Application of Machine Learning for eutrophication analysis and algal bloom prediction in an urban river: A 10-year study of the Han River, South Korea.

The increasing release of nutrients to aquatic environments has led to great concern regarding eutrophication and the risk of unwanted algal blooms. Based on observational data of 20 water quality parameters measured on a monthly basis at 40 stations from 2011 to 2020, this study applied different Machine Learning (ML) algorithms to suggest the best option for algal bloom prediction in the Han River, a large river in South Korea. Eight different ML algorithms were categorized into several groups of statistical learning, regression family, and deep learning, and were then compared for their suitability to predict the chlorophyll-derived trophic index (TSI-Chla). ML algorithms helped identify the most important water quality parameters contributing to algal bloom prediction. The ML results confirmed that eutrophication and algal proliferation were governed by the complex interplay between nutrients (nitrogen and phosphorus), organic contaminants, and environmental factors. Of the models tested, the adaptive neuro-fuzzy inference system (ANFIS) exhibited the best performance owing to its consistent and outperforming prediction both quantitatively (i.e., via regression) and qualitatively (i.e., via classification), which was evidenced by the lowest value of mean absolute error (MAE) of 0.09, and the highest F1-score, Recall and Precision of 0.97, 0.98 and 0.96, respectively. In a further step, a representative web application was constructed to assist common users to predict the trophic status of the Han River. This study demonstrated that ML techniques are not only promising for highly accurate water quality modeling of urban rivers, but also reduce time and labor intensity for experiments, which decreases the number of monitored water quality parameters, providing further insights into the driving factors of water quality deterioration. They ultimately help devise proactive strategies for sustainable water management.

Changes in adsorption mechanisms of radioactive barium, cobalt, and strontium ions using spent coffee waste biochars via alkaline chemical activation: Enrichment effects of O-containing functional groups.

The single adsorption of radioactive barium (Ba(II)), cobalt (Co(II)), and strontium (Sr(II)) ions using pristine (SCWB-P) and chemically activated spent coffee waste biochars with NaOH (SCWB-A) were thoroughly explored in order to provide deeper insights into the changes in their adsorption mechanisms through alkaline chemical activation. The greater removal efficiencies of SCWB-A (76.6-97.3%) than SCWB-P (45.6-75.2%) and the consistency between the adsorptive removal patterns (Ba(II) > Sr(II) > Co(II)) and oxygen bond dissociation enthalpies (BaO (562 kJ/mol) > SrO (426 kJ/mol) > CoO (397 kJ/mol)) of radioactive species supported the assumption that the adsorption removal of radioactive species with spent coffee waste biochars highly depended on the abundances of O-containing functional groups. The calculated R2 values of the pseudo-first-order (SCWB-P = 0.998-0.999; SCWB-A = 0.850-0.921) and pseudo-second-order kinetic models (SCWB-P = 0.988-0.998; SCWB-A = 0.935-0.966) are evident that the physisorption mainly controlled the adsorption of radioactive species toward SCWB-P and the chemisorption played a crucial role in their adsorptive removal with SCWB-A. From the calculated intra-particle diffusion, isotherm, thermodynamic parameters, it can be concluded that the intra-particle diffusion and monolayer adsorption primarily governed the adsorption of radioactive species using SCWB-P and SCWB-A, and their adsorption processes occurred spontaneously and endothermically. The dominant adsorption mechanism of spent coffee waste biochars was changed from physisorption (ΔH° of SCWB-P = 21.6-29.8 kJ/mol) to chemisorption (ΔH° of SCWB-A = 42.4-81.3 kJ/mol) through alkaline chemical activation. The distinctive M-OH peak in the O1s XPS spectra of SCWB-A directly corresponding to the decrease in the abundances of O-containing functional groups confirms again that the enrichment of O-containing functional groups markedly facilitated the adsorption removal of radioactive species by chemisorption occurred at the inner and outer surfaces of spent coffee waste biochars.

Evaluating physico-chemical influences on cyanobacterial blooms using hyperspectral images in inland water, Korea.

Understanding harmful algal blooms is imperative to protect aquatic ecosystems and human health. This study describes the spatial and temporal distributions of cyanobacterial blooms to identify the relations between blooms and environmental factors in the Baekje Reservoir. Two-year cyanobacterial cell data at one fixed station and four remotely sensed distributions of phycocyanin (PC) concentrations based on hyperspectral images (HSIs) were used to describe the relation between the spatial and temporal variations in the blooms and the affecting factors. An artificial neural network model and a three-dimensional hydrodynamic model were implemented to estimate the PC concentrations using remotely sensed HSIs and simulate the hydrodynamics, respectively. The statistical test results showed that the variations in the cyanobacterial biomass depended significantly on variations in the water temperature (slope = 0.13, p-value < 0.01), total nitrogen (slope = -0.487, p-value < 0.01), and total phosphorus (slope = 20.7, p-value < 0.05), whereas the variation in the biomass was moderately dependent on the variation in the outflow (slope = -0.0097, p-value = 0.065). Water temperature was the main factor affecting variations in the PC concentrations for the three months from August to October and was significantly different for the three months (p-value < 0.01). Hydrodynamic parameters also had a partial effect on the variations in the PC concentrations in those three months. Overall, this study helps to describe spatial and temporal variations in cyanobacterial blooms and identify the factors affecting the variation in the blooms. This study may play an important role as a basis for developing strategies to reduce bloom frequency and severity.

