment, and population stability (Bagenal and Tesch 1978). In this study, individuals were grouped into 10 mm intervals to derive distributions for each survey year.

2.4. Gonadosomatic index (GSI) analysis

The gonadosomatic index (GSI) was used as a physiological indicator of reproductive maturity and spawning season, and it was expressed as the percentage of gonad weight relative to body weight (Brown-Peterson et al. 2011). Individuals larger than 90 mm TL were anesthetized with MS-222, dissected, and gonads were extracted. Gonads were weighed using an analytical balance (0.001 g precision), and GSI was calculated following Delahunty and De Vlaming (1980) (Eq. 1).

2.5. Analysis of relative abundance between Z. koreanus and Z. platypus

To compare changes in population trends of Z. koreanus and Z. platypus, relative abundances of the two species at each survey time were calculated. Time-series graphs were produced, and linear regression analysis was performed to assess increasing and decreasing trends.

2.6. Statistical analysis

One-way ANOVA for the gonadosomatic index (GSI) was performed using IBM SPSS (ver. 21).

3. RESULTS AND DISCUSSION

3.1. Length-weight analysis of Z. koreanus

The results of the length-weight relationship and condition factor analyses for the Z. koreanus population collected from Bongseonsa Stream are shown (Fig. 2). These two indices are used as biological information to indirectly evaluate how healthily a population is growing, depending on environmental conditions at the survey area (Seo 2005;Choi et al. 2011). In general, when the regression coefficient b of the length-weight relationship is less than 3.0, the growth condition is considered poor, and when it is greater than 3.0, the growth condition is considered good (Han et al. 2007). In the case of the condition factor, the growth condition of the population can be inferred from the slope of the regression line. A positive slope indicates a corpulent growth condition through abundant feeding activity, whereas a negative slope indicates that food resources are not abundant and feeding activity is low, resulting in a relatively lean growth condition (Seo 2005;Han et al. 2007). In the 2015, 2021, and 2025 surveys, the regression coefficients (b) of the reintroduced Z. koreanus population in Bongseonsa Stream were 3.1696, 3.2129, and 3.2320, respectively. In all survey years, the b value was >3.0, indicating that body weight increased rapidly in proportion to increases in length. These results suggest that food resources may have been abundant for the Z. koreanus population in the Bongseonsa Stream, and that the well-maintained habitat environment may have positively contributed to the growth condition of the reintroduced population. However, since the survey period corresponds to the spawning season of Z. koreanus, the increase in body weight may have been influenced not only by somatic growth but also by gonadal development.

The slopes of the condition factor, which indicate the corpulence of the population, were 0.0015 (2015), 0.0020 (2021), and 0.0024 (2025) for each survey period, showing positive slopes in all surveys. This indicates that the population is maintaining a healthy state through vigorous feeding, and these results support that food resources are abundant within Bongseonsa Stream, the habitat environment is stably maintained, and the ecological conditions have improved compared to the period prior to the past local extinction.

Including the results of the prior study conducted in 2016 (Wang et al. 2017a), the regression coefficient (b) and the condition factor (K) were compared and analyzed across survey years in each stream for 2021 and 2025, yielding the following results (Table 1). The b value range for Bongseonsa Stream was 3.2389 (2016) to 3.2320 (2025), 3.4020 (2016) to 3.2416 (2025) for Jojong Stream, and 3.4404 (2016) to 3.3560 (2025) for Sudong Stream. Among these, the b value of the Bongseonsa Stream population showed the least temporal change, whereas Jojong and Sudong Streams showed large changes in b depending on the survey time. The K values ranged from 0.0028 (2016) to 0.0024 (2025) for Bongseonsa Stream, 0.0044 (2016) to 0.0029 (2025) for Jojong Stream, and 0.0042 (2016) to 0.0038 (2025) for Sudong Stream, confirming that the Jojong and Sudong Streams had large temporal fluctuations in K. For both indices, the reintroduced population maintained a relatively stable growth condition with small fluctuations, whereas Jojong and Sudong Streams showed sharp decreases in growth condition depending on the time period. In general, fish growth differs according to habitat environment, population density, food resources, and the timing of collection (Song et al. 2012), and populations with abundant food resources and vigorous feeding activity are known to be in good growth condition (Seo 2005;Han et al. 2007). It is thought that factors reducing the growth condition of the Z. koreanus population arise in the Jojong and Sudong Streams depending on the time period. In contrast, except for some upstream sections, most habitats in the Bongseonsa Stream are managed by the Korea National Arboretum, and thus habitat disturbances that may affect the population occur relatively infrequently (Byeon 2011). In contrast, the areas surrounding the Jojong and Sudong Streams are predominantly used as agricultural land, and when farmland is located adjacent to streams, direct and indirect disturbances, such as the inflow of non-point pollution sources and turbid water during rainfall events, are known to occur frequently (Yang 2006). Due to these watershed characteristics, the temporal variability in growth condition is considered to be greater in the source populations than in the reintroduced population.

