1) 발표자: 심태용

2) 학술대회명: 13th World Congress on Water Resources and Environment

3) 발표 주제: An Ecohydrological Waterfront Management Approach Associated with Civilian Science

4) 발표 내용

The demand for ecologically healthy rivers and water-friendly spaces from citizens is increasing over time (Poff et al. 2010; Lee et al. 2022). However, alterations in the hydrological regime may cause rapid decline in biodiversity of aquatic ecosystems which can lead to reduced services of waterfronts (Bunn and Arthington 2002; Dudgeon et al. 2006). Accordingly, the importance of regional scale application has increased (Dyson et al. 2003; Arthington et al. 2006; Poff et al. 2010), requiring management or decision support systems to consider regional heterogeneity. Thus, this study proposes an ELOHA (Ecological Limitations Of Hydrological Alteration; Poff et al. 2010)-based approach for supporting the determination of release of ecological/environmental flow and timing for streams, while incorporating civilian collaboration. The study area consists of 38 stations located regulated streams in the Han River Basin, South Korea, and 3 sites of interest were selected for evaluation. This study targeted 62 freshwater fish species that occur in the Han River Basin, and the occurrence data were collected from the Water Information System (http://www.water.nier.go.kr/). In this study, period was classified by past (2011-2016) and present (2017-2023). Additionally, civilian fish monitoring data (2014-2023) was acquired from the Suwon Sustainable City Foundation. Missing data of streamflow was acquired from a simulation result that uses the Tank model driven by observed climatic data (NSE=0.233 at downstream junction of evaluation sites). Hydrological alteration was quantified using IHA (Indicators of Hydrological Alteration) variables based on Richter et al. (1996). A total of 34 variables were derived that represents: Magnitude of monthly and annual hydrological condition (14), Magnitude and duration of annual extreme hydrological condition (10), Timing of annual extreme hydrological events (2), Frequency and duration of pulses (4), and Rate and frequency of hydrologic condition changes (4). Nine variables were selected using the variance inflation factor, while ensuring at least one IHA was chosen from each functional group. Ecological alteration was quantified using α- and β-diversity indices. α-diversity was defined as species abundance (number of occurred species), while Jaccard’s similarity coefficient was used as β-diversity (Eq. 1): β = j/(a+b-j) (1) where a = abundance at past, b = abundance at present, and j = number of common species. The ecohydrological response was simulated (Orange version 3.38.1) with multiple models including LR (Logistic Regression), RF (Random Forest), NB (Naive Bayes), kNN (k-Nearest Neighbors), NN (Neural Network), SVM (Support Vector Machine). Additionally, an ensembled model was applied, that follows the majority rule (n≥3). Model performance was evaluated by comparing α- and β-diversity between observed data and simulation results using R2 and %error. The prediction results for the Han River Basin generally showed a low %error in both α- and β-diversity, particularly for RF, NN, and the ensemble model (Figure 1a, b; R2 = 0.735？0.981). Meanwhile, the prediction results for the target site exhibited a higher %error in both α- and β-diversity indices (Figure 1c, d). The high error in β-diversity was mainly due to its sensitivity to hydrological alterations, resulting in low j values. Despite the differences in monitoring process, the results highlight the importance of external validation. Additionally, incorporating variables including water temperature, water quality (e.g., nitrogen, phosphorous, suspended solid), and vegetation may improve model performance.