Global biodiversity frameworks, such as the Kunming-Montreal Global Biodiversity Framework (K-M GBF), primarily rely on taxonomic indicators to monitor and mitigate biodiversity loss. However, this taxonomic focus may not fully capture ecological dynamics that functional diversity can reflect. Our findings -- particularly the trait reorganization observed across land-use gradients -- highlight the interpretive potential of functional metrics in detecting ecological responses to disturbance. This perspective supports the integration of trait-based indicators into biodiversity assessments, offering added value in guiding restoration priorities and conservation strategies aligned with the goals of the K-M GBF.
This study focused on the extensive Han River Basin, which stretches over 32,971 km² and represents one of South Korea's primary river systems, encompassing ~25% of the country (Fig. 4). This basin is located between the latitudes of 36.03° N and 38.55° N and longitudes of 126.24° E and 129.02° E. The rivers converge and proceed towards Seoul, the country's densely populated capital, with a population of ~10 million. The region experiences a four-season climate with notable monsoon rains from June to August, with annual precipitation ranging from 500 to 1500 mm. Seasonal temperature fluctuations are observed, with averages of 10.8 °C in spring, 23.6 °C in summer, 12.6 °C in autumn, and -2.9 °C in winter. This pattern contributes to a predominantly moist environment, with precipitation heavily skewed towards the summer months (June-August), according to the Korea Meteorological Administration. The average altitude of the Han River Basin's landscape is 404.7 m above sea level. Urban development accounted for 8.2% of the area, whereas forests and agricultural land accounted for 71.3% and 16.6%, respectively, including a small fraction (3.9%) of other uses.
Data on benthic macroinvertebrates within the Han River Basin were sourced from the National Aquatic Ecological Monitoring Program (NAEMP) database (http://water.nier.go.kr/). Biannual surveys were conducted in spring (April) and autumn (September) to coincide with river periods of water stability. These months were selected to mitigate the impact of severe summer flooding, which is commonly induced by monsoon weather patterns. September marks the onset of autumn and is particularly notable for the resurgence of aquatic life following flood disturbances. Benthic macroinvertebrates were collected using a Surber sampling net (30 × 30 cm, mesh size 1 mm) following the standardized protocol of the National Aquatic Ecological Monitoring Program (NAEMP) in Korea. While this mesh size may limit the collection of smaller organisms, it has been widely adopted for macroinvertebrate studies due to its suitability for long-term ecological monitoring and its alignment with the focus of this study. At each site, three replicate samples were collected from different locations within the riffle area to capture spatial variability during each sampling event. Sorting and species-level identification of the samples were subsequently performed, and the species abundance was quantified through direct observation or microscopic analysis. Most taxa were identified to the species level, except for certain groups such as Chironomidae (Diptera) and Oligochaeta (Clitellata), which were identified to higher taxonomic levels due to morphological constraints.
Our study incorporated 24 diverse environmental parameters, including meteorological, geographical, land cover, and water quality factors (Table 4). Meteorological data, including average annual temperature, maximum and minimum temperatures, and annual precipitation, were obtained from the Korea Meteorological Administration (http://www.climate.go.kr) using 30-year climate averages to ensure comprehensive and representative environmental assessments. Geographic details were accurately mapped using ArcGIS digital mapping tools (version 10.1) with elevation data derived from digital elevation models (DEM). Land-cover data were extracted within a 500-meter radius from each sampling site, using a circular buffer approach via the Environmental Spatial Information Service (https://egis.me.go.kr/). Water quality data, including dissolved oxygen (DO), biological oxygen demand (BOD), ammonia nitrogen (NH-N), nitrate nitrogen (NO-N), total nitrogen (TN), phosphate-phosphorus (PO-P), total phosphorus (TP), and chlorophyll-a (Chl-a) in the Han River Basin were sourced from Water Quality Monitoring Program database (http://water.nier.go.kr/).
We classified nine functional traits and 28 categories into three distinct groups: life history, morphology, and habit (Table 5). These traits were selected using existing studies and data accessibility.
The life history group comprises three traits: voltinism, life span, and aquatic stages. Voltinism, which refers to the number of generations a species produces annually, is associated with variation in size within a species and may influence intra-species competition and predation rates by altering life history strategies and temporal resource use patterns. Life span, which is indicative of the duration of a species' life cycle, correlates with its reproductive capabilities. Species exhibiting shorter life spans typically show greater resilience to disturbances. Aquatic stages, which describe the dispersal abilities of a species, suggest that species with non-aquatic flying adults possess superior dispersal capabilities.
Morphology is characterized by four traits: maximum size, respiratory organs, shape, and armoring. Maximum size is directly linked to an organism's fecundity, position within the food web, and mobility. The type of respiratory organ possessed by an organism indicates its adaptability to various environmental oxygen levels. The shape of an organism influences its mobility and is an adaptation to the velocity of the water flow it experiences, whereas its armoring level indicates its ability to endure physical and environmental challenges.
Finally, the habit category comprises locomotion and functional feeding habits. The mode of locomotion and its relationship with the substrate play crucial roles in habitat connectivity and ecosystem resilience, influencing microhabitat choice. Functional feeding groups highlight the role of organisms in trophic dynamics and their reactions to environmental changes.