3.2. Length frequency distribution

The results of the length frequency distribution analysis of the reintroduced Z. koreanus population by time period are as follows (Fig. 3). In the 2015 survey, among a total of 158 individuals, age 0⁺ and age 1⁺ individuals accounted for a high proportion, and individuals in the 60-80 mm range of full age 1⁺ were predominant, showing a population structure centered on relatively young age classes. Individuals of full age 3⁺ and older also appeared in small numbers, but their proportion was low. In the 2021 survey, a total of 270 individuals were collected; the number of individuals in the 1 to 2-year-old range (60-120 mm) increased markedly, and the frequency of occurrence of mature individuals of full age 3 and older over 130 mm also increased. In the 2025 survey, 292 individuals were collected, and all age classes (0⁺, 1⁺, 2⁺, and 3⁺) were distributed relatively evenly. In particular, individuals aged 1-3 years were most abundant, and many large individuals (>130 mm) were observed. According to Im et al. (2010), the length distribution by age class of a typical Z. koreanus population is as follows: age 0⁺ ≤60 mm, full age 1⁺ 60-80 mm, full age 2⁺ 90-120 mm, and full age 3⁺ and older mature individuals ≥130 mm. The 2025 population showed a length frequency distribution most similar to the results of the previous study (Im et al. 2010), suggesting that the Z. koreanus population in Bongseonsa Stream has entered a stabilization phase at present, and if a stable habitat environment without habitat disturbance continues, it is judged that a favorable life-history can be continuously maintained.

3.3. GSI analysis

To compare the reproductive ability between the reintroduced Z. koreanus population and the source populations, GSI analyses by stream were conducted for male and female Z. koreanus populations. The GSI comparison used individuals collected in June 2016, July 2021, and June 2025, and the analysis results are as follows (Fig. 4). For the male population, in 2016, the median GSI of the Bongseonsa Stream (BS) population was analyzed to be higher than those of the Jojong Stream (JJ) and the Sudong Stream (SD), and in 2021 and 2025, the median of the Jojong Stream (JJ) population was found to be the highest; however, the difference from the reintroduced population was confirmed to be small. For the female population, in all survey periods, the median GSI of the BS population was analyzed to remain higher than those of the Jojong and Sudong Stream populations.

The median GSI of males in 2016 was slightly higher in the Bongseonsa Stream (BS) population than in the Jojong (JJ) and Sudong (SD) populations, although no marked differences were observed among streams, and in 2021 and 2025 the highest median value was recorded in the Jojong (JJ) population; one-way ANOVA indicated no significant differences among rivers in 2016 and 2021, whereas in 2025 the Sudong (SD) population, which showed a relatively low median GSI, differed significantly from the Bongseonsa (BS) and Jojong (JJ) populations (p<0.05). In females, the Bongseonsa (BS) population consistently exhibited higher median GSI values than the source populations across all survey years, with significant differences detected between the reintroduced population and the source populations in 2016 (p<0.001) and 2025 (p<0.05), while in 2021 the Sudong (SD) population, characterized by a relatively low median value, showed significant differences compared with the Bongseonsa (BS) and Jojong (JJ) populations (p<0.05).

A higher gonadosomatic index (GSI) is generally associated with stronger habitat competition (Neat et al. 1998). Considering that Z. platypus, a sympatric competitor occupying a similar ecological niche, was the dominant species in the Bongseonsa Stream prior to the reintroduction of the Z. koreanus population, the relatively higher GSI observed in the reintroduced population compared with the source populations is considered to result from interspecific competition with Z. platypus. In addition, populations newly introduced into a habitat are generally reported to exhibit life-history strategies that involve investing greater amounts of energy into reproduction to promote population expansion (Grabowska et al. 2011), and it is therefore presumed that the reintroduced population in the Bongseonsa Stream showed such a life-history strategy for the purpose of population expansion.

3.4. Relative abundance of Z. koreanus and Z. platypus

The results of the analysis of changes in the numerical proportion of Z. koreanus and Z. platypus collected from Bongseonsa Stream from 2007 to 2025 are shown (Fig. 5). The proportion of the Z. koreanus population showed a gradual increasing trend over time after reintroduction, whereas the proportion of the Z. platypus population showed an inversely proportional decreasing trend following the introduction of the Z. koreanus population. In particular, around 2016, the numerical proportions of the two species were reversed, and in the 2025 survey, the Z. koreanus population accounted for more than 80% of the total, indicating a very high occupancy. Z. platypus is known to be a dominant species even in lentic or disturbed river sections with relatively slow flow (Hur et al. 2013;Lee and Choi 2015), whereas the Z. koreanus population tends to be dominant in mid- to upper-reach with low water temperature, fast flow, and good water quality (Hur and Seo 2011;Kim et al. 2022). In the past, Bongseonsa Stream experienced local extinction of the Z. koreanus population due to habitat disturbance and deterioration of water quality; however, currently, the habitat environment has improved through continuous water quality improvement and habitat management. These efforts have positively influenced the stable reproduction and settlement of the reintroduced Z. koreanus population, and the habitat environment-which contains many lotic sections with well-developed riffles has also likely acted as a factor enabling the population to gain an advantage in competition with the sympatric Z. platypus population.