Application of a combined three-stage system for reclamation of tunnel construction wastewater.

A combined three-stage system, (1) coagulation (2) zeocarbon filtration and (3) membrane filtration, a combination of microfiltration (MF) and reverse osmosis (RO), was investigated for reclamation of tunnel construction wastewater having a salinity of 10.8-12.9‰ and a concentration of suspended solids (SS) in the range of 264-1084 mg/L. The initial stages - coagulation, zeocarbon filtration and MF - served as a precursor to RO membrane filtration to successfully reduce water contaminants to less than 0.2 nephelometric turbidity units (NTU) of turbidity, thereby minimizing the potential for fouling. The RO system subsequently removed over 99% of remaining pollutants including ionic substances, resulting in less than 0.02 NTU turbidity, less than 0.04 mg/L total nitrogen (TN) and less than 0.01 mg/L total phosphorus (TP). Also, addition of an RO system markedly reduced high salt concentrations (high chloride (Cl(-)) concentrations) in the wastewater, exceeding 99% salt elimination. Thus, reclaimed water from our combined system met and exceeded currently regulatory quality standards for wastewater reuse (turbidity ≤ 2.0 NTU; TN ≤ 10 mg/L; TP ≤ 0.5 mg/L; Cl(-) ≤ 250 mg/L).

Characterization of membrane foulants in a pilot-scale tunnel construction wastewater treatment process.

A pilot-scale combination of a microfiltration (MF) and a reverse osmosis (RO) membrane system was applied on-site to treat tunnel construction wastewater. The MF membrane system initially removed contaminants (turbidity of less than 0.3 NTU) in the form of particulate materials in the feed water, thereby allowing the combined MF-RO system to efficiently remove more than 99% of known organic and inorganic contaminants and qualify the reclaimed water for reuse. The MF membrane autopsy analysis using X-ray fluorescence (XRF) and inductively coupled plasma-mass spectrometry (ICP-MS) revealed that the dominant foulants were inorganic deposits involving Si, Al and Fe, comprising the main components of cement materials, as well as deposits involving heavy metals such as Mn, Cu and Zn in the form of particles. Thus, thick cake contaminants shown by field emission scanning electron microscope (FE-SEM) images might be induced via suspended solids consisting of cement and clay materials and metals.

Evaluating causes of trends in long-term dissolved reactive phosphorus loads to Lake Erie.

Renewed harmful algal blooms and hypoxia in Lake Erie have drawn significant attention to phosphorus loads, particularly increased dissolved reactive phosphorus (DRP) from highly agricultural watersheds. We use the Soil and Water Assessment Tool (SWAT) to model DRP in the agriculture-dominated Sandusky watershed for 1970-2010 to explore potential reasons for the recent increased DRP load from Lake Erie watersheds. We demonstrate that recent increased storm events, interacting with changes in fertilizer application timing and rate, as well as management practices that increase soil stratification and phosphorus accumulation at the soil surface, appear to drive the increasing DRP trend after the mid-1990s. This study is the first long-term, detailed analysis of DRP load estimation using SWAT.

Interpretation of seasonal water quality variation in the Yeongsan Reservoir, Korea using multivariate statistical analyses.

The Yeongsan (YS) Reservoir is an estuarine reservoir which provides surrounding areas with public goods, such as water supply for agricultural and industrial areas and flood control. Beneficial uses of the YS Reservoir, however, are recently threatened by enriched non-point and point source inputs. A series of multivariate statistical approaches including principal component analysis (PCA) were applied to extract significant characteristics contained in a large suite of water quality data (18 variables monthly recorded for 5 years); thereby to provide the important phenomenal information for establishing effective water resource management plans for the YS Reservoir. The PCA results identified the most important five principal components (PCs), explaining 71% of total variance of the original data set. The five PCs were interpreted as hydro-meteorological effect, nitrogen loading, phosphorus loading, primary production of phytoplankton, and fecal indicator bacteria (FIB) loading. Furthermore, hydro-meteorological effect and nitrogen loading could be characterized by a yearly periodicity whereas FIB loading showed an increasing trend with respect to time. The study results presented here might be useful to establish preliminary strategies for abating water quality degradation in the YS Reservoir.