To compare the taxonomic and functional diversity of benthic macroinvertebrates among streams with different land cover types, we considered three categories: urban/built-up areas, agricultural areas, and forest areas. Urban and agricultural streams were defined as those where more than 60% of the buffer zone was urban or agricultural land, respectively, while forest streams were classified as those with more than 80% forest cover. A total of 121 sites near rivers were surveyed, including 39 urban, 34 agricultural, and 48 forested sites.
Then, we applied for five steps of data analysis.
First, representative functional diversity (FD) indices, including functional richness (Fric), functional evenness (Feve), functional divergence (Fdiv), and Rao's quadratic entropy indices (RaoQ), were computed. These indices were derived using the "dbFD" function within the 'FD' package in R, version 3.6.2. Fric measures the extent of functional traits present within species communities, with lower Fric values suggesting limited exploitation of available ecological niches and resources by the species. Feve quantifies how evenly species are distributed across the functional trait space, where a high Feve value denotes more thorough and efficient use of available resources by the species. Fdiv assesses how species with higher abundances are spread relative to the center of the functional space, indicating the level of ecological niche differentiation and competition among species. A greater Fdiv value implies high niche differentiation among species, reduced niche overlap, and less resource competition. Moreover, RaoQ combines aspects of Fric and Fdiv and can help evaluate the degree of redundancy or complementarity among these functional diversity indices.
Second, to assess the variations in community structures between the land use types, we applied the Permutational Multivariate Analysis of Variance (PERMANOVA). This analysis was performed using the adonis function of the vegan package in R. PERMANOVA uses distance matrices to analyze variances across different sources, focusing specifically on the spatial positioning of group centroids rather than their dispersion. This method is highly effective in detecting differences in community composition resulting from various conditions. By conducting PERMANOVA, we aimed to determine whether significant differences existed in the taxonomic and functional diversity of benthic macroinvertebrate assemblages across different land cover types based on the spatial distribution of centroids.
Third, to explore the distribution patterns of benthic macroinvertebrate species in relation to land-use types, we used NMDS with the Bray-Curtis dissimilarity index to quantify compositional differences among communities. The NMDS method is well-suited for ecological community data, which are often characterized by many zeros and non-normal distributions, and is commonly applied to taxa such as benthic macroinvertebrates. To identify the optimal configuration that reflects the true spatial distribution of species with minimal distortion, we employed the metaMDS function from the vegan package in R, targeting the solution with the lowest stress value as the best representation of the data. This allowed us to visually interpret the impact of land use on the community structure of benthic macroinvertebrates in a multidimensional space. Additionally, to understand the relationship between the observed patterns of benthic macroinvertebrate diversity and environmental variables within these land-use contexts, we conducted an analysis using the envfit function in the vegan package in R. This analysis mapped environmental gradients onto the NMDS ordination results, providing insights into how specific environmental factors, such as water quality, are associated with the distribution of benthic macroinvertebrate species across different land-use types. This comprehensive approach enabled us to decipher the intricate interactions between land use, environmental conditions, and diversity of benthic macroinvertebrate communities.
Fourth, to ascertain the level of agreement between taxonomic and functional assemblages within the Han River Basin, we applied Procrustes Rotation Analysis (PRA) as outlined by Peres-Neto and Jackson. This analysis used the NMDS ordination results obtained in the previous step, overlaying the NMDS ordination outputs of taxonomic and functional diversity. This analytical approach overlays two distinct sets of ordination data, adjusting one through various transformations, such as scaling, rotation, reflection, and expansion, to minimize the total squared deviations (m) between their ordination points. A reduced m value implies higher congruence between the two community datasets. Importantly, the statistical correlation coefficient (R) derived from this minimized m served as a quantitative indicator of the degree of community agreement, where a larger R-value signified stronger congruence between taxonomic and functional assemblages. Significance testing for PRA was conducted using the protest method, which uses a permutation test (n = 999) to evaluate whether the observed m is significantly lower than expected by random chance. This involves randomly shuffling one of the ordination matrices, recalculating the m² for each permutation, and assessing whether the proportion of permutations with an m² greater than or equal to the observed m² provides a statistically significant result, as indicated by the calculated p-value. This comprehensive approach statistically validates the extent to which taxonomic and functional community structures reflect similar spatial patterns, strengthening interpretations of biodiversity responses to land-use changes. Taxonomic data were log-transformed to minimize the influence of abundant species and subsequently standardized before NMDS analysis, which provided the ordination results for both the visualization of community-level patterns and subsequent Procrustes Rotation Analysis (PRA) to assess congruence between taxonomic and functional diversity. Functional diversity metrics were calculated based on standardized trait data without additional transformations.
Finally, we employed the Indicator Species Analysis (IndVal) method to extract indicator species and traits of benthic macroinvertebrates that were distinctively associated with different types of land use (forest, agriculture, and urban). The IndVal was calculated from the relative abundance and frequency of the species (or trait), producing a value scale ranging from 0 (no indication) to 100 (perfect indication). We considered species with an IndVal greater than 25, which also showed statistical significance (p < 0.05), as indicator species. We conducted a site randomization test to assess the significance by redistributing sites across groups with 9999 permutations. This approach allowed us to pinpoint specific trait categories that were indicative of certain land-use groups based on their unique ecological characteristics. Only trait categories that demonstrated a significant (p < 0.05) maximal indicator value for a particular land-use type were designated as indicator trait categories for that specific group of sites. This strategy was instrumental in identifying the ecological signatures of benthic macroinvertebrate communities across various landscapes, thereby providing insights into how land use influences community composition and trait distribution. The 'labdsv' package for the indicator species analysis was applied in R.