Early warning systems for harmful cyanobacterial blooms (HCBs) that enable precautional control measures within water bodies and in water works are largely based on inferential time-series modelling. Among deep learning techniques, convolutional neural networks (CNNs) are widely applied for recognition of pictorial, acoustic and thermal images. Time-frequency images of environmental drivers generated by wavelets may provide crucial signals for modelling of HCBs to be recognized by CNNs. This study applies CNNs for time-series modelling of HCBs of Microcystis sp. in four South Korean rivers between 2016 and 2022 by means of time-frequency images of environmental drivers within the lead time of HCBs. After estimating the cardinal dates of beginning, peak, and ending of HCBs, wavelet analysis identified key drivers by phase analysis and generated time-frequency images of the drivers within the cardinal dates for 3, 4 and 5 years. Performances of CNNs were compared in terms of four determinants of input images: methods of estimating critical timings, the number of segments, time-series continuity, and image size. The resulting CNNs predicted high or low intensities of HCBs with a mean accuracy of 97.79 ± 0.06% and F1-score 97.49 ± 0.06% for training dataset, and a mean accuracy of 95.01 ± 0.06% and F1-score 93.30 ± 0.07% for testing dataset. Predictions of Microcystis abundances by CNNs achieved a mean MSE of 2.58 ± 2.46 and a mean R² of 0.78 ± 0.20 for training, and a mean MSE of 2.76 ± 2.42 and a mean R² of 0.55 ± 0.20 for testing dataset. Precipitation and discharge appeared to be the best performing drivers for qualitative and quantitative predictions of HCBs pointing at the nonstationary nature of river habitats. This study highlights the opportunities of time-series modelling by CNNs driven by wavelet generated time-frequency images of key environmental variables for forecasting of HCBs.

Thermal stratification is an important driver shaping phytoplankton community and their habitat condition in freshwater ecosystems. However, studies on river stratification have been restricted to rivers below dams or reservoirs affected by their water release and lacked examination of diel stratification and its impact on phytoplankton, in particular. Therefore, this study aimed to determine the degree of thermal stratification, its environmental drivers, and the response of water quality and phytoplankton community against stratification in the mid-lower reach of the Nakdong River, whose morphology has been highly modified, including the construction of eight weirs. We implemented vertical temperature profiling at three study sites, both seasonally and diurnally. From this data, we calculated three stratification indices: relative water column stability (RWCS), Schmidt stability (S), and maximum temperature gradient (Max). These indices showed that most sites experienced diel stratification during summer (mean = RWCS 74.3, S 41.5 J m⁻², Max 0.9 °C m⁻¹). Principal component analysis showed that stratification significantly led to seasonal and diel variations in the water environment. Solar radiation and air temperature were positive controllers, while a negative controller (in this case, the river flow rate) existed only for diel variation in the stratification. The seasonal shifts in phytoplankton community structure were either insensitive or showed a limited response to the stratification indices. In summer, Microcystis cell abundance and accumulation into the surface water was positively affected by the diel variations in the stratification indices and thermocline instead of with other temperature and nutrient variables. Overall, the results suggest that the river has summer stratification, which is involved in amplifying cyanobacterial bloom intensity. Without a suppressing factor, summer stratification is expected to be recurrent in the river, and thus mitigating the developed stratification is needed by promptly regulating the river flow.

Freshwater ecosystems face escalating degradation, demanding real-time, scalable, and biodiversity-aware monitoring solutions. This review proposes an integrated framework combining artificial intelligence (AI), geographic information systems (GISs), and environmental DNA (eDNA) to overcome these limitations and support next-generation river health assessment. The AI-GIS-eDNA system was applied to four representative river basins—the Mississippi, Amazon, Yangtze, and Danube—demonstrating enhanced predictive accuracy (up to 94%), spatial pollution mapping precision (85–95%), and species detection sensitivity (+18–30%) compared to conventional methods. Furthermore, the framework reduces operational costs by up to 40%, highlighting its potential for cost-effective deployment in low-resource regions. Despite its strengths, challenges persist in the areas of regulatory acceptance, data standardization, and digital infrastructure. We recommend legal recognition of AI and eDNA indicators, investment in explainable AI (XAI), and global data harmonization initiatives. The integrated AI-GIS-eDNA framework offers a scalable and policy-relevant tool for adaptive freshwater governance in the Anthropocene.

Roadkill has been one of the most problematic issues with wildlife under urbanization. South Korea, one of the fastest industrializing countries in the world, has been affected by the same roadkill issues and has researched how to mitigate wildlife–vehicle collisions (WVCs). In the present study, we aimed to (i) review scientific studies concerning roadkill conducted in South Korea (51 domestic, 15 international papers, 19 reports, and 1 thesis), (ii) compare bibliographic networks between international and South Korean roadkill studies, and (iii) discuss future challenges. From a search in the core collection of Web of Science peer-reviewed papers published from 1992 to 2022, keywords were extracted to create a bibliometric visualization map, using VOSviewer. Among the 85 articles related to WVCs, those about eco-corridors have steadily increased since 1998, while mitigation system and roadkill articles have been published since the mid-2000s. With increased awareness, more systems have been established, and research related to WVCs has been increasingly established. Currently, there are two systematic roadkill monitoring systems in South Korea, and an increasing number of modeling studies have suggested where roadkill hotspots are located. How to mitigate roadkill incidents has also been explored. A comparison of associations between international and Korean research shows that the network structures of Korean society were more disconnected and less dispersed. In addition, the keywords were narrower than those used in studies from the international community. Although studies on landscape connectivity and road ecology are few, their number and scope in South Korea have increased and broadened. These studies could be more rapidly developed in the future as some systems are equipped.